Color profile: Disabled

Composite Default screen

11

REVIEW / SYNTHÈSE

Merck Frosst Award Lecture 1998 /

La conference Merck Frosst 1998

Molecular dissection of the human multidrug

resistance P-glycoprotein
Tip W. Loo and David M. Clarke

Abstract: The human multidrug resistance P-glycoprotein is an ATP-dependent drug pump that extrudes a broad range
of cytotoxic agents from the cell. Its physiological role may be to protect the body from endogenous and exogenous
cytotoxic agents. The protein has clinical importance because it contributes to the phenomenon of multidrug resistance
during chemotherapy. In this review, we discuss some of the results obtained by using molecular biology and protein
chemistry techniques for studying this important and intriguing protein.
Key words: P-glycoprotein, ABC transporters, drug transport, dibromobimane, mutagenesis, disulfide crosslinking,
metal-chelate chromatography, ATPase activity.
Résumé : La glycoprotéine P humaine de résistance multidrogue est une pompe ATP-dépendante qui expulse un large
éventail de substances cytotoxiques de la cellule. Son rôle physiologique pourrait être de protéger l’organisme contre
des substances cytotoxiques endogènes et exogènes. La protéine a une importance clinique car elle contribue au
phénomène de résistance à plusieurs médicaments utilisés en chimiothérapie. Dans cette revue, nous discutons de
résultats obtenus en appliquant des méthodes de biologie moléculaire et de chimie des protéines à l’étude de cette
protéine importante et fascinante.
Mots clés : glycoprotéine P, transporteurs ABC, transport de médicaments, dibromobimane, mutagenèse, liaison disul-
fure, chromatographie par chélation de métaux, activité ATPase.
[Traduit par la Rédaction] Loo and Clarke 23

Introduction (ATP-binding cassette) superfamily of proteins. The ABC

family of proteins is the largest class of transport proteins.
The human multidrug resistance P-glycoprotein (P-gp) is This minireview summarizes the clinical and physiological
clinically important because it contributes to the phenome- relevance of the protein and describes results from our labo-
non of multidrug resistance during cancer and AIDS chemo- ratory that contribute to our understanding of the structure,
therapy. P-gp is also interesting for biochemists because it mechanism and biosynthesis of this intriguing protein. More
uses the energy from ATP to transport a broad range of comprehensive reviews on P-gp have been published else-
structurally unrelated cytotoxic compounds out of the cell. It where (Chan et al. 1996; Gottesman et al. 1996; Sharom
also serves as a model protein for understanding the struc- 1997).
tures and mechanisms of other members of the ABC
Received December 18, 1998. Accepted January 21, 1999.
P-gp and multidrug resistance in cancer
Chemotherapy is a major form of treatment for Hodgkin’s
Abbreviations: dBBn, dibromobimane; NEM, disease, testicular cancer, and many childhood cancers. Combi-
N-ethylmaleimide; P-gp, P-glycoprotein; TM, transmembrane. nation chemotherapy with cytotoxic agents that have different
T.W. Loo and D.M. Clarke.1 MRC Group in Membrane intracellular targets has been particularly effective. Unfortu-
Biology, Department of Medicine and Department of nately, the majority of cancers are either resistant to chemother-
Biochemistry, University of Toronto, ON M5S 1A8, Canada. apy (renal, colon, etc.) or acquire resistance (such as in
1
Author to whom all correspondence should be sent at the lymphoma, lung and breast cancers) during treatment (Lehnert
following address: Department of Medicine, University of 1996). This intrinsic or acquired ability of tumor cells to be re-
Toronto, Room 7342, Medical Sciences Building, sistant to multiple chemotherapeutic agents (multidrug resis-
1 King’s College Circle, Toronto, ON M5S 1A8, Canada tance) is a major obstacle to successful cancer chemotherapy.
(e-mail: [email protected]). P-gp, the product of the human MDR1 gene, was the first

Biochem. Cell Biol. 77: 11–23 (1999) © 1999 NRC Canada

I:\bcb\bcb77\01\O99-014.vp
Thursday, May 27, 1999 10:07:08 AM
Color profile: Disabled
Composite Default screen

12 Biochem. Cell Biol. Vol. 77, 1999

Fig. 1. P-gp and multidrug resistance. (A) Cytotoxic drugs such modulated when the tumour cells are stressed (e.g., heat,
as vinblastine or colchicine rapidly diffuse into the lipid bilayer heavy metals) and by the recent observation that there is
of a cell. The drugs enter the cytoplasm, bind to an intracellular p53-dependent regulation of MDR1 expression (Thottassery
target (microtubules in the case of colchicine or vinblastine) and et al. 1997). Inhibition of wild-type p53 resulted in increased
cause cell death. Cells expressing P-gp, however, are protected expression of P-gp. Since deletion or mutation of the p53
because the protein extracts the drugs from the lipid bilayer and (tumour suppressor gene) gene is also a frequent finding in
pumps them out in an ATP-dependent manner. (B) Schematic human malignancies (Hollstein et al. 1991; Levine et al.
model of P-gp. It consists of four domains; two cytoplasmic 1991), it would be expected that tumours expressing the
ATP-binding domains, and two hydrophobic domains each MDR1 gene would be highly resistant to cytotoxic drugs.
containing six predicted TM segments. The protein is The anticancer drugs most effectively extruded from tu-
glycosylated at three sites in the first extracellular loop between mor cells by P-gp are of natural origin. Examples include
TM1 and TM2. the anthracyclines (e.g., doxorubicin, daunorubicin, and
mitoxantrone), vinca alkaloids (e.g., vincristine and
vinblastine), epipodophyllotoxins (e.g., etoposide and
teniposide), taxanes (e.g., taxol and taxotere), and
actinomycin D (Sarkadi and Muller 1997). In cancer chemo-
therapy, the goal is to inhibit P-gp-mediated extrusion of
these anticancer drugs to increase the effectiveness of treat-
ment. Considerable efforts have been expended on finding
chemosensitizers that will inhibit the function of P-gp and
thereby reverse multidrug resistance. Initially, it was found
that the calcium channel blocker verapamil greatly increased
the sensitivity of multidrug resistant leukemia cells to
cytotoxic agents in vivo and in vitro (Tsuruo et al. 1981,
1983). The most widely used compounds to inhibit P-gp
function in initial clinical trials were verapamil and
cyclosporin A. Unfortunately, the concentrations of these
compounds required to inhibit P-gp lead to significant side
effects. High verapamil levels cause cardiovascular toxicity
and cyclosporin A enhances myeloid, renal, neural and
hepatic toxicity. The use of short high-dose cyclosporin infu-
sions at the same time of administration of vincristine, how-
ever, has been successful in the treatment of retinoblastoma
(Chan et al. 1996). The side effects of verapamil and
important protein implicated in the phenomenon of multidrug cyclosporin may be lessened with the development of sec-
resistance (see Fig. 1) (Juliano and Ling 1976). Mammalian ond generation analogs such as R-verapamil and PSC 833.
cell lines overexpressing this protein were initially selected R-verapamil has less calcium channel inhibitory effect than
for resistance to a single cytotoxic drug, but were subse- the S-enantiomer of verapamil but with a similar ability to
quently found to display cross-resistance to other structur- inhibit P-gp (Gruber et al. 1988). PSC 833 is a cyclosporin
ally unrelated drugs (Biedler and Riehm 1970; Riordan and analog with virtually no immunosuppressive effect (Boesch
Ling 1985; Shen et al. 1986). Following this discovery, et al. 1991). A desirable goal is to develop a more specific
other proteins have been identified that may also confer and effective modulator. The major drawback in modulator
multidrug resistance. These include the ABC multidrug re- development has been the lack of structural information re-
sistance proteins, MRP1 and MRP2; as well as garding specific binding sites or intramolecular arrange-
topoisomerase II and the lung resistance protein (Lehnert ments during drug transport by P-gp.
1996). The clinical relevance of these other proteins in can-
cer multidrug resistance, however, is still under investiga- HIV-1 protease inhibitors and P-gp
tion. Shutting down P-gp during chemotherapy would also ben-
There is considerable evidence that many different tu- efit treatment of other diseases, such as AIDS. Protease in-
mours express P-gp (reviewed in Chan et al. 1996; Fisher et hibitors are potent agents that are in vogue in the therapy of
al. 1996; Goldstein 1996; Lehnert 1996). In renal (Fojo et al. HIV-1 infection. Oral absorption and penetration of these in-
1987) or colon cancers (Weinstein et al. 1991), P-gp is con- hibitors into the brain, however, are poor. It is now apparent
stitutively expressed in relatively high amounts. In other that poor oral absorption and brain penetration of these
cancers (lung, myeloma, breast, ovary, lymphoma, acute FDA-approved protease inhibitors are due to the presence of
myeloid leukemia), the tumour cells frequently express P-gp P-gp (Kim et al. 1998; Lee and Gottesman 1998; Lee et al.
only after exposure to chemotherapeutic drugs or during re- 1998). Kim et al. (1998) recently showed that the HIV-1
lapse (Chan et al. 1995). Several studies have shown P-gp protease inhibitors indinavir, nelfinavir, and saquinavir are
expression to be predictive of poor response to chemother- all substrates of P-gp. The high levels of P-gp expressed in
apy and decreased overall survival (Chan et al. 1995; Fisher the blood-brain barrier may decrease the efficacy of these
et al. 1996; Leighton and Goldstein 1995; Marie 1995). drug in the treatment of central nervous system infections in
An interesting observation is that expression of P-gp is AIDS patients. Therefore, strategies to shut down P-gp

© 1999 NRC Canada

I:\bcb\bcb77\01\O99-014.vp
Thursday, May 27, 1999 10:07:10 AM
Color profile: Disabled
Composite Default screen

Loo and Clarke 13

Fig. 2. Ball model of human P-gp.

could potentially benefit other diseases that involve a che- Charuk and Reithmeier 1992). Perhaps the most important
motherapy regimen. role of P-gp is to protect us from the numerous toxins pres-
ent in our diet.
Physiological role of P-gp Because of such a striking similarity in the secondary
The major physiological role of P-gp is probably to pro- structure of P-gp and CFTR, it was thought that P-gp may
tect the body from hydrophobic toxic agents. P-gp is found also function as a chloride channel. Indeed, it was reported
on the apical (or lumenal) surface of polarized epithelial that P-gp was a volume-regulated chloride channel
cells of the small and large intestine, in the biliary (Valverde et al. 1992), and that the chloride channel activity
canalicular membranes of hepatocytes, and on the apical sur- of P-gp could be separated from its drug transport activity
face of epithelial cells of the proximal cells of kidney (Baas (Gill et al. 1992). Several subsequent studies, however, have
and Borst 1988; Croop et al. 1989; Schinkel 1997; Thiebaut questioned this finding (Dong et al. 1994; Morin et al. 1995;
et al. 1987; Trezise et al. 1992; Van der Bliek et al. 1987). Rasola et al. 1994; Tominaga et al. 1995) but some still be-
P-gp is also found in the blood–brain and blood–testis bar- lieve that P-gp is a regulator of chloride channel activity
rier (Cordon-Cardo et al. 1989) and probably serves to pro- (Valverde et al. 1996).
tect vital organs from naturally occurring xenotoxins Roepe (1995) have postulated an “altered partitioning
ingested in food. Studies on P-gp knock-out mice show that model” for P-gp where altered sequestration of cytotoxic
the protein is not essential (Schinkel et al. 1994; Schuetz et compounds by the cell occurs indirectly as a result of pertur-
al. 1996). The mice are viable and fertile and do not display bation of the plasma membrane electrochemical potential by
obvious phenotypic abnormalities other than hypersensitivity P-gp. This has been reviewed quite comprehensively (Roepe
to drugs. Therefore P-gp can provide protection by exclusion et al. 1996).
of these toxins in the intestine or blood–tissue barriers or by
active excretion in the intestine, liver, or kidney. Evidence Structure–function analysis of the
for such a protective role for P-gp in eliminating exogenous multidrug resistance P-gp
compounds was the recent demonstration that compounds
found in the urine of rats and humans are substrates of P-gp. Cloning of P-gp
One such class of compounds excreted in human urine are In parallel work, three groups cloned and sequenced the
the nonylphenol ethoxylates (Charuk et al. 1994, 1998; gene responsible for multidrug resistance by P-gp from ham-

© 1999 NRC Canada

I:\bcb\bcb77\01\O99-014.vp
Thursday, May 27, 1999 10:07:12 AM
Color profile: Disabled
Composite Default screen

14 Biochem. Cell Biol. Vol. 77, 1999

ster (Gerlach et al. 1986), mouse (Gros et al. 1986b), and mally do not express SERCA1. In addition, the amount of
human (Chen et al. 1986) cell lines. The 1280 amino acids P-gp-A52 expressed in whole cells could be quantitated by
of human MDR1 (Chen et al. 1986) are organized in two an ELISA assay by using purified SERCA1 as a standard.
tandem repeats of 610 amino acids, joined by a linker region The SERCA1 Ca-ATPase is readily purified in large
of 60 amino acids (Fig. 2). Each repeat consists of an amounts from rabbit muscle.
NH2-terminal hydrophobic domain containing six potential A second modification of the MDR1 cDNA was the addi-
TM sequences followed by a hydrophilic domain containing tion of a polyhistidine tag at the COOH-terminal end of
a nucleotide-binding site. The organization of the domains is P-gp, and the development of a rapid expression, purifica-
characteristic of members of the ABC (ATP-binding cas- tion, and assay protocol for human P-gp (Loo and Clarke
sette) superfamily of transporters (Higgins 1995). 1995d). This involved transient expression of the
It was apparent that P-gp is a member of a multigene pgp histidine-tagged P-gp in HEK 293 cells followed by purifi-
or mdr family (Riordan and Ling 1985). The mdr/pgp gene cation by nickel-chelate chromatography and measurement
family is composed of three members in rodents (hamster of drug-stimulated ATPase activity. Expression, purification,
and mouse) and two members in humans (Ng et al. 1989). and assay of ATPase activity can now be completed in
The close linkage of mdr/pgp genes on the chromosome im- 2–3 days, while previous methods using stable cell lines or
plied that the gene family arose from one or more gene du- expression in insect cells often took months. P-gp was the
plications (Bell et al. 1987; de Bruijn et al. 1986; Van der first ABC transporter and, to our knowledge, the first
Bliek et al. 1988). A study on the intron–exon structure of eukaryotic transporter that could be purified in an active
human MDR1, however, suggests that P-gp arose by fusion state using this approach.
of genes (Chen et al. 1990). Although all P-gp isoforms to The third modification was the construction of an active
date reveal the same overall structure, evidence suggests that Cys-less P-gp (Loo and Clarke 1995b). Mutation of each of
they may have distinct functions. Expression vectors have the seven endogenous cysteines to alanines yielded a mutant
been used to show that mdr1 of mouse or human is capable that was about 70% as efficient as wild-type protein in con-
of conferring multidrug resistance when transfected and ferring resistance to various cytotoxic substrates. It is syn-
overexpressed in otherwise drug-sensitive, cultured cells thesized less efficiently than wild-type enzyme but this
(Gros et al. 1986a; Ueda et al. 1987). There is evidence that problem can be corrected by carrying out synthesis in the
suggests that mouse mdr3 is associated with multidrug resis- presence of non-toxic substrates such as cyclosporin A (Loo
tance, whereas human MDR3 and mouse mdr2 may not be and Clarke 1997a). This was a very fortunate development
associated with multiple drug resistance but act as phospha- because efforts to construct Cys-less versions of other
tidyl translocases (Ruetz and Gros 1994; Smit et al. 1993; eukaryotic transporters have usually failed because of prob-
Smith et al. 1994). The results indicate that mouse mdr1 and lems with expression or protein stability. The Cys-less P-gp
mdr3 have overlapping, but distinct, specificities (Raymond has been an invaluable tool in mapping the topology, study-
et al. 1990). The drug specificity of human P-gp differs from ing the contribution of the two nucleotide-binding sites, de-
both mouse enzymes (Tang-Wai et al. 1995). termining packing of the transmembrane (TM) segments and
in mapping the drug binding sites as described below.
Modification of the P-gp cDNA
Understanding the structure of P-gp is central to our un- Topology of the TM domains of P-gp
derstanding of its mechanism. It is often quite difficult to The TM domains appear to contain the drug-binding sites
obtain structural information on most polytopic membrane and likely form the translocation pathway through the mem-
proteins because of technical difficulties in crystallizing brane (Homolya et al. 1993; Raviv et al. 1990). The results
membrane proteins. from labeling studies with photolabeled analogs of drug sub-
Our approach to obtaining structural and functional infor- strates (Ambudkar et al. 1997; Bruggemann et al. 1989,
mation about P-gp was to combine molecular biology with 1992; Greenberger 1993; Greenberger et al. 1990; Morris et
classical protein chemistry. This required several modifica- al. 1994), suggested that the labeled sites are closely associ-
tions of the MDR1 cDNA. ated with TM6 and TM12. These photolabeling studies sug-
The first modification was the construction of a functional gest that there may be two distinct drug-labeling sites or that
epitope-tagged human P-gp. In initial studies, the mutants of the two labeled segments are part of a single drug-labeling
human MDR1 were characterized by making stable cell lines site. Therefore, knowledge about the topology of the TM
in mouse NIH 3T3 cells. A difficulty in selecting for stable segments is crucial for understanding how P-gp functions.
cell lines expressing the mutant P-gp was in distinguishing To study the topology of P-gp, a Cys-less mutant of
between clones expressing human MDR1 from those which P-gp was constructed. Cysteine residues were
could arise through induction of endogenous mouse mdr1 or re-introduced into the Cys-less P-gp containing an A52
mdr3. Accordingly, the human MDR1 cDNA (cloned from a epitope tag to create a series of single-Cys mutants that
human kidney cDNA library) (Loo and Clarke 1993b) was contained one cysteine in a predicted extracellular or cyto-
modified to encode for the epitope for monoclonal antibody plasmic loop. The mutants were transiently expressed in
A52 (derived from rabbit SERCA1 Ca2+-ATPase) at the HEK 293 cells and treated with membrane-permeant (bio-
COOH-terminal of the expressed protein. This modification tin-maleimide) and -impermeant (stilbenedisulfonate
did not alter its expression or drug transport activity (Loo maleimide) thiol-specific reagents. The rationale was that
and Clarke 1993b). This also allowed detection of very low treatment with biotin maleimide would biotinylate any Cys
levels of P-gp-A52 in Western Blots because the cells nor- residue. Biotinylation was monitored by immunoprecipitating

© 1999 NRC Canada

I:\bcb\bcb77\01\O99-014.vp
Thursday, May 27, 1999 10:07:14 AM
Color profile: Disabled
Composite Default screen

Loo and Clarke 15

the labeled protein with A52 monoclonal antibody followed Mutational analysis to define amino acids important for
by blot analysis with streptavidin conjugated to a reporter mol- drug transport
ecule. Pre-treatment of the cells with membrane-impermeant Reconstitution studies with purified P-gp have shown that
stilbenedisulfonate maleimide was used to identify extracellular transport of hydrophobic drug substrates against a concen-
cysteines because they were unreactive when the cells were tration gradient is coupled to ATP hydrolysis (Sharom et al.
subsequently treated with biotin maleimide. The topology ob- 1993).
tained was consistent with the predicted model of P-gp, which Also, evidence from photoaffinity labeling studies with
predicts six TM segments in each of the two homologous drug analogues and from sequencing of P-gp from cells with
halves of the enzyme (Loo and Clarke 1995b). Subsequent altered drug resistance profiles indicate that the predicted
mutational analyses (Loo and Clarke 1996c) and cross-linking TM domains play a critical role in recognition and transport
experiments (Loo and Clarke 1996a) also suggest that each of substrates. Therefore, one goal was to identify amino ac-
half of P-gp is symmetrically arranged in the membrane. ids critical for binding of drug substrates and for transport of
Epitope insertion studies on the full-length protein have drugs out of the cell. Initially, the mutations were introduced
also confirmed the predicted topology (Kast et al. 1995, into the MDR1 cDNA and then transfected into NIH 3T3
1996). This topology, however, still remains controversial cells. The cells were then assayed for their ability to confer
since studies using truncated molecules showed that some resistance to various cytotoxic drugs. Since there was little
putative cytosolic or TM segments are located extra- evidence to suggest which residues in the TM domains
cellularly (Skach et al. 1993; Zhang et al. 1996). It was not might be good candidates for drug binding, we used a sys-
clear, however, whether these truncated molecules were tematic approach and tested the functional consequences of
functional. A resonance energy transfer study indicates that mutating each of the prolines (Loo and Clarke 1993b),
the nucleotide-binding domains are close (3.1–3.5 nm) to the phenylalanines (Loo and Clarke 1993a), or glycines (Loo
membrane surface (Liu and Sharom 1998). Recently, a very and Clarke 1994a) located in TM regions.
low resolution structure (25 D; 1 D = 0.1 nm) for P-gp was The potential role of the proline residues in the TM seg-
described that suggested the TM segments surrounded a ments was investigated because of their unique structural
large aqueous pore (Rosenberg et al. 1997). and functional properties. The structural destabilization in-
duced by prolines located in the middle of alpha-helices and
the possibility of cis-trans isomerization of peptide bonds
The hydrophilic domains of P-gp between prolines and their proceeding residues make
The hydrophilic domains of P-gp containing the consen- prolines important for specific functions related to
sus nucleotide-binding folds bind ATP (Azzaria et al. 1989). conformational changes in proteins (Brandl and Deber
It has been demonstrated that P-gp possesses high levels of 1986). The putative TM segments of many membrane trans-
drug-stimulatable ATPase activity (al-Shawi and Senior port proteins contain a significantly higher number of
1993; Ambudkar et al. 1992; Sarkadi et al. 1992; Shapiro proline residues than do the TM segments of non-
and Ling 1994; Sharom et al. 1993). Al-Shawi et al. (1994) transporting membrane proteins, suggesting some functional
showed that the ATPase activity of P-gp is inhibited by role of membrane-buried prolines in the transport reaction
N-ethylmaleimide (NEM). Maximal inhibition occurred with (Brandl and Deber 1986).
labeling at two sites, with equal distribution of the label be- In initial studies (Loo and Clarke 1993b), 13 proline resi-
tween the NH2- and COOH-terminal halves of the molecule. dues in MDR1 were mutated individually (residues 32, 66,
ATP prevented inhibition by NEM. Therefore, it was pre- 233, 350, 373, 693, 694, 709, 726, 745, 807, 866, and 996)
dicted that the critical cysteines were located in the (see Fig. 2). Five residues are located in putative TM helices
homology A nucleotide-binding consensus sequences (TM1, TM4, TM6, TM7, and TM10); one is located in the
(GNSGCGKS and GSSGCGKS, respectively) in the two nu- extracellular loop connecting segments TM7 and TM8 and
cleotide binding domains of Chinese hamster P-gp. seven are located within intracellular loops. Each of the
To test the contribution of either nucleotide-binding do- prolines was separately replaced with alanine since it is a
main to P-gp function, the Cys-less mutant was mutated to small neutral amino acid. Drug transport characteristics were
reintroduce a single cysteine back into each nucleo- assessed by their ability to confer drug resistance after
tide-binding consensus sequence. The sensitivity of the transfection into drug-sensitive NIH 3T3 cells. A dramatic
ATPase activity of each mutant after covalent modification shift in the drug resistance profile was observed for mutant
by NEM was then tested. It was found that covalent modifi- P223A. Cells expressing the mutant were preferentially re-
cation of a single cysteine residue within either nucleo- sistant to vinblastine with almost negligible resistance to
tide-binding consensus sequence (Cys431 and Cys1074, colchicine. A similar, but less dramatic, shift in the
respectively) with NEM inhibited drug-stimulated ATPase drug-resistance profile was observed in cells expressing mu-
activity of P-gp. In both cases, inactivation of ATPase activ- tant P866A. The cells were preferentially resistant to
ity by NEM was prevented by ATP. These results suggest vinblastine when compared to cells containing parental
that both nucleotide-binding domains are essential and that MDR1, but were relatively less resistant to colchicine,
they probably operate in a cooperative manner. adriamycin, and actinomycin D. Therefore, Pro223 and
The need for both ATP-binding sites to be functional was Pro866 appear to play critical roles in the transport of
also suggested by studies in which mutations introduced in colchicine by P-gp.
either nucleotide-binding fold inactivated P-gp (Azzaria et No drug-resistant colonies were obtained when cells were
al. 1989; Loo and Clarke 1995d). transfected with mutant P709A. Immunoblots of cells ex-

© 1999 NRC Canada

I:\bcb\bcb77\01\O99-014.vp
Thursday, May 27, 1999 10:07:16 AM
Color profile: Disabled
Composite Default screen

16 Biochem. Cell Biol. Vol. 77, 1999

Fig. 3. Structure of dibromobimane. Fig. 4. Distribution of dibromobimane-sensitive amino acids in

TM6 and TM12. The results of cysteine-scanning mutagenesis
and treatment of the Cys mutants with dBBn are shown. The
black bars show the percentage of verapamil-stimulated ATPase
activity remaining after treatment with dBBn. The TM segments
are projected as helical wheels (3.6 amino acids per turn). The
helices are oriented to reflect their predicted structure in a cell
membrane; they start on the cytoplasmic side of the membrane
and project into the membrane. Residues with an asterisk gave
unstable proteins when the residue was mutated to Cys.
Subsequent disulfide crosslinking experiments showed that the
two mutation-sensitive faces of TM6 and TM12 are close to
each other (Loo and Clarke 1997b).
pressing this mutant revealed that the majority of the ex-
pressed protein product corresponded in molecular weight to
an incompletely glycosylated form of the protein. Therefore,
a mutation to Pro709, which is predicted to lie at or close to
the intracellular end of TM7, affected the structural integrity
of the protein. Mutations to the other 10 prolines yielded
protein products that conferred drug resistance to transfected
cells. Drug-resistant clones exhibited similar resistance pro-
files to cells expressing wild-type human MDR1.
Each of the 31 phenylalanine residues in predicted TM re-
gions was then systematically mutated to alanine (Loo and
Clarke 1993a). Most of the mutants resembled wild-type en-
zyme except for F335A in TM6 and F978A in TM12. Both
of these mutants exhibited drastically altered drug resistance
profiles. It was interesting that these residues are found in
identical positions when the homologous halves of the pro-
tein are aligned.
The glycine residues in the cytoplasmic loop regions were
also analyzed (Loo and Clarke 1994a). Glycine residues
have the potential to play important roles in the transport re-
action. The small side group of glycine makes this amino
acid of unique importance in the mediation of inter-domain
and protein–ligand contacts as well as in conferring the flex-
ibility to the peptide backbone that is required for
conformational changes. The lack of a side-chain on glycine
permits a relatively wide range of flexibility. This property
also allows bends to occur at glycine positions and may pro-
vide a hinge to impart flexibility to the polypeptide chain.
Another reason for choosing glycine was because a sponta-
neous Gly to Val mutation, G185V, had been shown to con-
fer an altered drug resistance profile when compared to
wild-type enzyme (Choi et al. 1988). Therefore, the effects
of introducing Gly to Val mutations at 20 other positions
were tested (Loo and Clarke 1994a). Four of the mutations
inhibited maturation while five other mutants exhibited large
changes in the drug resistance profile conferred by P-gp.
These results showed that the glycines in the cytoplasmic
loops play important roles in structure and function of P-gp.
Analysis of the mutations to prolines, phenylalanines, and
glycines was informative, but time consuming. This was be-
cause of the need to establish stable cell lines for each mu-
tant. In addition, one could never be certain that endogenous
drug transporters were expressed when the cells were sub-
jected to drug selection or that other cellular factors were in-
fluencing the drug resistance profiles. These potential tandem histidine residues at the COOH end of the molecule
problems required the development of a more reliable sys- were developed (Loo and Clarke 1995d). The mutants could
tem for assessing the mutants. Accordingly, a rapid expres- now be isolated by nickel chelate chromatography and as-
sion and purification method using P-gp containing 10 sayed for drug-stimulated ATPase activity. These types of

© 1999 NRC Canada

I:\bcb\bcb77\01\O99-014.vp
Thursday, May 27, 1999 10:07:19 AM
Color profile: Disabled
Composite Default screen

Loo and Clarke 17

Fig. 5. Proposed model for the drug rescue of P-gp processing recognized as defective and the core-glycosylated intermediate is
mutants. P-gp is first synthesized in the endoplasmic reticulum rapidly degraded. (C) Synthesis of the processing mutant in the
as a core-glycosylated intermediate. (A) The carbohydrates are presence of drug substrates (Rescue), however, induces proper
modified in the Golgi and the protein is the delivered to the cell folding so that the mutant matures into a functional enzyme and
surface (Normal). (B) Processing mutants (Defective) are is trafficked to the plasma membrane.

results are valid since it was recently shown that the turn-
over numbers for drug transport and for drug-stimulated
ATP hydrolysis are comparable (Ambudkar et al. 1997).
Therefore purification of histidine-tagged P-gp, together
with alanine-scanning mutagenesis, could be used to exam-
ine the importance of most of the residues located within the
predicted TM segments.
There may be potential problems, however, in using
alanine-scanning mutagenesis to study drug-protein interac-
tions. One problem is that drug-protein interactions likely in-
volve a large number of residues so that a single change may
not have a measurable effect. If a change in substrate speci-
ficity is observed, it is then difficult to judge whether the
change is due to local or global structural changes. Another
problem is the difficulty in photolabeling P-gp with radioac-
tive analogs of drug substrates. Usually, the concentration of
radioactive analog required to achieve stoichiometric label-
ing of P-gp makes it economically unfeasible. To overcome
these problems, a thiol-reactive compound, dibromobimane
(dBBn), was identified that was a potent stimulator of the
ATPase activity of Cys-less P-gp (Fig. 3). Cysteine-scanning
mutagenesis was combined with dBBn modification of the
mutant P-gp for studying the contribution of TM6 and TM12
residues towards coupling of drug binding and ATPase activ-
ity (Loo and Clarke 1997c). The rationale was that a
thiol-reactive substrate would enter the drug-binding site of
P-gp, covalently bind to a nearby cysteine residue, and in-
hibit drug-stimulated ATPase activity. Dibromobimane is a
particularly useful compound because it is a potent
stimulator of ATP hydrolysis and both its reactivity and abil-
ity to act as a substrate could be quenched with cysteine.
The results suggested that the interface between TM6 and
TM12 forms part of the potential drug-binding pocket in
P-gp (Fig. 4). Indeed, recent cross-linking experiments show
that these TM segments are close to each other and undergo
conformational changes during the reaction cycle (Loo and
Clarke 1996a; Loo and Clarke 1997b).

Folding and maturation of P-gp

P-gp is first synthesized in the endoplasmic reticulum as a
core-glycosylated intermediate with a molecular mass of
about 150 kDa. The carbohydrates are subsequently modi-
fied in the Golgi to yield a protein of about 170 kDa that is
subsequently delivered to the cell surface (Fig. 5). During
mutational studies, we found that about 10% of the point
mutations affected processing of P-gp. These mutants are re-
tained in the endoplasmic reticulum as core-glycosylated in-
termediates in association with the molecular chaperones
calnexin (Loo and Clarke 1994b) and Hsc70 (Loo and
Clarke 1995c) and are rapidly degraded. Recently, we found
that the presence of drug substrates or modulators (Loo and
Clarke 1997a) during biosynthesis prevented abnormal pro-
tein folding and trafficking (protein kinesis) of the
misprocessed mutants. Remarkably, the presence of drug

© 1999 NRC Canada

I:\bcb\bcb77\01\O99-014.vp
Thursday, May 27, 1999 10:07:21 AM
Color profile: Disabled
Composite Default screen

18 Biochem. Cell Biol. Vol. 77, 1999

Fig. 6. Effect of drug substrates on interaction between the Fig. 7. Drug substrates induce superfolding of the TM domains.
halves of a P-gp processing mutant expressed as separate The inset shows the four domain of P-gp. TMD1 contains TM
polypeptides. The diagram shows the P-gp half-molecule segments 1 to 6 (residues 1–379), and TMD2 contains TM
polypeptides that would be recovered by nickel-chelate segments 7 to 12 (residues 681–1025). To determine the domains
chromatography after coexpression of a histidine-tagged N-half that are influenced by drug substrates during biosynthesis, each
protein with a processing mutation (X) and an epitope-tagged domain was expressed as a separate polypeptide
(A52) C-half polypeptide. Only the N-half polypeptide is (quarter-molecules), and then tested for sensitivity to digestion
recovered on a nickel column when coexpressed without drug by trypsin. When TMD1 and TMD2 are expressed without drug
substrate (No Drug). In the presence of drug substrate, the two substrates, TMD2 remained sensitive to trypsin (No Drug). Upon
halves appear to adopt an interactive structure and the C-half coexpression in the presence of drug substrates however, the two
polypeptide copurifies with the N-half polypeptide (+ Drug polypeptides interact and TMD2 adopts a trypsin-resistant
Substrate). conformation (+ Drug Substrate).

are critical for function. These results also suggested that

P-gp activity must involve interaction of all four domains.
This was confirmed by expressing each domain as a separate
polypeptide and testing for associations using
coimmunoprecipitation assays (Loo and Clarke 1995c). Each
domain was also tested for interaction with the chaperones,
calnexin and Hsc70. It was found that calnexin associated
substrates during biosynthesis rescued nearly all processing with the TM domains, whereas Hsc70 associated only with
mutants including those with mutations in TM segments, the cytoplasmic domains.
intracellular or intracellular loops, the linker region and ei- It was possible that processing mutations interfere in fold-
ther nucleotide-binding domain. A possible explanation is ing of P-gp by preventing interactions between the various
that the hydrophobic drug substrates diffuse into the domains and that these deficiencies can be corrected by the
endoplasmic reticulum and induce correct folding by occu- presence of drug substrates. This was found to be the case
pying the drug-binding site (Fig. 5). This observation has for interactions between the homologous halves of P-gp ex-
very important implications for the potential treatment of pressed as separate polypeptides. Interactions between the
other diseases associated with defective protein kinesis, such half-molecules can be monitored by attaching a histidine tag
as cystic fibrosis and Alzheimer’s disease. to one half-molecule and an epitope tag to the other. After
coexpression, the epitope tagged half-molecule will copurify
Interaction of large domains with the histidine-tagged half-molecule using nickel-chelate
An important question was whether P-gp functions as a chromatography. It was found that such interactions are in-
monomer or oligomer, because the drug-binding site formed hibited by the presence of processing mutations. Interactions
by a monomer would be significantly different from that between half-molecules with processing mutations however,
formed by an oligomer. Although P-gp has been shown to be can be restored if the molecules are coexpressed in the pres-
present in the membranes as monomers and oligomers, it ence of drug substrates (Fig. 6). These experiments suggest
was not known if the monomers were functional that the drug-binding site requires contributions from both
(Poruchynsky and Ling 1994). The results from halves of P-gp and that the presence of drug substrates dur-
immunoprecipitation and nickel-chelate chromatography ing biosynthesis helps to pull the two halves together.
studies, however, showed that the minimum functional unit The TM domain alone was recently shown to be sufficient
was a monomer (Loo and Clarke 1996b). for drug binding in experiments using quarter-molecule
Drug-stimulated ATPase activity requires the interaction polypeptides. The N- and C-terminal hydrophobic domain
of both halves of the molecule (Loo and Clarke 1994c). In quarter-molecule polypeptides, TMD1 (containing TM seg-
addition, results obtained from mutational analyses (Loo and ments 1 to 6) and TMD2 (containing TM segments 7 to 12),
Clarke 1995d) and protein modification studies (Loo and are normally rapidly degraded when expressed in cells.
Clarke 1995a) indicated that both nucleotide-binding sites When expressed together in the presence of a drug substrate,

© 1999 NRC Canada

I:\bcb\bcb77\01\O99-014.vp
Thursday, May 27, 1999 10:07:23 AM
Color profile: Disabled
Composite Default screen

Loo and Clarke 19

Fig. 8. Proposed working model for P-gp-mediated drug efflux. tide-binding domains required? Is ATP hydrolyzed at alter-
The hydrophobic substrate enters the lipid bilayer and interacts nating sites? Are there multiple drug binding sites or a sin-
with residues in the TM domain that form the drug-binding gle site for accommodating all substrates? How are
domain (oval). Recent experiments with the thiol-reactive substrates extracted from the membrane? Is P-gp a flippase
substrate dBBn suggests that TM6 and TM12 are close to the that moves substrates from one leaflet in the lipid bilayer to
drug-binding domain. Upon ATP hydrolysis at alternating sites, the other? Future studies will answer these questions and
there is a conformational change relayed to the TM domain such help us to understand the mechanism of P-gp mediated drug
as TM6 and TM12 that reduces the affinity for the substrate and transport.
leads to drug efflux.

Acknowledgments
This research was supported by a grant as part of a group
grant from the Medical Research Council of Canada and by
a grant from the Canadian Cystic Fibrosis Foundation. We
thank Merck Frosst Canada for supporting the Merck Frosst
Prize awarded by the Canadian Society for Biochemistry
and Molecular & Cellular Biology. We thank all investiga-
tors for contributing to the understanding of P-glycoprotein.

References
al-Shawi, M. K., and Senior, A. E. 1993. Characterization of the
adenosine triphosphatase activity of Chinese hamster
P-glycoprotein. J. Biol. Chem. 268: 4197–4206.
al-Shawi, M. K., Urbatsch, I. L., and Senior, A. E. 1994. Covalent
however, the two polypeptides showed markedly increased inhibitors of P-glycoprotein ATPase activity. J. Biol. Chem. 269:
resistance to protease digestion. Drug substrates had no ef- 8986–8992.
fect if TMD1 or TMD2 were expressed alone. These results Ambudkar, S. V., Lelong, I. H., Zhang, J., Cardarelli, C. O.,
indicate that the nucleotide-binding domains are not required Gottesman, M. M., and Pastan, I. 1992. Partial purification and
for drug binding and that both the N- and C-terminal TM reconstitution of the human multidrug- resistance pump: charac-
domains are required to form the drug-binding site. It also terization of the drug-stimulatable ATP hydrolysis. Proc. Natl.
suggests that drug substrates induce superfolding of the TM Acad. Sci. U.S.A. 89: 8472–8476.
domains so that they adopt a more compact structure Ambudkar, S. V., Cardarelli, C. O., Pashinsky, I., and Stein, W. D.
(Fig. 7). 1997. Relation between the turnover number for vinblastine
transport and for vinblastine-stimulated ATP hydrolysis by hu-
Model of the mechanism of drug transport man P-glycoprotein. J. Biol. Chem. 272: 21 160 – 21 166.
Azzaria, M., Schurr, E., and Gros, P. 1989. Discrete mutations in-
The picture that has emerged is that drug transport by troduced in the predicted nucleotide-binding sites of the mdr1
P-gp requires a co-ordinated effort between the two nucleo- gene abolish its ability to confer multidrug resistance. Mol. Cell
tide-binding domains and the two TM domains. Inactivation Biol. 9: 5289–5297.
of one nucleotide-binding domain by mutagenesis or chemi- Baas, F., and Borst, P. 1988. The tissue dependent expression of
cal modification inhibits ATP hydrolysis at the other nucleo- hamster P-glycoprotein genes. FEBS Lett. 229: 329–332.
tide-binding domain. This co-ordinated interaction between Bell, D. R., Trent, J. M., Willard, H. F., Riordan, J. R., and Ling,
the two nucleotide-binding domains becomes uncoupled V. 1987. Chromosomal location of human P-glycoprotein gene
when they are expressed as separate half-molecule sequences. Cancer Genet. Cytogenet. 25: 141–148.
polypeptides. Each half-molecule will exhibit ATPase activ- Biedler, J. L., and Riehm, H. 1970. Cellular resistance to
ity when expressed alone. Drug substrates however, do not actinomycin D in Chinese hamster cells in vitro:
stimulate the ATPase activity of half-molecule polypeptides. cross-resistance, radioautographic, and cytogenetic studies. Can-
cer Res. 30: 1174–1184.
It is possible that binding of drug substrates by P-gp requires
Boesch, D., Muller, K., Pourtier-Manzanedo, A., and Loor, F.
the presence of both TM domains. It appears that ligands for
1991. Restoration of daunomycin retention in
drug binding are contributed from residues in TM segments multidrug-resistant P388 cells by submicromolar concentrations
from both halves of the molecule. TM segments 6 and 12 of SDZ PSC 833, a nonimmunosuppressive cyclosporin deriva-
seem to be particularly important since there is evidence that tive. Exp. Cell Res. 196: 26–32.
they are important for drug binding and undergo Brandl, C. J., and Deber, C. M. 1986. Hypothesis about the func-
conformational changes during the reaction cycle. These TM tion of membrane-buried proline residues in transport proteins.
segments seem to be ideally located for coupling drug efflux Proc. Natl. Acad. Sci. U.S.A. 83: 917–921.
to ATP hydrolysis since each directly links a TM domain to Bruggemann, E. P., Germann, U. A., Gottesman, M. M., and
a nucleotide-binding site. Accordingly, we have proposed a Pastan, I. 1989. Two different regions of P-glycoprotein [cor-
working model for drug transport (Fig. 8). rected] are photoaffinity- labeled by azidopine [published erra-
There are many questions to be addressed about the struc- tum appears in J. Biol. Chem. 1990, 265(7): 4172]. J. Biol.
ture and mechanism of P-gp. Why are two nucleo- Chem. 264: 15 483 – 15 488.

© 1999 NRC Canada

I:\bcb\bcb77\01\O99-014.vp
Thursday, May 27, 1999 10:07:26 AM
Color profile: Disabled
Composite Default screen

20 Biochem. Cell Biol. Vol. 77, 1999

Bruggemann, E. P., Currier, S. J., Gottesman, M. M., and Pastan, I. and chloride channel functions of the human multidrug resis-
1992. Characterization of the azidopine and vinblastine binding tance P-glycoprotein. Cell, 71: 23–32.
site of P- glycoprotein. J. Biol. Chem. 267: 21 020 – 21 026. Goldstein, L. J. 1996. MDR1 gene expression in solid tumours.
Chan, H. S., DeBoer, G., Haddad, G., Gallie, B. L., and Ling, V. Eur. J. Cancer, 6: 1039–1050.
1995. Multidrug resistance in pediatric malignancies. Hematol. Gottesman, M. M., Pastan, I., and Ambudkar, S. V. 1996.
Oncol. Clin. North Am. 9: 275–318. P-glycoprotein and multidrug resistance. Curr. Opin. Genet.
Chan, H. S., Grogan, T. M., DeBoer, G., Haddad, G., Gallie, B. L., Dev. 6: 610–617.
and Ling, V. 1996. Diagnosis and reversal of multidrug resis- Greenberger, L. M. 1993. Major photoaffinity drug labeling sites
tance in paediatric cancers. Eur. J. Cancer, 6: 1051–1061. for iodoaryl azidoprazosin in P- glycoprotein are within, or im-
Charuk, J. H., and Reithmeier, R. A. 1992. Interaction of mediately C-terminal to, transmembrane domains 6 and 12. J.
P-glycoprotein with a hydrophobic component of rat urine. Biol. Chem. 268: 11 417 – 11 425.
Biochem. Biophys. Res. Commun. 186: 796–802. Greenberger, L. M., Yang, C. P., Gindin, E., and Horwitz, S. B.
Charuk, J. H., Loo, T. W., Clarke, D. M., and Reithmeier, R. A. 1990. Photoaffinity probes for the alpha 1-adrenergic receptor
1994. Interaction of rat kidney P-glycoprotein with a urinary and the calcium channel bind to a common domain in
component and various drugs including cyclosporin A. Am. J. P-glycoprotein. J. Biol. Chem. 265: 4394–4401.
Physiol. 266: F66–F75. Gros, P., Ben Neriah, Y. B., Croop, J. M., and Housman, D. E.
Charuk, J. H., Grey, A. A., and Reithmeier, R. A. 1998. Identifica- 1986a. Isolation and expression of a complementary DNA that
tion of the synthetic surfactant nonylphenol ethoxylate: a P- confers multidrug resistance. Nature (London), 323: 728–731.
glycoprotein substrate in human urine. Am. J. Physiol. 274: Gros, P., Croop, J., and Housman, D. 1986b. Mammalian
F1127–F1139. multidrug resistance gene: complete cDNA sequence indicates
Chen, C. J., Chin, J. E., Ueda, K., Clark, D. P., Pastan, I., strong homology to bacterial transport proteins. Cell, 47:
Gottesman, M. M., and Roninson, I. B. 1986. Internal duplica- 371–380.
tion and homology with bacterial transport proteins in the mdr1 Gruber, A., Peterson, C., and Reizenstein, P. 1988. D-verapamil
(P-glycoprotein) gene from multidrug-resistant human cells. and L-verapamil are equally effective in increasing vincristine
Cell, 47: 381–389. accumulation in leukemic cells in vitro. Int. J. Cancer, 41:
Chen, C. J., Clark, D., Ueda, K., Pastan, I., Gottesman, M. M., and 224–226.
Roninson, I. B. 1990. Genomic organization of the human Higgins, C. F. 1995. The ABC of channel regulation. Cell, 82:
multidrug resistance (MDR1) gene and origin of 693–696.
P-glycoproteins. J. Biol. Chem. 265: 506–514. Hollstein, M., Sidransky, D., Vogelstein, B., and Harris, C. C.
Choi, K. H., Chen, C. J., Kriegler, M., and Roninson, I. B. 1988. 1991. p53 mutations in human cancers. Science (Washington,
An altered pattern of cross-resistance in multidrug-resistant hu- D.C.), 253: 49–53.
man cells results from spontaneous mutations in the mdr1 Homolya, L., Hollo, Z., Germann, U. A., Pastan, I., Gottesman, M.
(P-glycoprotein) gene. Cell, 53: 519–529. M., and Sarkadi, B. 1993. Fluorescent cellular indicators are ex-
Cordon-Cardo, C., O’Brien, J. P., Casals, D., Rittman-Grauer, L., truded by the multidrug resistance protein. J. Biol. Chem. 268:
Biedler, J. L., Melamed, M. R., and Bertino, J. R. 1989. 21 493 – 21 496.
Multidrug-resistance gene (P-glycoprotein) is expressed by en- Juliano, R. L., and Ling, V. 1976. A surface glycoprotein modulat-
dothelial cells at blood-brain barrier sites. Proc. Natl. Acad. Sci. ing drug permeability in Chinese hamster ovary cell mutants.
U.S.A. 86: 695–698. Biochim. Biophys. Acta, 455: 152–162.
Croop, J. M., Raymond, M., Haber, D., Devault, A., Arceci, R. J., Kast, C., Canfield, V., Levenson, R., and Gros, P. 1995. Membrane
Gros, P., and Housman, D. E. 1989. The three mouse multidrug topology of P-glycoprotein as determined by epitope insertion:
resistance (mdr) genes are expressed in a tissue-specific manner transmembrane organization of the N-terminal domain of mdr3.
in normal mouse tissues. Mol. Cell Biol. 9: 1346–1350. Biochemistry, 34: 4402–4411.
de Bruijn, M. H., Van der Bliek, A. M., Biedler, J. L., and Borst, P. Kast, C., Canfield, V., Levenson, R., and Gros, P. 1996.
1986. Differential amplification and disproportionate expression Transmembrane organization of mouse P-glycoprotein deter-
of five genes in three multidrug-resistant Chinese hamster lung mined by epitope insertion and immunofluorescence. J. Biol.
cell lines. Mol. Cell Biol. 6: 4717–4722. Chem. 271: 9240–9248.
Dong, Y., Chen, G., Duran, G. E., Kouyama, K., Chao, A. C., Kim, R. B., Fromm, M. F., Wandel, C., Leake, B., Wood, A. J.,
Sikic, B. I., Gollapudi, S. V., Gupta, S., and Gardner, P. 1994. Roden, D. M., and Wilkinson, G. R. 1998. The drug transporter
Volume-activated chloride current is not related to P-glycoprotein limits oral absorption and brain entry of HIV-1
P-glycoprotein overexpression. Cancer Res. 54: 5029–5032. protease inhibitors. J. Clin. Invest. 101: 289–294.
Fisher, G. A., Lum, B. L., Hausdorff, J., and Sikic, B. I. 1996. Lee, C. G., and Gottesman, M. M. 1998. HIV-1 protease inhibitors
Pharmacological considerations in the modulation of multidrug and the MDR1 multidrug transporter [editorial]. J. Clin. Invest.
resistance. Eur. J. Cancer, 6: 1082–1088. 101: 287–288.
Fojo, A. T., Shen, D. W., Mickley, L. A., Pastan, I., and Lee, C. G., Gottesman, M. M., Cardarelli, C. O., Ramachandra,
Gottesman, M. M. 1987. Intrinsic drug resistance in human kid- M., Jeang, K. T., Ambudkar, S. V., Pastan, I., and Dey, S. 1998.
ney cancer is associated with expression of a human HIV-1 protease inhibitors are substrates for the MDR1
multidrug-resistance gene. J. Clin. Oncol. 5: 1922–1927. multidrug transporter. Biochemistry, 37: 3594–3601.
Gerlach, J. H., Endicott, J. A., Juranka, P. F., Henderson, G., Lehnert, M. 1996. Clinical multidrug resistance in cancer: a
Sarangi, F., Deuchars, K. L., and Ling, V. 1986. Homology be- multifactorial problem. Eur. J. Cancer, 6: 912–920.
tween P-glycoprotein and a bacterial haemolysin transport pro- Leighton, J., Jr., and Goldstein, L. J. 1995. P-glycoprotein in adult
tein suggests a model for multidrug resistance. Nature (London), solid tumors. Expression and prognostic significance. Hematol.
324: 485–489. Oncol. Clin. North Am. 9: 251–273.
Gill, D. R., Hyde, S. C., Higgins, C. F., Valverde, M. A., Mintenig, Levine, A. J., Momand, J., and Finlay, C. A. 1991. The p53 tumour
G. M., and Sepulveda, F. V. 1992. Separation of drug transport suppressor gene. Nature (London), 351: 453–456.

© 1999 NRC Canada

I:\bcb\bcb77\01\O99-014.vp
Thursday, May 27, 1999 10:07:28 AM
Color profile: Disabled
Composite Default screen

Loo and Clarke 21

Liu, R., and Sharom, F. J. 1998. Proximity of the nucleotide bind- Morris, D. I., Greenberger, L. M., Bruggemann, E. P., Cardarelli,
ing domains of the P-glycoprotein multidrug transporter to the C., Gottesman, M. M., Pastan, I., and Seamon, K. B. 1994. Lo-
membrane surface: a resonance energy transfer study. Biochem- calization of the forskolin labeling sites to both halves of P-
istry, 37: 6503–6512. glycoprotein: similarity of the sites labeled by forskolin and
Loo, T. W., and Clarke, D. M. 1993a. Functional consequences of prazosin. Mol. Pharmacol. 46: 329–337.
phenylalanine mutations in the predicted transmembrane domain Ng, W. F., Sarangi, F., Zastawny, R. L., Veinot-Drebot, L., and
of P-glycoprotein. J. Biol. Chem. 268: 19 965 – 19 972. Ling, V. 1989. Identification of members of the P-glycoprotein
Loo, T. W., and Clarke, D. M. 1993b. Functional consequences of multigene family. Mol. Cell Biol. 9: 1224–1232.
proline mutations in the predicted transmembrane domain of Poruchynsky, M. S., and Ling, V. 1994. Detection of oligomeric
P-glycoprotein. J. Biol. Chem. 268: 3143–3149. and monomeric forms of P-glycoprotein in multidrug resistant
Loo, T. W., and Clarke, D. M. 1994a. Functional consequences of cells. Biochemistry, 33: 4163–4174.
glycine mutations in the predicted cytoplasmic loops of Rasola, A., Galietta, L. J., Gruenert, D. C., and Romeo, G. 1994.
P-glycoprotein. J. Biol. Chem. 269: 7243–7248. Volume-sensitive chloride currents in four epithelial cell lines
Loo, T. W., and Clarke, D. M. 1994b. Prolonged association of are not directly correlated to the expression of the MDR-1 gene.
temperature-sensitive mutants of human P- glycoprotein with J. Biol. Chem. 269: 1432–1436.
calnexin during biogenesis. J. Biol. Chem. 269: 28 683 – Raviv, Y., Pollard, H. B., Bruggemann, E. P., Pastan, I., and
28 689. Gottesman, M. M. 1990. Photosensitized labeling of a func-
Loo, T. W., and Clarke, D. M. 1994c. Reconstitution of tional multidrug transporter in living drug-resistant tumor cells.
drug-stimulated ATPase activity following co- expression of J. Biol. Chem. 265: 3975–3980.
each half of human P-glycoprotein as separate polypeptides. J. Raymond, M., Rose, E., Housman, D. E., and Gros, P. 1990. Physi-
Biol. Chem. 269: 7750–7755. cal mapping, amplification, and overexpression of the mouse
Loo, T. W., and Clarke, D. M. 1995a. Covalent modification of hu- mdr gene family in multidrug-resistant cells. Mol. Cell Biol. 10:
man P-glycoprotein mutants containing a single cysteine in ei- 1642–1651.
ther nucleotide-binding fold abolishes drug- stimulated ATPase Riordan, J. R., and Ling, V. 1985. Genetic and biochemical charac-
activity. J. Biol. Chem. 270: 22 957 – 22 961. terization of multidrug resistance. Pharmacol. Ther. 28: 51–75.
Loo, T. W., and Clarke, D. M. 1995b. Membrane topology of a Roepe, P. D. 1995. The role of the MDR protein in altered drug
cysteine-less mutant of human P-glycoprotein. J. Biol. Chem. translocation across tumor cell membranes. Biochim. Biophys.
270: 843–848. Acta, 1241: 385–405.
Loo, T. W., and Clarke, D. M. 1995c. P-glycoprotein. Associations Roepe, P. D., Wei, L. Y., Hoffman, M. M., and Fritz, F. 1996. Al-
between domains and between domains and molecular tered drug translocation mediated by the MDR protein: direct,
chaperones. J. Biol. Chem. 270: 21 839 – 21 844. indirect, or both? J. Bioenerg. Biomembr. 28: 541–555.
Loo, T. W., and Clarke, D. M. 1995d. Rapid purification of human Rosenberg, M. F., Callaghan, R., Ford, R. C., and Higgins, C. F.
P-glycoprotein mutants expressed transiently in HEK 293 cells 1997. Structure of the multidrug resistance P-glycoprotein to
by nickel-chelate chromatography and characterization of their 2.5 nm resolution determined by electron microscopy and image
drug-stimulated ATPase activities. J. Biol. Chem. 270: 21 449 – analysis. J. Biol. Chem. 272: 10 685 – 10 694.
21 452. Ruetz, S., and Gros, P. 1994. Phosphatidylcholine translocase: a
Loo, T. W., and Clarke, D. M. 1996a. Inhibition of oxidative physiological role for the mdr2 gene. Cell, 77: 1071–1081.
cross-linking between engineered cysteine residues at positions Sarkadi, B., and Muller, M. 1997. Search for specific inhibitors of
332 in predicted transmembrane segments (TM) 6 and 975 in multidrug resistance in cancer. Semin. Cancer Biol. 8: 171–182.
predicted TM12 of human P-glycoprotein by drug substrates. J. Sarkadi, B., Price, E. M., Boucher, R. C., Germann, U. A., and
Biol. Chem. 271: 27 482 – 27 487. Scarborough, G. A. 1992. Expression of the human multidrug
Loo, T. W., and Clarke, D. M. 1996b. The minimum functional resistance cDNA in insect cells generates a high activity
unit of human P-glycoprotein appears to be a monomer. J. Biol. drug-stimulated membrane ATPase. J. Biol. Chem. 267:
Chem. 271: 27 488 – 27 492. 4854–4858.
Loo, T. W., and Clarke, D. M. 1996c. Mutational analysis of the Schinkel, A. H. 1997. The physiological function of
predicted first transmembrane segment of each homologous half drug-transporting P-glycoproteins. Semin. Cancer Biol. 8:
of human P-glycoprotein suggests that they are symmetrically 161–170.
arranged in the membrane. J. Biol. Chem. 271: 15 414 – 15 419. Schinkel, A. H., Smit, J. J., van Tellingen, O., Beijnen, J. H.,
Loo, T. W., and Clarke, D. M. 1997a. Correction of defective pro- Wagenaar, E., van Deemter, L., Mol, C. A., van der Valk, M. A.,
tein kinesis of human P-glycoprotein mutants by substrates and Robanus-Maandag, E. C., te Riele, H. P., and et al. 1994. Dis-
modulators. J. Biol. Chem. 272: 709–712. ruption of the mouse mdr1a P-glycoprotein gene leads to a defi-
Loo, T. W., and Clarke, D. M. 1997b. Drug-stimulated ATPase ac- ciency in the blood-brain barrier and to increased sensitivity to
tivity of human P-glycoprotein requires movement between drugs. Cell, 77: 491–502.
transmembrane segments 6 and 12. J. Biol. Chem. 272: 20 986 – Schuetz, E. G., Schinkel, A. H., Relling, M. V., and Schuetz, J. D.
20 989. 1996. P-glycoprotein: a major determinant of
Loo, T. W., and Clarke, D. M. 1997c. Identification of residues in rifampicin-inducible expression of cytochrome P4503A in mice
the drug-binding site of human P- glycoprotein using a and humans. Proc. Natl. Acad. Sci. U.S.A. 93: 4001–4005.
thiol-reactive substrate. J. Biol. Chem. 272: 31 945 – 31 948. Shapiro, A. B., and Ling, V. 1994. ATPase activity of purified and
Marie, J. P. 1995. P-glycoprotein in adult hematologic malignan- reconstituted P-glycoprotein from Chinese hamster ovary cells.
cies. Hematol. Oncol. Clin. North Am. 9: 239–249. J. Biol. Chem. 269: 3745–3754.
Morin, X. K., Bond, T. D., Loo, T. W., Clarke, D. M., and Bear, C. Sharom, F. J. 1997. The P-glycoprotein efflux pump: how does it
E. 1995. Failure of P-glycoprotein (MDR1) expressed in transport drugs? J. Membr. Biol. 160: 161–175.
Xenopus oocytes to produce swelling-activated chloride channel Sharom, F. J., Yu, X., and Doige, C. A. 1993. Functional reconsti-
activity. J. Physiol. 486: 707–714. tution of drug transport and ATPase activity in proteoliposomes

© 1999 NRC Canada

I:\bcb\bcb77\01\O99-014.vp
Thursday, May 27, 1999 10:07:31 AM
Color profile: Disabled
Composite Default screen

22 Biochem. Cell Biol. Vol. 77, 1999

containing partially purified P-glycoprotein. J. Biol. Chem. 268: Tsuruo, T., Iida, H., Tsukagoshi, S., and Sakurai, Y. 1981. Over-
24 197 – 24 202. coming of vincristine resistance in P388 leukemia in vivo and in
Shen, D. W., Cardarelli, C., Hwang, J., Cornwell, M., Richert, N., vitro through enhanced cytotoxicity of vincristine and
Ishii, S., Pastan, I., and Gottesman, M. M. 1986. Multiple vinblastine by verapamil. Cancer Res. 41: 1967–1972.
drug-resistant human KB carcinoma cells independently se- Tsuruo, T., Iida, H., Nojiri, M., Tsukagoshi, S., and Sakurai, Y.
lected for high-level resistance to colchicine, adriamycin, or 1983. Circumvention of vincristine and Adriamycin resistance
vinblastine show changes in expression of specific proteins. J. in vitro and in vivo by calcium influx blockers. Cancer Res. 43:
Biol. Chem. 261: 7762–7770. 2905–2910.
Skach, W. R., Calayag, M. C., and Lingappa, V. R. 1993. Evidence Ueda, K., Cardarelli, C., Gottesman, M. M., and Pastan, I. 1987.
for an alternate model of human P-glycoprotein structure and Expression of a full-length cDNA for the human “MDR1" gene
biogenesis. J. Biol. Chem. 268: 6903–6908. confers resistance to colchicine, doxorubicin, and vinblastine.
Smit, J. J., Schinkel, A. H., Oude Elferink, R. P., Groen, A. K., Proc. Natl. Acad. Sci. U.S.A. 84: 3004–3008.
Wagenaar, E., van Deemter, L., Mol, C. A., Ottenhoff, R., van Valverde, M. A., Diaz, M., Sepulveda, F. V., Gill, D. R., Hyde, S.
der Lugt, N. M., van Roon, M. A., and et al. 1993. Homozygous C., and Higgins, C. F. 1992. Volume-regulated chloride channels
disruption of the murine mdr2 P-glycoprotein gene leads to a associated with the human multidrug- resistance P-glycoprotein.
complete absence of phospholipid from bile and to liver disease. Nature (London), 355: 830–833.
Cell, 75: 451–462. Valverde, M. A., Bond, T. D., Hardy, S. P., Taylor, J. C., Higgins,
Smith, A. J., Timmermans-Hereijgers, J. L., Roelofsen, B., Wirtz, C. F., Altamirano, J., and Alvarez-Leefmans, F. J. 1996. The
K. W., van Blitterswijk, W. J., Smit, J. J., Schinkel, A. H., and multidrug resistance P-glycoprotein modulates cell regulatory
Borst, P. 1994. The human MDR3 P-glycoprotein promotes volume decrease. EMBO J. 15: 4460–4468.
translocation of phosphatidylcholine through the plasma mem- Van der Bliek, A. M., Baas, F., Ten Houte de Lange, T., Kooiman,
brane of fibroblasts from transgenic mice. FEBS Lett. 354: P. M., Van der Velde-Koerts, T., and Borst, P. 1987. The human
263–266. mdr3 gene encodes a novel P-glycoprotein homologue and gives
Tang-Wai, D. F., Kajiji, S., DiCapua, F., de Graaf, D., Roninson, I. rise to alternatively spliced mRNAs in liver. EMBO J. 6:
B., and Gros, P. 1995. Human (MDR1) and mouse (mdr1, mdr3) 3325–3331.
P-glycoproteins can be distinguished by their respective drug re- Van der Bliek, A. M., Baas, F., Van der Velde-Koerts, T., Biedler,
sistance profiles and sensitivity to modulators. Biochemistry, 34: J. L., Meyers, M. B., Ozols, R. F., Hamilton, T. C., Joenje, H.,
32–39. and Borst, P. 1988. Genes amplified and overexpressed in hu-
Thiebaut, F., Tsuruo, T., Hamada, H., Gottesman, M. M., Pastan, I., man multidrug-resistant cell lines. Cancer Res. 48: 5927–5932.
and Willingham, M. C. 1987. Cellular localization of the Weinstein, R. S., Jakate, S. M., Dominguez, J. M., Lebovitz, M.
multidrug-resistance gene product P- glycoprotein in normal hu- D., Koukoulis, G. K., Kuszak, J. R., Klusens, L. F., Grogan, T.
man tissues. Proc. Natl. Acad. Sci. U.S.A. 84: 7735–7738. M., Saclarides, T. J., Roninson, I. B., and Coon, J.S. 1991. Rela-
Thottassery, J. V., Zambetti, G. P., Arimori, K., Schuetz, E. G., and tionship of the expression of the multidrug resistance gene prod-
Schuetz, J. D. 1997. p53-dependent regulation of MDR1 gene uct (P-glycoprotein) in human colon carcinoma to local tumor
expression causes selective resistance to chemotherapeutic aggressiveness and lymph node metastasis. Cancer Res. 51:
agents. Proc. Natl. Acad. Sci. U.S.A. 94: 11 037 – 11 042. 2720–2726.
Tominaga, M., Tominaga, T., Miwa, A., and Okada, Y. 1995. Vol- Zhang, M., Wang, G., Shapiro, A., and Zhang, J. T. 1996. Topolog-
ume-sensitive chloride channel activity does not depend on en- ical folding and proteolysis profile of P-glycoprotein in mem-
dogenous P-glycoprotein. J. Biol. Chem. 270: 27 887 – 27 893. branes of multidrug-resistant cells: implications for the drug-
Trezise, A. E., Romano, P. R., Gill, D. R., Hyde, S. C., Sepulveda, transport mechanism. Biochemistry, 35: 9728–9736.
F. V., Buchwald, M., and Higgins, C. F. 1992. The multidrug re-
sistance and cystic fibrosis genes have complementary patterns
of epithelial expression. EMBO J. 11: 4291–4303.

Notes about the author Notes sur l’auteur

This article is based on the Merck Frosst Award Lecture Cet article est basé sur la conférence d’acceptation du prix
of the Canadian Society of Biochemistry and Molecular & Merck Frosst de la Société canadienne de biochimie et de
Cellular Biology presented by Dr. David M. Clarke at the biologie moléculaire et cellulaire décerné à Dr David M.
41st Annual Meeting of the Canadian Federation of Biologi- Clarke lors du 41e congrès annuel de la Fédération
cal Societies in Edmonton, in June 1998. Dr. Clarke was canadienne des sociétés de biologie tenu à Edmonton, Al-
born in Windsor and raised in Sarnia, Ontario and obtained berta, en juin 1998. Dr Clarke est né à Windsor et a grandi à
his Bachelor’s degree at the University of Windsor. He ob- Sarnia en Ontario. Il a obtenu un B.Sc. de l’Université de
tained his M.Sc. on “Post-Translational Modification of Pep- Windsor. Puis, il a étudié « Les modifications
tides in the Brain” under the supervision of Dr. W.W.C. post-traductionnelles de peptides dans le cerveau » sous la
Chan at McMaster University. He completed his Ph.D. at the direction de Dr W.W.C. Chan à l’Université McMaster à
University of British Columbia in Vancouver under the su- Hamilton, Ontario, et a obtenu un M.Sc. Par la suite, il a
pervision of Dr. P.O. Bragg, working on “Pyridine Nucleo- obtenu un Ph.D. de l’Université de la Colombie-Britannique
tide Transhydrogenase of E. coli : Nucleotide Sequence of à Vancouver après avoir déterminé « La séquence
the pnt Gene and Characterization of the Enzyme Complex.” nucléotidique du gène pnt et caractérisé le complexe
Following his Ph.D., he did postdoctoral work at the enzymatique de la transhydrogénase des nucléotides
Univeristy of British Columbia with Dr. Shirley Gillam on pyridiniques de E. coli » sous la direction de Dr P.O. Bragg.
the “Cloning, Nucleotide Sequence and In vitro Expression Après avoir obtenu un Ph.D., Dr Clarke a fait un stage post-

© 1999 NRC Canada

I:\bcb\bcb77\01\O99-014.vp
Thursday, May 27, 1999 10:07:33 AM
Color profile: Disabled
Composite Default screen

Loo and Clarke 23

of the Rubella Virus Subgenomic RNA.” He also did post- doctoral avec Dr Shirley Gillam à l’Université de la
doctoral work at the University of Toronto with Dr. David Colombie-Britannique au cours duquel il a cloné, déterminé
H. MacLennan, on “The Mechanism of Ca2+ Transport by « La séquence nucléotidique et mesuré l’expression in vitro
the SERCAI Ca2+- ATPase of Muscle.” After completing his de l’ARN subgénomique du virus de la rubéole. » Il a
postdoctoral studies he went to the Department of Medicine, également effectué un stage postdoctoral dans le laboratoire
University of Toronto in 1990 as Assitant Professor and is de Dr David H. MacLennan à l’Université de Toronto pour
now an Associate Professor in the same department. étudier « Le mécanisme de transport du Ca2+ par la
Ca2+-ATPase SERCA1 du muscle. » En 1990, Dr Clarke a
joint le département de médecine de l’Université de Toronto
en tant que professeur adjoint. Il est maintenant professeur
agrégé dans le même département.

© 1999 NRC Canada

I:\bcb\bcb77\01\O99-014.vp
Thursday, May 27, 1999 10:07:34 AM