Repression of Human Reduced Folate Carrier Gene Expression by Wild Type p53*

The relationship between loss of functional p53 and human reduced folate carrier (hRFC) levels and function was examined in REH lymphoblastic leukemia cells, which express wild type p53, and in p53-null K562 cells (K562pTet-on/p53) engineered to express wild type p53 under control of a tetracycline-inducible promoter. Activation of p53 in REH cells by treatment with daunorubicin was accompanied by decreased (∼5-fold) levels of hRFC transcripts and methotrexate transport. Treatment of K562pTet-on/p53 cells with doxycycline resulted in a dose-dependent expression of p53 protein and transcripts, increased p21 protein, decreased dihydrofolate reductase, and G1 arrest with decreased numbers of cells in S-phase. p53 induction was accompanied by up to 3-fold decreases in hRFC transcripts transcribed from the upstream hRFC-B promoter and similar losses of hRFC protein and methotrexate uptake capacity. Expression of p15 in an analogous inducible system in K562 cells resulted in a nearly identical decrease of S-phase cells and dihydrofolate reductase without effects on hRFC levels or activity. When the hRFC-B promoter was expressed as full-length and basal promoter-luciferase reporter constructs in K562pTet-on/p53 cells, induction of p53 with doxycycline resulted in a 3-fold loss of promoter activity, which was reversed by cotransfection with a trans-dominant-negative p53. These studies show that wild type p53 acts as a repressor of hRFC gene expression, via a mechanism that is independent of its effects on cell cycle progression.

Although one of the major roles of p53 is to stimulate apoptosis, the relationships between wild type p53 and sensitivity to chemotherapeutic drugs are frequently complex and contradictory. A number of studies suggest that loss of wild type p53 function leads to resistance to chemotherapy and radiation therapy due to decreased susceptibility to apoptosis (24), however, this has not always been found to be the case. For example, loss of wild type p53 has been reported to sensitize mammalian cells to taxol (25) and cisplatin (26). Furthermore, the anti-purine inhibitor, AG2034, was selectively cytotoxic toward cells with nonfunctional p53 that were lacking a functional G 1 checkpoint (27). In H35 rat hepatoma cells expressing a transdominant-negative p53, levels of P-glycoprotein and mdr1a transcripts were markedly elevated, resulting in resistance to P-glycoprotein substrates (16). Mtx sensitivity was reported to be increased in H35 cells characterized by loss of wild type p53 function (16). However, the basis for this finding was not considered further.
Mtx and related antifolates (Tomudex, 5,10-dideaza-5,6,7,8tetrahydrofolate) are transported into cells by the RFC (28,29). For Mtx, RFC-mediated transport is a key element in antitumor activity, reflecting its role in generating intracellular unbound Mtx for complete inhibition of DHFR and for Mtx polyglutamate synthesis (28,29). Given our interest in the mechanisms of transcriptional control of RFC, we began a systematic investigation to explore the possibility that wild type p53 may regulate RFC. Accordingly, we expressed wild type p53 in p53-null K562 cells under control of a tetracyclineinducible promoter. As described herein, in this system, wild type p53 acts as a potent repressor of hRFC gene expression, via a mechanism that seems to be independent of its wellestablished effects on cell cycle progression. Rather, decreased hRFC levels and function upon p53 induction are likely due to an inactivation of transcription from the upstream hRFC-B promoter (30 -32). We also show that induction of endogenous wild type p53 in REH leukemia cells is accompanied by a dramatic decrease in hRFC expression and Mtx transport. Cell Culture-K562 human erythroleukemia cells and REH B-lymphoblastic leukemia cells were purchased from the American Type Culture Collection (Rockville, MD). K562 and REH cells were grown in RPMI 1640 medium (Sigma Chemical Co.) containing 10% heat-inactivated iron-supplemented calf serum (Hyclone), 2 mM L-glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin. Cells were cultured in a 37°C humidified atmosphere containing 5% CO 2 /95% air. For cytotoxicity determinations, cells were cultured in complete RPMI 1640 medium containing 10% dialyzed fetal bovine serum. Cell numbers were determined after equivalent numbers of doublings by direct microscopic counting with a hemacytometer.

Chemicals and Reagents-[␣-
Expression of an Inducible p53 Construct in K562 Cells-A commercial "Tet-on" (CLONTECH) kit was used to express wild type p53 in p53-null K562 cells under control of a tetracycline/Dox-inducible promoter (37). The K562 pTet-on cell line was developed from wild type K562 cells by transfection with the regulator pTet-on plasmid (5 g) using Lipofectin (Life Technologies, Inc.), as instructed by the manufacturer. Colonies were isolated in soft agar in the presence of 1 mg/ml G418 (38). G418-resistant clones were expanded and screened by transient transfection with a pTRE-luc reporter plasmid. For reporter gene assays, cells were transfected with a 2.5 g of pTRE-luc plasmid and treated with and without Dox (2 g/ml) for 24 h, followed by additional incubation for 48 h. The cell lysates were prepared and tested for luciferase activity, as described below.
To construct the pTRE/p53 plasmid, wild type p53 was cloned into a pTRE response plasmid at the BamHI site. The plasmid was then cotransfected into the K562 pTet-on cell line with the pcDNA3.1/Zeo plasmid (0.25 g, Invitrogen). Colonies were isolated in soft agar, expanded in the presence of 1 mg/ml G418 and 150 g/ml zeocin, and tested for p53 expression in the presence and absence of Dox (2 g/ml) by Western blotting, as described below. A "double-stable" resistant clone (designated K562 pTet-on/p53 ), which exhibited no background and a high level of Dox-induced p53 expression, was identified and used for further analysis. K562 pTRE "vector control" cells were generated by transfecting K562 pTet-on cells with an empty pTRE vector. An analogous p15 "Tet-on"-inducible cell line (K562p15 Tet-on -8) was prepared by transfecting p15-and p16-null K562 pTet-on cells with a pTRE/p15 plasmid construct. 2 Western Blot Analysis of p53, p21, DHFR, and hRFC-Soluble protein extracts were prepared by sonication in hypotonic buffer (10 mM Tris-Cl, pH 7.0) containing 1% SDS and proteolytic inhibitors and subjected to SDS-PAGE (39). p53, p21, and DHFR proteins were electrophoretically transferred to PVDF membranes (DuPont (40)) and immunoblotted with anti-p53 monoclonal (DO-1) antibody (Oncogene Research Products), p21 waf1 antibody (Oncogene Research Products), and antiserum to purified recombinant human DHFR, respectively. Immunoreactive proteins were visualized by enhanced chemiluminescence (Pierce, Rockford, IL).
For hRFC protein expression, sucrose-gradient-purified plasma membranes were prepared as previously described (41)(42)(43). Membrane proteins were solubilized in 10 mM Tris-Cl (pH 7.0) hypotonic buffer and 1% SDS in the presence of proteolytic inhibitors (41)(42)(43). Aliquots (100 g) were subjected to a 7.5% SDS-PAGE and transferred to PVDF membranes followed by immunoblotting. hRFC proteins were detected by chemiluminescence (Lumi Light, Roche Molecular Biochemicals) using protein A-purified glutathione S-transferase-hRFC antibody (43). All antibody incubations were performed in Tween 20/Tris-buffered saline that was supplemented with 0.5% nonfat milk. Light emission was recorded on x-ray film with various exposure times, and the signal was quantitated on a Kodak Digital Science 440CF Image Station.
Northern Analysis-Total RNAs were isolated from cells with TRIzol reagent (Life Technologies, Inc.). Equal amounts (20 g) of total RNAs were loaded onto a 0.7% formaldehyde/agarose gel and transferred to nylon membranes (GeneScreen, PerkinElmer Life Sciences), followed by baking the membranes at 80°C until dry. The 32 P-labeled probes (i.e. hRFC, p53, and p15) were prepared by random-priming (Roche Molecular Biochemicals) using [␣-32 P]dCTP, and blots were autoradiographed after hybridization. Densitometry of autoradiographs on x-ray films was performed on the Kodak Digital Science 440CF Image Station. Equal loading of the RNAs was established by ethidium bromide staining of 28 S and 18 S RNAs or by probing with 32 P-labeled GAPDH cDNA.
Membrane Transport Methodology-Exponentially growing cells were treated with Dox over a range of concentrations. After 48 h of incubation, ϳ1 ϫ 10 7 cells were harvested for [ 3 H]Mtx (0.5 M) transport assays (41,43,44). Cells were washed twice with D-PBS and resuspended in 2 ml of Hanks' balanced salts. Cells were incubated with [ 3 H]Mtx at 37°C for 180 s. The reactions were quenched with ice-cold D-PBS, and the cells were washed three times with D-PBS. The cell pellets were dissolved in 1 ml of 0.5 N NaOH. The levels of intracellular radioactivity were expressed in units of pmol/mg of protein, calculated from direct measurements of radioactivity and protein contents of the cell homogenates as determined by the method of Lowry et al. (45).
Flow Cytometry Analysis-Approximately 1-2 ϫ 10 6 of K562 pTet-on/p53 , K562p15 Tet-on -8, or pTRE vector cells were harvested after treatment with and without Dox (2 g/ml) for 48 h, washed twice with D-PBS, and fixed in 70% ethanol. For DNA fluorescence staining, cells were stained in propidium iodide by the method of Crissman and Steinkamp (46) and analyzed by flow cytometry (Becton Dickinson FACScan). Apoptosis was assayed by annexin V staining and flow cytometry using a commercial kit (R&D Systems).

RESULTS AND DISCUSSION
Effects of Induction of Endogenous p53 on hRFC Levels in REH Cells-Loss of functional wild type p53 was originally reported to lead to enhanced Mtx sensitivity in H35 rat hepatoma cells (16). Although this was independent of changes in DHFR in this model (16), it could involve any combination of other determinants of drug response (e.g. Mtx polyglutamylation, Mtx membrane transport by RFC, or levels of thymidylate synthase (28)) and/or secondary effects on cell cycle progression, which manifest as decreased sensitivity to Mtx.
To further explore the relationship between induction of endogenous p53 and hRFC, we used the REH B-lymphoblastic leukemia subline, which expresses high levels of hRFC transcripts and protein, and wild type p53 (47). REH cells were treated with 0.5 M daunorubicin to induce p53. p53 levels increased over time and were maximal by 6 h (Fig. 1A). Notably, hRFC transcripts also decreased in a time-dependent fashion (ϳ5-fold by 9 h; Fig. 1B) and were accompanied by a similar loss of [ 3 H]Mtx transport (not shown). However, there were no changes in the levels of transcripts for the housekeeping gene, GAPDH (Fig. 1C), suggesting that the loss of hRFC in REH cells is comparatively specific. These data strongly suggest a suppression of hRFC levels and function by wild type p53.
Generation and Characterization of Double-stable K562 pTet-on/p53 Cells-To better study the relationships between p53 and hRFC levels and function, we generated a double-stable transfectant (designated K562 pTet-on/p53 ) in p53-null K562 cells, using a commercial Tet-on system (37) to express wild type p53 under control of a tetracycline/doxycy-cline-inducible promoter. We initially measured the time course induction of p53 expression following exposure to a maximum dose of Dox (2 g/ml). In the absence of Dox, p53 was undetectable in K562 pTet-on/p53 cells on Western ( Fig. 2A) or Northern blots (not shown). Following Dox treatment, we found that p53 transcripts and protein were detectable as early as 6 h and that induction by Dox was maximal after 48 h of treatment. The induction of p53 protein was dose-dependent over a wide range of Dox concentrations (0.015-2 g/ml; Fig. 2A), and p21 expression paralleled the induction of p53 on Western blots (Fig. 2B), indicating that p53 was transcriptionally active.
At 2 g/ml Dox, there was obvious G 1 arrest with a 3-fold loss in the number of cells in S-phase (Table I) and an ϳ30% decrease in growth rate (data not shown), compared with cells without Dox treatment. By annexin V staining and flow cytometry analysis, there was no detectable increase in the numbers of apoptotic cells after 48 h of Dox treatment (not shown), indicating that expression of high levels of p53 in K562 pTet-on/p53 cells did not stimulate apoptosis. Collectively, our results establish the tightly regulated nature of p53 expression and downstream pathways in K562 pTet-on/p53 cells treated with Dox.
In growth inhibition assays, preinduction of p53 (with 2 g/ml Dox) for 48 h was accompanied by insignificant differences in Mtx sensitivities over uninduced cells (IC 50 values of 8.83 Ϯ 1.64 and 7.70 Ϯ 0.70 nM (S.E.; n ϭ 3) with and without Dox, respectively). However, unlike H35 cells (16), DHFR levels were significantly decreased (ϳ3-fold) in K562 pTet-on/p53 cells in the presence of p53 (Fig. 2F). This was, presumably, due FIG. 1. Effects of induction of endogenous wild type p53 on hRFC levels in REH cells. REH cells were treated with 0.5 M daunorubicin (DNR) to induce endogenous p53. Soluble protein extracts were prepared, and total RNAs were extracted at the indicated times for Western and Northern blots analysis, respectively. A, expression of p53 protein in REH cells was assayed on Western blots with anti-p53 monoclonal antibody (DO-1). B, levels of hRFC transcripts on Northern blots were probed with 32 P-labeled hRFC cDNA. C, equal loading of the RNAs in B was established by probing with 32 P-labeled GAPDH cDNA.

FIG. 2. Induction of wild type p53 in K562 pTet-on/p53 cells decreases the levels of hRFC transcripts and increases DHFR protein.
K562 pTet-on/p53 cells were treated with various Dox concentrations (0 -2 g/ml) for 48 h. A, the results of a Western blot analysis of p53 protein with increasing doses of Dox are shown. B, induction of p21 in response to increasing p53 was assayed on Western blots with anti-p21 waf antibody. C, total RNAs were subjected to Northern blot analysis and probed with a 32 P-labeled hRFC cDNA. The arrow indicates the major 3.1-kb hRFC transcript. D, ethidium bromide staining of 28 S and 18 S RNAs for the blot in C were used to assess equal RNA loading. E, total RNAs from both induced (ϩDox) and uninduced (ϪDox) vector control cells were subjected to Northern analysis to assess the levels of hRFC transcripts. 28 S and 18 S ribosomal RNAs were stained with ethidium bromide as controls for equal loading. F, soluble proteins (20 g) from induced (ϩDox) and uninduced (ϪDox) cells were analyzed on Western blots with anti-DHFR antiserum.
hRFC Expression and Function in K562 pTet-on/p53 Cells following p53 Induction-The lack of a significant change in Mtx sensitivity despite significantly reduced DHFR (Fig. 2F), initially suggested that decreased levels of this enzyme target must somehow be offset by alterations in drug uptake. To directly explore the possibility that p53 may regulate hRFC, K562 pTet-on/p53 cells were treated with a wide range of Dox concentrations (0.015-2 g/ml). After 48 h of incubation, total RNAs were prepared and probed on Northern blots with 32 Plabeled full-length hRFC cDNA. hRFC transcripts decreased accompanying expression of p53 and in a dose-dependent manner with increasing concentrations of Dox (Fig. 2C). At a maximum dose of Dox (2 g/ml), there was an ϳ70% loss in hRFC transcripts. Likewise, there was a significant loss of immunoreactive hRFC protein, detected with hRFC-specific antibody on Western blots (Fig. 3). Furthermore, there was a nearly identical dose-dependent decrease in [ 3 H]Mtx influx capacity (Fig. 4A) in K562 pTet-on/p53 cells treated with Dox. At 2 g/ml Dox, hRFC transport activity was only ϳ30% of the level for the untreated control. By contrast, neither hRFC transcripts (Fig.  2E), hRFC protein (not shown), nor Mtx transport activity (Fig.  4B) significantly differed in pTRE vector control cells in the presence and absence of 2 g/ml Dox.
Effects of p15 Induction in K562p15 Tet-on -8 Cells on S-phase Fraction and hRFC Levels-hRFC levels and function have been linked to cell proliferation (49). Thus, the changes in hRFC levels and function upon p53 induction (Figs. 2C, 3, and 4A) could potentially be due to indirect effects of elevated p21, resulting in the loss of S-phase fraction (Table I) and S-phasedependent hRFC expression. In such a case, an identical effect should result from elevated levels of p15, another inhibitor of cyclin-dependent kinases-4 and -6 (34). However, for the analogous p15 Tet-on-inducible model (K562p15 Tet-on -8), hRFC mRNAs were unchanged on Northern blots accompanying p15 induction with 2 g/ml Dox (Fig. 5), even though there was a nearly identical ϳ3-fold loss of S-phase cells (Table I). Likewise, Mtx transport by hRFC was completely unaffected upon p15 induction (1.23 versus 1.31 pmol/mg of protein with and without Dox pretreatment, after 180-s incubation with 0.5 M Mtx). Thus, a simple cell cycle arrest does not appear to explain the decreased hRFC levels in cells that up-regulate wild type p53.
It was of interest that DHFR also decreased (ϳ2.5 fold) upon p15 induction in K562p15 Tet-on -8 cells (Fig. 5D), yet this was not accompanied by a corresponding increased Mtx sensitivity (6.85 Ϯ 0.15 and 9.00 Ϯ 1.00 nM, in the presence and absence of Dox, respectively). The basis of this paradoxical result is not entirely certain, however, it may relate to additional downstream effects of p15 (and also p53) on other folate-dependent pathways that impact Mtx response. For instance, elevated p15 (and p21) might result in decreased thymidylate synthase (50), analogous to DHFR, however, this would increase Mtx resist-ance (51). Clearly, changes in levels of p15 or p53 can potentially cause a multiplicity of effects in different cells and tissues, involving hRFC, DHFR, and, possibly, thymidylate synthase, which are not easily reconciled in terms of their Mtx sensitivities.
Effects of p53 Induction in K562 pTet-on/p53 Cells on hRFC Transcript Stability-p53 has been reported to exhibit RNA binding activity (52), which could conceivably affect mRNA stabilities. Accordingly, the loss of hRFC transcripts in K562 pTet-on/p53 cells treated with Dox could possibly reflect downstream effects of p53 on hRFC transcript stability. To directly assess this possibility, K562 pTet-on/p53 cells were pretreated with Dox (2 g/ml) for 24 h and subsequently with 10 g/ml actinomycin D. Changes in hRFC transcript levels were followed over 12 h on Northern blots (data not shown). Under these conditions, the rates of hRFC transcript turnover were identical and independent of p53 induction (half-lives of 5.2 and 5.9 h in the presence and absence of Dox, respectively).
Effects of p53 on hRFC Involve Transcriptional Repression of  3. Expression of wild type p53 in K562 pTet-on/p53 cells is associated with a loss of hRFC protein. K562 pTet-on/p53 cells were treated with 2 g/ml of Dox for 48 h, and sucrose gradient-purified plasma membranes of Dox-induced cells (ϩ) were prepared in parallel with the uninduced cells (Ϫ). Aliquots of membrane proteins (100 g) were electrophoresed on a 7.5% SDS-PAGE gel, transferred to PVDF membrane, and probed with a protein A-purified hRFC-specific antibody. Enhanced chemiluminescence was used for visualization. The broadly migrating ϳ85-kDa hRFC protein band (38,43,44), detectable with hRFC antiserum, is indicated. the Upstream hRFC-B Promoter-The hRFC gene encodes transcripts characterized by heterogeneous 5Ј-UTRs, reflecting multiple transcription starts and variable splicing of alternative exons, transcribed from dual TATA-less and CAAT-less promoters (30 -32, 44). By 5Ј-RACE (rapid amplification of cDNA ends) analysis, hRFC transcripts in K562 cells contain predominantly (ϳ93%) the KS43 5Ј-UTR from exon 1, transcribed from the upstream hRFC-B ("Pro43") promoter (53). When total RNAs from K562 pTet-on/p53 cells were reverse-transcribed and PCR-amplified to identify the unique 5Ј-UTRs (previously designated KS6 and KS43, and the KS32 splice forms I, II, and III (32,45)), transcripts with the KS43 5Ј-UTR again predominated (Ͼ90%; Fig. 6A). When normalized to ␤-actin (Fig. 6C), both the KS43 5Ј-UTR and open reading frame PCR products were significantly decreased (ϳ70%) in cells treated with Dox (ϩDox; panel B) compared with untreated cells (ϪDox; panel A). Thus, the loss of hRFC transcripts in K562 pTet-on/p53 cells upon p53 induction is completely due to effects on KS43 transcripts transcribed from the hRFC-B promoter.
To further examine the mechanism(s) involved with this effect, full-length (positions Ϫ2016 to Ϫ959) and basal (positions Ϫ1088 to Ϫ960) hRFC-B promoter constructs (in pGL3-Basic) (32), containing a luciferase reporter gene, were transiently expressed in K562 pTet-on/p53 cells in the presence and absence of Dox (2 g/ml). For both full-length and basal promoter constructs, induction of p53 resulted in an ϳ3to 4-fold decrease in relative luciferase activity (Fig. 7, A and B, respectively). p53 repression of the low residual levels of basal activity was completely abolished in a promoter construct in which a highly conserved GC-box element (cccgccc; positions Ϫ1081 to Ϫ1075 (32)) was mutated (to cccgaac; data not shown). Luciferase activity was completely unaffected by Dox for the vector control cells (not shown). When a TDN-p53 expression construct was cotransfected with the full-length hRFC-B construct, the repressive effects of p53 were largely abolished (Fig.  7A). A similar effect of TDN-p53 on basal hRFC-B activity was also observed (not shown). Fig. 7C shows that induction of p53 results in a potent stimulation of luciferase activity in K562 pTet-on/p53 cells transfected with the p53-responsive p53CON reporter plasmid (36).
These results show that induction of p53 suppresses hRFC-B promoter activity and that this can be reversed with a transdominant-negative form of p53. Furthermore, the inhibitory FIG. 5. Effects of p15 induction in K562p15 Tet-on -8 cells on hRFC levels. K562p15 Tet-on -8 cells were induced with 2 g/ml Dox for 48 h followed by total RNA extraction. When these cells were subjected to flow cytometry analysis, a G 1 arrest and an ϳ3-fold loss of S-phase fraction were detected (Table I). A, total RNAs were fractionated on a 0.7% formaldehyde/agarose gel and hybridized with 32 P-labeled hRFC cDNA to detect the major 3.1-kb hRFC transcript. B, the blot was then stripped and reprobed with 32 P-labeled p15 cDNA to assess the levels of p15 transcripts in the presence (ϩ) and absence (Ϫ) of Dox. C, 28 S and 18 S ribosomal RNAs were stained with ethidium bromide as controls for equal loading. D, soluble proteins (20 g) from induced (ϩDox) and uninduced (ϪDox) cells were analyzed on Western blots with anti-DHFR antiserum. PCR products were fractionated on a 2% agarose gel and detected with ethidium bromide staining. A 100-bp DNA ladder (M) and the sizes of the major PCR products are indicated. C, a commercial primer set was used to amplify ␤-actin (838 bp) to normalize the relative levels of hRFC PCR products in the presence and absence of Dox.

FIG. 7. Effects of p53 on hRFC levels in K562 pTet-on/p53 cells involve transcriptional repression of both full-length and basal hRFC-B promoters.
Full-length (A) and basal (B) hRFC-B promoter constructs linked to a luciferase reporter gene in pGL3-Basic, were transiently expressed in K562 pTet-on/p53 cells with (ϩ) and without (Ϫ) 2 g/ml Dox. Five micrograms of each hRFC promoter construct was cotransfected with 25 ng of pRL-SV40 plasmid for normalization. Also shown in A are the effects of cotransfections with the TDN-p53 construct (2.5 g) on luciferase activity of the full-length hRFC-B promoter construct (5 g). C shows the effects of p53 induction on promoter activity of the p53-responsive luciferase reporter construct, p53CON, as a positive control. All data are presented mean values (ϮS.E.) from three experiments in duplicate. effects of p53 can be localized to the hRFC-B minimal promoter region. The finding that transcriptional repression by p53 can be abolished by mutating the highly conserved GC-box suggests a direct or indirect effect of p53 on Sp1/Sp3 transactivation of the hRFC-B promoter.
Mechanisms of p53 Transcriptional Repression of the Basal hRFC-B Promoter-The mechanism by which p53 typically regulates cellular functions involves its role as a transcriptional regulatory activator/repressor (1-3) and may involve binding to specific consensus sequences (11,12,14) and/or direct binding with individual transcription factors (18 -21). Although the p53 effects on hRFC transcription can be localized to the basal hRFC-B promoter, no p53 response element in this region can be identified.
An important role for the Sp1 family of transcription factors in activating basal hRFC-B transcription (via the conserved GC-box) has been established by gel shift analyses of nuclear extracts prepared from HepG2 and HT1080 cells and by transient cotransfections in Drosophila SL2 cells with a pPac-Sp1 expression vector and hRFC-B promoter constructs in pGL3-Basic vector (32). p53 has been reported to bind with Sp1 (20), and its repressive effects on Sp1-regulated promoters (MRP1 (17) and thymidylate synthase (18)) have been suggested to occur by direct association and functional antagonism of promoter transactivation. Consistent with this notion, in Drosophila cells cotransfected with Sp1, p53, and the hRFC-B promoter, a potent suppressive effect on promoter activity by p53 was observed (data not shown).
However, on gel shifts with a 46-bp oligonucleotide (positions Ϫ1088 to Ϫ1043) encompassing the Sp1 binding site (Ϫ1081 to Ϫ1075) (32) and nuclear extracts prepared from K562 pTet-on/p53 cells, treated with and without Dox (2 g/ml), there were no notable differences in the levels or migrations of the DNA⅐Sp1 or DNA⅐Sp3 complexes (not shown). The DNA⅐Sp1 and DNA⅐Sp3 complexes could be supershifted with antibodies specific for Sp1 and Sp3, respectively, however, p53 antibody did not effect a supershift (not shown).
Of course, putative Sp1⅐p53 interactions may not necessarily be detectable on gel shifts, due to a lack of sensitivity and/or use of suboptimal binding conditions. Alternatively, the repressive effects of p53 on hRFC transcription may include downstream effects on Sp1 or Sp3 levels or structure (e.g. phosphorylation or O-glycosylation (54,55)) not detectable on gel shifts or effects of p53 on accessory transcription proteins. Indeed, any repressive effect of p53 can best be interpreted within the overall context of hRFC-B promoter structure and function. The nature of the putative inhibitory complex that forms ultimately depends not only on the levels of individual transcription factors but on the formation of the transcription initiation complex, and assorted other elements such as transcription factor accessibility and post-translational modifications (see above), the presence of upstream inhibitory or activating elements, and the overall promoter architecture.
Conclusions-In summary, our results document a suppression of hRFC levels and function by wild type p53, suggesting that hRFC is an important cellular target for this tumor suppressor. Wild type p53 represses hRFC expression at the level of transcription, and it appears to involve direct or indirect effects on the transactivation of the hRFC-B minimal promoter. The mechanism of p53-mediated transcriptional repression of hRFC and its similarity to the p53 repression of other TATAless promoters is not well established. However, for hRFC, it seems to be independent of its effects on cell cycle progression.
Although our findings suggest that treatment with DNAdamaging agents (e.g., chemotherapy, irradiation) could possibly attenuate sensitivities to antifolates due to decreased hRFC for cells that express wild type p53, the ramifications of this likely depend on the intracellular enzyme target and may vary in different tumors and tissues. Likewise, decreased uptake of reduced folates by this mechanism may exacerbate the effects of DNA-damaging agents, due to perturbations of nucleotide pools required for DNA repair. Clearly, a better understanding of the regulation of hRFC expression by wild type p53 should provide insights into the chemotherapeutic efficacy of antifolates in combination with other anticancer drugs and into the relationships between nutritional folate deficiency, genetic instability, and malignancy.