Multidrug resistance proteins MRP3, MRP1, and MRP2 in lung cancer: correlation of protein levels with drug response and messenger RNA levels.

Previously (L. C. Young et al., Clin. Cancer Res., 5: 673-680, 1999), we found, in a panel of 23 lung cancer cell lines that had not been selected for in vitro drug resistance, that the mRNA levels of MRP3 and MRP1, two members of the ATP-binding cassette superfamily of transport proteins, correlated with resistance to doxorubicin, vincristine, VP-16, and cis-diamminedicholoroplatinum(II). To extend these studies, we measured multidrug resistance protein (MRP)1, MRP2, and MRP3 protein levels in a panel of 30 lung cancer cell lines that included the original 23 cell lines as well as an additional 7 unselected lung cancer cell lines. In the case of MRP3, a polyclonal antibody was developed that was found to be a sensitive reagent for the detection of MRP3 by Western blot analysis. We found good agreement in the original 23 cell lines between the cognate mRNA and protein levels for MRP1, MRP2, and, especially, MRP3 (r, 0.852), supporting the use of semiquantitative PCR to predict MRP1, MRP2, and MRP3 protein levels in patient samples. There were also strong correlations between the mRNA and protein levels of MRP3 and MRP1, which suggested that these genes might be expressed in a coordinate manner. MRP3, MRP1, and MRP2 protein levels were higher in the non-small cell lung cancer (NSCLC) than in the SCLC cell lines and, in addition, MRP3 and MRP2 were detected almost exclusively in the NSCLC cell lines. Finally, we found that both MRP3 and MRP1, but not MRP2, protein levels correlated with decreased sensitivity of these lung cancer cell lines to doxorubicin, VCR, VP-16, and cis-diamminedicholoroplatinum(II). These findings are consistent with our hypothesis that both MRP3 and MRP1 are components of the multifactorial multidrug resistance phenotype of lung cancer and that MRP3 contributes to the intrinsic resistance of NSCLC cells.

Simultaneous quantitation of topoisomerase II alpha and beta isoform mRNAs in lung tumor cells and normal and malignant lung tissue.

Certain drugs used in the treatment of lung cancer and other human malignancies are cytotoxic because of their ability to interact with the two isoforms of topoisomerase II (topo II), topo IIalpha and topo IIbeta. As part of an effort to evaluate the contribution of topo II alterations to drug sensitivity and resistance in lung cancer, we have developed a semi-quantitative reverse transcriptase-polymerase chain reaction (RT-PCR) assay to measure levels of topo II alpha and beta mRNAs simultaneously using a single pair of primers with sequences common to both isoforms. The PCR products derived from the topo II alpha and beta mRNAs are both 446 bp but have different electrophoretic mobilities in a nondenaturing polyacrylamide gel, allowing sensitive, rapid quantitation when the products are radiolabeled with [35S]-dATP. Using this RT-PCR method, poly(A+) RNA from 13 non-small cell lung cancer (NSCLC) cell lines was analyzed. The results obtained indicated that the cell lines express a wide range of topo II alpha mRNA levels (12-fold) and topo IIbeta mRNA levels (5.5-fold). Tumor and normal lung tissues from 25 patients with NSCLC were also examined. In the tumor samples, the levels of the topo II alpha and beta mRNAs were similar. However, mean topo IIalpha mRNA levels in the tumors were approximately 7-fold higher than those of the paired normal lung tissues. In contrast, topo IIbeta mRNA levels were similar in both tumor and normal lung. Topo II alpha and beta mRNA levels were both significantly lower in the squamous cell tumors than in the adenocarcinoma samples. Topo IIbeta mRNA levels in the squamous cell tumors were also significantly lower than those in paired normal lung tissue. The RT-PCR method described is reliable and convenient, and for the first time, makes the rapid simultaneous direct comparison of topo IIalpha and topo IIbeta mRNA levels feasible in large numbers of clinical samples.

Sequence determinants of nuclear localization in the alpha and beta isoforms of human topoisomerase II.

The alpha and beta isoforms of DNA topoisomerase II (topo II) are targets for several widely used chemotherapeutic agents, and resistance to some of these drugs may be associated with reduced nuclear localization of the alpha isoform. Human topo IIalpha contains a strong bipartite nuclear localization signal (NLS) sequence between amino acids 1454 and 1497 (alphaNLS(1454-1497)). In the present study, we show that human topo IIalpha tagged with green fluorescence protein is still detectable in the nucleus when alphaNLS(1454-1497) has been disrupted. Seven additional regions in topo IIalpha containing overlapping potential bipartite NLSs were evaluated for their nuclear targeting abilities using a beta-galactosidase reporter system. A moderately functional NLS was identified between amino acids 1259 and 1296. When human topo IIbeta was examined in a similar fashion, it was found to contain two strongly functional sequences betaNLS(1522-1548) and betaNLS(1538-1573) in the region of topo IIbeta comparable to the region in topo IIalpha that contains the strongly functional alphaNLS(1454-1497). The third, betaNLS(1294-1332), although weaker than the other two beta sequences, is significantly stronger than the analogous alphaNLS(1259-1296). Differences in the NLS sequences of human topo II isoforms may contribute to their differences in subnuclear localization.

Expression of multidrug resistance protein-related genes in lung cancer: correlation with drug response.

Recently, cDNAs have been identified that encode four human proteins (MRP2-5) with structural similarity to the multidrug resistance protein (MRP). Preliminary studies have shown that levels of mRNAs encoding MRP2, MRP3, and MRP5, are increased in some drug-selected cell lines, but the correlation of MRP2-5 mRNA levels with drug resistance has not been examined. Using a collection of small cell lung cancer (SCLC) and non-SCLC patient samples and unselected cell lines established from patients at various stages of treatment, we examined the expression of MRP2, MRP3, MRP4, and MRP5, as well as MDR1 and MRP, by PCR. The levels of individual mRNAs were correlated with the sensitivity of these cell lines to doxorubicin (DOX), vincristine, VP-16, and cis-diamminedichloroplatinum(II), as determined by a modified MTT assay. Using both SCLC and non-SCLC cell lines, we confirmed the previously observed correlation of MRP mRNA levels with resistance to DOX (B. G. Campling et al., Clin. Cancer Res., 3:115-122, 1997) and found a strong correlation of MRP3 mRNA levels with resistance of the cell lines to DOX. In addition, the mRNA levels of both MRP and MRP3 correlated with resistance of the cell lines to vincristine, VP-16, and cis-diamminedichloroplatinum(II). These findings are consistent with the suggestion that MRP3, like MRP, may contribute to the drug resistance phenotype of lung cancer cells.

Structural organization of the human TOP2A and TOP2B genes.

Eucaryotic topoisomerase II is an essential nuclear enzyme involved in processes such as chromosome condensation, chromatid separation, and in the relief of torsional stress that occurs during DNA transcription and replication. In cells from vertebrate species, there are two forms of the enzyme, designated alpha and beta. Human topoisomerase IIalpha (TOP2A) is encoded by the TOP2A gene on chromosome 17q21-22, and human topoisomerase IIbeta (TOP2B) is encoded by the TOP2B gene on chromosome 3p24. The protein products of these two genes are important cellular targets of several drugs widely used in the treatment of many human cancers, and a variety of mutations in TOP2A have been associated with the development of drug resistance. In the present study, we have defined the intron-exon structures of TOP2A and TOP2B. TOP2A is approx. 30kb whereas TOP2B is at least 49kb. TOP2A and TOP2B contain 35 and 36 exons, respectively, and both genes contain a high proportion of class 0 introns. Alignment of the amino-acid sequences of the two proteins indicates that the intron-exon organization of the two genes is highly conserved, except for the regions encoding the extreme NH2 and COOH termini of the proteins. These findings suggest strongly that the vertebrate isoforms evolved by duplication of an ancestral gene. Mutations in TOP2A associated with drug resistance show clustering in exons 12, 13, 19-21 and 34-35. Knowledge of the genomic organization of TOP2A and TOP2B will be useful for detection of mutations in clinical samples from patients with drug-resistant malignant disease.

Multidrug resistance genes (MRP) and MDR1 expression in small cell lung cancer xenografts: relationship with response to chemotherapy.

Intrinsic or acquired drug resistance is a major limiting factor of the effectiveness of chemotherapy. Increased expression of either the MRP gene or the MDR1 gene has been demonstrated to confer drug resistance in vitro. In this study, we examined MRP and MDR1 gene expression in a panel of 17 small cell lung cancers (SCLC) xenografted into nude mice from treated and untreated patients using an RT-PCR technique. For some of them, the outcome of the corresponding patients was known and we related MDR1/MRP expression with the xenograft response to C'CAV (cyclophosphamide, cisplatin, adriamycin and etoposide) combined chemotherapy. Fifteen (88%) of the 17 cases of SCLC were found to be positive for either MDR1 or MRP. MRP gene expression was present in 12 (71%) of 17 cases, whereas MDR1 gene expression was detected in eight (50%) of 16 cases. For six SCLC, the survival duration of patients differed, with three patients surviving for more than 30 months after therapy. Among these six turnours, five expressed MRP and/or MDR1. These six xenografts responded to the C'CAV treatment but a significant rate of cure was obtained in only three cases. No obvious relationship was observed between the response to this treatment and MRP or MDR1 expression. However, the remarkably high levels and frequency of MRP expression in some SCLC samples indicate that future developments in chemotherapy of this tumour type should anticipate that drugs which are substrates of MRP may be of limited effectiveness.

Membrane topology of the multidrug resistance protein (MRP). A study of glycosylation-site mutants reveals an extracytosolic NH2 terminus.

Multidrug resistance protein, MRP, is a 190-kDa integral membrane phosphoglycoprotein that belongs to the ATP-binding cassette superfamily of transport proteins and is capable of conferring resistance to multiple chemotherapeutic agents. Previous studies have indicated that MRP consists of two membrane spanning domains (MSD) each followed by a nucleotide binding domain, plus an additional extremely hydrophobic NH2-terminal MSD. Computer-assisted hydropathy analyses and multiple sequence alignments suggest several topological models for MRP. To aid in determining the topology most likely to be correct, we have identified which of the 14 N-glycosylation sequons in this protein are utilized. Limited proteolysis of MRP-enriched membranes and deglycosylation of intact MRP and its tryptic fragments with PNGase F was carried out followed by immunoblotting with antibodies known to react with specific regions of MRP. The results obtained indicated that the sequon at Asn354 in the middle MSD is not utilized and suggested approximate sites of N-glycosylation. Subsequent site-directed mutagenesis studies established that Asn19 and Asn23 in the NH2-terminal MSD and Asn1006 in the COOH-terminal MSD are the only sites in MRP that are modified with N-linked oligosaccharides. N-Glycosylation of Asn19 and Asn23 provides the first direct experimental evidence that MRP has an extracytosolic NH2 terminus. This finding, together with those of previous studies, strongly suggests that the NH2-terminal MSD of MRP contains an odd number of transmembrane helices. These results may have important implications for the further understanding of the interaction of drugs with MRP.

Expression of the MRP and MDR1 multidrug resistance genes in small cell lung cancer.

Acquired multidrug resistance is a major obstacle to a cure for small cell lung cancer (SCLC). Overexpression of the MDR1 gene occurs infrequently in multidrug-resistant SCLC cell lines. The multidrug resistance protein (MRP) can confer multidrug resistance, but its role in clinically acquired drug resistance is unknown. The purpose of this study was to measure expression of MRP and MDR1 mRNA in cell lines and clinical samples from SCLC patients and to correlate the results with drug sensitivity profiles. Twenty-three SCLC cell lines and 10 tumor samples from SCLC patients were examined. Samples expressing MRP and MDR1 were identified by reverse transcription-PCR, and levels of MRP mRNA in the cell lines were measured by quantitative reverse transcription-PCR. One of 23 cell lines (4%) expressed MDR1 mRNA, whereas MRP expression was detected in 19 of 23 cell lines (83%). There was a significant correlation between doxorubicin resistance and MRP expression levels (r = 0.422; P = 0.045). Of the 10 clinical samples, 3 expressed only MRP, 2 expressed only MDR1, and 4 expressed both drug resistance genes. In summary, MRP is frequently expressed in clinical samples and cell lines from SCLC patients, and the levels correlate with doxorubicin resistance in unselected SCLC cell lines. Expression of MDR1 can be detected in clinical samples of SCLC but is rarely found in cell lines from drug-resistant patients. These multidrug resistance proteins may contribute to the multifactorial problem of clinically acquired drug resistance in SCLC.

Bipartite nuclear localization signals in the C terminus of human topoisomerase II alpha.

DNA topoisomerase II alpha is the intracellular target for several important chemotherapeutic agents, and drug-resistant human tumor cell lines have been described in which deletions in the C-proximal region of this enzyme are associated with its cytoplasmic localization. We have identified multiple potential bipartite nuclear localization signal (NLS) sequences in this region using a modified definition of the motif, and in the present study, we have expressed five of these as fusion proteins with beta-galactosidase. Only one sequence (spanning amino acids 1454 to 1497) was sufficient to cause strong nuclear localization. Subsequent mutation analyses indicated that this NLS sequence was bipartite and that both domains contain more than two basic amino acids. Substitution of the lysine residue at position 1492 in the second basic domain with glutamine resulted in a fusion protein that localized inefficiently to the nucleus, indicating that all three basic residues in this domain are necessary. Our results confirm that a broader definition is required to detect all potential bipartite NLS motifs in a polypeptide sequence, although functional tests are still essential for identification of those sequences actually capable of directing nuclear localization.

Structure and expression of the messenger RNA encoding the murine multidrug resistance protein, an ATP-binding cassette transporter.

In vitro, overexpression of the human multidrug-resistance protein (MRP) causes a form of multidrug resistance similar to that conferred by P-glycoprotein, although the two proteins are only very distantly related. Studies with MRP-enriched membrane vesicles have demonstrated that the protein can bind and transport cysteinyl leukotrienes, as well as some other glutathione conjugates, with high affinity. In contrast, there is no direct evidence of the ability of MRP to bind or transport unmodified forms of the drugs to which it confers resistance. To facilitate studies of the physiological function(s) of MRP and its ability to cause multidrug resistance in vivo, we cloned and characterized the mRNA specifying its murine homolog. The murine MRP mRNA encodes a protein of 1528 amino acids that is 88% identical to human MRP. Although detectable by Northern blotting at variable levels in a wide range of tissues, in situ hybridization experiments revealed that MRP mRNA expression in some tissues is cell-type specific. High levels of the mRNA were detected in epithelia lining bronchi and bronchioles, as well as stage-specific expression in the seminiferous epithelium of the testes. Comparison of the predicted hydropathy profiles of human and murine MRP suggests a highly conserved membrane topology, the most distinctive feature of which is an extremely hydrophobic NH2-terminal region containing five or six potential transmembrane sequences. This structural feature is shared with the sulfonylurea receptor and the yeast cadmium factor 1 but is not present in members of the superfamily, such as the cystic fibrosis transmembrane conductance regulator and P-glycoproteins. Finally, we used overlapping cDNAs to construct an episomally replicating murine MRP expression vector that was stably transfected into HeLa cells. MRP-Transfected cell populations expressed markedly elevated levels of a 180-190-kDa protein that cross-reacted with a polyclonal antiserum raised against a peptide that is completely conserved in murine and human MRPs. The MRP transfectants also displayed increased resistance to vincristine (5-6-fold) and doxorubicin (< 2-fold).

Benzo[a]pyrene-resistant MCF-7 human breast cancer cells. A unique aryl hydrocarbon-nonresponsive clone.

Wild-type MCF-7 human breast cancer cells were cultured for 3 months in 1 microM benzo[a]pyrene (BaP), and resistant clones were screened for inducibility of CYP1A1 gene expression by 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). One of the BaP-resistant (BaPR) clones exhibited unique genotypic expression which distinguished it from both wild-type and drug-resistant (AdrR) variant MCF-7 cells. Glutathione levels, glutathione S-transferase activities, estrogen receptor levels, estrogen responsiveness, and expression of the multidrug-resistant MDR1 and MRP mRNA levels were similar in the wild-type and BaPR cells, whereas these parameters were reported to be altered in AdrR cells. In contrast, TCDD induced CYP1A1 gene expression and inhibited selected estrogen-induced responses in wild-type but not BaPR MCF-7 cells. Treatment of wild-type and BaPR cells with [3H]TCDD resulted in formation of the radiolabeled aryl hydrocarbon (Ah) 6 S nuclear receptor complex in both cell lines. The loss of Ah responsiveness in the BaPR variant cells correlated with the failure of the nuclear or transformed cytosolic Ah receptor complex to bind genomic dioxin-responsive elements as determined in gel retardation assays.

Quantitative polymerase chain reaction analysis of mdr1 mRNA in multiple myeloma cell lines and clinical specimens.

We have designed a new polymerase chain reaction (PCR) protocol for the quantitation of mdr1 mRNA in cell lines and clinical specimens. This protocol uses an in vitro-generated RNA molecule as an internal standard. This synthetic RNA contains the same mdr1 primer sequences as the cellular mRNA, but yields a different-sized PCR product after amplification. Since a single primer set is used in quantitation, differences in primer efficiency are not a concern. We have used this assay to measure mdr1 expression in a multiple myeloma cell line, 8226/S, its drug resistant variants 8226/dox6 and 8226/dox40, and tumor samples from 10 patients with B-cell malignancies (9 multiple myeloma, 1 chronic lymphocytic leukemia). 8226/S does not express mdr1 mRNA. 8226/dox6 is 10-fold resistant to doxorubicin, and expresses 32 mdr1 mRNA/10 pg cellular RNA. 8226/dox40 is 140-fold resistant to doxorubicin, and expresses 890 mdr1 mRNA/10 pg cellular RNA. Seven of the 10 patients had levels of mdr1 mRNA expression below that seen in the multidrug-resistant, human multiple myeloma cell line, 8226/dox6. Three patients had levels of mdr1 expression comparable to those seen in 8226/dox6. No patient had levels of mdr1 expression close to that seen in 8226/dox40. Sample RNA integrity is assured by PCR analysis of a different, ubiquitous, cell cycle independent, histone variant, H3.3. This assay will be useful for studying low level mdr1 expression in cell lines and clinical specimens.

Overexpression of a transporter gene in a multidrug-resistant human lung cancer cell line.

The doxorubicin-selected lung cancer cell line H69AR is resistant to many chemotherapeutic agents. However, like most tumor samples from individuals with this disease, it does not overexpress P-glycoprotein, a transmembrane transport protein that is dependent on adenosine triphosphate (ATP) and is associated with multidrug resistance. Complementary DNA (cDNA) clones corresponding to messenger RNAs (mRNAs) overexpressed in H69AR cells were isolated. One cDNA hybridized to an mRNA of 7.8 to 8.2 kilobases that was 100- to 200-fold more expressed in H69AR cells relative to drug-sensitive parental H69 cells. Overexpression was associated with amplification of the cognate gene located on chromosome 16 at band p13.1. Reversion to drug sensitivity was associated with loss of gene amplification and a marked decrease in mRNA expression. The mRNA encodes a member of the ATP-binding cassette transmembrane transporter superfamily.

Non-P-glycoprotein-mediated multidrug resistance in a small cell lung cancer cell line: evidence for decreased susceptibility to drug-induced DNA damage and reduced levels of topoisomerase II.

Data obtained from clinical samples suggest that non-P-glycoprotein mechanisms of multidrug resistance are likely to be important in small cell lung cancer. The H69AR cell line was derived from the H69 small cell lung cancer cell line by selection in doxorubicin (adriamycin) and does not overexpress P-glycoprotein as detected by monoclonal antibody C219 (S.E.L. Mirski et al., Cancer Res., 47:2594, 1987). In the present study, we have used the polymerase chain reaction to verify that H69AR cells do not overexpress P-glycoprotein. Further, transport studies with radiolabeled daunomycin, VP-16, and vinblastine demonstrate that differences in net drug accumulation or efflux are not part of the resistance phenotype of H69AR cells. To determine if H69 and H69AR cells differ in their susceptibility to drug-induced DNA damage, DNA single-strand breaks (SSB) generated by VP-16 and Adriamycin were measured using the alkaline filter elution assay. Readily detectable SSB were produced in intact H69 cells by 5 microM VP-16, but 100 microM drug was required to cause similar damage in H69AR cells. H69AR cells were also resistant to SSB induction by Adriamycin. The formation of SSB by VP-16 was similarly reduced in isolated H69AR nuclei, indicating that resistance to this drug resides, at least in part, in the nucleus. No significant differences were observed in the rate or extent of repair of VP-16-induced DNA SSB in H69 and H69AR cells. The reduced susceptibility to drug-induced SSB may result from alterations in topoisomerase II, since less immunoreactive topoisomerase II was found in H69AR cells compared to H69 cells. However, changes in topoisomerase II cannot explain the resistance of H69AR cells to such drugs as the Vinca alkaloids and gramicidin D, indicating that multiple mechanisms contribute to drug resistance in this small cell lung cancer cell line.

Structure and function of P-glycoprotein.

P-glycoprotein remains the best understood molecule that has been implicated in the multidrug resistance phenomenon. As I have attempted to show, our understanding of this molecule is still in its infancy, with many questions as yet unanswered. It is anticipated that the answers to these questions will contribute to our understanding of drug resistance, as it relates to cancer, and also to our understanding of what appears to be a fundamental transport system.

Simultaneous expression of two P-glycoprotein genes in drug-sensitive Chinese hamster ovary cells.

Overexpression of P-glycoprotein is characteristic of multidrug-resistant cells. We analyzed four P-glycoprotein transcripts that are simultaneously expressed in a drug-sensitive Chinese hamster ovary cell line. We concluded that these transcripts are encoded by two distinct members of a P-glycoprotein multigene family, each of which has two alternative polyadenylation sites. A comparison of the two hamster sequences with the single reported human and mouse P-glycoprotein cDNA sequences demonstrates that P-glycoprotein is a highly conserved protein, that the hamster multigene family is undergoing concerted evolution, and that differences between gene family members are maintained across species. These conserved differences suggest that there may be functional differences between P-glycoprotein molecules.

P-glycoprotein in human sarcoma: evidence for multidrug resistance.

Overexpression of an immunologically conserved, cell-surface glycoprotein (P-glycoprotein) is consistently associated with multidrug resistance in cell lines in vitro. A preliminary survey of specimens from 12 solid tumor types in our laboratories indicates significant overexpression of P-glycoprotein in some sarcomas. When tested by immunoblotting with monoclonal antibodies directed against P-glycoprotein; tumors from six of 25 sarcoma patients displayed elevated levels of P-glycoprotein. Three of the sarcoma patients exhibiting P-glycoprotein had not previously been exposed to chemotherapy, implying that overexpression of this marker and possible concomitant multidrug resistance may not depend only on selection during prior drug treatments. The P-glycoprotein overexpression in the sarcoma specimens is evidence for the presence of multidrug resistant cells in these tumors; thus, our data suggest that this mode of resistance may have clinical significance in sarcoma patients.

Multidrug resistance in a human small cell lung cancer cell line selected in adriamycin.

A multidrug resistant variant (H69AR) of the human small cell lung cancer cell line NCI-H69 was obtained by culturing these cells in gradually increasing doses of Adriamycin up to 0.8 microM after a total of 14 months. H69AR expresses the multidrug resistant phenotype because it is cross-resistant to anthracycline analogues including daunomycin, epirubicin, menogaril, and mitoxantrone as well as to acivicin, etoposide, gramicidin D, colchicine, and the Vinca alkaloids, vincristine and vinblastine. H69AR is also similar to other multidrug resistant cell lines in that it displays little or no cross-resistance to bleomycin, 5-fluorouracil, and carboplatin. It has a slight collateral sensitivity to 1-dehydrotestosterone and lidocaine. H69AR has increased cell-cell adhesiveness compared to H69, but a similar growth rate in vitro and tumorigenicity in nude mice. When cultured in the absence of Adriamycin, there is a 40% decrease in resistance by 35 days of culture, compared to cells in continuous culture in drug, but no further decrease in resistance up to 181 days. Monoclonal antibodies to P-glycoprotein have no detectable reactivity with H69AR cells as determined by enzyme-linked immunosorbent assay and immunoblotting techniques. Thus, unlike most multidrug resistant cell lines, H69AR does not appear to express enhanced levels of P-glycoprotein. H69AR will provide a useful model for the study of multidrug resistance in human small cell lung cancer.

Homology between P-glycoprotein and a bacterial haemolysin transport protein suggests a model for multidrug resistance.

Increased expression of P-glycoprotein, a plasma membrane glycoprotein of relative molecular mass (Mr) 170,000 (170K), occurs in a wide variety of cell lines that exhibit pleiotropic resistance to unrelated drugs. The presence of P-glycoprotein in human cancers refractory to chemotherapy suggests that tumour cells with multidrug resistance can arise during malignant progression. We have discovered striking homology between P-glycoprotein and the HlyB protein, a 66K Escherichia coli membrane protein required for the export of haemolysin (protein of Mr 107K). P-glycoprotein can be viewed as a tandem duplication of the HlyB protein. The hydropathy profiles of the two proteins are similar and reveal an extensive transmembrane region resembling those found in pore-forming plasma membrane proteins. The C-terminal region of P-glycoprotein and the HlyB protein contain sequences homologous to the nucleotide-binding domains of a group of closely related bacterial ATP-binding proteins. We propose a model for multidrug resistance in which P-glycoprotein functions as an energy-dependent export pump to reduce intracellular levels of anticancer drugs.