Eur. J. Biochem.

268, 1687±1697 (2001) q FEBS 2001

Drug transport by reconstituted P-glycoprotein in proteoliposomes

Effect of substrates and modulators, and dependence on bilayer phase state

Peihua Lu, Ronghua Liu and Frances J. Sharom

Guelph-Waterloo Centre for Graduate Work in Chemistry and Biochemistry, Department of Chemistry and Biochemistry,
University of Guelph, Ontario, Canada

The P-glycoprotein multidrug transporter (Pgp) is an active hydrolysis, and the Km for TMR was 0.3 mm. TMR
efflux pump for chemotherapeutic drugs, natural products transport was inhibited in a concentration-dependent
and hydrophobic peptides. Pgp is envisaged as a fashion by verapamil and cyclosporin A, and activated
`hydrophobic vacuum cleaner', and drugs are believed to (probably by a positive allosteric effect) by the transport
gain access to the substrate binding sites from within the substrate colchicine. TMR transport by Pgp reconstituted
membrane, rather than from the aqueous phase. The into proteoliposomes composed of two synthetic phospha-
intimate association of both Pgp and its substrates with tidylcholines showed a highly unusual biphasic temperature
the membrane suggests that its function may be regulated dependence. The rate of TMR transport was relatively high
by the biophysical properties of the lipid bilayer. Using the in the rigid gel phase, reached a maximum at the melting
high affinity fluorescent substrate tetramethylrosamine temperature of the bilayer, and then decreased in the fluid
(TMR), we have monitored, in real time, transport in liquid crystalline phase. This pattern of temperature
proteoliposomes containing reconstituted Pgp. The TMR dependence suggests that the rate of drug transport by
concentration gradient generated by Pgp was collapsed by Pgp may be dominated by partitioning of drug into the
the addition of either the ATPase inhibitor, vanadate, or Pgp bilayer.
modulators. TMR transport by Pgp obeyed Michaelis±
Menten kinetics with respect to both of its substrates. The Keywords: fluorescence; multidrug resistance; phospholipid
Km for ATP was 0.48 mm, close to the Km for ATP bilayer; phase transition; tetramethylrosamine.

Overexpression of the P-glycoprotein multidrug transporter Pgp function. Two commonly used modulators are the
(Pgp) in the plasma membrane is believed to be a major Ca21-channel blocker verapamil and the immuno-
cause of resistance to multiple chemotherapeutic drugs suppressive agent cyclosporin A. The latter compound has
(multidrug resistance; MDR) in human cancers. The proved highly effective in clinical trials in overcoming
biochemistry and biology of Pgp have been the subject of MDR in paediatric cancers [12].
several recent reviews [1±4]. The protein, which is a Despite the advances of recent years, we still have only a
member of the ATP-binding cassette superfamily of mem- very nebulous view of the molecular mechanism by which
brane proteins [5], operates as an ATP-dependent efflux Pgp transports such a wide diversity of compounds across
pump, exporting a wide variety of hydrophobic drugs, the membrane, and how this process is powered by ATP
natural products and peptides from the cell. Purified Pgp hydrolysis. Some insights have come from in vitro studies
exhibits a high level of constitutive ATPase activity [6±8] on the purified protein. Vanadate is a highly effective
which is modulated further by drug and chemosensitizers, inhibitor of Pgp ATPase activity [6±8]. Senior and
and the reconstituted protein has been shown to transport coworkers observed that the addition of vanadate to Pgp
several drugs and a hydrophobic peptide in proteoliposomes in the presence of ATP trapped the complex ADP´Vi´Mg21
[9±11]. Many compounds, known as modulators, chemo- (a transition state analog) in the active site of one of the
sensitizers or reversers, can overcome MDR by blocking nucleotide-binding (NB) domains of Pgp, following a
single catalytic turnover [13]. Vanadate trapping at one
active site results in complete inactivation of the ATPase
activity of both NB domains, which led to the proposal that
Correspondence to F. J. Sharom, Department of Chemistry and Pgp operates by an alternating sites type of mechanism
Biochemistry, University of Guelph, Guelph, Ontario N1G 2W1, [14].
Canada. Fax: 11 519 766 1499, Tel.: 11 519 824 4120 extn 2247, The process of drug transport by Pgp differs from that of
E-mail: sharom@chembio.uoguelph.ca many other ATP-driven transporters in that its substrates,
Abbreviations: MDR, multidrug-resistance/resistant; MIANS, the majority of which are hydrophobic, have a relatively
2-(4-maleimidoanilino)-naphthalene-6-sulfonic acid; Myr2GroPCho, high rate of passive diffusion across the membrane
dimyristoyl-phosphatidylcholine; NB, nucleotide binding; compared with ions and small polar molecules. As a
PamMyrGroPCho, 1-palmitoyl-2-myristoyl-phosphatidylcholine; consequence, the observed (net) rate of drug transport by
Pgp, P-glycoprotein; PtdCho, phosphatidylcholine; TMR, Pgp represents a balance between two opposing processes;
tetramethylrosamine. active pumping by Pgp up a concentration gradient, and
(Received 15 November 2000, revised 11 January 2001, accepted passive diffusion of drug across the membrane down the
18 January 2001) concentration gradient. Pgp substrates probably gain access
1688 P. Lu et al. (Eur. J. Biochem. 268) q FEBS 2001

to the protein after partitioning into the membrane rather transport of TMR was characterized with respect to
than from the aqueous phase. The hydrophobic vacuum kinetics, inhibition by MDR modulators, and allosteric
cleaner model [15] envisions the transporter expelling drugs activation by a substrate. The rate of TMR transport in
directly from the bilayer. This proposal is supported by liposomes of two synthetic phosphatidylcholines showed an
recent results from our laboratory showing that the apparent unusual biphasic response as the temperature was
affinity of a drug for binding to purified Pgp is highly increased, reaching a maximum at the phase transition, or
correlated with its lipid-water partition coefficient [16]. melting temperature of the bilayer. This work represents the
Evidence is also accumulating that Pgp may act as a plasma first report, using proteoliposomes containing reconstituted
membrane flippase [15], moving drug molecules from the Pgp, of real-time measurements of the rate of substrate
cytoplasmic leaflet to the extracellular leaflet [17,18]. It has transport and its modulation by membrane phase state.
been demonstrated already that the bacterial multidrug
transporter LmrP expels drugs from the inner leaflet of the
M AT E R I A L S A N D M E T H O D S
membrane [19].
Reconstitution of membrane transport proteins into
Materials
defined lipid bilayer systems, and measurement of the
kinetic characteristics of substrate transport, is a powerful Dimyristoylphosphatidylcholine (Myr2GroPCho) and 1-
approach to enhancing our understanding of their function palmitoyl-2-myristoyl-phosphatidylcholine (PamMyrGroP-
and mechanism [20,21]. In previous studies of drug Cho) were obtained from Avanti Polar Lipids. ATP, AMP,
transport by reconstituted Pgp in proteoliposomes, we sodium ortho-vanadate, rhodamine 123, verapamil,
employed radiolabelled substrates in a rapid filtration colchicine and Chaps were from Sigma. Cyclosporin A
technique [9,11]. As drug uptake was very fast (it became was from Pfizer Central Research (Groton, CT, USA). TMR
nonlinear in , 1 min, and reached steady state in 3±4 min), was from Molecular Probes. Creatine kinase and creatine
it was possible to determine uptake only at steady state, phosphate were from Roche Diagnostics.
rather than the rate of transport. Methods using continuous
monitoring of fluorescent dyes can allow measurement of
Plasma membrane preparation and Pgp purification
transport rates in real time. In native plasma membrane
vesicles from MDR cells, rhodamine 123 [22], Hoechst Plasma membrane vesicles from the colchicine-selected
33342 [17], and LDS-751 [18] were used to measure rates MDR Chinese hamster ovary cell line CHRB30 were
of transport. However, rate measurements of this type have isolated as described previously [24,25]. Plasma membrane
not yet been achieved for Pgp in a reconstituted system. In vesicles were stored at 270 8C for no longer than 3 months
the only report to date of fluorescent substrate transport by before use. Pgp was isolated from CHRB30 plasma
reconstituted Pgp, Shapiro and Ling used the dye Hoechst membrane and reconstituted into proteoliposomes of
33342 to monitor transport by Pgp in proteoliposomes [10]. Myr2GroPCho or PamMyrGroPCho by gel filtration
This dye has negligible fluorescence in aqueous solution, chromatography on a Sephadex G-50 column [23]. Pgp
and displays enhanced fluorescence on partitioning into the made up . 85% of the reconstituted protein in the
lipid bilayer, so that the monitored decrease in fluorescence proteoliposomes as indicated by SDS/PAGE. In a typical
intensity corresponded to Pgp-mediated pumping of the dye preparation, 0.4±0.5 mg Pgp was reconstituted into 5 mg
out of the bilayer. Hoechst 33342 efflux was extremely lipid. The final lipid/protein ratio of the proteoliposomes
slow, and as the vesicles were not sealed, it was not possible was in the range 7±10 : 1. Dynamic light scattering
to establish a substrate concentration gradient. In addition, measurements to assess vesicle size were carried out as
because the dye concentration in the lipid phase was described previously [23]. The protein content of plasma
unknown, analysis of the kinetic parameters of substrate membrane was determined by the method of Bradford [26],
transport could not be carried out. and the protein content of Pgp preparations and proteolipo-
The intimate association of both Pgp and its drug somes was measured by a modified Lowry method [27].
substrates with the membrane suggests that the biochemical
and biophysical nature of the lipid bilayer will probably
Real-time fluorescence measurement of drug transport in
have strong modulatory effects on its function. A previous
proteoliposomes
study using Pgp reconstituted into bilayers of synthetic
phospholipids showed that both ATP binding and ATPase Fluorescence measurements were carried out using either a
activity were modulated by bilayer phase state [23]. More Spex Fluorolog-2 or a PTI QuantaMaster C-61 high
recently, we found that drug binding affinity was also sensitivity steady-state fluorimeter, equipped with a
highly dependent on the nature of the lipid environment temperature-controlled sample compartment. A 450-mL
[16]. In the present work, one of our goals was to explore aliquot of proteoliposomes containing 10±20 mg Pgp was
the effect of lipid environment and membrane phase state preincubated with the appropriate concentration of TMR in
on the rate of ATP-driven drug transport by Pgp in a transport buffer. After a 5-min incubation at the desired
reconstituted system. temperature, the sample was transferred to a thermostatted
In the present study, we have used a continuous quartz cuvette, and allowed to equilibrate for about 300 s to
fluorescence assay with tetramethylrosamine (TMR, a allow stabilization of the fluorescence intensity. Transport
rhodamine dye) to quantify drug transport by purified Pgp of TMR was initiated by addition of a 50-mL aliquot of
in proteoliposomes of defined phospholipids. Both the buffer containing ATP (final concentration 1 mm) and an
formation of a substrate concentration gradient and its ATP regenerating system (30 mg´mL21 creatine kinase,
collapse by inhibitors were observed, and the rate of 3.5 mm creatine phosphate) followed by mixing for 5±10 s.
drug transport was measured in real time. Pgp-mediated Excitation at 550 nm was carried out with a slit width of
q FEBS 2001 Drug transport by reconstituted P-glycoprotein (Eur. J. Biochem. 268) 1689

of proteoliposomes and TMR in the presence of various

concentrations of sucrose. For control experiments using
denatured Pgp, the protein was heated at 100 8C for 10 min
and then cooled on ice. In experiments to determine the
temperature dependence of TMR transport, the same
reconstituted preparation was used for all of the kinetic
runs spanning the temperature range under study.

Measurement of substrate binding affinity by

fluorescence quenching
The affinity of binding of rhodamine 123 and TMR to
highly purified Pgp labelled with the fluorophore MIANS
was determined using quenching titrations, as described
previously [16,28,29]. Kd values were determined by fitting
of the fluorescence quenching data to an equation
describing binding to a single site.

R E S U LT S

Drug gradient formation and collapse in

proteoliposomes: a continuous fluorescence assay using
tetramethylrosamine
We reported previously the purification of Pgp from the
MDR Chinese hamster ovary cell line CHRB30, and its
reconstitution into proteoliposomes of the defined
phospholipids Myr2GroPCho and PamMyrGroPCho using
a gel filtration technique [23,30]. Pgp retained both its
Fig. 1. Binding of (A) rhodamine 123 and (B) TMR to purified ATPase activity [23,30] and transport function [11]
MIANS-labelled Pgp as assessed by fluorescence quenching. following reconstitution. In the present study, Pgp was
Increasing concentrations of dye were added to 50 mg´mL21 reconstituted into proteoliposomes at a lipid/protein ratio of
MIANS-labelled Pgp at 22 8C in the presence of 0.5 mg´mL21 7±10 : 1 (w/w). Dynamic light scattering showed that the
asolectin, and the fluorescence emission at 420 nm was recorded. proteoliposomes comprised a bimodal population of
The percent quenching of fluorescence (DF/Fo  100) was calculated vesicles of mean diameters 0.17 and 0.75 mm. We reported
relative to MIANS-labelled Pgp in the absence of dye. The continuous previously bimodal proteoliposome populations for Pgp
line represents the best computer-generated fit of the data points reconstituted into Myr2GroPCho [30] and a mixture of egg
(shown by the symbols) to an equation describing interaction with a phosphatidylcholine (PtdCho) and dipalmitoyl-PtdCho [9].
single type of binding site. Data represent the mean ^ range for Several fluorescent compounds are known to be transport
duplicate determinations. substrates for Pgp in intact MDR cells, including the
acetoxymethyl esters of Ca21-and pH-sensitive dyes, such
as calcein, fura-2, and BCECF [31,32], fluorescent
1.8 nm to reduce photobleaching, and fluorescence derivatives of verapamil [33], and various rhodamine dyes
emission was monitored continuously at 575 nm (slit [34]. Eytan and coworkers examined seven rhodamine dyes
width 3.6 nm). Fluorescence intensity was measured at for their interaction with Pgp, and concluded that TMR was
1 s intervals for < 150 s, until a steady state was a much superior substrate to rhodamine 123, which has
approached. The traces of fluorescence intensity vs. time been widely used in dye exclusion studies in intact MDR
were normalized to the intensity measured immediately cells. Based on this information, we chose TMR as a
before addition of a particular reagent, which was taken as suitable substrate for measuring Pgp-mediated transport in
100%. Linear regression analysis of the data collected vitro. We have shown previously that quenching of purified
during the first 20 s was carried out, and rate data were Pgp labelled with the fluorophore MIANS can be used to
subsequently fitted to the Michaelis±Menten equation, quantitate the affinity of binding of substrates and
using sigmaplot (SPSS Inc.). modulators [28,29,35,36]. Similar experiments carried out
In other experiments, one or more 10±50 mL aliquots of using a series of rhodamine dyes showed that TMR
AMP, Chaps, TMR, vanadate, the Pgp modulators vera- interacted with the transporter with substantially higher
pamil and cyclosporin A, or the Pgp substrate colchicine, affinity than the others, including rhodamine 123 (Fig. 1).
were added at various time points during the transport The Kd values for binding of rhodamine 123 and TMR to
process, and fluorescence intensity data were collected for a Pgp, obtained by fitting of the quenching curves to an
further 100±150 s. Drug substrates were prepared as stock equation for a single binding site, were 12.8 mm and
solutions in dimethylsulfoxide, which never exceeded 0.1% 0.70 mm, respectively.
(v/v). Osmotic sensitivity transport experiments were Addition of 1 mm ATP and an ATP regenerating system
initiated by addition of ATP after a 10-min preincubation to reconstituted Myr2GroPCho proteoliposomes containing
1690 P. Lu et al. (Eur. J. Biochem. 268) q FEBS 2001

mixing) by intensity measurements made every 1 s for

about 20 s (Fig. 2D). Only background changes in
fluorescence intensity were observed when ATP was
replaced with AMP, and denaturation of Pgp also abolished
the change in TMR fluorescence (Fig. 2F). Thus the
process of TMR transport into the proteoliposomes is
dependent on ATP and requires functionally active Pgp.
Addition of 100 mm vanadate, which inhibits Pgp
ATPase activity [39], would be predicted to block Pgp-
mediated transport of TMR into the proteoliposome
interior. Addition of vanadate (Fig. 2A) to the proteolipo-
somes following establishment of the new steady state
resulted in rapid restoration of the TMR fluorescence,
indicating that the TMR gradient had collapsed. Addition of
the MDR modulators verapamil (Fig. 2B) and cyclosporin A
(Fig. 2C) also led to a similar increase in fluorescence. All
three of these agents inhibit drug transport by Pgp, and will
thus lead to efflux by passive diffusion of TMR accumu-
lated in the proteoliposome lumen. It was possible to
estimate the rate of fluorescence increase after addition of
inhibitor or modulator, which represents the rate of passive
diffusion of TMR out of the proteoliposome lumen, driven
by the concentration gradient previously built up by Pgp
(Fig. 2E).

Fig. 2. Real-time fluorescence measurements of TMR transport TMR accumulates in the proteoliposome lumen during
into reconstituted Myr2GroPCho proteoliposomes containing Pgp. transport
A 450-mL aliquot of proteoliposomes containing 10±20 mg of Pgp
was equilibrated with 1 mm TMR in transport buffer at 27 8C. TMR
TMR accumulates in the vesicle interior during the
transport was initiated (indicated by the arrow) by addition of 50 mL
transport process, as indicated by the results of the
of buffer containing ATP (final concentration 1 mm) and an ATP following experiments. As shown in Fig. 3A, addition of
regenerating system (30 mg´mL21 creatine kinase, 3.5 mm creatine 4 mm Chaps after TMR transport and establishment of a
phosphate), followed by mixing. Excitation was carried out at 550 nm, steady state, resulted in restoration of the TMR
and the intensity of fluorescence emission at 575 nm was measured at fluorescence, indicating the release of accumulated TMR
1-s intervals for < 150 s, until a steady state was approached. At this from the interior of the proteoliposomes. Chaps addition
time, an aliquot of buffer was added (indicated by the arrow) did not result in full restoration of TMR fluorescence,
containing either vanadate (A; final concentration 100 mm), verapamil compared with that achieved by addition of vanadate or
(B; final concentration 20 mm), or cyclosporin A (C; final concentra- modulators. Further investigation using a range of detergent
tion 1.6 mm) and after mixing, further fluorescence intensity data were concentrations revealed that release of TMR was
collected. Fluorescence data collected immediately after addition of concentration dependent, and that Chaps concentrations
ATP or vanadate are shown in (D) and (E), respectively. (F) up to 8 mm were insufficient to fully permeabilize the
Fluorescence traces for: addition to active proteoliposomes at time proteoliposomes. Detergent concentrations higher than this
zero of 1 mm ATP or 1 mm AMP; addition of 100 mm vanadate prior could not be used, as micelle formation interfered with
to 1 mm ATP; and addition of 1 mm ATP to heat-inactivated fluorescence measurements. Additional experiments using
proteoliposomes. Each addition of reagent to the cuvette resulted in SDS in place of Chaps resulted in complete restoration of
a 10% dilution of the sample, resulting in a 10% decrease in the TMR fluorescence, suggesting that the lower value
overall fluorescence. observed for 4 mm Chaps was due to incomplete
permeabilization of the vesicles. Addition of Chaps before
ATP resulted in no transport-associated fluorescence
Pgp in the presence of 1 mm TMR led to a rapid drop in decrease, consistent with the inability to build up a TMR
TMR fluorescence, until a new lower steady-state level was gradient in permeabilized proteoliposomes (Fig. 3B). Pre-
reached (Fig. 2A±C). Previous work suggested that for vious work in our laboratory indicated that Chaps is not a
colchicine, such a steady state represents a balance between substrate for Pgp; MDR cells are not cross-resistant to it
active inward pumping of drug by Pgp and outward passive [40], and it does not stimulate Pgp ATPase activity [41].
diffusion through the membrane [25]. The decrease in Pgp proteoliposomes were capable of carrying out more
fluorescence intensity arises from transport of TMR into the than one round of TMR transport, and this was also both
vesicle (see below), where its fluorescence is decreased, disrupted and prevented by detergent permeabilization.
probably as a result of self quenching due to its Following the initial round of TMR transport at a
accumulation in the lumen. This phenomenon was noted concentration of 0.5 mm, a second portion of TMR was
previously for rhodamine 123, and has been applied to the added, corresponding to a final concentration of 3 mm. This
measurement of dye transport in native plasma membrane resulted in an immediate second round of TMR transport,
vesicle systems in yeast and mammalian cells [37,38]. The as indicated by a rapid fluorescence decrease, and
rate of the fluorescence decrease could be estimated (after establishment of a new steady state (Fig. 3C). Addition of
q FEBS 2001 Drug transport by reconstituted P-glycoprotein (Eur. J. Biochem. 268) 1691

Fig. 3. Effect of detergent permeabilization on TMR transport into

reconstituted Myr2GroPCho proteoliposomes containing Pgp. (A)
TMR transport (1 mm TMR, 1 mm ATP) was initiated by addition of
buffer containing ATP (indicated by the arrow), followed by mixing,
and the intensity of fluorescence mission was measured at 1-s intervals,
until a steady state was approached. At this time, buffer containing Fig. 4. Osmotic sensitivity of TMR transport into reconstituted
Chaps was added to a permeabilizing level (indicated by the arrow; Myr2GroPCho proteoliposomes containing Pgp. (A) TMR transport
final concentration 4 mm) and after mixing, further fluorescence was initiated by the addition of 1 mm ATP to Pgp proteoliposomes,
intensity data were collected. (B) As in (A), but Chaps was added prior pre-equilibrated in the presence of 1 mm TMR and increasing
to addition of buffer containing ATP. (C) TMR transport (0.5 mm TMR, concentrations of sucrose. (B) Plot of the fluorescence decrease as a
1 mm ATP) was initiated (indicated by the arrow) by addition of buffer function of the reciprocal of the sucrose concentration. Data points
containing ATP. Buffer containing additional TMR was then added to represent the mean of duplicate determinations for each sucrose
initiate a second round of transport (indicated by the arrow; final concentration.
concentration 3 mm), and after mixing, fluorescence intensity data
were collected until a new steady state was established. At this time,
buffer containing Chaps was added (indicated by the arrow; final
plot of the fluorescence decrease against the reciprocal of
concentration 4 mm) and after mixing, further fluorescence intensity
the sucrose concentration (Fig. 4B) showed that TMR
data were collected. (D) As in (C), but Chaps was added prior to
transport was osmotically sensitive and therefore represents
addition of the second portion of TMR. Each addition of reagent to the accumulation of TMR in the aqueous phase of the
cuvette resulted in a 10% dilution of the sample, resulting in a 10% proteoliposome lumen.
decrease in the overall fluorescence.
Kinetics of TMR transport in Pgp proteoliposomes
TMR transport was carried out with increasing amounts of
Chaps again restored the fluorescence to a higher level, proteoliposomes containing Pgp. Results indicated that the
indicating release of TMR from the proteoliposome lumen. rate of TMR transport was directly proportional to the
If the proteoliposomes were permeabilized with Chaps after amount of Pgp present (data not shown). Relative rates of
the first round of transport, the fluorescence intensity TMR transport varied linearly from 0.15 at 6.5 mg of
increased to a higher level on addition of the second portion protein, to 0.82 at 32.5 mg of protein (units ˆ % change in
of TMR, but no transport-associated fluorescence decrease fluorescence intensity´s21). The rate of TMR transport was
was seen, indicating that the second round of TMR determined at a fixed TMR concentration of 1 mm, and
transport had been prevented (Fig. 3D). increasing concentrations of ATP up to 1.0 mm (Fig. 5A).
Osmotic sensitivity experiments provided further The resulting data were fitted to the Michaelis±Menten
evidence that TMR is sequestered in the aqueous phase in equation, and led to an estimate of 0.48 mm for the Km of
the interior of the proteoliposomes during transport. When ATP for drug transport. This value is very close to that of
transport of TMR was carried out in the presence of both the Km for hydrolysis of ATP by this Pgp preparation
increasing concentrations of sucrose, the size of the (< 0.4 mm, see [8,23]), and the Kd for binding of ATP to
fluorescence decrease declined, until virtually no change the catalytic site, as determined by fluorescence quenching
in fluorescence was observed at 2.0 m sucrose (Fig. 4A). A of MIANS-labelled Pgp (0.4±0.5 mm, see [23,28]). The
1692 P. Lu et al. (Eur. J. Biochem. 268) q FEBS 2001

Thus verapamil both inhibited the net rate of transport of

TMR, and prevented the generation of a substrate concen-
tration gradient. Similar results were obtained for the
modulator cyclosporin A, which inhibited TMR transport
almost completely at 4 mm (Fig. 6C). A small amount of
stimulation of TMR transport (< 15%) was reproducibly
observed at very low cyclosporin A concentrations (100±
200 nm). This is similar to the effect of the drug on the
ATPase activity of Pgp [42], and is assumed to arise from
complex allosteric interactions affecting substrate binding
and transport (see below).
Many Pgp substrates and modulators inhibit each other's
transport. However, we noted previously that Pgp-mediated
transport of [3H]colchicine in plasma membrane vesicles
was activated by the linear tripeptide NAc-LLY-amide, and
some longer members of the series NAc-LnY-amide [11].
Others have also reported positive allosteric effects of some
compounds on the transport of Hoechst 33342 and
rhodamine 123 [38,43]. We now report that colchicine
activates transport of TMR in Pgp proteoliposomes, as
assessed using the real-time fluorescence assay (Fig. 6D).
The stimulatory effect was maximal in the range 12.5±
50 mm colchicine, substantially less than the Kd for
colchicine binding to Pgp of 158 mm [28], which suggests
that transport activation arises from allosteric effects. As
more colchicine was added, the rate of TMR transport
returned to the control 100% value; inhibition of transport
was not observed, even at concentrations as high as
200 mm.
Fig. 5. Dependence of the rate of TMR transport on ATP and TMR
concentrations in reconstituted Myr2GroPCho proteoliposomes
containing Pgp. (A) TMR transport was measured at a fixed TMR Effect of lipid bilayer phase state on TMR transport by
concentration of 1 mm and varying concentrations of ATP. (B) TMR Pgp
transport was measured at a fixed ATP concentration of 1 mm, and
Given that the functioning of Pgp is intimately connected
varying concentrations of TMR. Two independent measurements
with the membrane, we might expect the transport
(X,V), expressed as the change in fluorescence (arbitrary units)
properties of Pgp to be strongly influenced by the physical
per s, were conducted. Data points were fitted to the Michaelis±
Menten equation, as shown by the solid line, and the values of Km for
properties of the bilayer, such as its phase state. In the
ATP and TMR were estimated.
present study, we examined TMR transport in
proteoliposomes of two synthetic phosphatidylcholines,
Myr2GroPCho and PamMyrGroPCho, whose melting
rate of TMR transport was measured at increasing TMR temperatures are compatible with performing experiments
concentrations up to 2.5 mm, at a fixed ATP concentration in both the gel and liquid crystalline phases without risk of
of 1 mm (Fig. 5B). The resulting data fitted well to the Pgp denaturation. The rate of TMR transport was examined
Michaelis±Menten equation, and the Km for TMR transport over a temperature range spanning the phase transition (the
was estimated to be 0.3 mm. Thus, reconstituted Pgp melting temperature, Tm) for both lipids. The Tm values for
obeys Michaelis±Menten kinetics with respect to both PamMyrGroPCho and Myr2GroPCho under the
ATP and the transported substrate, TMR. Many of these experimental conditions of this study are 30.4 and
experiments were also carried out using Pgp reconsti- 24.0 8C, respectively [23]. Most membrane proteins show
tuted into the defined lipid PamMyrGroPCho, with very an increased rate of transport with increasing temperature,
similar results. with a discontinuity or change in slope in the Arrhenius plot
at Tm, and are often completely inactive in the gel phase.
The slope of the Arrhenius plot, which is negative, can be
used to obtain the activation energy, Eact for transport. As
Inhibition and activation of TMR transport by Pgp
shown in Fig. 7A,B reconstituted Pgp displayed unusual
modulators and substrates
behaviour. In PamMyrGroPCho bilayers, as the temperature
When TMR transport was measured in the presence of the was increased from 20 to 28 8C, the Arrhenius plot was
modulator verapamil, both the rate of the fluorescence relatively flat, with a slope close to zero. Around Tm, there
decrease and the magnitude of the total decrease seen at was an abrupt discontinuity, and the rate of transport then
steady state were lowered, in a concentration-dependent decreased as the temperature increased, so that the plot
fashion (Fig. 6A). The rate of transport declined with exhibited a positive slope. A similar pattern was seen in
increasing concentration of the modulator, with almost Myr2GroPCho bilayers, where the slope of the Arrhenius
complete inhibition observed at 20 mm verapamil (Fig. 6B). plot was small and negative from 16.5 to 24 8C, followed
q FEBS 2001 Drug transport by reconstituted P-glycoprotein (Eur. J. Biochem. 268) 1693

Fig. 6. Inhibition and activation of

Pgp-mediated TMR transport into
reconstituted Myr2GroPCho
proteoliposomes containing Pgp.
(A) Fluorescence traces showing changes in
TMR fluorescence resulting from inward
transport (1 mm TMR, 1 mm ATP),
measured following preincubation of
Pgp proteoliposomes with increasing
concentrations of verapamil, as indicated.
The rate of TMR transport was measured at
various concentrations of verapamil (B),
cyclosporin A (C), and colchicine (D). The
data points represent the mean of two
independent measurements of the rate of
TMR transport, expressed as the percentage
change in fluorescence compared with the
rate measured in the absence of added drug.

by a discontinuity at Tm, with a progressive decrease in

transport rate above Tm, once again leading to a positive
DISCUSSION
slope. It was not possible to estimate activation energies for The observed rate of drug transport by Pgp is proposed to
transport above Tm from the Arrhenius plots shown in represent a balance between two opposing processes; active
Fig. 7, as the positive slopes for both lipid systems give pumping by Pgp up a concentration gradient, and passive
formally negative values of Eact. Below Tm, the slope of the diffusion of drug across the membrane down the
Arrhenius plot for PamMyrGroPCho was essentially zero, concentration gradient. We previously inferred the presence
while that for Myr2GroPCho was negative, leading to a of a substrate concentration gradient indirectly from
value for Eact of < 24 kJ´mol21. This is in good agreement calculations of differential association of radiolabelled
with the activation energy of 25 kJ´mol21 estimated by drug with plasma membrane vesicles and proteoliposomes
Tarasiuk and Garnier-Suillerot for Pgp-mediated efflux of a in the presence and absence of ATP. The gradient was
doxorubicin derivative by MDR K562 cells [44]. It should estimated at 18- and 10-fold for colchicine and vinblastine,
be noted that the decrease in the rate of TMR transport respectively, in plasma membrane vesicles [25], and 5.6-
above Tm is not due to protein denaturation as the and 25-fold for colchicine [9] and NAc-LLY-amide [11],
temperature is increased, as we showed previously that respectively, in reconstituted proteoliposomes. It has been
ATP hydrolysis by Pgp in Chaps solution exhibited a argued that indirect measurements of relatively small
normal linear Arrhenius plot with a negative slope [23]. concentration gradients such as these are not evidence for
There was no evidence of loss of catalytic activity over a drug pumping up a gradient by Pgp, and may be subject to
very wide temperature range that spanned the Tm for both potential artefacts such as drug aggregation within the
PamMyrGroPCho and Myr2GroPCho. vesicle lumen, as well as nonspecific association of drug
Many membrane transporters show very low, or with the membrane [45]. In the present study, we
negligible, rates of transport when the host lipid bilayer is demonstrate directly for the first time that Pgp generates
in the rigid gel phase, i.e. below Tm. In contrast, the rate of a substrate concentration gradient by inward-pumping of
TMR transport by Pgp in PamMyrGroPCho bilayers was the fluorescent substrate TMR into reconstituted proteo-
actually at its highest in the gel phase, below 28 8C liposomes. The subsequent collapse of this gradient by
(Fig. 7A). For Myr2GroPCho bilayers, the rate of TMR addition of vanadate to the vesicle exterior can be followed
transport declined slightly as the temperature was by monitoring the fluorescence in real time. The ATPase
decreased below the Tm of 24 8C, but overall, the activity of inward-facing Pgp molecules is inhibited, and
transport rate also remained very high in the gel phase. the ADP´Vi ´Mg21 complex is trapped at one active site
We reported previously that Pgp catalysed ATP hydro- after a single catalytic turnover [13]. Thus, the presence of
lysis quite well in the gel phase, and the value of Eact vanadate outside the proteoliposomes blocks inward
was substantially lower than that measured in liquid pumping by Pgp, leading to collapse of the previously
crystalline bilayers [23]. Overall, Pgp-mediated transport established gradient.
of TMR displayed highly unusual temperature-dependence The present work establishes that the fluorescent dye
characteristics. TMR is a high affinity substrate for Pgp, and shows that it
1694 P. Lu et al. (Eur. J. Biochem. 268) q FEBS 2001

can be used to carry out real-time transport measurements Pgp-mediated transport of TMR resulted in its
in proteoliposomes containing Pgp. The high sensitivity of sequestration inside the vesicle lumen, as shown by
this technique is suitable for reconstituted systems, and it Chaps permeabilization and osmotic sensitivity data.
has several advantages over methods used previously to TMR transport by Pgp obeyed Michaelis±Menten kinetics
examine transport in proteoliposomes. Measurements can with respect to both ATP and TMR. There was no evidence
be made in real time using a single proteoliposome sample, of positively cooperative kinetics, as reported previously for
a substantial improvement over the fixed time-point rapid transport of daunorubicin in intact cells [47] and DNA-
filtration technique, where multiple samples must be used loaded plasma membrane vesicles [48]. As it is difficult to
to obtain a single value of the transport rate, and time make true kinetic measurements in these cases, the previous
resolution is poor (first data point at 15 s, then 15 s observations may be an artefact arising from the complexity
intervals) [9,11]. The rate of transport can be measured over of the systems under study.
a wide range; relative values ranged from 0.03 to 1.0 (% The dynamic nature of the steady state maintained by
change in fluorescence intensity´s21). Unlike the Hoechst Pgp is exemplified by experiments in which TMR is added
33342 method, transport kinetics can be examined. In to the system more than once (Fig. 3C,D). Addition of ATP
principle, this real-time approach could be extended to to proteoliposomes in the presence of 0.5 mm TMR led to
other Pgp transport substrates or their derivatives, e.g. an immediate round of transport, with establishment of a
fluorescent verapamil, fluorescent peptides. TMR gradient. Addition of TMR up to 3 mm resulted in a
We have shown recently that a variety of NBD-labelled second round of transport and establishment of a new
phospholipids are flipped from one membrane leaflet to the steady state, with a concentration gradient that could also
other by Pgp in reconstituted proteoliposomes [46]. be collapsed by detergent permeabilization. Both the
Therefore, it is likely that normal unmodified membrane transport rate and the size of the new gradient established
lipids are also substrates for Pgp flippase activity, although for the second round of TMR transport were much larger
they are likely to be of relatively low affinity compared than those observed for the first round of transport, in
with compounds such as TMR. The measured affinity of keeping with the higher substrate concentration.
TMR binding may reflect its ability to displace bilayer The rate of TMR transport in proteoliposomes was
lipids from the substrate binding site(s) of the transporter. dependent on the concentrations of both ATP and TMR.
The Km value estimated for ATP by Michaelis±Menten
analysis was 0.48 mm, close to the Km for ATP hydrolysis
(< 0.4 mm) and the Kd for binding of ATP estimated by
fluorescence quenching measurements (0.4±0.5 mm)
(summarized in [35,36]). Analysis of kinetic measurements
showed that TMR is a high affinity transport substrate, with
a Km value of 0.3 mm. This is the first time that Km values
have been measured for drug transport by Pgp in a simple
reconstituted system. The value of Km obtained for TMR
compares favourably with its Kd for binding to purified
MIANS-labelled Pgp, which was estimated to be 0.70 mm
(Fig. 1).
The TMR gradient was collapsed by the modulators
cyclosporin A and verapamil, which compete for TMR
transport at the drug binding site(s) of the protein, and by
vanadate, which interacts with the NB domains of Pgp to
inhibit ATPase activity. Both the rate of TMR transport and
the size of the generated substrate concentration gradient
were inhibited by verapamil in a concentration-dependent
manner. We also report the novel finding that colchicine is
able to activate Pgp-mediated transport of TMR in proteo-
liposomes. Additional experiments revealed that TMR acts
reciprocally, activating transport of [3H]colchicine 5.5-fold
(P. Lu and F. J. Sharom, unpublished data). This mutual
activation of transport probably takes place via positive
allosteric effects, and reinforces the concept of multiple
binding sites, where occupancy of one site by substrate
influences the other site. Thus, if we assume (as proposed
by Neyfakh and coworkers [49]) the presence of a flexible
`binding region' within the transporter, more than one
Fig. 7. Arrhenius plot showing the effect of temperature and lipid substrate molecule can apparently occupy this region
phase state on Pgp-mediated transport of TMR into reconstituted simultaneously. We also noted mutual activation of
proteoliposomes containing Pgp. (A) PamMyrGroPCho proteo- transport for colchicine and the tripeptide NAc-LLY-
liposomes; (B) Myr2GroPCho proteoliposomes. The gel-to-liquid amide [11], suggesting that the peptide and TMR may
crystalline phase transition temperature, Tm, for each lipid is indicated occupy the same site within this binding region. Previous
by the arrow; Tm ˆ 30.4 8C for PamMyrGroPCho, and 24.0 8C experiments demonstrating allosteric activation of transport
for Myr2GroPCho [23]. were carried out in native plasma membrane vesicles. The
q FEBS 2001 Drug transport by reconstituted P-glycoprotein (Eur. J. Biochem. 268) 1695

present work indicates that such allosteric effects also take TMR. However, it seems reasonable to assume that the
place in a simple reconstituted system, which may allow temperature dependence of Plip observed by Rogers and
more detailed dissection of the phenomenon. Davis would also apply to some other lipophilic com-
The temperature dependence of TMR transport in pounds. ATP-driven membrane transport by Pgp is a
reconstituted bilayers of two synthetic phospholipids multistep process, involving binding and release of the
showed several very unusual attributes. First, transport transported substrate, binding and enzymatic hydrolysis of
rates were relatively high in the gel state. The behaviour of ATP, release of ADP and Pi, and protein conformational
Pgp is unusual in this respect, as transport by many changes. In the case of Pgp, we can add to this the
membrane proteins in a rigid gel phase lipid environment is partitioning of transport substrate into the membrane, which
often low, or nonexistent, compared with activity in the is probably an essential prerequisite for transport to take
fluid liquid crystalline phase. Gel phase lipid is rigid and place. The rates of all of these steps are likely to be
poorly compressible, and it is thought that a rate-limiting temperature dependent. The results of the present study
protein conformational change is hindered in such an suggest that the temperature dependence of drug transport
environment. Pgp does not appear to have a conformational by Pgp may be dominated by partitioning effects on the
barrier of this type. As the temperature was increased, the local drug concentration within the bilayer, providing
rate of TMR transport reached a maximum at Tm, and then further support for the hydrophobic vacuum cleaner
declined substantially in the fluid liquid crystalline phase. model.
Again, this type of behaviour is very different from that
observed for other membrane transport proteins.
Overall, the temperature dependence of Pgp-mediated ACKNOWLEDGEMENT
TMR transport displayed highly atypical characteristics.
This research was supported by a research grant to F. J. S. from the
This unusual behaviour is unlikely to be restricted to TMR National Cancer Institute of Canada, with funds provided by the
as a substrate. For example, we previously noted that uptake Canadian Cancer Society.
of [3H]colchicine into reconstituted proteoliposomes of
Myr2GroPCho (determined by rapid filtration) was more
than twofold higher at the Tm of 24 8C than in the fluid REFERENCES
liquid crystalline phase at 32 8C [50]. Sinicrope et al.
reported that addition of chemical fluidizing agents led to a 1. Bosch, I. & Croop, J. (1996) P-glycoprotein multidrug resistance
decrease in Pgp-mediated transport of vinblastine and and cancer. Biochim. Biophys. Acta 1288, F37±F54.
daunorubicin in canalicular membrane vesicles [51], and 2. Germann, U.A. (1996) P-Glycoprotein ± a mediator of multidrug
Eytan and coworkers also found that membrane fluidizers resistance in tumour cells. Eur. J. Cancer 32A, 927±944.
inhibited calcein-AM efflux in MDR cells [52]. Both of 3. Sharom, F.J. (1997) The P-glycoprotein efflux pump: how does it
these reports are consistent with the results of the present transport drugs? J. Membr. Biol. 160,
161±175.
study.
4. Ambudkar, S.V., Dey, S., Hrycyna, C.A., Ramachandra, M.,
What could account for this unusual behaviour? Accord-
Pastan, I. & Gottesman, M.M. (1999) Biochemical, cellular, and
ing to the hydrophobic vacuum cleaner model, drugs must pharmacological aspects of the multidrug transporter. Annu. Rev.
partition into the membrane before gaining access to the Pharmacol. Toxicol. 39, 361±398.
transport site on Pgp. Because the rate of transport by Pgp 5. Holland, I.B. & Blight, M.A. (1999) ABC-ATPases, adaptable
will depend directly on the drug concentration within the energy generators fuelling transmembrane movement of a variety
bilayer, and this in turn is related to the degree of of molecules in organisms from bacteria to humans. J. Mol. Biol.
membrane partitioning, it seems likely that the rate of 293, 381±399.
drug transport depends directly on the lipid/water partition 6. Urbatsch, I.L., al-Shawi, M.K. & Senior, A.E. (1994)
coefficient, Plip. The temperature dependence of Plip may Characterization of the ATPase activity of purified Chinese
thus be reflected directly in the temperature dependence of hamster P-glycoprotein. Biochemistry 33,
drug transport. Rogers and Davis [53] showed that Plip for 7069±7076.
partitioning of various p-alkylphenols between water and 7. Shapiro, A.B. & Ling, V. (1994) ATPase activity of purified and
Myr2GroPCho bilayers displayed an interesting biphasic reconstituted P-glycoprotein from Chinese hamster ovary cells. J.
temperature dependence. The value of Plip showed a Biol. Chem. 269, 3745±3754.
maximum at Tm, declined substantially at temperatures 8. Sharom, F.J., Yu, X., Chu, J.W.K. & Doige, C.A. (1995)
higher than Tm, and remained constant, or declined slightly Characterization of the ATPase activity of P-glycoprotein from
at temperatures lower than Tm. The pattern of temperature multidrug-resistant Chinese hamster ovary cells. Biochem. J. 308,
381±390.
dependence of Plip seen on an Arrhenius plot is strikingly
9. Sharom, F.J., Yu, X. & Doige, C.A. (1993) Functional
similar to that seen for TMR transport by Pgp in bilayers of
reconstitution of drug transport and ATPase activity in
PamMyrGroPCho and Myr2GroPCho in the present study.
proteoliposomes containing partially purified P-glycoprotein. J.
We attempted to measure Plip for TMR in Myr2GroPCho Biol. Chem. 268, 24197±24202.
liposomes at different temperatures, as we have done 10. Shapiro, A.B. & Ling, V. (1995) Reconstitution of drug transport
previously for chemotherapeutic drugs at 22 8C [16], using by purified P-glycoprotein. J. Biol. Chem. 270, 16167±16175.
TMR fluorescence intensity as means of quantification. 11. Sharom, F.J., Yu, X., DiDiodato, G. & Chu, J.W.K. (1996)
Unfortunately, this proved technically impossible. First, Synthetic hydrophobic peptides are substrates for P-glycoprotein
TMR fluorescence emission is itself temperature depen- and stimulate drug transport. Biochem. J. 320, 421±428.
dent, and second, the synthetic liposome properties changed 12. Chan, H.S., Grogan, T.M., DeBoer, G., Haddad, G., Gallie, B.L. &
substantially at higher temperatures, so they could not be Ling, V. (1996) Diagnosis and reversal of multidrug resistance in
separated readily from an aqueous solution containing paediatric cancers. Eur. J. Cancer 32A, 1051±1061.
1696 P. Lu et al. (Eur. J. Biochem. 268) q FEBS 2001

13. Urbatsch, I.L., Sankaran, B., Weber, J. & Senior, A.E. (1995) 33. Lelong, I.H., Guzikowski, A.P., Haugland, R.P., Pastan, I.,
P-glycoprotein is stably inhibited by vanadate-induced trapping Gottesman, M.M. & Willingham, M.C. (1991) Fluorescent
of nucleotide at a single catalytic site. J. Biol. Chem. 270, 19383± verapamil derivative for monitoring activity of the multidrug
19390. transporter. Mol. Pharmacol. 40, 490±494.
14. Senior, A.E., al-Shawi, M.K. & Urbatsch, I.L. (1995) The catalytic 34. Eytan, G.D., Regev, R., Oren, G., Hurwitz, C.D. & Assaraf, Y.G.
cycle of P-glycoprotein. FEBS Lett. 377, 285±289. (1997) Efficiency of P-glycoprotein-mediated exclusion of rhoda-
15. Higgins, C.F. & Gottesman, M.M. (1992) Is the multidrug mine dyes from multidrug-resistant cells is determined by their
transporter a flippase? Trends Biochem. Sci. 17, 18±21. passive transmembrane movement rate. Eur. J. Biochem. 248,
16. Romsicki, Y. & Sharom, F.J. (1999) The membrane lipid 104±112.
environment modulates drug interactions with the P-glycoprotein 35. Sharom, F.J., Liu, R.H. & Romsicki, Y. (1998) Spectroscopic and
multidrug transporter. Biochemistry 38, 6887±6896. biophysical approaches for studying the structure and function of
17. Shapiro, A.B. & Ling, V. (1997) Extraction of Hoechst 33342 from the P-glycoprotein multidrug transporter. Biochem. Cell Biol. 76,
the cytoplasmic leaflet of the plasma membrane by P-glycoprotein. 695±708.
Eur. J. Biochem. 250, 122±129. 36. Sharom, F.J., Liu, R.H., Romsicki, Y. & Lu, P.H. (1999) Insights
18. Shapiro, A.B. & Ling, V. (1998) Transport of LDS-751 from the into the structure and substrate interactions of the P-glycoprotein
cytoplasmic leaflet of the plasma membrane by the rhodamine- multidrug transporter from spectroscopic studies. Biochim.
123-selective site of P-glycoprotein. Eur. J. Biochem. 254, 181± Biophys. Acta 1461, 327±345.
188. 37. Kolaczkowski, M., van der Rest, M., Cybularz-Kolaczkowska, A.,
19. Bolhuis, H., Van Veen, H.W., Brands, J.R., Putman, M., Poolman, Soumillion, J.P., Konings, W.N. & Goffeau, A. (1996) Anticancer
B., Driessen, A.J.M. & Konings, W.N. (1996) Energetics and drugs, ionophoric peptides, and steroids as substrates of the yeast
mechanism of drug transport mediated by the lactococcal multi- multidrug transporter Pdr5p. J. Biol. Chem. 271, 31543±31548.
drug transporter LmrP. J. Biol. Chem. 271, 24123±24128. 38. Shapiro, A.B. & Ling, V. (1997) Positively cooperative sites for
20. McDonald, T.P., Walmsley, A.R., Martin, G.E. & Henderson, P.J. drug transport by P-glycoprotein with distinct drug specificities.
(1995) The role of tryptophans 371 and 395 in the binding of Eur. J. Biochem. 250, 130±137.
antibiotics and the transport of sugars by the d-galactose-H1 39. Doige, C.A., Yu, X. & Sharom, F.J. (1992) ATPase activity of
symport protein (GalP) from Escherichia coli. J. Biol. Chem. 270, partially purified P-glycoprotein from multidrug-resistant Chinese
30359±30370. hamster ovary cells. Biochim. Biophys. Acta 1109, 149±160.
21. Martin, G.E., Rutherford, N.G., Henderson, P.J. & Walmsley, A.R. 40. Loe, D.W. & Sharom, F.J. (1993) Interaction of multidrug-resistant
(1995) Kinetics and thermodynamics of the binding of forskolin to Chinese hamster ovary cells with amphiphiles. Br. J. Cancer 68,
the galactose-H1 transport protein, GalP, of Escherichia coli. 342±351.
Biochem. J. 308, 261±268.
41. Doige, C.A., Yu, X. & Sharom, F.J. (1993) The effects of lipids
22. Shapiro, A.B. & Ling, V. (1998) Stoichiometry of coupling of
and detergents on ATPase-active P-glycoprotein. Biochim. Bio-
rhodamine 123 transport to ATP hydrolysis by P-glycoprotein. Eur.
phys. Acta 1146, 65±72.
J. Biochem. 254, 189±193.
42. Sharom, F.J., DiDiodato, G., Yu, X. & Ashbourne, K.J. (1995)
23. Romsicki, Y. & Sharom, F.J. (1998) The ATPase and ATP binding
Interaction of the P-glycoprotein multidrug transporter with
functions of P-glycoprotein: modulation by interaction with
peptides and ionophores. J. Biol. Chem. 270, 10334±10341.
defined phospholipids. Eur. J. Biochem. 256, 170±178.
43. Shapiro, A.B., Fox, K., Lam, P. & Ling, V. (1999) Stimulation of
24. Doige, C.A. & Sharom, F.J. (1991) Strategies for the purification
P-glycoprotein-mediated drug transport by prazosin and proges-
of P-glycoprotein from multidrug-resistant Chinese hamster ovary
terone. Evidence for a third drug-binding site. Eur. J. Biochem.
cells. Protein Expr. Purif. 2, 256±265.
259, 841±850.
25. Doige, C.A. & Sharom, F.J. (1992) Transport properties of
P-glycoprotein in plasma membrane vesicles from multidrug- 44. Tarasiuk, J. & Garnier-Suillerot, A. (1992) Kinetic parameters for
resistant Chinese hamster ovary cells. Biochim. Biophys. Acta the uptake of anthracycline by drug-resistant and drug-sensitive
1109, 161±171. K562 cells. Eur. J. Biochem. 204, 693±698.
26. Bradford, M.M. (1976) A rapid and sensitive method for the 45. Roepe, P.D. (1998) The P-glycoprotein efflux pump: How does it
quantitation of microgram quantities of protein utilizing the transport drugs? J. Membr. Biol. 166, 71±72.
principle of protein-dye binding. Anal. Biochem. 72, 248±254. 46. Romsicki, Y. & Sharom, F.J. (2001). Phospholipid flippase activity
27. Peterson, G.L. (1977) A simplification of the protein assay method of the reconstituted, P-glycoprotein multidrug transporter. Bio-
of Lowry et al. which is more generally applicable. Anal. Biochem. chemistry, in press.
83, 346±356. 47. Spoelstra, E.C., Westerhoff, H.V., Dekker, H. & Lankelma, J.
28. Liu, R. & Sharom, F.J. (1996) Site-directed fluorescence labeling (1992) Kinetics of daunorubicin transport by P-glycoprotein of
of P-glycoprotein on cysteine residues in the nucleotide binding intact cancer cells. Eur. J. Biochem. 207, 567±579.
domains. Biochemistry 35, 11865±11873. 48. Guiral, M., Viratelle, O., Westerhoff, H.V. & Lankelma, J. (1994)
29. Sharom, F.J., Lu, P., Liu, R. & Yu, X. (1998) Linear and cyclic Cooperative P-glycoprotein mediated daunorubicin transport into
peptides as substrates and modulators of P-glycoprotein: peptide DNA-loaded plasma membrane vesicles. FEBS Lett. 346,
binding and effects on drug transport and accumulation. Biochem. 141±145.
J. 333, 621±630. 49. Zheleznova, E.E., Markham, P., Edgar, R., Bibi, E., Neyfakh, A.A.
30. Romsicki, Y. & Sharom, F.J. (1997) Interaction of P-glycoprotein & Brennan, R.G. (2000) A structure-based mechanism for drug
with defined phospholipid bilayers: a differential scanning binding by multidrug transporters. Trends Biochem. Sci. 25,
calorimetric study. Biochemistry 36, 9807±9815. 39±43.
31. Homolya, L., Hollo, Z., Germann, U.A., Pastan, I., Gottesman, 50. Sharom, F.J. (1997) The P-glycoprotein multidrug transporter:
M.M. & Sarkadi, B. (1993) Fluorescent cellular indicators are interactions with membrane lipids, and their modulation of
extruded by the multidrug resistance protein. J. Biol. Chem. 268, activity. Biochem. Soc. Trans. 25, 1088±1096.
21493±21496. 51. Sinicrope, F.A., Dudeja, P.K., Bissonnette, B.M., Safa, A.R. &
32. Hollo, Z., Homolya, L., Davis, C.W. & Sarkadi, B. (1994) Calcein Brasitus, T.A. (1992) Modulation of P-glycoprotein-mediated drug
accumulation as a fluorometric functional assay of the multidrug transport by alterations in lipid fluidity of rat liver canalicular
transporter. Biochim. Biophys. Acta 1191, 384±388. membrane vesicles. J. Biol. Chem. 267, 24995±25002.
q FEBS 2001 Drug transport by reconstituted P-glycoprotein (Eur. J. Biochem. 268) 1697

52. Regev, R., Assaraf, Y.G. & Eytan, G.D. (1999) Membrane fluid- 53. Rogers, J.A. & Davis, S.S. (1980) Functional group contributions
ization by ether, other anesthetics, and certain agents abolishes to the partitioning of phenols between liposomes and water.
P-glycoprotein ATPase activity and modulates efflux from Biochim. Biophys. Acta 598, 392±404.
multidrug-resistant cells. Eur. J. Biochem. 259, 18±24.