Wild-type p53-mediated induction of rat mdr1b expression by the anticancer drug daunorubicin.

The expression of P-glycoproteins encoded by the mdr gene family is associated with the emergence of the multidrug resistance phenotype in animal cells. mdr expression can be induced by many extracellular stimulants including cytotoxic drugs and chemical carcinogens. However, little is known about the mechanisms involved. Here, we report that the expression of the rat mdr1b can be induced by anticancer drug daunorubicin. Further analysis identified a bona fide p53-binding site spanning from base pairs -199 to -180 (5'-GAACATGTAGAGACATGTCT-3') in the rat mdr1b promoter that is essential for basal and daunorubicin-inducible promoter activities. In addition, our results show that wild-type p53 can up-regulate not only the promoter function but also endogenous expression of the rat mdr1b. To the best of our knowledge, this is the first report showing that a specific p53-binding site is involved in the transcriptional regulation of mdr gene by wild-type p53. Since p53 is a sensor for a wide variety of genotoxic stresses, our finding has broad implications for understanding the mechanisms involved in the inducible expression of mdr gene by anticancer drugs, chemical carcinogens, UV light, and other DNA-damaging agents.

Multidrug resistance (MDR), 1 a major obstacle to the effective chemotherapy of many human malignancies, is characterized by the increased survival of cells in the presence of cytotoxic drugs with unrelated structures. A major mechanism for the development of MDR phenotype is overexpression of Pglycoproteins which are encoded by the MDR gene family (for reviews, see Refs. 1 and 2). The MDR gene family contains two members in humans and three in rodents. However, only one human (MDR1) and two rodent (mdr1a and mdr1b) mdr genes are functionally related to the MDR phenotype. High mdr mRNA levels are seen in certain tumor types before chemotherapy and, in some cases, are associated with relapse following chemotherapy (for reviews, see Refs. 1 and 3).
Increased mdr gene expression occurs in cultured cells selected by continuous exposure to both anticancer drugs and other cytotoxic agents, in which gene amplification is believed to be often associated with the overexpression of mdr genes (4,5). However, increased mdr gene expression preceding gene amplification has been observed in early passages of drugselected cells (6). Transient exposure of cells to different cytotoxic agents such as antitumor drugs (7)(8)(9)(10), chemical carcinogens (11)(12)(13)(14)(15)(16)(17)(18)(19), and UV irradiation (20), etc. is also able to activate mdr expression, indicating that increased mdr expression is mediated by complex mechanisms.
The precise mechanisms of the induction of mdr gene expression by anticancer drugs, chemical carcinogens, UV, and other DNA-damaging agents remain unknown. It has been suggested that both post-transcriptional and transcriptional mechanisms are involved (7). A possible role for the cytoskeleton in posttranscriptional stabilization of mdr1 mRNA in rat hepatocytes treated with certain agents was suggested (21). On the other hand, in rat liver cells, it was found that doxorubicin-mediated mdr1 mRNA induction was fully inhibited by actinomycin D, suggesting that transcriptional regulation is involved (10). Nuclear run-off and transfection analyses showed that AAF-, methylcholanthrene-, aflatoxin B1-, methyl methanesulfonate-, or mitoxantrone-induced mdr1 expression is also associated with increased rates of transcription (9,11,15).
Here, we show that the expression of the rat mdr1b can be induced by anticancer drug daunorubicin. Further analysis demonstrates that a bona fide p53-binding site (5Ј-GAACATG-TAGAGACATGTCT-3Ј) located within bp Ϫ199 to Ϫ180 of rat mdr1b promoter is essential for not only basal but also daunorubicin-inducible promoter functions. We also provide evidence indicating that both the promoter activity and endogenous expression of the rat mdr1b could be modulated by wild-type p53. Although the modulation of mdr expression by either mutant or wild-type p53 has been noted, no p53-binding sites have been identified previously (22)(23)(24)(25)(26)(27). The present report represents the first evidence that a specific p53-binding site is involved in the transcriptional regulation of the mdr gene. Since p53 is responsive to a variety of genotoxic stresses (for reviews, see Refs. 28 and 29), which also induce mdr gene expression, our finding has important implications for understanding mechanisms involved in the inducible expression of drug-resistant genes by DNA-damaging agents.
Cell Culture, DNA Transfection, and Chloramphenicol Acetyltransferase (CAT) Assay-The rat hepatoma H-4-II-E cells were purchased from the American Type Culture Collection (ATCC 1548). Human osteosarcoma SAOS-2 cells, low-passage rat embryonic fibroblasts (REFs), A1-5 cells, and T101-4 cells were generously provided by Dr. G. Lozano. All the cell lines were maintained in Dulbecco's modified Eagle's medium (Sigma) supplemented with 10% fetal calf serum (Life Technologies, Inc.), 1 mM glutamine, and 50 g of neomycin/ml in a humidified incubator containing 5% CO 2 . Prior to treatment, cells were grown in the medium to 70 -80% confluence. Then cells were treated with daunorubicin (7 g/ml) for various periods of time and harvested for the preparation of nuclear extracts and RNA.
The calcium phosphate precipitation method (33) was used to transfect cells with DNA. In brief, 2 h before transfection, cells in the exponential growth phase (approximately 70 -80% confluence) were plated in Corning six-well plates. DNA-CaPO 4 precipitate was added to the medium and incubated for 5-6 h. After cells were shocked with 15% glycerol for 30 s, washed with phosphate-buffered saline, cells were grown in Dulbecco's modified Eagle's medium containing 10% fetal calf serum. Stable transfectant A1-5 and H-4-II-E cells were established by co-transfecting the cells with the reporter constructs and pcDNA3 (Introgen, Carlsbad, CA) at 5:1 ratio followed by selection with G418 (400 g/ml, Life Technologies, Inc.). Pools of G418-resistant cells were collected and used for further analysis. In transient transfection assays, cells were directly treated with daunorubicin (7 g/ml) 24 h after transfection. After 20 -24 h of drug exposure, cells were harvested. CAT activities in the cell extracts were measured by a previously described method (34) using total protein extract (measured by the Bio-Rad protein assay kit) as a reference. Relative CAT activity levels were calculated by a PhosphorImager (model 400S, Molecular Dynamics) in terms of the conversion of [ 14 C]chloramphenicol into acetylated chloramphenicol.
Preparation of Nuclear Extracts and GMSA-Nuclear extracts were prepared from H-4-II-E cells by the method of Digman et al. (35) with modifications as described previously (31). GMSAs were performed with approximately 5 g of nuclear proteins in a total volume of 20 l of binding mixture containing 10 mM Tris-HCl, pH 7.5, 50 mM NaCl, 0.5 mM dithiothreitol, 10% glycerol, 0.2% Nonidet P-40, 3 g of poly(dI-dC)⅐poly(dI-dC), and radiolabeled DNA probe at room temperature for 20 min.
RNase Protection Assay-A 162-nucleotide antisense RNA probe (Ϫ37 to ϩ125) was synthesized using T7 RNA polymerase as described previously (30). Of total RNA from cells, either 20 g (for the mdr1b probe) or 1 g (for the 18 S rRNA probe) was hybridized with 32 Plabeled antisense RNA probes (2 ϫ 10 5 cpm) and subjected to RNase protection assays as described previously (17,30). The protected RNA products were analyzed on a 7% denaturing polyacrylamide gel and quantified using a Personal Densitometer SI (Molecular Dynamics).
Reverse Transcriptase-PCR Amplification and DNA Sequencing-Two micrograms of total RNA isolated from H-4-II-E cells was used for reverse transcriptase reaction. On completion of the reverse transcriptase reaction, the enzyme was inactivated by heating to 94°C for 56 min. Ten picomoles of each 5Ј primer (5Ј-CCTGAAGACTGGATA-ACTGTCATGGAGGAT) and 3Ј primer (5Ј-AGAGGGGGCCGAGTAC-TATCTACAAGGTAA) were used in a PCR to amplify rat p53 cDNA (30 cycles of 1 min at 94°C, 2 min at 45°C, and 3 min at 72°C). PCR products were electrophoresed on an agarose gel, purified, and subjected to automated sequencing (ABI PRISM).

Daunorubicin Induces mdr1b Expression in Rat Hepatoma
Cells-To investigate whether the expression of the rat mdr1b gene expression is regulated by the anticancer drug daunorubicin, we treated rat hepatoma H-4-II-E cells with daunorubicin (7 g/ml). At various time intervals, cells were harvested and mdr1b mRNA levels were measured by the RNase protection assay. As shown in Fig. 1, the steady-state mdr1b mRNA levels in these cells were elevated after exposure to daunorubicin for 12 and 24 h. Increases of about 3-5-fold were seen in three independent experiments. Similar results were obtained in cells treated with adriamycin and chemical carcinogen 2-AAAF (data not shown). These results demonstrated that rat mdr1b expression can be induced by these cytotoxic agents in rodent cells.
Rat mdr1b Promoter Responds to Daunorubicin Treatment-To investigate the possible involvement of transcriptional regulation in the induction of the rat mdr1b gene expression by daunorubicin and, if so, to identify DNA sequences responsible for the daunorubicin induction of mdr1b expression, we generated a set of 5Ј deletion mutant CAT constructs and transfected them into H-4-II-E cells following treatment with or without daunorubicin. When Ϫ1288 RMICAT, Ϫ243 RMICAT, and Ϫ214 RMICAT reporter constructs containing 1288, 243, and 214 bp of the rat mdr1b upstream sequences, respectively, plus 125 bp downstream from the transcription start site, were transiently transfected into H-4-II-E cells, CAT activities increased an approximately 1.6 -1.9-fold in daunorubicin treated versus untreated cells (p Ͻ 0.05) (Fig. 2). However, when Ϫ163 RMICAT, which contains additional deletion to Ϫ163 bp was transfected, basal transcriptional activities were reduced more than 80%. More importantly, the deletion also abolished daunorubicin inducibility (Fig. 2). Together, RNase protection analyses of rat mdr1b transcripts in H-4-II-E cells. Cells were treated with daunorubicin (7 g/ml) for different times as indicated. RNA was extracted and then subjected to RNase protection assays as described under "Materials and Methods." An 18 S rRNA probe was used as a reference. The autoradiograph is representative of results from three independent experiments.

p53-mediated Regulation of Rat mdr1b Expression
these results indicated that the rat mdr1b promoter can respond to daunorubicin treatment and that the sequence from bp Ϫ214 to Ϫ163 is essential for the promoter's daunorubicin responsiveness.
We previously identified a NF-B-binding site (bp Ϫ167 to Ϫ158) involved in basal and insulin-induced promoter function (31). Since it was reported that daunorubicin can induce the NF-B activity in human fibrosarcoma HT1080 cells and HL-60 promyelocytes (36,37), it is possible that NF-B was also responsible for the inducible promoter activity of the rat mdr1b by daunorubicin in the H-4-II-E cells. To test this possibility, we transfected Ϫ243 RMICAT-m, in which the NF-B-binding site was mutated (31), into H-4-II-E cells following daunorubicin treatment. As shown in Fig. 2, although basal activity was reduced when compared with the wild-type Ϫ243 RMICAT, Ϫ243 RMICAT-m still retained daunorubicin responsiveness. Besides, we did not observe an obvious increase of NF-B binding activity after daunorubicin treatment using GMSAs (data not shown). These results suggested that NF-B may not be directly involved in the daunorubicin-inducible promoter function of the rat mdr1b in H-4-II-E cells.
Ϫ214 to Ϫ177 bp Is Sufficient to Confer mdr1b Promoter Inducibility by Daunorubicin-To further substantiate the above observations, we generated two additional constructs by inserting sequences from bp Ϫ214 to Ϫ127 (containing NF-B site) or Ϫ214 to Ϫ177 (containing no NF-B site), respectively, into a pBLCAT 2 vector containing the tk basal promoter and a CAT reporter gene. These constructs were then transiently transfected into H-4-II-E cells, and CAT expression was measured. As shown in Fig. 3, both constructs are capable of responding to daunorubicin treatment, giving rise to comparable levels of induction, whereas daunorubicin did not have effects on the tk promoter. These results suggested that NF-B site is dispensable for the inducibility of mdr1b promoter, and that the sequence from bp Ϫ214 to Ϫ177 may contain important cis-acting elements responsible for the induction of mdr1b promoter activity by daunorubicin.
Daunorubicin Induces Formation of a Specific Protein-DNA Complex Within bp Ϫ201 to Ϫ177-To test whether daunorubicin treatment could induce protein DNA binding at sequences within bp Ϫ214 to Ϫ177, we prepared nuclear extracts from H-4-II-E cells treated with or without daunorubicin and performed GMSAs. As shown in Fig. 4A, a major DNA-protein complex was formed in the daunorubicin-untreated nuclear extracts when a double-stranded oligonucleotide spanning bp Ϫ214 to Ϫ177 was used as the probe (lane 1, C1). The binding activity of this complex remained largely unchanged after daunorubicin treatment (lanes 2-5 versus lane 1). However, a slow migrating protein-DNA complex was induced 1 h after treatment (C2, lane 2). The binding activity of this induced complex remained elevated but gradually reduced throughout the 12-h induction period (lanes 2-5).
To further characterize the sequence specificity of this daunorubicin-induced DNA binding activity, double-stranded oligonucleotides covering the left and right regions of bp Ϫ214 to Ϫ177 (fragments A and B, Fig. 4C) were used in competition GMSA. Fig. 4B shows that both the unlabeled probe (lane 2) and fragment B (bp Ϫ201 to Ϫ177) (lane 4) could compete for daunorubicin-induced DNA-protein complex, indicating that the induced protein binding required the sequence residing within fragment B.
To further define the binding sequence of the induced protein complex, two site-directed mutant oligonucleotides (M1 and M2) containing mutations on 5Ј-or 3Ј-end of fragment B, respectively ( Fig. 4C), were used as competitors. However, neither mutated oligonucleotides could compete for daunorubicininduced DNA-protein complex (Fig. 4B, lanes 5-6), suggesting that the daunorubicin-induced protein binding required both the 5Ј-and 3Ј-sequences of fragment B (bp Ϫ201 to Ϫ177).
Daunorubicin-induced DNA-binding Protein Is p53-In examining the DNA sequence of bp Ϫ201 to Ϫ177 (fragment B), we found within it a sequence (bp Ϫ199 to Ϫ180) strikingly similar to the p53-binding consensus sequence 5Ј-PuPuPuC(A/ T)(A/T)GPyPyPy-3Ј (38), with only 2 base pair mismatches. A comparison of the putative mdr1b p53-binding site with the p53 consensus sequence and the p53-binding site from the gadd45 third intron (39) is shown in Fig. 5B. To determine whether the sequence located between bp Ϫ199 and Ϫ180 was indeed a p53-binding site, we carried out GMSAs using a double-stranded oligonucleotide spanning bp Ϫ214 to Ϫ177 as the probe and nuclear extracts prepared from daunorubicintreated H-4-II-E cells in the presence of various unlabeled oligonucleotides as competitors. As shown in Fig. 5A To further strengthen this observation, antibodies were used in GMSAs. As shown in Fig. 5A, the daunorubicin-induced protein-DNA complex was supershifed by p53 antibody PAb421 (lane 6), whereas c-Jun and NF-B p65 antibodies did not affect the formation of induced DNA-protein complexes (lanes 7-8). Taken together, these results strongly suggested that the rat mdr1b promoter sequence located between bp Ϫ199 and Ϫ180 (5Ј-GAACATGTAGAGACATGTCT-3Ј) is a p53-binding site.
To determine whether the rat p53 in H-4-II-E cells is a wild-type or mutant form, we amplified cDNA copies of the rat p53 by reverse transcriptase-PCR and sequenced it directly (see "Materials and Methods"). The result showed that H-4-II-E cells has a wild-type rat p53 mRNA (data not shown) with a sequence consistent with that published previously (42). p53-binding Site Is Required for the Daunorubicin-inducible promoter Activities-To characterize the functional role of the FIG. 3. Sequences from bp ؊214 to ؊177 confer the inducibility by daunorubicin. H-4-II-E cells were transfected with Ϫ214/Ϫ127 tk-CAT and Ϫ214/Ϫ177 tk-CAT constructs, and CAT activity was assayed as described under "Materials and Methods." Values shown are the averages for three representative experiments in which each transfection was performed in duplicate. S.D. values are represented by the bars. 2 g of DNA was used in each transfection in the absence or presence of daunorubicin (7 g/ ml), and the extracts were normalized to the same protein concentration. In the schematic diagram, the position of NF-B-binding site is indicated.

FIG. 4. Induction of a protein-DNA binding complex by daunorubicin.
A, GMSA using 5 g of nuclear extracts from H-4-II-E cells treated with or without daunorubicin (7 g/ml) for the indicated times. Extracts were assayed for binding to the labeled double-strand oligonucleotide (bp Ϫ214 to Ϫ177) shown in panel C. Note that a new DNA-protein complex (C2) was induced in daunorubicin-induced nuclear extracts (lanes 2-5). B, GMSA using nuclear extracts prepared from H-4-II-E cells treated with daunorubicin (7 g/ml) for 6 h. Extracts were assayed for binding to the labeled double-stranded Ϫ214 to Ϫ177 oligonucleotide in the presence or absence of an 100-fold molar excess of the indicated competitors shown in panel C. The autoradiographs are representative of the results from three independent experiments. C, the mdr1b promoter sequence from bp Ϫ214 to Ϫ177 and oligonucleotides used for competition in panel B. In oligonucleotides, mutated bases are indicated in boldface.

FIG. 5. Identification of a p53-binding site in the rat mdr1b
promoter. A, GMSA using nuclear extracts prepared from H-4-II-E cells treated with daunorubicin (7 g/ml) for 6 h. Extracts were assayed for binding to the labeled double-strand oligonucleotide (bp Ϫ214 to Ϫ177) in the presence or absence of an 100-fold molar excess of the indicated competitors (lanes 2-5), p53 (PAb421), anti-c-Jun, and anti-p65 antibodies (lanes 6 -8), respectively. Note that the slower migrating complex was completed by gadd45 p53-binding sequence (lane 2) and supershifted by PAb421 (lane 6). The autoradiograph is representative of the results from three independent experiments. B, comparison of the sequence from bp Ϫ199 to Ϫ180 with the p53 consensus sequence and the p53-binding site from the gadd45 third intron. Two mismatch base pairs are indicated in lowercase. In the schematic diagram, P represents G or A; W represents A or T; Y represents T or C.

p53-mediated Regulation of Rat mdr1b Expression
identified p53-binding site, the same mutations in Fig. 4C were introduced within the context of the wild-type Ϫ214 RMICAT construct, and resultant recombinants were designated Ϫ214 RMICAT-m1 and Ϫ214 RMICAT-m2. These mutant constructs were then transfected into H-II-4-E cells following treatment with or without daunorubicin. As shown in Fig. 6A, both mutations abolished the daunorubicin responsiveness. Similar results were obtained when the same mutations were introduced into heterologous (tk) promoter constructs (Fig. 6B). These results suggested that the integrity of p53 binding is essential for the daunorubicin inducible-promoter function of the rat mdr1b.
It has been reported that promoters containing p53-binding sites, in some cases, essentially showed no obvious DNA damage responsiveness in transient transfection assays after the treatment of UV or other DNA-damaging agents, whereas higher levels of the induction of the same reporters were seen in stable transfectants (43). Consistent with these observed only low levels of inductions of mdr1b CAT activities by daunorubicin were observed in our transient transfection assays (Figs. 2, 3, 6, A and B). To test whether the rat mdr1b promoter can respond to daunorubicin more dramatically in stable transfectants than in transient transfected cells, we stably transfected both Ϫ214 RMICAT and Ϫ214 RMICAT-m1 into H-4-II-E cells. As expected, wild-type CAT reporter (Ϫ214 RMICAT) exhibited more significant responsiveness to daunorubicin (4-fold, Fig. 6C), which is comparable with the induction levels of mdr1b mRNA by daunorubicin (Fig. 1). As a control, Ϫ214 RMICAT-m1 in stably transfected H-4-II-E cells failed to respond to daunorubicin (Fig. 6C). These results further strengthened the notion that the p53-binding site is required for the promoter's daunorubicin responsiveness. Why the fold induction is different between transient and stable transfectants is unclear but could be due to the participation of chromatin proteins or structure in p53-mediated gene expression, since studies have indicated that transiently transfected DNA, unlike stably transfected templates, are not efficiently packed into chromatin (44). Consistent with this finding, a recent report showed that high mobility group protein-1, an important component of chromatin, is a coactivator of p53 (45).
In another set of experiments, wild-type p53 (pCMVp53) or mutant p53 (pCMVp53 248 ) expression vectors were co-transfected with reporter constructs into SAOS-2 cells, which contain a homozygous deletion at the p53 gene locus and do not produce a p53 protein (47). As shown in Fig. 7B, co-transfection   FIG. 6. Requirement of p53-binding site for both basal and daunorubicininducible promoter activities. A, CAT assays of wild-type Ϫ214 RMICAT, mutants Ϫ214 RMICAT-m1, and Ϫ214 RMI-CAT-m2 which contain mutant p53-binding sites (see schematic at left) were transiently transfected into H-4-II-E cells following treatments with or without daunorubicin (7 g/ml). In the schematic diagram, mutated bases are indicated in boldface. B, CAT assays of Ϫ214/Ϫ177 tk-CAT wild-type and mutant constructs after transient transfection into H-4-II-E cells following treatment with or without daunorubicin (7 g/ml). Mutated bases (indicated by X) are the same as shown in panel A. 2 g of DNA was used in each transfection. Results shown are the averages for three representative experiments after normalization to the protein concentration of the cellular extracts. S.D. values are represented by the bars. C, CAT assay of Ϫ214 RMICAT and Ϫ214 RMI-CAT-m1 after stable transfection into H-4-II-E cells in the presence (ϩ) or absence (Ϫ) of daunorubicin (Dau) (7 g/ml). The autoradiograph shown is a representative of the results from one of three independent pools for each stable Ϫ214 RMICAT and Ϫ214 RMICAT-m1 cell line. Fold induction refers to that in transfectants not treated with daunorubicin.
p53-mediated Regulation of Rat mdr1b Expression of pCMVp53 trans-activated CAT activity in cells transfected with the wild-type reporter (Ϫ214 RMICAT) but not in cells co-transfected with reporters containing mutated p53 site (Ϫ214 RMICAT-m1 or Ϫ214 RMICAT-m2). Moreover, co-transfection of mutant p53 expression vector also failed to activate wild-type as well as mutant reporters (Fig. 7B). Similar results were obtained when heterologous reporter constructs (Ϫ214/ Ϫ177 tk-CAT and mutant Ϫ214/Ϫ177 tk-CAT-m2) were used in co-transfection assays (Fig. 7C). These results, collectively, demonstrated that wild-type p53 can trans-activate rat mdr1b promoter activity specifically via the identified p53-binding site.
Endogenous mdr1b Expression Is Modulated by Wild-type p53-To assess the regulation of endogenous mdr1b expression by p53, we examined mdr1b mRNA levels following temperature shift in A1-5 cells. We reasoned that if the mdr1b is a true p53 target gene, its expression should increase following wildtype p53 induction after temperature shift. As a control, we also measured mRNA levels in REFs and T101-4 cells following temperature shift. REFs has an endogenous wild-type p53, whereas T101-4 cells, like A1-5 cells, are derived from REFs but carry a non-temperature-sensitive p53 mutant (46). RNase protection assays revealed mdr1b mRNA levels increased a 3-6-fold in A1-5 cells after temperature shift from 37°C (mutant p53) to 32.5°C (wild-type p53) (Fig. 8, compare lanes 3 and  4 to 1 and 2). This induction was unlikely due to a nonspecific phenomenon by the temperature change, because no induction of mdr1b expression was observed in control REFs or T101-4 cells (compare lanes 7 and 8 to 5 and 6, and 11 and 12 to 9 and 10). These results indicated that the up-regulation of the mdr1b in A1-5 cells after temperature shift was induced by wild-type p53, and that p53 is indeed capable of modulating endogenous mdr1b gene expression.

DISCUSSION
In this study, we identified an authentic p53-binding site located from bp Ϫ199 to Ϫ180 of the rat mdr1b that is important for both basal and daunorubicin-inducible promoter activities. We also provided evidence showing that both the promoter function and endogenous expression of the rat mdr1b can be modulated by wild-type p53. A bona fide p53 response gene should fit the following criteria (28): (i) the existence of p53-binding sites that can be specifically recognized by p53; (ii) the ability of these sites to act as a p53 response element, activating basal transcription in a wild-type p53-dependent manner; (iii) the response of the element to p53 in the endogenous genomic promoter context; and (iv) the induction of the target genes after cellular stress, such as DNA damage, in cells containing wild-type but not mutant forms of p53. The results presented in this study suggested that the rat mdr1b meet all these criteria. Therefore, like p21/WAF1, mdm-2, GADD45, cyclin G, bax, and IGF-BP3, etc., the rat mdr1b can be considered as a genuine p53 response gene.
Studies on the role of p53 in the regulation of the mdr gene family have been quite controversial. Previous studies have demonstrated that co-transfection with several mutant p53 expression vectors activated the human MDR1 and hamster pgp-1 promoters, whereas co-transfection of a wild-type p53 expression vector had no effect or repressed the promoter activity (22)(23)(24)(25)(26). Yet, no bona fide p53-binding sites were elucidated in these studies. It has been shown that p53 can indeed repress activities of many promoters without specific p53-binding sites (for review, see Ref. 28). The repression usually involves promoters containing the TATA box, presumably through sequestering TATA-binding protein, transcription activating factors, or interacting with other transcriptional activators by p53. Paradoxically, the human MDR1 promoter is a TATA-less promoter, therefore mechanisms involved in the repression of the MDR1 promoter by wild-type p53 are un- FIG. 7. Activation of mdr1b promoter by wild-type p53 but not mutant p53. A, CAT assay of Ϫ214 RMICAT and Ϫ214 RMICAT-m1 after stable transfection into A1-5 cells. Cells cultured at 37°C were either shifted to 32.5°C or continuously cultured at 37°C for 24 h and then harvested for CAT assays. The autoradiograph shown is representative of the results from one of three independent pools for each cell line. Fold induction refers to that in cells cultured at 37°C. B and C, p53-null SAOS-2 cells were transfected with 2 g of wild-type (Ϫ214 RMICAT or Ϫ214/Ϫ177 tk-CAT), p53-binding site-mutated (Ϫ214 RMI-CAT-m1, -m2, or Ϫ214/Ϫ177 tk-CAT-m2) mdr1b promoter reporter alone or in combination with 1 g of wild-type p53 (pCMVp53) or mutant p53 expression vector (pCMVp53 248 ) as indicated. Empty control vector (pCMV) was used to normalize the amounts of the transfected DNA to a total 3 g of DNA in each transfection reaction. Each column represents the mean of relative CAT activities from three independent experiments after normalization to the protein concentration of the cellular extracts. S.D. values are represented by the bars. p53-mediated Regulation of Rat mdr1b Expression known (26). Similarly, it is also unclear how the p53 mutants gain the functions to activate the human MDR1 promoter (24). In addition to repressing it, wild-type p53 was also shown to stimulate the MDR1 promoter in p53-null cell line in a transfection assay (27). The reasons for the discrepancies among these studies are still unknown but there are many plausible explanations: (i) p53 is a multiple functional protein whose functions are regulated by a complex network (48), its regulation of gene expression may differ not only among cell types but also among physiological conditions under which assays are performed; (ii) p53 can also bind transcriptional coactivators such as CBP/p300 (49 -51), which interacts with a battery of other transcriptional regulators such as NF-B, Jun/Fos, nuclear receptors, and their coactivators (for review, see Ref. 52). The abundance of these transcriptional regulators may differ among different cell settings and thereby influence the overall expression of transfected genes; (iii) different p53 expression vectors, mdr reporter constructs, and time of analysis, may affect the overall results. It should also be noted that even the transfection procedures themselves may perturb endogenous p53 levels (53), affecting results of transient transfection assays. These considerations, taken together, may explain the discrepancy of the transfection results described above. In this regard, the identification of an authentic p53-binding site in the mdr1b promoter region as described herein is of particular importance, since it is the first time a specific p53-binding site was elucidated to be implicated in the transcriptional regulation of mdr gene expression.
Our observation of the involvement of wild-type p53 but not mutant p53 in the regulation of the rat mdr1b expression may be relevant to the increased expression of the mdr1b during hepatocarcinogenesis. Although the expression of mdr1 is highly activated, mutation of p53 does not always occur during hepatocarcinogenesis, at least in its early stage of liver tumor development (54). In addition, it has been known that the mdr1b expression in rat liver can be rapidly activated by chemical carcinogens such as 2-AAF and aflatoxin B1 (12,13), however, in rat hepatocellular carcinomas induced by these carcinogens, p53 mutations do not always occur (55,56). More importantly, van Gijssel et al. (57) recently reported that p53 activity can be also induced by 2-AAF and aflatoxin B1 in rat liver. When rat hepatoma H-4-II-E cells (contain wild-type p53) were treated with 2-AAAF, p53 activity was also been induced. 2 These results, taken together, suggested that the activation of the rat mdr1b during chemical hepatocarcinogenesis may be due to the elevated wild-type p53 activities.
In broader prospects, it has been known that p53 is a universal sensor of genotoxic stress (58), and can be induced by a wide variety of DNA-damaging agents such as UV, ␥-irradiation, carcinogens, and cytotoxic drugs (for reviews, see Refs. 28 and 29). Strikingly, many of these agents are also known inducers of mdr gene expression, suggesting that p53 may contribute to the induction.
The p53-binding site identified in this study lies in the previously identified murine mdr1b enhancer region (59). It overlaps a palindromic sequence recognized by two peptides (41 and 49 kDa) (30), and adjoins a downstream NF-B-binding site which is also important for the promoter function (31). It is believed that most inducible cis-acting elements contain multiple, distinct transcription factor-binding sites that are part of a combinatorial mechanism that relies on cooperative binding, interaction of transcriptional activator proteins, and transcriptional synergy (60). Our study of site-directed mutations demonstrated that the full promoter activity of the rat mdr1b requires the integrity of both the p53-binding site (bp Ϫ199 to Ϫ180) and NF-B-binding site (bp Ϫ167 to Ϫ158) (Fig. 2) (31), suggesting that cooperative mechanisms between these two cis-acting elements are implicated in the regulation of the rat mdr1b expression. More recently, coactivator CBP/p300 was shown to interact with both p53 (49 -51) and NF-B (61,62), and enhance p53-and NF-B-dependent transactivation, respectively. The activity of the rat mdr1b promoter was also found to be enhanced by CBP/p300. 2 Taken together, these may suggest that the binding of p53 and NF-B to the mdr1b promoter may recruit CBP/p300 and basal transcriptional machinery to form a higher order transcription enhancer complex, similar to that proposed in interferon-␤ and E-selectin promoters (61,63), which modulates inducible expression of the rat mdr1b. However, since our knowledge is rather limited at this moment, the validity of this model still needs to be further tested.
Finally, we would like to stress that, although our present results clearly demonstrated the direct involvement of p53 in the rat mdr1b gene regulation, the roles of p53 in the evolution of drug resistance in cancers remain to be critically evaluated.
In clinical setting, the loss of functional p53 has been reported to be well correlated with de novo resistance to radiation and anticancer drugs, and some tumors with wild-type p53 respond well to chemotherapeutic drugs (Refs. 64 -66, for review, see Ref. 29). However, it is unknown whether the correlation of drug resistance and p53 mutations is directly due to the activation of mdr by mutant p53, or other mechanisms such as alterations in drug targets, transporters, metabolisms, or the expression of genes regulating cell death and/or survival. Further studies are required to elucidate the molecular insights into how p53 regulates clinical drug sensitivity in cancer chemotherapy.