Báo cáo khoa học: Perturbation of membrane microdomains in GLC4 multidrug-resistant lung cancer cells ) modification of ABCC1 (MRP1) localization and functionality

Perturbation of membrane microdomains in GLC4 multidrug-resistant lung cancer cells ) modification of ABCC1 (MRP1) localization and functionality Carole Marbeuf-Gueye, Ve´ rene Stierle, Paiwan Sudwan, Milena Salerno and Arlette Garnier-Suillerot

Laboratoire Biophysique Mole´ culaire, Cellulaire et Tissulaire (BioMoCeTi), UMR CNRS 7033, Universite´ Paris 13 et Paris 6, Bobigny, France

Keywords ABCC1 functionality; ABCC1 localization; membrane cholesterol level; multidrug resistance; raft

Correspondence M. Salerno, Laboratoire BioMoCeTi, CNRS UMR 7033, UFR SMBH, 74 rue Marcel Cachin, 93017 Bobigny, Cedex France Fax: +33 1 48 38 88 88 Tel: +33 1 48 38 77 48 E-mail: m.salerno@smbh.univ-paris13.fr

(Received 20 November 2006, revised 20 December 2006, accepted 10 January 2007)

doi:10.1111/j.1742-4658.2007.05688.x

its

The multidrug resistance-associated protein transporter ABCC1 (MRP1) is an integral plasma membrane protein involved in the multidrug resistance phenotype. It actively expels a number of cytotoxic molecules from cells. To gain insight into the modulation of the functional properties of this integral membrane protein by cholesterol, a main component of the lipid bilayer, we used multidrug-resistant GLC4 ⁄ ADR cells, which overexpress MRP1. Upon altering the plasma membrane cholesterol content of these cells, membrane localization and the activity of MRP1 were analyzed. A detergent-free methodology was used to separate ‘light’ and ‘heavy’ plasma membrane fractions. Our data show that MRP1 was exclusively found in ‘light’ fractions known as L0 phase membrane microdomains, together with (cid:2) 23% of gangliosides GM1 and 40% of caveolin-1. Depletion of the mem- brane cholesterol level to 40% by treatment with the cholesterol-chelating agent methyl-b-cyclodextrin did not modify MRP1 activity, as evidenced either by the rate of efflux of pirarubicin or that of glutathione. Further cholesterol depletion below 40% yielded both a partial shift of MRP1 to the high-density fraction and a decrease of functionality. Taken together, these data suggest that MRP1 funtionality depends on its local- ization in cholesterol-rich membrane microdomains.

supposed to share

A major obstacle to the success of chemotherapy is multidrug resistance (MDR) [1]. The classical MDR phenotype is characterized by cross-resistance to a wide variety of structurally unrelated chemotherapeutic agents of natural origin after exposure to only one. in vitro, of ABCB1 (P-glycoprotein, P-gp) The roles, and ABCC1 (multidrug resistance-associated protein, MRP1) in MDR are uncontested [2–5]. Both proteins are members of the ATP-binding-cassette (ABC) super- family of transport proteins, which lower the intracel- lular drug content through active drug efflux.

The well-documented membrane protein sensitivity to a lipid environment has led to the hypothesis of functional cross-talk between membrane proteins and membrane microdomains such as rafts and caveolae, inspiring numerous studies over the past 10 years [6]. Both types of microdomain, rich in cholesterol and glycosphingolipids, are similar physicochemical properties, in particular a Lo phase, less fluid than a liquid crystal phase [7]. Raft domains, and by extension caveolae, are generally isolated from intact cells on the basis of their insolubility in cold,

Abbreviations ABC, ATP-binding cassette; COase, cholesterol oxidase; CTB, cholera toxin B; GSH, glutathione; GST, glutathione S-transferase; HRP, horseradish peroxidase; MbCD, methyl-b-cyclodextrin; MCB, monochlorobimane; MDR, multidrug resistance; MRP1, multidrug resistance- associated protein; P-gp, p-glycoprotein; PIRA, pirarubicin.

FEBS Journal 274 (2007) 1470–1480 ª 2007 The Authors Journal compilation ª 2007 FEBS

1470

C. Marbeuf-Gueye et al.

ABCC1 localization in light membrane microdomains

nonionic detergents such as Triton X-100, Brij 96 or 98, Tween-20 or Chaps octylglucoside. In fact, the widely different sensitivities of membrane proteins to extraction by various detergents is well documented, and, barring possible artefacts, suggests that lipid rafts represent a heterogeneous collection of domains show- ing differences in both lipid and protein content [8,9]. Other protocols without detergent have also been used, and present the advantage of not making a preselec- tion of the microdomain protein or lipidic components by the detergent [10].

mediated efflux of the anthracycline derivative, PIRA, and of GSH. Fifteen independent experiments were performed on 15 different days. During that time, the resistance factor varied slightly, yielding a rate of MRP1-mediated efflux of PIRA, Va, which varied within the range 0.4–0.8 · 10)18 molÆcell)1Æs)1. Concom- itantly, the amount of cholesterol present exhibited some variation within the range 1.3–2.0 · 10)14 molÆcell)1. For this reason, the data are presented in term of ratios between the rate of PIRA efflux in the presence of modulator to that in the absence of modulator; the same holds for the rate of GSH efflux and for the cholesterol concentration. In non-MbCD-treated cells, no correlation was found between the membrane cholesterol content and the MRP1-mediated efflux of PIRA.

Effect of MbCD on cellular cholesterol content

Recent studies of the colocalization of ABC transport- ers such as P-gp in raft or caveolae microdomains have been interpreted in different ways [11]. Similar studies on MRP1 are scarce, and have only used methods with the detergents Lubrol or Triton X-100 [12–15]. Hinrichs et al. [12] found that MRP1 was predominantly located in the Lubrol-based fraction of detergent-insoluble membrane domains in the colchicine-selected cell line HT29-col. Although there was a cocalization of caveo- lin-1 and MRP1 in the low-density fraction, they pro- posed a different localization within the microdomains, on the basis of immunoprecipitation experiments with sequential Triton X-100 and Lubrol extraction.

Cells were incubated with 15 mm MbCD for various times ranging from 30 s to 20 min, and the cholesterol content was determined. As can be seen in Fig. 1, cellular cholesterol depletion was fast, 50% depletion being observed after less than 1 min, and 80% deple- tion being observed after 10 min of incubation; this was not modified by a longer time of incubation.

its

Effect of cholesterol depletion on the rate of MRP1-mediated efflux of PIRA

þ

The aim of this work was to determine the localiza- tion of MRP1 in the plasma membrane and to deter- mine whether localization the modification of would modify its functional properties. Therefore, using a detergent-free method to isolate the low-den- sity microdomains of the MDR lung cancer cell line GLC4 ⁄ ADR, we studied the effect of cholesterol depletion by MbCD (a) on the localization of MRP1 and (b) on the functionality of MRP1, i.e. its ability to pump out an anthracycline derivative, pirarubicin (PIRA) on the one hand and reduced glutathione (GSH) on the other hand. Our data show that MRP1 is totally localized in the low-density membrane frac- tion, together with some GM1 and caveolin-1. The reduction of cholesterol to 40% did not affect the anthracycline transport by MRP1. However, when more cholesterol was removed, we observed a shift of low-density to the high-density MRP1 from the its fraction membrane, paralleled by a decrease of functionality.

Results

Cells were incubated for 10 min with various MbCD concentrations, ranging from 0 to 12 mm. This incuba- tion yielded an increase in the cell membrane permeab- ility to PIRA. In other words, the rate of passive uptake of the drug in the presence of MbCD, V MbCD , is higher than the rate, V+, in its absence; for instance, a 1 h incubation of cells with 15 mm MbCD yielded a three-fold increase of the rate of PIRA passive uptake. In order to determine an eventual impact of MbCD on the rate of MRP1-mediated efflux of PIRA, experi- ments were performed with energy-depleted resistant cells, which were incubated with MbCD as previously described. At steady state, the incorporation of PIRA was the same as in sensitive cells [16–20]. At this stage, the addition of 5 mm glucose yielded an increase in the fluorescent signal due to the MRP1-mediated efflux of the drug only, and as there was no concentration gradient across the plasma membrane, the effect of MbCD on passive PIRA diffusion did not need to be taken into account (Fig. 2).

The GLC4 small lung cancer cell line, and its multi- drug-resistant counterpart GLC4 ⁄ ADR, were used to study the influence of cholesterol on the activity of MRP1 in its membrane microenvironment.

The amount of intracellular nonesterified cholesterol together with the rate of MRP1-

was determined,

The cholesterol content of the cells was also deter- mined on the same samples. Figure 3 shows the plots ⁄ Va(cid:2) as a function of [Chol]MbCD ⁄ [Chol](cid:2). of V MbCD a

FEBS Journal 274 (2007) 1470–1480 ª 2007 The Authors Journal compilation ª 2007 FEBS

1471

C. Marbeuf-Gueye et al.

ABCC1 localization in light membrane microdomains

a

Fig. 1. Time course of cholesterol depletion in GLC4 ⁄ ADR cells by incubation with MbCD. Cells (106 ⁄ mL) were incubated in Hepes buffer in the presence of 15 mM MbCD for various times, ranging from 0 to 30 min. Data points are from a representative experiment (n ¼ 3).

a

Fig. 3. Variation of the MRP1-mediated efflux of PIRA as a function of the fraction of cholesterol present in the cells. V MbCD ⁄ Va(cid:2) is plot- ted as a function of [Chol]MbCD ⁄ [Chol](cid:2). Cholesterol depletion was obtained by 10 min of incubation of cells with different MbCD con- centrations. Va(cid:2) and V MbCD are the rates of MRP1-mediated efflux of PIRA, before and after, respectively, treatment with MbCD. [Chol](cid:2) and [Chol]MbCD are the cellular cholesterol contents before and after treatment with MbCD. They are the mean ± SE of five independent experiments performed on different days.

a

cholesterol depletion to 40% of the initial content did not give rise to modification of the ratio V MbCD ⁄ Va(cid:2) that is characteristic of the MRP1-mediated efflux of PIRA. However, further depletion yielded a decrease of the ratio.

Effect of cholesterol depletion on the rate of MRP1-mediated efflux of GSH

VGSH was determined after incubation of cells with MbCD for 5 min. Measurement was also performed after incubation of cells for 30 min with 5 mm MbCD; in both cases, no modification of the rate of MRP1- mediated efflux of GSH was observed.

Effect of Triton derivatives on the MRP1-mediated efflux of PIRA

Fig. 2. Incorporation of PIRA in energy-depleted GLC4 ⁄ ADR cells and determination of the active efflux rate (Va). Cells, 106 ⁄ mL, were incubated for 15 min with MbCD at concentrations equal to 0 mM (a), 10 mM (b). Cells were then centrifuged as explained under Experimental procedures, and incubated with 1 lM PIRA. The fluorescence intensity at 590 nm (kex ¼ 480 nm) was recorded as a function of the time of incubation of cells with PIRA. The act- ive efflux rate (Va) was determined from dF ⁄ dt after the addition of glucose.

a

Va(cid:2) and V MbCD are the rates of MRP1-mediated efflux of PIRA before and after treatment with MbCD, [Chol](cid:2) and [Chol]MbCD are the cellular respectively. cholesterol contents before and after treatment with MbCD. They are the mean of five independent experi- ments performed on different days. As can be seen,

Triton X-45 (n ¼ 5) and Triton X-165 (n ¼ 16) were used at nonpermeabilizing concentrations. The active efflux was measured: energy-depleted cells were incuba- ted with 1 lm PIRA in the presence of various concen- tration of Triton, either X-45 or X-165. At steady state, glucose was added and the rate, Va, of pump- mediated efflux of drug was measured. Figure 4 shows the plot of V T a ⁄ Va(cid:2) as a function of Triton concentra- tion, where V T a and Va(cid:2) are the rates of efflux in the

FEBS Journal 274 (2007) 1470–1480 ª 2007 The Authors Journal compilation ª 2007 FEBS

1472

C. Marbeuf-Gueye et al.

ABCC1 localization in light membrane microdomains

A

B

Fig. 4. Effect of Triton on MRP1-mediated efflux of PIRA. V T a ⁄ Va(cid:2) is plotted as function of the Triton X-45 (d) or Triton X-165 (h) added to the cells. V T a and Va(cid:2) are the rates of MRP1-mediated efflux of PIRA in the presence or in the absence, respectively, of Triton.

β

C

absence and presence of Triton, respectively. Fifty per cent inhibition of PIRA efflux was observed with 16 ± 2 lm Triton, either X-45 or X-165. Similar experiments were also performed with Triton X-100 (data not shown), and strictly analogous data were obtained.

MRP1, GM1 and caveolin-1 distribution within membrane fractions

Fig. 5. Detection of MRP1 in GLC4 ⁄ ADR cell lysates. GLC4 ⁄ ADR cells were lysed by sonication, before (empty square) or after (full circle) treatment with MbCD. The lysate was separated by density centrifugation, and collected from the top in 1 mL fractions. Frac- tion 1 is from the top of the gradient. (A) MRP1 expression was analyzed with IMAGE J v1.30 software. In the absence of treatment with MbCD, fraction 4 contains MRP1 (100%). After treatment with MbCD, MRP1 is found in fraction 4 (70%) and in fraction 6 (30%). Immunodetection of MRP1 in the absence (B) and in the presence (C) of treatment with MbCD. S and R stand for GLC4 (sensitive) and GLC4 ⁄ ADR (resistant) cells.

6–9. The amount of caveolin-1 present in fraction 4 also decreased slightly.

Discussion

Membrane lipids do not form a homogeneous phase consisting of glycerophospholipids and cholesterol, but a mosaic of domains with unique biochemical composi- tions. Among these domains, those containing sphingol- ipids and cholesterol, referred to as lipid rafts, have received much attention in the past few years [21]. Tight interactions between the sterol and the sphingolipids result in the formation of domains that are resistant to solubilization in detergents at low temperature [22–24] and are destabilized by cholesterol-depleting and sphingolipid-depleting agents. Caveolae can be consid- ered as a functional specialized raft, because they con- tain several specific lipids and proteins typical of

In order to determine the effect of MbCD on the integrity of rafts and the localization of MRP1, GM1 and caveolin-1, GLC4 ⁄ ADR cells were treated or not treated with MbCD. Low-density membrane fractions were then isolated, and the distributions of MRP1, GM1 and caveolin-1 were assessed by western blotting and dot blotting of the analytical density gradients (Figs 5 and 6). Fraction 1 represents the top of the gradient, and fraction 11 is the bottom of the gradient. Fractions 3 and 4 and fractions 6 and 7 contain the 10% ⁄ 22% and 22% ⁄ 35% sucrose interface, respect- ively. In untreated cells, 100% of MRP1 was found in fraction 4, 40% of GM1 was found in fraction 4 and (cid:2) 10–15% in fractions 5–9, fractions 8 and 9 corres- ponding to the cytoplasm and intracellular membranes, and caveolin-1 was present in all the fractions but was slightly more abundant in fraction 4 than in the others. After mild treatment with MbCD, the amount of MRP1 present in fraction 4 decreased to 70%, with a shift of 30% to fraction 6; GM1 present in fractions 4 decreased to (cid:2) 25%, whereas in fraction 6 it increased to 20%, and it remained equal to 15% in fractions

FEBS Journal 274 (2007) 1470–1480 ª 2007 The Authors Journal compilation ª 2007 FEBS

1473

C. Marbeuf-Gueye et al.

ABCC1 localization in light membrane microdomains

A

organism, as reflected by the relatively weak phenotype seen in the knockout mice.

Caveolae have recently been shown to be involved in MDR. However, reports on the effect of caveolins on the development of MDR are controversial: Caveo- lin-1 expression has been shown to be upregulated in MDR phenotypes in a number of human cell lines [27,28], but expression of caveolin-1 and caveolin-2 has lines that not been detected in several MDR cell express high levels of P-gp [29–31], suggesting that caveolin-1 expression is not associated with that of P-gp protein or MDR1 genes [32]. Increased levels of glucosylceramide have been observed in many MDR tumor cells [33–35].

B

Fig. 6. Detection of GM1 and caveolin-1 in GLC4 ⁄ ADR cell lysates. GLC4 ⁄ ADR cells were lysed by sonication, before (empty square) or after (full circles) treatment with MbCD. The lysate was separ- ated by density centrifugation, and collected from the top in 1 mL fractions. Fraction 1 is from the top of the gradient. (A) GM1 con- tent and (B) caveolin-1 expression were analyzed with IMAGE J v1.30 software.

These conflicting data can be explained by the use of different methods to isolate rafts and ⁄ or caveolae. Actually, most biochemical purifications of lipid rafts are based on an operational definition, namely that they are insoluble in Triton X-100 and have low buoy- ant density. A simplified definition of rafts is the 1% Triton X-100-insoluble material that floats at the inter- face of a 5% ⁄ 30% sucrose step gradient [36]. Thus, Triton X-100-resistant lipid rafts are distinguished from bulk plasma membrane because they are enriched in cholesterol and sphingolipids, but are relatively to depleted in glycerolphospholipids. Subsequent the identification of low-density detergent-resistant domains in Triton X-100 cell extracts, a variety of other detergents, including Lubrol WX, Lubrol PX, Brij 58, Brij 96, Brij 98, Nonidet P40, Chaps, and octylglucoside, have been employed at different con- centrations to prepare detergent-resistant membrane domains [37–41]. Unsurprisingly, use of these different detergents in the preparation of rafts yielded mem- brane domains with different lipid compositions from those of standard, Triton X-100-resistant membranes [42]. It is not clear whether this heterogeneity pre- existed in rafts or was induced by the application of the detergent.

fraction of

However, nondetergent methods that do not involve the solubilization properties of membranes can be used to isolate rafts. These methods largely obviate prob- such as membrane mixing and the selective lems extraction of lipids. In addition, these preparations seem to retain a greater inner-leaflet- membrane lipids [43] than detergent-extracted rafts do, and may therefore yield domains in which the coupling between raft leaflets is maintained [44,45]. For these reasons, rafts prepared by nondetergent methods seem more likely to reproduce the in vivo composition of these microdomains accurately.

detergent-resistant-membrane components. They are typical flask-shaped invaginations of the plasma mem- brane that originate from the presence of caveolin fam- ily proteins. Among the three isoforms, caveolin-1, caveolin-2 and caveolin-3, caveolin-1 is the predominant isoform responsible for this structure and is used as a biochemical marker of caveolae [25,26]. It now seems clear that caveolae are stable membrane domains that are kept in place by the actin cytoskeleton. Rather than having a specific function, caveolae might be considered to be multifunctional organelles with a physiologic role that varies according to cell type and cellular needs [25,26]. Importantly, the various possible functions of caveolae do not seem to be of vital importance for the

Several studies have been performed to determine localized with raft or caveolae;

whether P-gp is

FEBS Journal 274 (2007) 1470–1480 ª 2007 The Authors Journal compilation ª 2007 FEBS

1474

C. Marbeuf-Gueye et al.

ABCC1 localization in light membrane microdomains

LC phase [47,48]. For these reasons, it is not surprising that MRP1 funtionality is not affected by cholesterol depletion down to 40%, as the microdomains are probably still present.

however, very few studies have been done with MRP1 [12–14]. In this work, after cholesterol extraction from the membrane, we measured both the functionality of MRP1 and its localization in the plasma membrane. We report that membrane cholesterol is a central ele- ment in the control of both MRP1 functionality and localization in the GLC4 ⁄ ADR cell line.

In this study we used a nondetergent method to iso- late rafts. In this isolation procedure, the light and heavy fractions were found to be derived from the plasma membrane, whereas the extra-heavy third frac- tion originated mainly from intracellular components. We examined the distribution of signaling molecules as well as plasma membrane markers in each fraction. Our data show that MRP1 is exclusively localized in light membrane fraction 4 (Fig. 5). Hinrichs et al. also found, with the detergent-based method, that MRP1 was localized in microdomains in human colon carci- noma cells (HT29col). Nevertheless, MRP1 localization in microdomains was partial [12].

At this stage, it is interesting to compare the present data with those that we have previously obtained with K562 ⁄ ADR cells overexpressing P-gp. The transport functionality vs. the cholesterol content, obtained after different MbCD treatments, shows different profiles for P-gp and MRP1 (Fig. 3 and [49]). In K562 ⁄ ADR cells, which do not express caveolin-1 protein and therefore do not possess caveolae, the progressive removal of cholesterol with MbCD yields a progressive inhibition of P-gp functionality, whereas MRP1 func- tionality does not depend on the cholesterol content being reduced to 40%. Given that in K562 ⁄ ADR cells, P-gp is not localized in the microdomain, whereas we found that MRP1 is exclusively localized in the micro- domain, these different profiles could be explained by the fact that cholesterol could be more easily removed from nonmicrodomain regions than from the LO phase in the microdomain, where it is more tightly packed.

Caveolin-1, the marker of caveolin, and GM1, the marker of rafts, are also found in microdomain frac- tion 4. It should be emphazised that it is difficult to distinguish between caveolar and noncaveolar rafts, given the cofractionation properties that they have in common. Caveolin-1 is present not only in light frac- tion 4, but in all the membrane cell fractions (Fig. 6). This is not surprising, as it is now clear that multiple locations for caveolin-1 exist throughout the cell, and in the caveolin-1 has been reported to be present plasma membrane and in a number of other cellular sites, including mitochondria, the endoplasmic reticu- lum lumen, and secretory vesicles [26]. GM1 distribu- tion was similar to that of caveolin-1 in ‘light’ and ‘heavy’ membranes.

Now let us compare the impact of the membrane fluidity on P-gp and MRP1 transport activity. We have checked (a) that the passive influx of PIRA was the same in both sensitive and energy-depleted cells, mean- ing that there is no variation of passive influx during the acquisition of MDR in the two cell lines [20], and (b) that the initial rate of passive PIRA uptake was not modified by the addition of fluidizing agents such as Triton derivatives at the low concentrations used in this study [50]. The effect of Triton derivatives on P-gp and MRP1 functionality was measured in K562 ⁄ ADR and GLC4 ⁄ ADR cells, respectively, It appears that 6 ± 2 lm Triton X-45 yields 50% of P-gp function- ality inhibition [50], whereas a concentration of 16 ± 2 lm was required to inhibit 50% of MRP1 functionality, corroborating the observation that as MRP1 is localized in a more tightly packed fraction than P-gp, more Triton is required to modify the fluid- ity of the membrane around the transporter. Second, the whole membrane fluidity of these two cell lines can be compared using the rate of passive influx of various molecules through the plasma membrane. We have previously measured the rate of passive influx of anthracycline derivatives in these two cell lines [20], and we have observed that, systematically, for a given molecule the rate of its influx in GLC4 ⁄ ADR cells was 3–4 times higher than in K562 ⁄ ADR cells, showing that, as a whole, the GLC4 ⁄ ADR membrane is more fluid than the K562 ⁄ ADR membrane.

In summary, our

show that MRP1 is results localized in microdomains and that its functionality,

In a first set of experiments, in order to determine whether cholesterol affects MRP1 function, we used MbCD to extract cholesterol from the lipid phase of intact living cells. MbCD is a highly hydrophilic cyclic oligosaccharide that specifically binds sterol, rather than other membrane lipids, to form water-soluble complexes [45], without causing further membrane per- turbation by insertion [46]. Depletion of the membrane cholesterol level down to 40% by treatment with the cholesterol-chelating agent MbCD did not modify MRP1 activity, as followed either by the rate of efflux of PIRA or that of GSH. However, further cholesterol depletion below 40% yields both a partial shift of MRP1 to the high-density fraction and a decrease of its functionality (Figs 3 and 5). Indeed, membrane lipid models made of binary or ternary cholesterol mixtures can exhibit a bell-shaped phase diagram where the cholesterol-rich L0 phase coexists with the

FEBS Journal 274 (2007) 1470–1480 ª 2007 The Authors Journal compilation ª 2007 FEBS

1475

C. Marbeuf-Gueye et al.

ABCC1 localization in light membrane microdomains

Determination of the nonesterified cholesterol content of GLC4 cells

is con- measured as its ability to pump out PIRA, served when it is localized in rafts only. In addition, cholesterol perturbation is more difficult to achieve in rafts than in the other membrane regions. Altogether, these data suggest that MRP1-related functions can be regulated by the microdomain membrane.

Experimental procedures

Cell lines and culture

50 lm Amplex Red, 0.5 UÆmL)1 HRP,

GLC4 cells and MRP1-overexpressing GLC4 ⁄ ADR cells [51] were cultured in RPMI 1640 (Sigma Chemical Co., St Louis, MO) medium supplemented with 10% fetal bovine serum (Gibco Cergy Pointoise, France) at 37 (cid:2)C in a humidified incubator with 5% CO2. The resistant GLC4 ⁄ ADR cells were cultured with 1.2 lm doxorubicin until 1–4 weeks before the experiments. Cell cultures used for experiments were split 1 : 2 1 day before use in order to ensure logarithmic growth. Cells (106 ⁄ mL; 2 mL per cu- vette) were energy-depleted by preincubation for 30 min in Hepes buffer with sodium azide but without glucose [52]. We have previously checked that no P-gp was expressed in the resistant cells and that anthracycline efflux is due only to MRP1 [52].

Drugs and chemicals

The nonesterified cholesterol assay was adapted from a spectrofluorometric method used in the kit assay from Molecular Probes [53]. Briefly, cells, 2.5 · 106 mL)1, were washed once with NaCl ⁄ Pi and suspended in 1 mL of reac- tion buffer. The reaction buffer at pH 7.4 contained 0.1 m NaCl ⁄ Pi, 0.05 m NaCl, and 0.1% Triton X-100. Samples were incubated at 37 (cid:2)C for 15 min, and this was followed by sonication on ice (three times for 30 s) and then one additional hour of incubation at 37 (cid:2)C under continuous stirring. Unless indicated otherwise, 160 lL of this sample was added to the reaction buffer (total volume 1.6 mL), and the fluorescence signal at 560 nm (kex ¼ 585 nm) was monitored continuously when the following reactants were and added: 0.1 UÆmL)1 cholesterol oxidase (COase). The concentration of cholesterol in the solution was proportional to the differ- ence of the fluorescence signal DF ¼ FCOase ) FHRP, where FHRP and FCOase are the fluorescence signal intensities before and after the addition of HRP and COase, respect- ively. Because the interaction of the cell suspension with HRP yielded a small change in the fluorescence signal (not shown), a control standard was systematically carried out in the presence of 160 lL of sonicated cell suspension to which 0–10 lm cholesterol solution was added. The curve DF ¼ FCOase ) FHRP against the exogenous cholesterol con- centration was linear within the 0–5 lm range. Cholesterol titration was not affected by the presence of MbCD at the concentrations used in this study.

Treatment of cells with MbCD

The methylated derivatives of b-cylodextrin are known to preferentially trap membrane cholesterol in comparison to other cyclodextrins, which also show affinity for phospho- lipids and proteins [45]. We therefore used a methylated b-cylodextrin (MbCD) with a substitution degree of 10.5– 14.7 to study the effect of cholesterol depletion on GLC4 ⁄ ADR and GLC4 cells. Cells were grown as des- cribed. The standard culture medium was replaced with He- pes buffer, to which 2–20 mm MbCD had been added. Unless stated otherwise, cells were then incubated for 10 min at 37 (cid:2)C. Because MbCD interacted with anthra- cycline and modified the fluorescence signal, cells were washed with Hepes buffer, and the transport activity was then measured as described below.

Cellular anthracycline accumulation

FEBS Journal 274 (2007) 1470–1480 ª 2007 The Authors Journal compilation ª 2007 FEBS

1476

The rationale and validation of our experimental set-up for measuring the kinetics of the transport of anthracyclines in tumor cells has been extensively described and discussed Doxorubicin and PIRA were kindly provided by Pharma- cia-Upjohn (St Quentin Yualines, France). Concentrations were determined by diluting working solutions to approxi- mately 10)5 m with e480 ¼ 11 500 m)1Æcm)1. Working solu- tions were prepared just before use. MbCD (mean degree of substitution: 10.5–14.7), horseradish peroxidase (HRP), GSH, dithiothreitol and compounds of the polyoxyethylene series were purchased from Sigma and were dissolved in water. Trade names of polyoxyethylene (where n is the number of ethylene oxide units) are as follows: Triton X-45 (n ¼ 5), Triton X-100 (n ¼ 9.6) and Triton X-165 (n ¼ 165). 10-Acetyl-3,7-dihydroxyphenoxazine (Amplex Red) and monochlorobimane (MCB) were supplied by Molecular Probes (Eugene, OR). Before the experiments, the cells were counted, centrifuged at 5000 g for 30 s and resuspended in solutions containing 20 mm Hepes plus Hepes buffer 132 mm NaCl, 3.5 mm KCl, 1 mm CaCl2 and 0.5 mm MgCl2 at pH 7.3, with or without 5 mm glucose. Other chemicals were of the highest grade available. Deionized double-distilled water was used throughout the experiments. MRPm5 anti-MRP1 mouse serum was purchased from Alexis Biochemical (San Diego, CA, USA), anti-caveolin-1 (N-20 sc 894) rabbit serum was supplied by Santa Cruz Biotechnology (Santa Cruz, CA, USA). Poly(vinylidene difluoride) membrane was purchased from Hybond-P, Amersham Pharmacia Biotech (Orsay, France).

C. Marbeuf-Gueye et al.

ABCC1 localization in light membrane microdomains

The MCB and GST concentrations were kept constant at 100 lm and 0.5 UÆmL)1, respectively. The fluorescence sig- nal recorded over a short time (50 s), which was used as a measure of the initial rate of MCB–GSH formation, is directly proportional to the concentration of GSH at least within the range 0–20 lm (this corresponded to the concen- trations expected when 106 cells ⁄ mL were lysed, the intra- cellular GSH concentrations being within the 0–20 mm range). We have checked that oxidized glutathione did not give rise to any modification of the fluorescence signal.

before [16–19]. It is based on the continuous spectrofluoro- metric monitoring (Perkin Elmer LS50B spectrofluorometer Perkin Elmer, Courtabouef, France) of the decrease in the fluorescence signal of anthracycline at 590 nm (kex ¼ 480 nm) after incubation with the cells in a 1 cm quartz cuvette (Fig. 2). The decrease in fluorescence occurring during incubation with cells is due to the quenching of fluorescence after intercalation of anthracycline between the base pairs of DNA. We have previously shown that this methodology allows the accurate measurement of the free cytosolic concentration of anthracyclines under steady-state conditions, their initial rates of uptake, and the kinetics of active efflux [16–20].

For the intracellular GSH determination, cells (2 · 106) suspended in 2 mL of buffer were disrupted by sonication on ice (3 · 10 s, power 2). The rate of MCB–GSH forma- tion was followed after addition of MCB 100 lm and GST 0.5 UÆmL)1, as described above.

Determination of the MRP1-mediated efflux of PIRA

For the determination of GSH released by the cells, they were resuspended in Hepes buffer (106 ⁄ mL) in the absence or in the presence of the appropriate concentration of MbCD. After specified time intervals, 2 mL aliquots con- taining 2 · 106 cells were centrifuged at 5000 g for 30 s and washed twice, and the GSH in the extracellular medium, and therefore released from the cells, and the GSH present in the pellet were determined. The extracellular concentra- tion of GSH was not affected by 250 lm acivicin, indicating negligible activity of c-glutamyltransferase in the membrane of GLC4 ⁄ ADR cells.

Isolation of ‘light’ and ‘heavy’ membrane fractions

Cells (1 · 106 ⁄ mL; 2 mL per cuvette) were preincubated for 30 min in Hepes buffer with sodium azide, but without glu- cose (energy-deprived cells). Depletion of ATP in these cells was 90%, as checked with the luciferin–luciferase test [54]. The cells remained viable throughout the experiment, as checked with Trypan blue and calcein vital stain (not shown). After addition of PIRA, the decrease in the signal was monit- ored until steady state was reached. As the pH of the medium was chosen to equal the intracellular pH, at steady state, the extracellular free drug concentration (Ce) was equal to the cy- tosolic free drug concentration (Ci). Glucose was then added, which led to the restoration of control ATP levels within 2 min and to an increase in the fluorescence signal due to the efflux of PIRA. ATP-dependent PIRA efflux was determined from the slope of the tangent of the curve F ¼ f(t), where F is the fluorescence intensity at the time of addition of glucose (Fig. 2). As under these conditions, at the moment of addi- tion of glucose Ci ¼ Ce, the passive influx and efflux were equal, the net initial efflux represents the MRP1-mediated active efflux only [52].

GSH measurements

FEBS Journal 274 (2007) 1470–1480 ª 2007 The Authors Journal compilation ª 2007 FEBS

1477

We used a detergent-free procedure for purification of membrane domains. ‘Light’ and ‘heavy’ fractions were iso- lated according to previous methods [10,45,57], with the following modifications. GLC4 ⁄ ADR cells, 10–15 · 106, were washed twice with NaCl ⁄ Pi and suspended in this buf- fer (2 · 108 cells ⁄ mL). They were then incubated for 2 h in the absence or presence of 2.5 mm MbCD at 37 (cid:2)C with stirring. Cells were then again washed twice with NaCl ⁄ Pi. The pellet was suspended in 500 lL of sodium carbonate buffer at pH 9–10, and sonicated in a cold bath three times 30 s (50 Hz, 117 V, and 80 W) (Vibra cell; Sonics & Mate- rials Inc., Danbury, CT). The lysate was mixed with 80% sucrose to yield 2 mL of 40% sucrose solution. This mix- ture was transferred to the bottom of the ultracentrifuga- tion tube (Beckman Instruments, Palo Alto, CA) and was overlaid with 3 mL of 35% sucrose, 3 mL of 22% sucrose, and 3 mL of 10% sucrose solution. The ultracentrifugation was performed at 160 000 g for 14 h at 4 (cid:2)C with an SW41 the swinging rotor. Light-scattering bands confined at 10–22% sucrose and 22–35% sucrose interfaces, respect- ively, were observed. Eleven 1 mL fractions were collected by suction with a syringe from the top to the bottom. The velocity of aspiration was kept low in order to avoid distur- bance of the sucrose layers. The first fraction was called F1, and the last fraction was called F11. The total protein In order to quantify free GSH, either inside the cells or released in the extracellular medium, an enzymatic essay was used [55] MCB, itself nonfluorescent, is conjugated to GSH by glutathione S-transferase (GST) to yield a fluores- cent adduct [56] We have used this property to develop a very rapid and sensitive fluorometric method for GSH measurement [55]. Briefly, a 10 mm stock solution of MCB was prepared in ethanol, and aliquots were stored at ) 80 (cid:2)C in the dark. The nonenzymatic reaction that occurred between GSH and MCB was very slow. However, when GST was added, the increase of the fluorescent signal characteristic of MCB–GSH derivative formation was very fast. The initial rate of MCB–GSH formation was deter- mined as the increase in the fluorescent signal between 100 and 150 s after the addition of GST to MCB plus GSH.

C. Marbeuf-Gueye et al.

ABCC1 localization in light membrane microdomains

the manuscript. This work was

concentration was measured in each fraction using the Bradford reagent (Bio-Rad, Mames La Coquette, France).

supported by of grants from the Centre National de la Recherche Scientifique and l’Universite Paris XIII.

Western blotting measurement of MRP1 and caveolin-1 expression

References

1 Leslie EM, Deeley RG & Cole SP (2005) Multidrug resistance proteins: role of P-glycoprotein, MRP1, MRP2, and BCRP (ABCG2) in tissue defense. Toxicol Appl Pharmacol 204, 216–237. 2 Cole SP & Deeley RG (1998) Multidrug resistance

mediated by the ATP-binding cassette transporter pro- tein MRP. Bioessays 20, 931–940. 3 Nielsen D & Skovsgaard T (1992) P-glycoprotein as

multidrug transporter: a critical review of current multi- drug resistant cell lines. Biochim Biophys Acta 1139, 169–183.

anti-caveolin-1 (N-20 sc 894) rabbit

4 Fojo AT, Shen DW, Mickley LA, Pastan I & Gottes- man MM (1987) Intrinsic drug resistance in human kidney cancer is associated with expression of a human multidrug-resistance gene. J Clin Oncol 5, 1922–1927. 5 Leier I, Jedlitschky G, Buchholz U, Cole SP, Deele RG & Keppler D (1994) The MRP gene encodes an ATP- dependent export pump for leukotriene C4 and structu- rally related conjugates. J Biol Chem 269, 27807–27810. 6 Pike LJ (2004) Lipid rafts: heterogeneity on the high seas. Biochem J 378, 281–292. Equal volumes (12 lL) of membrane fractions were mixed with concentrated SDS reducing buffer (final concentrations are 0.75% SDS, 45 mm Tris, pH 6.8, 75 mm dithiothreitol). The samples were then incubated for 1 h at 50 (cid:2)C for MRP1 detection, or for 5 min at 95 (cid:2)C for caveolin-1 detection. Protein samples were separated on 7.5% (MRP1) or 12% (caveolin) SDS ⁄ PAGE, and then transferred to poly(vinylidene difluoride) membrane for 2 h and 20 min for MRP1 in transfer buffer (25 mm Tris-base, 192 mm gly- cine, 0.1% SDS) and for 1 h and 45 min for caveolin-1 in transfer buffer (25 mm Tris-base, 192 mm glycine, 0.1% SDS, 10% methanol). The membrane was blocked with 5% nonfat dry milk in 0.1% Tween ⁄ NaCl ⁄ Pi overnight at 4 (cid:2)C and treated with 1 lgÆmL)1 MRPm5 anti-MRP1 mouse serum (Alexis Biochemical) overnight at 4 (cid:2)C, or with 1 lgÆmL)1 serum (Santa Cruz) for 2 h at room temperature. Detection by HRP-linked was performed according to the manufacturer’s protocol (ECL plus kit with mouse IgG, HRP-linked whole antibody; Amersham Pharmacia Biotech). MRP1 and cave- olin expression were evaluated after densitometric scanning of film and analysis with image j 1.30 software. 7 Pike LJ (2003) Lipid rafts: bringing order to chaos. J Lipid Res 44, 655–667. 8 Pizzo P & Viola A (2003) Lymphocyte lipid rafts: struc-

Localization of GM1 ganglioside

ture and function. Curr Opin Immunol 15, 255–260. 9 Radeva G, Perabo J & Sharom FJ (2005) P-Glycopro-

tein is localized in intermediate-density membrane microdomains distinct from classical lipid rafts and caveolar domains. FEBS J 272, 4924–4937.

10 MacDonald JL & Pike LJ (2005) A simplified method for the preparation of detergent-free lipid rafts. J Biol Chem 280, 26796–26804.

11 Orlowski S, Martin S & Escargueil A (2006) P-glycopro- tein and ‘lipid rafts’: some ambiguous mutual relation- ships (floating on them, building them or meeting them by chance?) Cell Mol Life Sci 63, 1038–1059.

12 Hinrichs JWJ, Klappe K, Hummel I & Kok JW (2004) ATP-binding cassette transporters are enriched in non- caveolar detergent-insoluble glycosphingolipid-enriched membrane domains (DIGs) in human multidrug-resis- tant cancer cells. J Biol Chem 279, 5734–5738. 13 Hinrichs JWJ, Klappe K, van Riezen M & Kok JW GM1 is commonly found in high concentrations in rafts [57], and can be labeled using fluorescein isothiocyanate- conjugated cholera toxin B (CTB). CTB [45] binds to the glycosphingolipid GM1, and we used it as a marker for raft localization in ‘light’ and ‘heavy’ plasma membrane frac- tions. We employed a CTB–HRP conjugate for simple, rapid detection of ganglioside GM1. Briefly, after activation of a PVDF membrane with methanol and washing with water and 0.1% Tween ⁄ NaCl ⁄ Pi, 4 lL of each gradient was dotted onto the wet membrane. The air-dried mem- brane was reactivated and blocked with 5% nonfat dry milk in 0.1% Tween ⁄ NaCl ⁄ Pi. After being washed with NaCl ⁄ Pi, the membrane was incubated with HRP-conju- gated CTB (dilution 1 : 5000, Sigma) in 0.1% Tween ⁄ NaCl ⁄ Pi for 90 min, rinsed several times with NaCl ⁄ Pi, and then detected by enhanced chemiluminescence (Amer- sham Pharmacia Biotech) and analyzed with image j 1.30 software.

Acknowledgements

(2005) Drug resistance-associated changes in sphingo- lipids and ABC transporters occur in different regions of membrane domains. J Lipid Res 46, 2367–2376. 14 Klappe K, Hinrichs JWJ, Kroesen B-J, Sietsma H &

Kok JW (2004) MRP1 and glucosylceramide are coordi- nately over expressed and enriched in rafts during

and thank Professor Laurence Le Moyec We Dr Catherine Herve´ du Penhoat for critical review

FEBS Journal 274 (2007) 1470–1480 ª 2007 The Authors Journal compilation ª 2007 FEBS

1478

C. Marbeuf-Gueye et al.

ABCC1 localization in light membrane microdomains

multidrug resistance acquisition in colon cancer cells. Int J Cancer 110, 511–522.

28 Pang A, Au WY & Kwong YL (2004) Caveolin-1 gene is coordinately regulated with the multidrug resistance 1 gene in normal and leukemic bone marrow. Leuk Res 28, 973–977.

15 Omidi Y, Campbell L, Barar J, Connell D, Akhtar S & Gumbleton M (2003) Evaluation of the immortalised mouse brain capillary endothelial cell line, b. End3, as an in vitro blood–brain barrier model for drug uptake and transport studies. Brain Res 990, 95–112.

29 Haber M, Burkhart CA, Regi DL, Madafiglio J, Norris MD & Horwitz SB (1995) Altered expression of M beta 2, the class II beta-tubulin isotype, in a murine J774.2 cell line with a high level of taxol resistance. J Biol Chem 270, 31269–31275.

16 Fre´ zard F & Garnier-Suillerot A (1991) Determination of the osmotic active drug concentration in the cyto- plasm of anthracycline-resistant and -sensitive K562 cells. Biochim Biophys Acta 1091, 29–35.

17 Fre´ zard F & Garnier-Suillerot A (1991) Comparison of the membrane transport of anthracycline derivatives in drug-resistant and drug-sensitive K562 cells. Eur J Bio- chem 196, 483–491.

30 Greenberger LM, Lothstein L, Williams SS & Horwitz SB (1988) Distinct P-glycoprotein precursors are over- produced in independently isolated drug-resistant cell lines. Proc Natl Acad Sci USA 85, 3762–3766. 31 Parolini I, Sargiacomo M, Galbiati F, Rizzo G, Grignani F, Engelman JA, Okamoto T, Ikezu T, Scherer PE, Mora R et al. (1999) Expression of caveo- lin-1 is required for the transport of caveolin-2 to the plasma membrane. Retention of caveolin-2 at the level of the golgi complex. J Biol Chem 274, 25718–25725. 32 Davidson B, Goldberg I, Givant-Horwitz V, Nesland JM, Berner A, Bryne M, Risberg B, Koplovic J, Kris- tensen GB, Trope CG et al. (2002) Caveolin-1 expres- sion in ovarian carcinoma is MDR1 independent. Am J Clin Pathol 117, 225–234. 33 Lavie Y, Cao H, Bursten SL, Giuliano AE & Cabot 18 Mankhetkorn S, Dubru F, Hesschenbrouck J, Fiallo M & Garnier-Suillerot A (1996) Relation among the resis- tance factor, kinetics of uptake, and kinetics of the P- glycoprotein-mediated efflux of doxorubicin, daunorubi- cin, 8-(S)-fluoroidarubicin, and idarubicin in multidrug- resistant K562 cells. Mol Pharmacol 49, 532–539. 19 Borrel MN, Pereira E, Fiallo M & Garnier-Suillerot A (1994) Mobile ionophores are a novel class of P-glyco- protein inhibitors. The effects of ionophores on 4¢-O- tetrahydropyranyl-adriamycin incorporation in K562 drug-resistant cells. Eur J Biochem 233, 125–133.

MC (1996) Accumulation of glucosylceramides in multi- drug-resistant cancer cells. J Biol Chem 271, 19530– 19536. 34 Kok JW, Veldman RJ, Klappe K, Koning H, Filipeanu

CM & Muller M (2000) Differential expression of sphingolipids in MRP1 overexpressing HT29 cells. Int J Cancer 87, 172–178. 20 Marbeuf-Gueye C, Ettori D, Priebe W, Kozlowski H & Garnier-Suillerot A (1999) Correlation between the kinetics of anthracycline uptake and the resistance fac- tor in cancer cells expressing the multidrug resistance protein or the P-glycoprotein. Bochim Biophys Acta 1450, 374–384. 35 Veldman RJ, Klappe K, Hinrichs J, Hummel L, van der 21 Fantini J, Garmy N, Mahfoud R & Yahi N (2002)

Lipid rafts: structure, function and role in HIV, Alzhei- mers and prion diseases. Expert Rev Mol Med 2002, 1–22.

Schaaf G, Sietama H & Kok JW (2002) Altered sphingolipid metabolism in multidrug-resistant ovarian cancer cells is due to uncoupling of glycolipid biosynth- esis in the Golgi apparatus. FASEB J 16, 582–592. 36 Hope HR & Pike LJ (1996) Phosphoinositides and 22 Brown DA & London E (2000) Structure and function of sphingolipid- and cholesterol-rich membrane rafts. J Biol Chem 275, 17221–17224. 23 Edidin M (2003) The state of lipid rafts: from model phosphoinositide-utilizing enzymes in detergent-insolu- ble lipid domains. Mol Biol Cell 7, 843–851.

membranes to cells. Annu Rev Biophys Biomol Struct 32, 257–283.

37 Madore N, Smith KL, Graham CH, Jen A, Brady K, Hall S & Morris R (1999) Functionally different GPI proteins are organized in different domains on the neuronal surface. EMBO J 18, 6917–6926. 24 Shogomori H & Brown DA (2003) Use of detergents to study membrane rafts: the good, the bad, and the ugly. Biol Chem 384, 1259–1263. 25 Van Deurs B, Roepstorff K, Hommelgaard AM &

38 Drevot P, Langlet C, Guo X-J, Bernard A-M, Colard O, Chauvin J-P, Lasserre R & He H-T (2002) TCR signal initiation machinery is pre-assembled and acti- vated in a subset of membrane rafts. EMBO J 21, 1899–1908. Sandvig A (2003) Caveolae: anchored, multifunctional platforms in the lipid ocean. Trends Cell Biol 13, 92–100.

26 Quest AFG, Leyton L & Parraga M (2004) Caveolins, caveolae, and lipid rafts in cellular transport, signaling, and disease. Biochem Cell Biol 82, 129–144.

FEBS Journal 274 (2007) 1470–1480 ª 2007 The Authors Journal compilation ª 2007 FEBS

1479

39 Ilangumaran S, Arni S, van Echten-Deckert G, Borisch B & Hoessli DC (1999) Microdomain-dependent regula- tion of Lck and Fyn protein-tyrosine kinases in T lym- phocyte plasma membranes. Mol Biol Cell 10, 891–905. 40 Roper K, Corbeil D & Huttner WB (2000) Retention of prominin in microvilli reveals distinct cholesterol-based 27 Lavie Y, Fiucci G & Liscovitch M (1998) Up-regulation of caveolae and caveolar constituents in multidrug-resis- tant cancer cells. J Biol Chem 273, 32380–32383.

C. Marbeuf-Gueye et al.

ABCC1 localization in light membrane microdomains

lipid micro-domains in the apical plasma membrane. Nat Cell Biol 2, 582–592. 41 Gumbleton M, Hollins AJ, Omidi Y, Campbell L & MbetaCD and filipin and lack of effect induced by cholesterol oxidase indicate that this transporter is not located in rafts. J Bioenerg Biomembr 36, 533–543. 50 Borrel MN, Fiallo M, Veress I & Garnier-Suillerot A Taylor G (2003) Targeting caveolae for vesicular drug transport. J Control Release 87, 139–151. 42 Schuck S, Honsho M, Ekroos K, Shevchenko A &

(1995) The effect of crown ethers, tetraalkylammonium salts, and polyoxyethylene amphiphiles on pirarubicin incorporation in K562 resistant cells. Biochem Pharma- col 50, 2069–2076. Simons K (2003) Resistance of cell membranes to differ- ent detergents. Proc Natl Acad Sci USA 100, 5795– 5800.

51 Zijlstra JG, De Vries EGE & Mulder NH (1987) Multi- factorial drug resistance in an adriamycin-resistant human small cell lung carcinoma cell line. Cancer Res 47, 1780–1784.

43 Pike LJ, Han X, Chung K-N & Gross R (2002) Lipid rafts are enriched in arachidonic acid and plasmeny- lethanolamine and their composition is independent of caveolin-1 expression: a quantitative electrospray ioniza- tion ⁄ mass spectrometric analysis. Biochemistry 41, 2075–2088.

52 Marbeuf-Gueye C, Broxterman H, Dubru F, Priebe W & Garnier-Suillerot A (1998) Kinetics of anthracycline efflux from multidrug resistance protein-expressing cancer cells compared with P-glycoprotein-expressing cancer cells. Mol Pharmacol 53, 141–147. 44 Janes PW, Ley SC & Magee AI (1999) Aggregation of lipid rafts accompanies signaling via the T cell antigen receptor. J Cell Biol 147, 447–461. 45 Ohvo H & Slotte P (1996) Cyclodextrin-mediated 53 Amundson DM & Zhou M (1999) Fluorometric method for the enzymatic determination of cholesterol. J Bio- chem Biophys 38, 43–52.

removal of sterols from monolayers: effects of sterol structure and phospholipids on desorption rate. Bio- chemistry 35, 8018–8024.

46 Gimpl G, Burger K & Fahrenholz F (1997) Cholesterol as modulator of receptor function. Biochemistry 36, 10959–10974.

54 Kimmich G, Randles J & Brand J (1975) Assay of pico- mole amounts of ATP, ADP, and AMP using the luci- ferase enzyme system. Anal Biochem 69, 187–192. 55 Salerno M & Garnier-Suillerot A (2001) Kinetics of glutathione and daunorubicin efflux from multidrug resistance protein overexpressing small-cell lung cancer cells. Eur J Pharmacol 421, 1–9. 47 Mouritsen OG & Zuckermann MJ (2004) What’s so special about cholesterol? Lipids 39, 1101–1113. 48 Veatch SL & Keller SL (2003) Separation of liquid

phases in giant vesicles of ternary mixtures of phospho- lipids and cholesterol. Biophys J 85, 3074–3093. 56 Fernandez-Checa JC & Kaplowitz N (1990) The use of monochlorobimane to determine hepatic GSH levels and synthesis. Anal Biochem 190, 212–219. 57 Monneron A & d’Alayer J (1978) Isolation of plasma

FEBS Journal 274 (2007) 1470–1480 ª 2007 The Authors Journal compilation ª 2007 FEBS

1480

and nuclear membranes of thymocytes. II. Biochemical composition. J Cell Biol 77, 232–245. 49 Reungpatthanaphong P, Marbeuf-Gueye C, Le Moyec L, Salerno M & Garnier-Suillerot A (2004) Decrease of P-glycoprotein activity in K562 ⁄ ADR cells by