letters to nature

dynamics calculations to find the lowest energy conformation followed by vinblastine and magnesium-ATP were indistinguishable. Blockers
another energy minimization to provide the final structure. of P-glycoprotein-mediated multidrug resistance also inhibited
Received 23 June; accepted 8 October 1997. LmrA-dependent drug resistance. Kinetic analysis of drug dis-
1. Stemmer, W. P. C. DNA shuffling by random fragmentation and reassembly: in vitro recombination sociation from LmrA expressed in plasma membranes of insect
for molecular evolution. Proc. Natl Acad. Sci. USA 91, 10747–10751 (1994). cells revealed the presence of two allosterically linked drug-
2. Stemmer, W. P. C. Searching sequence space. Bio/Technology 13, 549–553 (1995).
3. Stemmer, W. P. C. Rapid evolution of a protein in vitro by DNA shuffling. Nature 370, 389–391 (1994).
binding sites indistinguishable from those of P-glycoprotein.
4. Zhang, J., Dawes, G. & Stemmer, W. P. C. Evolution of a fucosidase from a galactosidase by DNA These findings have implications for the reversal of antibiotic
shuffling and screening. Proc. Natl Acac. Sci. USA 94, 4504–4509 (1997).
5. Crameri, A., Whitehorn, E., Tate, E. & Stemmer, W. P. C. Improved green fluorescent protein by
molecular evolution using DNA shuffling. Nature Biotech. 14, 315–319 (1996).
6. Crameri, A., Dawes, G., Rodriguez, E., Silver, S. & Stemmer, W. P. C. Molecular Evolution of an
arsenate detoxification pathway by DNA shuffling. Nature Biotech. 15, 436–438 (1997).
resistance in pathogenic microorganisms. Taken together, they
demonstrate that bacterial LmrA and human P-glycoprotein are
functionally interchangeable and that this type of multidrug-
resistance efflux pump is conserved from bacteria to man.
8
7. Moore, J. C. & Arnold, F. H. Directed evolution of a para-nitrobenzyl esterase for aqueous-organic
solvents. Nature Biotech. 14, 458–467 (1996). Using the polymerase chain reaction (PCR), the bacterial lmrA
8. Zhao, H. & Arnold, F. H. Optimization of DNA shuffling for high fidelity recombination. Nucleic coding sequence3 was cloned into a pCI-neo mammalian expression
Acids Res. 25, 1307–1308 (1997).
9. Lindberg, F. & Normark, S. Sequence of the Citrobacter freundii OS60 chromosomal ampC b- vector under the control of the human cytomegalovirus immedi-
lactamase gene. Eur. J. Biochem. 156, 441–445 (1986). ate–early enhancer/promoter region. A Kozak sequence9 was intro-
10. Galleni, M. et al. Sequence and comparative analysis of three Enterobacter cloacae ampC b-lactamase
genes and their products. Biochem. J. 250, 753–760 (1988).
duced at the ATG initiation codon of lmrA to enhance translational
11. Leiza, M. G. et al. Gene sequence and biochemical characterization of FOX-1 from Klebsiella efficiency. For control experiments, a transport-inactive LmrA
pneumoniae, a new AmpC-type plasmid-mediated b-lactamase with two molecular variants. \ita- protein was generated in the same vector by introducing a lysine-
Antimicrob. Agents Chemother. 38, 2150–2157 (1994).
12. Seoane, A., Francia, M. V. & Garcia Lobo, J. M. Nucleotide sequence of the ampC-ampR region from to-methionine substitution at position 388 (K388M)10 in the Walker
the chromosome of Yersinia enterocolitica. Antimicrob. Agents Chemother. 36, 1049–1052 (1992). A motif of the nucleotide-binding domain of the protein by site-
13. Stemmer, W. P. C., Crameri, A., Ha, K. D., Brennan, T. M. & Heyneker, H. L. Single-step PCR assembly
of a gene and a whole plasmid from large numbers of oligonucleotides. Gene 164, 49–53 (1995).
directed mutagenesis. Hexa-histidine tags were also added to the
14. Lobkovsky, E. et al. Evolution of enzyme activity: crystallographic structure at 2 Å resolution of amino termini of both the wild-type and K388M forms of LmrA.
cephalosporinase from the ampC gene of Enterobacter cloacae P99 and comparison with a class A
penicillinase. Proc. Natl Acad. Sci. USA 90, 11257–11261 (1993).
15. Kauffman, S. The Origins of Order (Oxford University Press, Oxford, 1993).
16. Eigen, M. Steps Towards Life: a Perspective on Evolution (Oxford University Press, Oxford, 1992).

Acknowledgements. We thank G. Dawes, J. Kieft, S. DelCardayre and M. Tobin and R. Howard,

C. Yanofsky, P. Schultz, F. Arnold and A. Kornberg for useful comments on the manuscript.

Correspondence should be addressed to W.P.C.S. (e-mail: pim_stemmer@maxygen.com).

A bacterial antibiotic-
resistance gene that
complements the human
multidrug-resistance
P-glycoprotein gene
Hendrik W. van Veen*, Richard Callaghan†,
Loredana Soceneantu†, Alessandro Sardini‡,
Wil N. Konings* & Christopher F. Higgins†
* Department of Microbiology, Groningen Biomolecular Sciences and
Biotechnology Institute, University of Groningen, Kerklaan 30, 9751 NN Haren,
The Netherlands
† Imperial Cancer Research Fund Laboratories and Cancer Research Campaign
Drug Resistance Group, Nuffield Department of Clinical Biochemistry,
Institute of Molecular Medicine, John Radcliffe Hospital, University of Oxford,
Oxford OX3 9DS, UK
‡ Department of Physiology, King’s College London, The Strand,
London WC2R 2LS, UK
.........................................................................................................................

Bacteria have developed many fascinating antibiotic-resistance

mechanisms1,2. A protein in Lactococcus lactis, LmrA, mediates
antibiotic resistance by extruding amphiphilic compounds from
the inner leaflet of the cytoplasmic membrane3,4. Unlike other
known bacterial multidrug-resistance proteins, LmrA is an ATP-
binding cassette (ABC) transporter5. The human multidrug-resis- Figure 1 Expression of LmrA in GM0637 human lung fibroblast cells. a, Western
tance P-glycoprotein6, encoded by the MDR1 gene, is also an ABC blot of total cell protein (30 mg per lane) using the anti hexa-histidine tag antibody.
transporter, overexpression of which is one of the principal causes Control, wild type and mutant refer to mock-transfected cells, and cells
of resistance of human cancers to chemotherapy7,8. We expressed expressing the wild-type and K388M forms of LmrA, respectively. The migration
lmrA in human lung fibroblast cells. Surprisingly, LmrA was of molecular mass markers is indicated. Arrowhead indicates LmrA protein. b,
targeted to the plasma membrane and conferred typical multi- Distribution of wild-type LmrA expressed in fibroblast cells 48 h after transfection.
drug resistance on these human cells. The pharmacological The three frames show cross-sections through a cell, from top to bottom along
characteristics of LmrA and P-glycoprotein-expressing lung fibro- the z-axis. The fluorescence intensity is shown colour-coded on a scale from blue
blasts were very similar, and the affinities of both proteins for (low) to white (high). Scale bar, 25 mm.

Nature © Macmillan Publishers Ltd 1998

NATURE | VOL 391 | 15 JANUARY 1998 291
letters to nature
LmrA was transiently expressed in the GM0637 lung fibroblast in lung fibroblast cells: (1) calcium-channel blockers such as
cell line that had previously been used for molecular characteriza- verapamil and its analogue CP100-356; (2) 1,4-dihydropyridines
tion of P-glycoprotein11. Transfection of both the wild-type and such as nicardipine; (3) indolizine sulphones such as SR33557; (4)
mutant lmrA genes into the fibroblast cells resulted in the expression antimalarials such as quinine and quinidine; and (5) immunosup-
of equivalent amounts of a 66K polypeptide, which could be pressants such as cyclosporin A (Fig. 2c, d).
detected by using an anti hexa-histidine tag monoclonal antibody To demonstrate that drug extrusion across the plasma membrane
(Fig. 1a). Confocal fluorescence microscopy indicated that, as for is the underlying mechanism of drug resistance in LmrA-expressing
human P-glycoprotein7, LmrA was primarily localized in the plasma
membrane of the fibroblast cells (Fig. 1b). Transfected GM0637
lung fibroblast cells contained undetectable levels of P-
glycoprotein11 or the multidrug-resistance-associated protein
lung fibroblast cells, we did transport assays using both vinblastine
and Hoechst 33342 (Fig. 3a, b). The accumulation of these com-
pounds was significantly reduced in cells expressing wild-type
LmrA. LmrA-mediated drug efflux was inhibited by the Rauwolfia
8
MRP112 (data not shown). alkaloid reserpine and by the phenylalkylamine verapamil, two
Lung fibroblast cells expressing wild-type LmrA protein showed a classic inhibitors of drug transport by P-glycoprotein7. These
10–60-fold increased resistance to a variety of natural product drugs inhibitors did not affect the accumulation of vinblastine or Hoechst
and synthetic chemotherapeutic drugs which are typical substrates 33342 in cells expressing the K388M mutant form of LmrA (data
of P-glycoprotein7: (1) anthracyclines such as daunomycin and not shown).
doxorubicin; (2) vinca-alkaloids such as vinblastine and vincristine; To explore the pharmacology of wild-type LmrA and to compare
and (3) cytotoxic agents such as ethidium bromide, rhodamine 123 it with that of P-glycoprotein, larger quantities of both proteins were
and colchicine (Fig. 2a, b). Mock-transfected control cells, and cells expressed in recombinant baculovirus-infected Spodoptera frugi-
expressing equivalent amounts of the K388M mutant LmrA, perda insect cells13 (data not shown). LmrA and P-glycoprotein
showed no increase in resistance, demonstrating that LmrA itself displayed a basal ATPase activity which was stimulated up to three-
is responsible for the drug-resistance profile. fold by 50 mM verapamil and completely inhibited by 2 mM
P-glycoprotein function can be inhibited by a large group of vanadate, and which were comparably dependent on the concen-
structurally unrelated modulators7. All of these modulators also tration of magnesium-ATP (LmrA: half-maximal ATPase activity at
reversed drug resistance generated by expression of wild-type LmrA 0:24 6 0:04 mM Mg-ATP; P-glycoprotein: half-maximal ATPase

Figure 2 Drug-resistance phenotypes generated by LmrA expression in GM0637 refers to the half-maximal inhibitory concentration. c, Daunomycin resistance of cells
human lung fibroblast cells. a, Survival of cells expressing the wild-type (squares) expressing wild-type LmrA in the absence (A) and presence (L) of the P-glycoprotein
and K388M (circles) forms of LmrA, at increasing concentrations of ethidium modulator CP100 2 356 and the K388M form of LmrA in the absence (X) and
bromide. b, Drug-survival characteristics of mock-transfected cells (grey bars), and presence (O) of CP100 2 356. d, Effect of individual modulators on the daunomycin-
cells expressing wild-type (white bars) and K388M (black bars) forms of LmrA. IC50, resistance of cells expressing the wild-type (A) and K388M (B) forms of LmrA.

Nature © Macmillan Publishers Ltd 1998

292 NATURE | VOL 391 | 15 JANUARY 1998
 letters to nature
activity at 0:13 6 0:03 mM Mg-ATP) (Fig. 4a). To our knowledge, the bacterial antibiotic-resistance protein
Remarkably, LmrA and P-glycoprotein bound vinblastine with LmrA is the first prokaryotic membrane protein to be functionally
similar affinity (LmrA: K d ¼ 46 6 6 nM vinblastine; P-glycopro- expressed in mammalian cells. The correct sorting of LmrA to the
tein: K d ¼ 39 6 8 nM vinblastine) (Fig. 4b). Vinblastine binding to plasma membrane in mammalian cells is surprising in itself. Most
LmrA was inhibited by P-glycoprotein modulators, including the remarkably, the drug-resistance phenotype of LmrA- and P-glyco-
1,4-dihydropyridine nicardipine and the indolizine sulphone protein-expressing human lung fibroblast cells, and the range of
SR33557 (Fig. 4c), which reverse LmrA-mediated drug resistance reversing agents able to modulate LmrA and P-glycoprotein-depen-
in human lung fibroblast cells (Fig. 2d).
P-glycoprotein possesses 1,4-dihydropyridine- and indolizine
sulphone-selective drug-binding sites which are allosterically
coupled to a vinca-alkaloid-selective binding site14–20. The pharma-
dent drug efflux were similar. Thus, the bacterial LmrA protein and
the human multidrug-resistance P-glycoprotein are functionally
identical. It has been proposed that P-glycoprotein participates in
the protection of human cells against hydrophobic xenobiotics by
8
cological interactions between these modulators and LmrA were active excretion of these compounds from the membrane into bile,
examined by drug-dissociation kinetics15 (Fig. 4d). Nicardipine urine or the intestinal lumen, preventing their accumulation in
caused an acceleration of vinblastine dissociation from LmrA in a critical organs such as the brain21. Likewise, the principal physio-
similar fashion to that observed for P-glycoprotein. The verapamil logical role of LmrA and related multidrug transporters in bacteria
analogue CP100-356 did not affect the kinetics of vinblastine could involve the efflux of hydrophobic compounds that are
dissociation from wild-type LmrA. These results show that generated intracellularly by metabolism or which are encountered
CP100-356 is a competitive inhibitor of vinblastine binding, and in the extracellular environment. Alternatively, there may be a
can only gain access to the vinca-alkaloid-binding site of LmrA common endogenous substrate, such as a lipid22,23, in both bacterial
when vinblastine has dissociated. In contrast, nicardipine acts as a and mammalian cells, which has yet to be identified.
non-competitive inhibitor that interacts with LmrA at a binding site Functional, heterologous expression of LmrA in eukaryotic cells
distinct from, but allosterically linked to, the vinca-alkaloid-binding strongly implies that its ability to confer drug resistance is independent
site. Thus, not only are the substrate and modulator specificities of of any auxiliary proteins. LmrA is a half-molecule version of P-
LmrA and P-glycoprotein very similar, but the binding sites for glycoprotein3 and, by analogy with P-glycoprotein, is likely to be active
substrates and allosteric non-competitive inhibitors are also as a homodimer. LmrA transports amphiphilic drugs from the inner
conserved between P-glycoprotein and LmrA. leaflet of the phospholipid bilayer of L. lactis4, providing strong support
for the hydrophobic vacuum cleaner mode4 and the flippase model24
for P-glycoprotein. The remarkable conservation of function, substrate
specificity and the pharmacology of allosteric sites between LmrA
and P-glycoprotein suggests a fundamental biological role for this
type of ABC transporter in prokaryotic and eukaryotic cells. M
.........................................................................................................................

Methods
Genetic manipulations. The lmrA gene was amplified from pGKLmrA3 by
PCR using the oligonucleotide 59-CTAGTCTAGACCACCATGCATCACCATC
ACCATCACGATGACGATGACAAAGCC-39 to introduce an XbaI site, Kozak
sequence9, hexa-histidine tag, and enterokinase cleavage site at the 59-end of
lmrA, and the oligonucleotide 59-ATGGCCGACGTCGACTTATTATTGACC
AACAGTCAATTGTTCTGAAACATATTTAGC-39 to introduce a SalI site at the
39-end of lmrA. The K388M mutant form of LmrA was constructed by site-
directed mutagenesis of lmrA using the Altered Sites System (Promega) and the
oligonucleotide 59-GCCTTTGCTGGTCCTTCTGGTGGTGGTATGTCAAC-
CATCTTCTCACTTTTAGAACG-39. The wild-type and mutant lmrA genes
were cloned into the mammalian expression vector pCIneo (Promega) as
1,796-base-pair (bp) XbaI–SalI fragments, giving pCHLmrA and
pCHLmrAK388M, respectively, and sequenced to ensure that only the intended
changes had been introduced.
The wild-type lmrA gene and human MDR1 gene11 were inserted into the
genome of the Autographa californica nuclear polyhedrosis virus under the
control of the polyhedrin promoter using the BacPAK baculovirus expression
system (Clontech). The wild-type lmrA gene was subcloned from pCIneo into
the transfer vector pBacPAK9 as a 1,807-bp XbaI–NotI fragment. The human
MDR1 gene, together with a Kozak sequence9, was subcloned from pKSMDR1
(D. Gill, unpublished data) into the transfer vector pBacPAK9 using flanking
BamHI and XhoI restriction sites. To generate recombinant baculovirus
carrying the lmrA or MDR1 gene, the permissive host insect cell line Sf913
was cotransfected with modified transfer vector and linearized Bsu36 I-digested
BacPAK6 viral DNA by lipofection (Clontech). After isolation of recombinant
baculovirus DNA, the lmrA and MDR1 genes were sequenced to confirm their
Figure 3 Wild-type LmrA protein functions as a multidrug transporter in GM0637 proper insertion into the baculovirus genome.
human lung fibroblast cells. a, [3H]Vinblastine accumulation by cells expressing Cell culture and membrane preparations. The human lung fibroblast cell
the wild-type (A) and K388M mutant (X) forms of LmrA. After 10 min of vinblastine line GM0637 was cultured in DMEM medium supplemented with 10% fetal
accumulation, verapamil was added to wild-type LmrA-expressing cells (K) to calf serum11. For transfection, lung fibroblast cells were seeded at a density of
inhibit LmrA-mediated vinblastin extrusion. b, Transport of Hoechst 33342 in cells 2 3 105 in a 35-mm dish and grown for 18 h to ,40% confluence. Cells were
expressing wild-type LmrA (blue). Fluorescence intensity is presented in arbitrary washed three times with serum-free Opti-MEM (Life Technologies). 5 mg of
units (a.u.). At the time indicated by the arrow, reserpine was added to one of the plasmid DNA in 1 ml Opti-MEM was gently mixed with 25 mg of Lipofectin
incubations (red) to inhibit LmrA-mediated extrusion of Hoechst 33342. reagent (Life Technologies) in 1 ml of Opti-MEM, in a polystyrene tube. After

Nature © Macmillan Publishers Ltd 1998

NATURE | VOL 391 | 15 JANUARY 1998 293
letters to nature
15 min of incubation at 20 8C, the DNA–lipid mixture was added to the washed When used, modulators (inhibitors of P-glycoprotein-mediated drug trans-
lung fibroblast cells. Following an incubation of 5 h, the opti-MEM was port) were tested at 0.1 mM. After 48 h incubation, the number of viable cells in
replaced by DMEM supplemented with 10% fetal calf serum. Subsequently, each well was estimated by methylene blue staining. Drug survival is expressed
cells were incubated for 4 h. as the IC50 (the concentration necessary to reduce the plating efficiency of the
Transfected lung fibroblast cells were lysed in PBS containing 2% SDS. Sf9 cells by 50% after 48 h incubation). Data represent the mean and standard error
insect cells were grown in TC100 medium (Life Technologies) supplemented of at least two independent experiments, each carried out in duplicate.
with 10% fetal calf serum, and infected with wild-type or recombinant Drug transport assays. Vinblastine accumulation by attached lung fibroblast
baculovirus at a multiplicity of infection of 10. Plasma membranes were
prepared from baculovirus-infected insect cells, 48 h post-infection. Cells were
disrupted by nitrogen cavitation at 1,500 p.s.i. The membranes were purified by
sucrose density centrifugation25 and stored at −70 8C in 10 mM Tris-Cl (pH 7.4)
cells in transport buffer27 was initiated by addition of [G-3H]vinblastine
(0:5 3 1012 Bq mmol 2 1 Amersham) at 100 nM. Following incubation at
room temperature and washing with ice-cold PBS, cell extracts in 0.4 ml
0.4 M NaOH were neutralized with 0.8 ml 0.25 M ammonium acetate (pH
8
supplemented with 0.25 M sucrose. 6.6). The incorporated radioactivity was estimated by liquid scintillation
Immunochemistry. Protein was analysed by SDS–PAGE (10%) followed by counting. Hoechst 33342 (Molecular Probes) accumulation by trypsinized
western blot analysis using the anti-hexa histidine tag monoclonal antibody lung fibroblast cells in 50 mM potassium HEPES (pH 7.0), 2 mM MgSO4 and
DIA900 (Dianova). Expression of P-glycoprotein and MRP1 was assayed using 0.9% NaCl, was measured at a drug concentration of 5 mM by fluorescence
the monoclonal antibody C219 and the polyclonal antibody K5, respectively. spectrometry28 using excitation and emission wavelengths of 355 and 457 nm,
For confocal fluorescence microscopy, DIA900 antibody labelling of LmrA in and slit widths of 5 nm each. In drug accumulation assays, verapamil and
cold acetone–methanol fixed fibroblast cells was visualized with anti-mouse reserpine were used at 1 mM.
IgG1 antibody conjugated to fluorescein (Sigma). Cells were imaged with a ATPase assay. The ATPase activity in plasma membranes was based on a
MRC-600 confocal microscope using a Nikon ×60 PlanoApo 1.4NA oil colorimetric ascorbic acid/ammonium molybdate assay29 to measure the
immersion objective lens. liberation of inorganic phosphate from ATP. Assays were performed in the
Cell cytotoxicity assays. For cell cytotoxicity assays26, trypsinized transfected presence of 0.1 mM EGTA and 2 mM ouabain to inhibit P-type ATPases.
lung fibroblast cells from various 35-mm dishes were pooled, and seeded in 24- Drug-binding and dissociation assays. For equilibrium binding of
well dishes at a density of 3 3 104 cells per well and incubated for 12 h. vinblastine15 to LmrA, plasma membranes were incubated with [3H]vinblastine
Cytotoxic drugs were then added to each well at the indicated concentrations. at the indicated concentrations at 22 8C in the dark. After 90 min, 3 ml ice-cold

Figure 4 Pharmacological properties of wild-type LmrA and human P-glycopro- containing (K) and control (X) plasma membranes, as a function of the free
tein expressed in insect cell membranes. a, Effect of the magnesium-ATP concentration of [3H]vinblastine. c, Effect of P-glycoprotein modulators on the
concentration on the basal ATPase activity of LmrA in the absence (A) or equilibrium binding of [3H]vinblastine to wild-type LmrA. d, Dissociation kinetics
presence of 50 mM verapamil (B) or 2 mM sodium vanadate (X), and of P- of [3H]vinblastine from wild-type LmrA in the absence (A) and presence of
glycoprotein in the absence (K) or presence of 50 mM verapamil (O). b, Equilibrium CP100 2 356 (K) or nicardipine (W).
binding of [3H]vinblastine to wild-type LmrA-containing (A), P-glycoprotein-

Nature © Macmillan Publishers Ltd 1998

294 NATURE | VOL 391 | 15 JANUARY 1998
 letters to nature
20 mM Tris-Cl (pH 7.4) containing 2 mM MgSO4 was added. Samples were
filtered through Whatman GF/F filters. Filter-retained radioactivity was mea-
sured by liquid scintillation counting. Nonspecific binding (usually ,35% The candidate tumour
total) was defined as the amount of [3H]vinblastine bound in the presence of
10 mM unlabelled vinblastine and was subtracted from all values. When used, suppressor p33ING1
modulators were included at a final concentration of 5 mM.
In drug-dissociation kinetic experiments15,17, plasma membranes were cooperates with p53
equilibrated for 120 min at 22 8C in the dark, in the presence of 92 nM
[3H]vinblastine. Subsequently, the plasma membranes were taken to 12 8C
after which non-labelled vinblastine was added to a final concentration of
10 mM in the presence or absence of either 3 mM nicardipine or 3 mM
in cell growth control
Igor Garkavtsev*¶k, Irina A. Grigorian†¶,
8
Valeria S. Ossovskaya‡#, Mikhail V. Chernov§,
CP100 ¼ 356. Samples were filtered at the indicated times as described. In
Peter M. Chumakov‡ & Andrei V. Gudkov†
Fig. 4d, dissociation of [3H]vinblastine was plotted as the natural logarithm of
* Department of Medical Biochemistry and Southern Alberta Cancer Research
the ratio of the amount of drug bound at time t (Bt) over the amount at
Center, University of Calgary, Calgary, Alberta, Canada T2N 4N1
equilibrium (Be), as a function of time t.
† Department of Molecular Genetics, College of Medicine, University of Illinois at
Received 28 May; accepted 24 September 1997. Chicago, Chicago, Illinois 60607, USA
1. Van Veen, H. W. & Konings, W. N. Drug efflux proteins in multidrug resistant bacteria. Biol. Chem. ‡ Engelhardt Institute of Molecular Biology, Moscow, Russia
378, 769–777 (1997).
§ Cleveland Clinic Foundation, Cleveland, Ohio 44195, USA
2. Nikaido, H. Prevention of drug access to bacterial targets: permeability barriers and active efflux.
Science 264, 382–388 (1994). ¶ These authors contributed equally to this work.
.........................................................................................................................
3. Van Veen, H. W. et al. Multidrug resistance mediated by a bacterial homolog of the human multidrug
transporter MDR1. Proc. Natl Acad. Sci. USA 93, 10668–10672 (1996).
4. Bolhuis, H. et al. Multidrug resistance in Lactococcus lactis: evidence for ATP-dependent drug
The candidate tumour-suppressor gene ING1 has been identified
extrusion from the inner leaflet of the cytoplasmic membrane. EMBO J. 15, 4239–4245 (1996). by using the genetic suppressor element (GSE) methodology1.
5. Higgins, C. F. ABC transporters: from microorganisms to man. Annu. Rev. Cell Biol. 8, 67–113 (1992). ING1 encodes a nuclear protein, p33ING1, overexpression of which
6. Chen, C. et al. Internal duplication and homology with bacterial transport proteins in the mdr1 (P-
glycoprotein) gene from multidrug resistance human cells. Cell 47, 381–389 (1986). inhibits growth of different cell lines. The properties of p33ING1
7. Endicott, J. A. & Ling, V. The biochemistry of P-glycoprotein mediated multidrug resistance. Annu. suggest its involvement in the negative regulation of cell pro-
Rev. Biochem. 58, 137–171 (1989).
8. Gottesman, M. M., Hrycyna, C. A., Schoenlein, P. V., Germann, U. A. & Pastan, I. Genetic analysis of
liferation and in the control of cellular ageing, anchorage depen-
the multidrug transporter. Annu. Rev. Genet. 29, 607–649 (1995). dence and apoptosis1–3. These cellular functions depend largely on
9. Kozak, M. Point mutations define a sequence flanking the AUG initiator codon that modulates the activity of p53, a tumour-suppressor gene that determines the
translation by eukaryotic ribosomes. Cell 44, 283–292 (1986).
10. Gill, D. R. et al. Separation of drug transport and chloride channel functions of the human multidrug cellular response to various types of stress4. Here we report that
resistance P-glycoprotein. Cell 71, 23–32 (1992). the biological effects of ING1 and p53 are interrelated and require
11. Goodfellow, H. R. et al. Protein kinase C-mediated phosphorylation does not regulate drug transport
by the human multidrug resistance P-glycoprotein. J. Biol. Chem. 271, 13668–13674 (1996).
the activity of both genes: neither of the two genes can, on its own,
12. Cole, S. P. C. et al. Overexpression of a transporter gene in a multidrug-resistant human lung cancer cause growth inhibition when the other one is suppressed.
cell line. Science 258, 1650–1654 (1992). Furthermore, activation of transcription from the p21/WAF1
13. Vaughn, J. L., Goodwin, R. H., Tompkins, G. J. & McCawley, P. The establishment of two cell lines
from the insect Spodoptera frugiperda (Lepidoptera: Noctuidae). In Vitro 13, 213–217 (1977). promoter, a key mechanism of p53-mediated growth control,
14. Tamai, I. & Safa, A. R. Azidopine non-competitively interacts with vinblastine and cyclosporin A depends on the expression of ING1. A physical association
binding to P-glycoprotein in multidrug resistant cells. J. Biol. Chem. 266, 16796–16801 (1991).
15. Ferry, D. R., Russell, M. A. & Cullen, M. H. P-glycoprotein possesses a 1,4-dihydropyridine-selective
between p33ING1 and p53 proteins has been detected by immuno-
drug acceptor site which is allosterically coupled to a vinca-alkaloid-selective binding site. Biochem. precipitation. These results indicate that p33ING1 is a component
Biophys. Res. Commun. 188, 440–445 (1992). of the p53 signalling pathway that cooperates with p53 in the
16. Ferry, D. R. & Kerr, D. J. in New Molecular Targets For Cancer Chemotherapy (eds Kerr, D. J. &
Workman, P.) 177–193 (CRC Press, London, 1994). negative regulation of cell proliferation by modulating p53-
17. Malkhandi, J. et al. Dexniguldipine-HCl is a potent allosteric inhibitor of [3H]vinblastine binding to dependent transcriptional activation.
P-glycoprotein of CCRF ADR 5000 cells. Eur. J. Pharmacol. 288, 105–114 (1994).
18. Shao, Y.-M., Ayesh, S. & Stein, W. D. Mutually co-operative interactions between modulators of P-
Expression of ING1 is upregulated in senescent human
glycoprotein. Biochim. Biophys. Acta 1360, 30–38 (1997). fibroblasts2, and ectopic expression of ING1 cDNA leads to arrest
19. Martin, C., Berridge, G., Higgins, C. F. & Callaghan, R. The multidrug resistance reversal agent in the G1 phase of the cell cycle or induces apoptosis in several cell
SR33557 modulates vinca alkaloid binding to P-glycoprotein by an allosteric interaction. Br. J.
Pharmacol. 122, 765–771 (1997). types1,3. Inhibition of ING1 expression by antisense RNA promotes
20. Kostenis, E. & Mohr, K. Two-point kinetic experiments to quantify allosteric effects on radioligand anchorage-independent growth in mouse breast epithelial cells,
dissociation. Trends Pharmacol. Sci. 17, 280–283 (1996).
21. Schinkel, A. H. et al. Disruption of the mouse mdr1a P-glycoprotein gene leads to a deficiency in the
increases the frequency of focus formation in NIH 3T3 cells, and
blood-brain barrier and to increased sensitivity to drugs. Cell 77, 491–502 (1994). prolongs the lifespan of diploid human fibroblasts in culture1. ING1
22. Van Helvoort, A. et al. MDR1 P-glycoprotein is a lipid translocase of broad specificity, while MDR3 P- is mapped to human chromosome 13q34 (refs 5, 6), and loss of
glycoprotein specifically translocates phosphatidylcholine. Cell 87, 507–517 (1996).
23. Ruetz, S., Brault, M., Dalton, W. & Gros, P. Functional interactions between synthetic alkyl heterozygosity in this chromosomal locus has been reported in
phospholipids and the ABC transporters P-glycoprotein, Ste-6, MRP, and Pgh-1. Biochemistry 36, squamous-cell carcinomas of the head and neck7. Limited analysis
8180–8188 (1997).
24. Higgins, C. F. & Gottesman, M. M. Is the multidrug transporter a flippase? Trends Biochem. Sci. 17,
of tumour cell lines revealed mutation of ING1 in the neuroblas-
18–21 (1992). toma cell line and reduced expression in several breast carcinoma
25. Lever, J. E. Active amino acid transport in plasma membrane vesicles from Simian virus 40- cell lines1.
transformed mouse fibroblasts. Characteristics of electrochemical Na+ gradient stimulated uptake.
J. Biol. Chem. 252, 1990–1997 (1977). Association of ING1 with growth suppression, replicative senes-
26. Cano-Cauci, D. F. & Riordan, J. R. Action of calcium antagonists on multidrug resistant cells. Biochem. cence, anchorage dependence, and apoptosis raised the question of
Pharmacol. 36, 2115–2123 (1987).
27. Callaghan, R. & Riordan, J. R. Synthetic and natural opiates interact with P-glycoprotein in multidrug
its relationship with p53, a tumour-suppressor gene involved in the
resistant cells. J. Biol. Chem. 268, 16059–16064 (1993). control of all these cellular functions4. To test this possibility, we
28. Shapiro, A. B. & Ling, V. Reconstitution of drug transport by purified P-glycoprotein. J. Biol. Chem. compared the effects of overexpression or suppression of p33ING1 in
270, 16167–16175 (1995).
29. Chifflet, S., Chiesa, U. T. R. & Tolosa, S. A method for the determination of inorganic phosphate in the cells differing in their p53 status. Modulation of endogenous p33ING1
presence of labile organic phosphate and high concentrations of protein: application to lens ATPases. expression was achieved by retroviral transduction of either full-
Analyt. Biochem. 168, 1–4 (1988).
length ING1 cDNA or an antisense-oriented ING1 cDNA fragment
Acknowledgements. We are grateful to Sahofi Recherche for supplying SR33557. We thank M. Müller for (anti-ING1 GSE) acting as a potent inhibitor of p33ING1 expression1.
anti-MRP1 antibody K5, D. Gill for plasmid pKSMDR1, and J. Taylor, K. Linton, E. Blott, G. Berridge,
C. Martin, G. Begley, R. Horvath, C. Nastrucci, S. Hyde, D. Gill and P. A. McNaughton for discussions. We found that the suppression of colony formation by ING1
This research was funded by the Biotechnology program of the Commission of the European Com-
munities, the Dutch Cancer Society, the Cancer Research Campaign (UK), and the Imperial Cancer
Research Fund. H.W.V.V. was the recipient of a short-term EMBO fellowship, and is a fellow of the Royal
Present address: k Department of Functional Genomics, Genome Therapeutics Corporation, Waltham,
Netherlands Academy of Arts and Sciences. C.F.H. is a Howard Hughes international Research Scholar.
Massachusetts 02154, USA; # Department of Pathology, University of California, San Francisco,
Correspondence and requests for materials to H.W.V.V. (e-mail: h.w.van.veen@biol.rug.nl). California 94143, USA.

Nature © Macmillan Publishers Ltd 1998

NATURE | VOL 391 | 15 JANUARY 1998 295