Peptide transport by the multidrug resistance pump.

The membrane P-glycoprotein (P170) is an ATP-hydrolyzing transmembrane pump, and elevated levels of P170, due to higher expression with or without amplification of the multidrug resistance gene (mdr1), result in resistance to a variety of chemotherapeutic agents in mammalian cells. The function of the P170 pump has been proposed as a protection against toxic substances present in animal diets. Here we describe a Chinese hamster ovary cell line that was selected for resistance to a synthetic tripeptide, N-acetyl-leucyl-leucyl-norleucinal (ALLN). This ALLN-resistant variant shows the classical multidrug resistance (MDR) phenotype, including overexpression and amplification of the mdr1 gene. Additionally, a mouse embryo cell line overexpressing the transfected mdr1 gene is likewise resistant to ALLN. Our results demonstrate that P170 is capable of transporting peptides and raise the possibility that the mdr1 gene product or other MDR-like genes, present in the genome of mammalian cells, may be involved in secretion of peptides or cellular proteins as is the case with the structurally similar hylB and ste6 gene products of Escherichia coli and yeast, respectively.

The membrane P-glycoprotein (P170) is an ATPhydrolyzing transmembrane pump, and elevated levels of P170, due to higher expression with or without amplification of the multidrug resistance gene ( m d r l ) , result in resistance to a variety of chemotherapeutic agents in mammalian cells. The function of the P170 pump has been proposed as a protection against toxic substances present in animal diets. Here we describe a Chinese hamster ovary cell line that was selected for resistance to a synthetic tripeptide, N-acetyl-leucylleucyl-norleucinal (ALLN). This ALLN-resistant variant shows the classical multidrug resistance (MDR) phenotype, including overexpression and amplification of the mdrl gene. Additionally, a mouse embryo cell line overexpressing the transfected mdrl gene is likewise resistant to ALLN. Our results demonstrate that P170 is capable of transporting peptides and raise the possibility that the mdrl gene product or other MDRlike genes, present in the genome of mammalian cells, may be involved in secretion of peptides or cellular proteins as is the case with the structurally similar hylB and ste6 gene products of Escherichia coli and yeast, respectively.
There exist in any number of species, from bacteria to man, a class of ATP-hydrolyzing transmembrane exit pumps that transport various intracellular compounds to the outside. These pumps bear varying degrees of structural or sequence homology. Many bacterial examples suggest a role in preventing accumulation of toxic agents (1,2). A similar role is proposed for the mammalian P170 protein (mdrl gene product). Elevated expression of mdrl gene was observed in kidney, liver, and colon epithelia (3), and P170 protein was present on the secretory epithelium of mouse uterus (4) and human placenta (5). Overexpression of P170 results in resistance to a variety of unrelated chemotherapeutic drugs, ie. the multidrug resistance (MDR)' phenotype (6, 7). Certain members of this transmembrane pump family of genes/proteins * This work was supported in part by National Institutes of Health Grants HL 26502 (to R. D. S.) and CA 16318 (to R. T. S.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
4 To whom correspondence should be addressed.
The genomes of mammals contain, in addition to the mdrl gene, various numbers of mdr-like genes of unknown function (11,12). The question, thus, is raised as to whether the mdrl gene itself (generally associated with exclusion of toxic agents) or other members of this gene family in mammals may have a physiological function in outward transport of endogenous peptides (see Ref. 13 for a review of peptide or protein transport by such pumps). Here we provide evidence that the synthetic hydrophobic tripeptide, N-acetyl-leucyl-leucyl-norleucinal (ALLN) is a substrate for the P170 protein.
The reason for studying ALLN relates to its properties as an inhibitor of various intracellular proteases, including calpain I and I1 and cathepsins B, D, and E (14). Inoue et al.
(15) reported that ALLN blocks the physiological degradation of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase in CHO cells. In an attempt to further understand the role of ALLN-sensitive protease(s) in the degradation of HMG-CoA reductase, we derived by stepwise selection CHO-K1 cells highly resistant to ALLN with the anticipation of amplifying gene(s) for the protease(s). The resistance phenotype obtained was related to overexpression and amplification of the mdrl gene.

EXPERIMENTAL PROCEDURES
Cell Culture, Drug Selection, and Cytotoxicity Assay-CHO-K1 cells, stably transfected with the chimeric gene HMGal (16), were grown in minimal essential medium containing dialyzed fetal bovine serum (lo%), glutamine (2 mM), gentamicin (10 pg/ml), and G418 (250 pglml) at 37 "C in 5% C02. For stepwise selection to ALLN resistance, exponentially growing CHO-K1 cells were treated with 5 pg/ml ALLN. The medium containing ALLN was replaced every 2-3 days until cells grew to about 70% confluency, and at this stage cells were trypsinized and replated in the same medium. After about 1 month CHO-K1 cells became resistant to 5 pg/ml ALLN. These ALLN-resistant cells were replated as above with 50-100% incrementa in ALLN concentration. The stepwise selection protocol was terminated at 50 pg/ml ALLN. These ALLN-resistant cells are hereafter designated ALLN'" cells. Cytotoxicity assays (by clonogenic method) were performed as described earlier (17).
Measurement of P-glycoprotein and MDR Gene Copy Number-To quantitate the amount of P-glycoprotein by immunoblotting, cells were trypsinized, collected, and lysed in lysis buffer (1% deoxycholate, 1% Nonidet P-40, 2 mM phenylmethanesulfonyl fluoride, 0.1 mM leupeptin, 5 p~ ALLN, 5 mM EDTA, 5 mM EGTA in phosphatebuffered saline). Nuclei were removed by centrifugation (3,000 X g for 15 min), and proteins were resolved by 5-15% gradient SDSpolyacrylamide gel electrophoresis. Proteins were electroblotted onto nitrocellulose membrane, and the membranes were fixed with 0.2% Ponceau S in 5% trichloroacetic acid. The membranes were blocked with phosphate-buffered saline containing 5% nonfat milk for 1 h at room temperature. The membranes were then probed overnight with monoclonal anti-P170 antibody C219 (Centocor, Malvern, PA) or anti-calpain I1 antibody at 4 "C, washed, and incubated with horseradish peroxidase-conjugated anti-mouse Ig antibody for 1 h at room temperature. The membranes were washed again, and the blots were developed using an enhanced chemiluminescence Western blotting detection kit (Amersham Corp.) and exposed to x-ray film for 2-30 s. Genomic DNA was isolated from parental CHO-K1, a MDR cell line CHO CH'C-5, and CHO-K1 ALLNrS0 cells by the standard method of phenol-chloroform extraction. MDR gene copy was estimated as described earlier (17).

RESULTS AND DISCUSSION
In order to investigate the MDR phenotype of ALLW' O cells, we determined the cytotoxicity of ALLN and the pleiotropic drugs (colchicine, doxorubicin, etoposide) to ALLN'" cells ( Fig. 1). When compared with parental CHO-K1 cells, ALLN'" cells were 50-, 50-, 20-, and 14-fold more resistant to ALLN, doxorubicin, colchicine, and etoposide, respectively (Fig. 1). These data show that ALLN'" cells are highly resistant to the pleiotropic drugs involved in MDR. The ALLN'" cell line did not show any cross-resistance to either trimetrexate or methotrexate (data not shown) suggesting that the ALLN resistance phenotype does not overlap with antifolate resistance (18). We do not know why ALLN is toxic to cells. However, it is reasonable that its toxicity relates to its inhibition of endogenous protease(s).
The MDR phenotype results from overexpression of P170 which is encoded by the mdrl gene. To estimate cellular levels of P-glycoprotein, postnuclear extracts from parental CHO-K1 and ALLN'" cells were analyzed by immunoblotting with specific antibodies (Fig. 2A). When the C219 monoclonal antibody to P-glycoprotein was used, a band with a M, of 170,000 corresponding to P-glycoprotein is detected in extracts from ALLN'" cells (lane 2 ) . A longer exposure showed a band with the same molecular weight in the extract from wild type CHO-K1 cells (lane 3). The intensity of the band, as determined densitometrically, is more than 10-fold higher in the ALLN'" cell extract than in wild type CHO-K1 cell extract. When monoclonal anti-rabbit calpain I1 antibody was used, a band with a M, of 80,000 is detected (lanes 5 and 6). The intensity of this band was the same in both cell extracts, indicating that the amount of calpain I1 did not change in ALLN-resistant cells. When antibodies against HMG-CoA reductase or annexin I1 were used to probe immunoblots, the same amounts of these proteins were detected in both cell extracts (data not shown). These results demonstrate that P170 is specifically overexpressed in ALLN'O cells.
The increase in the level of P170 is not always accompanied by MDR gene amplification. However, most of the highly resistant stable MDR cell lines do show MDR gene amplification accompanied by a higher level of P170 (19). To determine whether the mdrl gene is amplified in a ALLN-resistant cell line, we isolated genomic DNA from parental CHO-K1, ALLN'", and a well characterized MDR cell line CHO CHC-5 (20, 21), subjected genomic DNA to digestion with EcoRI, and analyzed it by Southern blot (Fig. 2B). Amplification of the mdrl gene was detected in ALLN-resistant as well as in CHO CHC-5 cell lines (note: 5-fold less DNA was loaded for this latter cell line). As a control, the nylon membrane was stripped by washing in 0.1 X SSPE, 0.1% SDS at 85-95 "C and reprobed with c-myc DNA (22). Wild-type CHO-K1 and ALLN'" cells showed similar hybridization to c-myc (data not shown). The amount of mdrl DNA present in these cells was quantitated by the slot-blot technique (23) using c-myc DNA as a control gene. Densitometric scanning of exposed films show about 10-fold amplification of mdrl gene in ALLW50 cells (data not shown). Thus, the elevated levels of P170 could be accounted for solely by mdrl gene amplification.
The typical MDR phenotype which includes: (i) crossresistance to pleiotropic drugs like doxorubicin, colchicine, etoposide; (ii) increased expression of membrane-associated P-glycoprotein; (iii) higher expression of P-glycoprotein with mdrl gene amplification, is found in the ALLN'" cells. Characteristically, the MDR phenotype is reversed by various agents, including verapamil (24) and quinidine (25). Verapamil binds to P-glycoprotein and reverses the MDR phenotype by a mechanism that involves competitive inhibition of drug transport by P-glycoprotein (26-28). Fig. 3 shows that verapamil (10 p~) reverses the ALLN resistance phenotype of ALLN'" cells by increasing their sensitivity (35-fold) to ALLN. Verapamil results in a 2-fold increase in sensitivity to ALLN in parental CHO-K1 cells. This is consistent with results obtained with other wild-type cell lines and pleiotropic drugs (25). Verapamil alone (up to 30 p M ) has no effect on the viability of wild-type or ALLN r50 cells (data not shown). These observations are consistent with the conclusion that ALLN is a substrate for P-glycoprotein.
The above results suggest that the apparent resistance to ALLN is, in fact, a manifestation of lower steady-state levels of ALLN in ALLN'" cells due to the outward pumping activity of P-glycoprotein. If this is the case, one would predict lower intracellular steady-state levels of ALLN in ALLN'" cells to affect the mevalonate-accelerated degradation of HMGal. To test this possibility, we determined the effect of ALLN on the mevalonate-accelerated degradation of HMGal in the parental CHO-K1 and the ALLW50 cell lines. Fig. 4 shows that the activity of HMGal in both cell lines drops to -15% of control when the cells are given excess mevalonate. This decrease in activity is due to accelerated degradation of HMGal (15,29). Addition of ALLN to the parental CHO-K1 cells causes a dose-dependent increase in HMGal activity to more than 90% of control levels, indicating that degradation was effectively blocked by ALLN with an IC5, of 6 pg/ml (see also Ref. 15). In the ALLW5' cells, however, ALLN inhibited the mevalonate-accelerated degradation of HMGal with an  K1 (lanes I , 3, and 5 ) and ALLN'" cells (lanes 2, 4, and 6 ) were resolved by SDS-polyacrylamide gel electrophoresis and were electroblotted onto nitrocellulose membrane. The membranes were then probed overnight with monoclonal anti-P170 antibody C219 (Centocor, Malvern, PA) (lanes [1][2][3][4] or anti-calpain I1 antibody (lanes 5 and 6) a t 4 "C, washed, and incubated with horseradish peroxidase-conjugated anti-mouse Ig antibody for 1 h a t room temperature. The membranes were washed, and blots were developed using an enhanced chemiluminescence Western blotting detection kit (Amersham Corp.) and exposed to x-ray film for 2-30 s. Lunes 3 and 4 are longer exposures of laws I and 2. Molecular mass markers (in kDa) are indicated on the left. B, to estimate MDR genomic DNA by Southern blot analysis, high molecular weight genomic DNA (10 pg/lane, except for lane CH'C-5, which contained 2 pg of DNA) was digested to completion with EcoRI, fractionated on an 0.8% agarose gel, and blotted onto a Zetabind membrane. Blots were then hybridized with a "P-oligolabeled CHO MDR probe (0.66kilobase pair EcoRI fragment of the MDR cDNA clone p-CHP-1). Hybridization was carried out a t 42 "C for 48 h in 50% formamide, 5 X saline/sodium phosphate/EDTA (SSPE), 1% SDS, and 1 X Denhardt's solution. The filter was given two 30-min high stringency washes in 0.1 X SSPE, 0.1% SDS at 65 "C and exposed for autoradi- ography (lanes 4-6). Ethidium bromide staining of the fractionated DNA is shown to confirm the amount of DNA loaded on the gel  ( l a w s 1-3)
Colonies were fixed and counted as described for Fig. 1. 0, 0, ALLN alone; 0   cells transfected with a wild-type MDRl cDNA (3T3 pHaMI>Rl/fiA CI 2-1) (W) were obtained from Michael M. Gottesman. S I H . Cells were grown in minimal essential medium supplemented with 10"; fetal calf serum, 10 pg/ml gentamicin. and 2 mM glutamine. Exponentially growing cells were trypsinized and plated (1 X 10' cells/f% cm dish) in 5 ml of complete media containing various amounts of ALLN. After 6-8 days of incubation, colonies were fixed and counted as described in Fig. 1. IC,,, of 90 pg/ml, and almost full inhibition was achieved at approximately 200 pg/ml (Fig. 4). At concentrations higher than 200 pg/ml ALLN starts precipitating. Since in the absence of ALLN, HMGal is degraded to a similar extent in both cell lines, these results indicate that the machinery responsible for the regulated degradation of HMGal was unaffected during selection for ALLN resistance. Therefore, the higher IC,, for inhibition of HMGal degradation reflects lower steady-state concentrations of ALLN in ALLN'"" cells. This