Human and Mouse Fas (APO-1/CD95) Death Receptor Genes Each Contain a p53-responsive Element That Is Activated by p53 Mutants Unable to Induce Apoptosis*

p53 is a tumor suppressor protein that induces apoptosis at least in part through its ability to act as a sequence-specific transactivator. This work reports that intron 1 of the mouse Fas death receptor gene contains a p53-responsive element (p53RE) that matches the p53 consensus sequence and that is located between nucleotides +1704 and +1723 from the transcription initiation site. This element is specifically bound by p53 and functions as a p53-dependent enhancer in mammalian or in yeast reporter gene assays. Contrary to bax, another known pro-apoptotic p53-target gene, both mouse and human FASp53REs are still activated by the discriminatory p53 mutants Pro-175 and Ala-143, a class of mutants unable to induce apoptosis. We propose that p53-dependent up-regulation of Fas does not induce apoptosis per se but sensitizes the cell to other pro-apoptotic signal(s). The functional conservation of p53-dependent Fas up-regulation argues strongly in favor of its biological importance and suggests that murine models may be used to study further the in vivo role of Fas in the p53 response.

Inactivation of the p53 tumor suppressor gene occurs in over half of all human tumors, implying that loss of the functional protein represents a key event in promoting tumoral pathogenesis (1,2). p53 is an inducible phosphoprotein acting as a "guardian of the genome" (3), mainly by mediating cell cycle checkpoints and/or apoptosis after genotoxic stresses, thereby allowing reparation or elimination of altered cells (4).
The biochemical mechanisms through which wild type (wt) 1 p53 acts have not yet been completely identified. However, an essential aspect of p53 function depends on its activity as a transcription factor (5); p53 contains a strong transcriptional activation domain at its N terminus (6) and binds DNA in a sequence-specific manner through an inter-species conserved central "core" domain (7,8). Most of the mutations in the p53 gene found in tumors affect this core domain, underscoring the importance of DNA binding and transactivation in its role as a tumor suppressor (9,10). The consensus sequence for p53binding sites consists of at least two copies of the sequence 5Ј-PuPuPuC(A/T)(T/A)GPyPyPy-3Ј (where Pu indicates purine and Py indicates pyrimidine) separated by 0 -13 bp (11)(12)(13)(14).
One of the well characterized p53 target genes is WAF1, which encodes the 21-kDa inhibitor (p21) of cyclin-dependent kinases and appears to be an essential effector of p53-mediated G 1 cell cycle arrest after genotoxic stress (15)(16)(17). Pathways leading to p53-dependent apoptosis are less well understood. Several p53 target genes potentially implicated in this process have been characterized, namely h-BAX, h-IGFBP3, h-PIG3, and more recently h-FAS (18 -21). The protein products of h-BAX and h-FAS are directly implicated in apoptosis; BAX acts as an antagonist to the pro-survival protein Bcl2, and FAS is a cell surface receptor that contains a "Death Domain" in its cytoplasmic part and is able to trigger an apoptotic program after activation. For both proteins, a partial requirement for full p53-dependent apoptosis has been shown in some models (21)(22)(23)(24). In other models, BAX or FAS are dispensable for p53-dependent induction of cell death (25)(26)(27). Implication of PIG3 and IGFBP3 in p53-dependent apoptosis has been proposed but not demonstrated. It is therefore difficult to assign a precise role for each of these genes during p53-dependent apoptosis.
Several p53 mutants that have lost apoptotic but not cellcycle arrest function have been described (28,29). Interestingly, these mutants, so-called discriminatory mutants, present a differential ability to transactivate target cellular promoters; they retain the ability to activate the expression of WAF1 but fail to activate BAX or IGFBP3 p53RE (29 -31). Therefore, these mutants are unique molecular tools to study the implication of p53 target genes in apoptosis. We were interested in studying the involvement of Fas during p53-dependent apoptosis. At the beginning of this work, Fas was not yet characterized as a direct target of p53, and we therefore looked for a p53-responsive element (p53RE) in the mouse Fas gene. We then employed p53 discriminatory mutants to characterize further the p53-dependent regulation of Fas. We report here that, like its human counterpart, the murine Fas gene contains a functional p53RE located in the first intron. In addition, we show that p53 discriminatory mutants are able to activate Fas p53RE in contrast to p53REs derived from other pro-apoptotic genes, suggesting a distinct function of the Fas gene in the course of p53-dependent apoptosis. * This work was supported in part by a grant from the Association pour la Recherche sur le Cancer (ARC). 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.
§ Supported by a fellowship from the Commissariat à l'Energie Atomique et l'Electricité De France.

MATERIALS AND METHODS
Library Screening-A C3H/He mouse genomic DNA library (32) was screened under high stringency conditions (33) using a DNA fragment carrying the mouse Fas cDNA. A recombinant DNA (MF8S) that contains the 5Ј part of the Fas gene was subcloned into pBluescript II (Stratagene). Two plasmids were further used as follows: pMFS8-1X that contains an XhoI DNA fragment encompassing exon 1 (approximately from nucleotide Ϫ4500 to ϩ3900 starting from the transcription initiation site) and pMFS8-2X that contains the 3Ј adjacent part of the intron 1 (from nucleotide ϩ3900 to ϩ10,000).
RNA Extraction and Northern Blot Analysis-Total cellular RNAs were extracted using the Trizol LS reagent (Life Technologies) according to the manufacturer's recommendations. Electrophoresis and Northern blotting were performed as described previously (34). A 0.9-kb EcoRI-BamHI DNA fragment corresponding to the cDNA of the mouse Fas gene and a 1.3-kb PstI cDNA fragment corresponding to the rat GAPDH gene were used as probes for Northern blot analysis. The relative level of the hybridization signal of specific mRNAs was evaluated with a Bio-Rad GS 363 molecular imager.
DNA Manipulations and Sequence Determination-Plasmids were constructed using standard procedures (33). DNA sequences were determined using a T7 sequencing kit (Amersham Pharmacia Biotech) according to the instructions of the manufacturer.
Yeast p53RE Functional Assays-The transcription reporter plasmid used in this study is pLG⌬178 (35). It contains the URA3 gene, the yeast 2-m replication origin, and the lacZ gene under control of a truncated CYC1 promoter. A unique XhoI site located just upstream of the promoter allows the cloning of DNA fragments to test them for p53-dependent transactivation. The following pairs of oligonucleotides were used to clone the core of the different putative p53REs: for p53RE A: Aa, TCGATGACATGTTTGTACATGCCC and Ab, TCAGGGGCATG-TACAAACATGTAC; for p53RE B: Ba, TCGAGAGCTTGCTCTGTCTC-GCTTGTCC and Bb, TCGAGGACAAGCGAGACAGAGCAAGCTC; for p53RE C: Ca, TCGAAACCAAGCTTCTGACTTGACT and Cb, TCGAA-GTCAAGTCAGAAGCTTGGTT; for p53RE D1: D1a, TCGAGGGCAAG-TCCAGCTTCTGAGTAAGAACACGTCT and D1b, TCGAAGACGTGT-TCTTACTCAGAAGCTGGACTTGCCC; and for p53RE D2: D2a, TCG-AGGACTTGCCCAACACCATGCCT and D2b, TCGAAGGCATGGTGT-TGGGCAAGTCC. Oligonucleotides of each pair were annealed, leading to double strand oligonucleotides with extremities compatible with an XhoI digestion product and ligated into pLG⌬178 digested with XhoI leading, respectively, to the pFA, pFB, pFC, pFD1, and pFD2 reporter plasmids. p53 was expressed using the pLS76 family vector (36). These vectors contain the LEU2 gene, the Yeast ARS-CEN replication origin, and the human p53 coding sequence under the control of the ADH1 constitutive promoter. pLS76-WT allows expression of the wt p53. The LEU2 ARS-CEN YCplac111 was used as a negative control plasmid that does not express p53 (37). Yeast strain W303 (kindly provided by E. Boy-Marcotte) was co-transformed with an appropriate reporter plasmid and with a p53 expression vector (38). For each experiment, ␤-galactosidase activity of the transformants was determined in three independent cultures as described (39).
Transfection of Mammalian Cells and Reporter Gene Assays-pSVE-hump53 and pCMV-hump53 are expression vectors of human wt p53 under the control of the SV40 early promoter (SVE) and the cytomegalovirus immediate-early enhancer (CMV) promoter (14). p53 mutants derived from these vectors are indicated in lowercase (i.e. pSVE-hump53-P175 allows expression of the Pro-175 mutant). pSVE-Renluc is an expression vector of the Renilla luciferase under the control of the SVE promoter (Promega). pGL3-basic (Promega) reporter plasmid contains the luciferase coding sequence downstream of a multiple cloning site and allows measurement of promoter activity. pPy and pGL3-E1bTATA are the reporter plasmids used to measure enhancer activity in mammalian cells. They contain a truncated polyoma virus early promoter (7) and an E1B minimal promoter (41), respectively, that drive expression of the luciferase coding sequence. pPymFA contains the p53RE A (from nucleotide ϩ1603 to ϩ1812) cloned upstream of the truncated promoter in pPy. pE1B-mF4 and pE1B-mF2 contain, respectively, the four decamers of the p53RE A (from nucleotide ϩ1674 to ϩ1723) or only the two distal decamers of p53RE A (from nucleotide ϩ1704 to ϩ1723) cloned upstream of the E1B minimal promoter in pGL3-E1bTATA. Oligonucleotides TCGACAACATGGTGCAGAGACT-GTTTGCTTGGCAGGGCATGTACAAACATGTCA and CTAGTGACAT-GTTTGTACATGCCCTGCCAAGCAAACAGTCTCTGCACCATGTTG were annealed and cloned into pGL3-E1bTATA digested with NheI and XhoI to obtain pE1B-mF4. To obtain pE1B-mFas, oligonucleotides Aa and Ab (see under "Yeast p53RE Functional Assays") were annealed and cloned into pGL3-E1bTATA digested with XhoI in the same orientation as pE1B-mF4. pE1B-hWAF1, pE1B-hFAS, pE1B-hBAX, and pE1B-hPIG3 contain p53RE derived from the indicated human gene cloned upstream of the E1B minimal promoter in pGL3-E1bTATA between the NheI and the XhoI restriction sites. Pairs of oligonucleotides used are as follows: TCGAGAACATGTCCCAACATGTTG and CTAGC-AACATGTTGGGACATGTTC for pE1B-WAF1, TCGAGGACAAGCCC-TGACAAGCCA and CTAGTGGCTTGTCAGGGCTTGTCC for pE1B-hFAS, TCGATCACAAGTTAGAGACAAGCCTGGGCGTGGGCTATA-TTG and CTAGCAATATAGCCCACGCCCAGGCTTGTCTCTAACTTG-TGA for pE1B-bax, and TCGACAGCTTGCCCACCCATGCTC and CTA-GGAGCATGGGTGGGCAAGCTG for pE1B-hPIG3. pWWP-luc and pbax-luc contain, respectively, 2400 bp of the human WAF1 promoter and 370 bp of the human BAX promoter cloned upstream of the luciferase (42) (18).
Cells seeded in 6-well (35-mm) plates at 2 ϫ 10 5 cells per well were transfected by the calcium phosphate method (43). For each well, 200 l of precipitate containing 1 g of reporter gene plasmid, 20 ng of pSVE-Renluc plasmid, either pSVE-hump53 (0 or 40 ng) or pCMV-hump53 (0 or 80 ng) and sonicated calf thymus DNA to obtain a total of 2 g of DNA were added and left on the cells 6 h, rinsed twice with Dulbecco's modified medium, and incubated for 24 h at the indicated temperature prior to measuring the reporter gene expression. To activate p53 in the SAOS-Val-135 cell lines, cells were incubated for 8 h at 32.5°C prior to the luciferase assay.
Luciferase assay was performed using the Dual-luciferase TM reporter assay system (Promega), and the activity of the reporter gene was measured with a Microlumat LB 96P (EG & G Berthold). The Renilla luciferase was used to normalize the transfection efficiency. For each reporter vector, fold activation was calculated as the ratio between the experimental luciferase activity and the corresponding value obtained without p53-expressing vector. Results shown are the mean of at least three experiments, and the standard deviations are indicated.

Putative p53REs Are Located in the Intron 1 of the Mouse Fas
Gene-To search for a p53RE in the mouse Fas gene, we have employed a multistep approach, based on sequence analysis and reporter gene assay. Sequence analysis of the 800 bp upstream of the transcription initiation site did not reveal any p53 consensus sequence as previously defined. Furthermore, when inserted upstream of the luciferase gene in the pGL3basic vector, this fragment was not able to induce expression of the reporter gene in a p53-dependent manner in H1299 or in SAOS-Val-135 cells (data not shown). To test whether a p53RE could be present further upstream of the transcription initiation site, an EcoRI fragment spanning from Ϫ3250 to Ϫ385 (starting from the transcription initiation site) was inserted in both orientations in pPy upstream of the polyoma virus minimal promoter driving the expression of the luciferase gene. When transfected either in H1299 or in SAOS-Val-135 cells, this fragment did not lead to a p53-dependent up-regulation of the reporter gene (data not shown). From these results, we conclude that the mouse Fas promoter region does not contain a p53RE. We therefore decided to look for the presence of a p53RE downstream of the transcription initiation site. The sequence between nucleotide ϩ1 and ϩ4150 (encompassing exon 1 and part of intron 1) was determined and searched for a p53 consensus sequence, two PuPuPuC(A/T)(A/T)GPyPyPy decamers separated by 0 -13 nucleotides (11). Compared with the above consensus sequence, some mismatches do not alter the functionality of the p53-binding site (11)(12)(13)(14). According to this notion we have accepted a maximum of three variations compared with the consensus sequence proposed by El-Deiry et al. (11) with no more than one error in a single decamer if it is located in one of the two central nucleotides of a decamer. The C in position 4 and the G in position 7 were considered as invariant positions. Fig. 1 presents the position and sequence of the four putative p53REs found, named A, B, C, and D. These p53REs are located between nucleotides ϩ1600 and ϩ4000 from the transcriptional initiation site. Putative p53RE D and p53RE A contain one and two additional decamers, respectively.
A Functional p53RE Is Located between Nucleotides ϩ1704 and ϩ1723 within the Mouse Fas Intron 1-To explore the functionality of the putative p53REs found in intron 1 of the mouse Fas gene, we first used a yeast reporter gene assay. Pairs of decamers (corresponding to the p53 DNA-binding site (11)) found in the four putative Fas p53REs were subcloned into pLG⌬178 upstream of a truncated CYC1 promoter regulating lacZ reporter gene expression (35). The reporter plasmids obtained were used to co-transform yeast strain W303 with empty vector (YCplac111) or with a plasmid expressing human wild type p53 (pLS76-WT), and expression of ␤-galactosidase was analyzed. Only decamers of p53RE A gave rise to a p53-dependent expression of the reporter gene (Table I). This p53RE was activated 3500-fold by wt p53. To confirm the functionality of the p53RE A, this element was tested in a luciferase reporter gene assay in mammalian cells. A Sau3A DNA fragment, spanning from nucleotide ϩ1603 to ϩ1812 and encompassing the decamers found in p53RE A, was cloned into the pPy luciferase reporter plasmid upstream of the truncated polyoma virus promoter. This plasmid was transfected into the SAOS-Val-135 human cell line (expressing a ts mutant of p53) and into the H1299 p53-deficient human cell line. In SAOS-Val-135 cells, activation of p53 by temperature shift resulted in a p53-dependent luciferase reporter gene expression ( Fig. 2A). In H1299 cells, co-transfection with a plasmid encoding for wt p53 results in a p53-dependent expression of the luciferase reporter gene (Fig. 2A). To explore further the functionality of this p53RE composed of four decamers matching the p53 consensus sequence, it was subcloned into a reporter plasmid containing the E1B minimal promoter (21). Constructs containing either the four decamers or only the two distal decamers were tested by transfection into H1299 cells. Both constructs gave rise to a comparable p53-dependent activation of the reporter gene (Fig. 2B). Thus, a minimal functional p53RE is located in intron 1 between nucleotides ϩ1704 and ϩ1723 of the murine Fas gene.
p53 Specifically Binds to the Fas p53RE in Vitro-Gel retardation assay was performed to test whether p53 specifically binds to the p53RE A. The p53 used in this experiment was produced in Sf9 cells using recombinant baculovirus and partially purified on a Q-Sepharose column (44). When p53 was incubated with a double strand oligonucleotide corresponding to the Fas p53RE (nucleotides ϩ1704 to ϩ1723), complexes leading to band shifts were detected (Fig. 3, lane 2). To identify the specific protein-DNA complex in lane 2, competition assays were performed. Addition of a 100-fold excess of non-radioactive oligonucleotide devoid of a p53-binding site did not affect the presumably specific complex (lane 4), whereas the formation of this complex was inhibited by 100-fold excess of unlabeled homologue probe or an alternative p53 consensus sequence ( lanes 5 and 6). Finally, the mobility of this specific complex was retarded in the presence of an antibody directed against p53 (lane 3), confirming the presence of p53 in the complex. Taken together these results demonstrate that p53 specifically binds to the Fas p53RE.
p53 Activates the Transcription Rate of the Mouse Fas Gene   Fig. 2 for the corresponding DNA sequence) was cloned into the pLG⌬I78 reporter plasmid upstream of the truncated CYC1 promoter and analyzed in the yeast reporter assay in the presence or absence of wild type p53. ␤-Galactosidase activities indicated are expressed in Miller units.  (45). Our laboratory has recently shown that activation of p53 in the M1-LTR13 cells leads to an accumulation of the Fas mRNA. 3 We tested whether this accumulation resulted from transcriptional activation or mRNA stabilization. To this end, the transcription inhibitor actinomycin D was employed to measure the half-life of Fas mRNA induced by p53. M1-S6 cells were used to obtain the control Fas mRNA population. Since the basal level of Fas mRNA is not easily detectable in M1derived cell lines, we induced Fas transcription with IL-6. 3 IL-6 action is mediated through the transcription factor NF-IL-6 which induces Fas transcription without modifying the stability of the Fas mRNA (46,47). To analyze Fas mRNA stability, RNAs were sequentially harvested from M1-LTR13 cells cultured at 32.5°C and from M1-S6 cells cultured in presence of IL-6 at 32.5°C after addition of actinomycin D. As shown in Fig. 4, the kinetics of Fas mRNA degradation in cells treated with actinomycin D was quite similar in both cell lines, whether Fas was induced by p53 or by IL-6. Thus, as expected from the characterization of a p53RE in intron 1 of the mouse Fas gene, p53 activates the transcription rate of this gene. Fas p53RE Is Activated by p53 Discriminatory Mutants-Concomitantly with the loss of apoptotic but not growth arrest functions, p53 discriminatory mutants retain the ability to activate expression of the cell cycle inhibitor WAF1 gene but fail to activate p53-responsive sequences derived from the proapoptotic BAX or IGFBP3 genes (30,31). To investigate the role of Fas in p53-mediated apoptosis, we tested whether the mouse Fas p53-responsive sequence is activated by such p53 mutants. In H1299 cells we confirmed that the Pro-175 discriminatory mutant transactivated the WAF1 p53-responsive promoter as well as wt p53, although this mutant failed to transactivate the BAX p53-responsive promoter (30) (Fig. 5A). Under the same experimental conditions, a 210-bp DNA fragment containing the mouse Fas p53RE placed in front of a polyoma virus minimal promoter was significantly activated by the Pro-175 p53 discriminatory mutant (Fig. 5A). This result indicates that the mouse Fas p53RE may have a unique property compared with other known p53REs found in pro-apoptotic genes. To analyze specifically the differential regulation by p53 of its target genes, we compared the activation of p53REs derived from h-WAF1 (42), h-BAX (18), m-Fas, and h-FAS (21) in the same minimal promoter context by discriminatory Pro-175 and ts Ala-143 mutants. Since h-PIG3 was suggested to be implicated in p53-mediated apoptosis (20), we also tested in this system the response of its p53RE. Surprisingly, under the conditions used in our reporter system, wt p53 was unable to activate the p53RE derived from h-PIG3 (data not shown). Pro-175 and Ala-143 p53 discriminatory mutants were able to activate efficiently p53REs from h-WAF1, m-Fas, and h-FAS genes but did not, or very weakly, activate p53RE derived from h-BAX gene (Fig. 5B). We conclude that mouse and human FAS p53REs have a unique feature among p53RE derived from pro-apoptotic genes; they can be activated by p53 mutants unable to induce cell death. 3 C. Choisy-Rossi et al., submitted for publication.   FIG. 2. The mouse Fas gene contains a p53RE located between nucleotides ؉1704 and ؉1723. A, the DNA fragments of Fas intron 1 indicated below the graph were subcloned upstream of a truncated promoter in the pPy luciferase reporter gene plasmid and transfected into SAOS-Val-135 cells, or co-transfected into H1299 cells with the human p53 expression vector pSVE-hump53. p53-dependent fold activation was calculated for both cell types as described under "Materials and Methods." B, DNA fragments of p53RE A indicated below the graph were subcloned upstream of the E1B minimal promoter in the pGL3-E1bTATA luciferase reporter plasmid and co-transfected into H1299 cells with pSVE-hump53. p53-dependent fold activation was calculated as described under "Materials and Methods. "   FIG. 3. p53 binds in vitro to the Fas p53RE. Binding of purified baculovirus-produced human p53 (44) to the Fas p53RE is shown by EMSA. A DNA probe corresponding to the core of the p53RE A (from nucleotide ϩ1703 to ϩ1723) was incubated with p53 and, when indicated, with anti-p53 VJO1 (a monoclonal antibody directed against the N-terminal part of human p53, E. May and P. May, personal communication.). Competitions were performed by using 100-fold molar excess of unlabeled double-stranded oligonucleotides: either nonspecific (containing no p53 consensus sequence: CGGTATCCACCAGGTCTGCGAC-AACGATGAAGCC) or homologous (the p53RE A core) or specific (containing a synthetic p53 consensus sequence: GTCGACGGACATGCCC-GGGCATGTCC).

DISCUSSION
Mouse Fas Gene Contains a p53RE-In this report we show that the mouse Fas gene contains a functional p53RE. The minimal element is located in the first intron between nucleotides ϩ1704 and ϩ1723 (from the transcription initiation site) and is composed of two decamers that match the p53 consensus sequence (score 18/20). Evidence for p53-dependent activity of this cis-acting sequence is based on EMSA experiments, functional assays with reporter gene constructs and on the analysis of Fas gene transcription in p53-overexpressing M1 cells. Two additional decamers that show limited homology to the p53 consensus sequence (score 16/20) are located eight nucleotides upstream of this minimal element. These two decamers are not necessary for the activity of the minimal p53RE at least under the conditions used in our reporter systems. Whether they are necessary for the p53-dependent regulation of the mouse Fas gene in the chromosomal context cannot be completely ruled out.
It has been shown recently that the human FAS gene also contains a p53RE composed of two decamers located within the first intron (21). To check whether the murine and human FAS p53RE are homologous, the DNA sequence of the mouse intron 1 was aligned with the published human p53RE DNA sequence fragment. One region of significant homology with the human sequence was found in the mouse intron 1, but this mouse sequence did not contain any counterpart of the human p53RE (result not shown). In addition, sequence alignment of the human and mouse p53REs did not reveal a high degree of similarity. These results suggest that both p53RE may have a distinct origin. A more conclusive analysis would require a complete sequencing of the human FAS intron 1. The conservation of the p53-dependent regulation of Fas between the two species argues against a coincidental event and emphasizes the requirement for this regulation in the control of mammalian homeostasis. Indeed, it has been shown in several human cell lines that increased Fas expression is necessary for a full p53dependent apoptotic response following genotoxic stress (21,24). Since this regulation is conserved, the murine model is a relevant experimental system to study the precise in vivo role of Fas in p53-dependent apoptosis.
Fas p53REs Are Activated by p53 Mutants Unable to Induce Apoptosis-Several lines of evidence have suggested that the ability of p53 to induce apoptosis is separable from its role in cell growth arrest. Thus, studying the mechanism(s) of p53induced apoptosis would require specific tools, and such a tool is provided by a particular set of p53 mutants, the so-called discriminatory mutants. These mutants are able to induce growth arrest but not apoptosis (28,31). Concomitantly, the discriminatory mutants are capable of activating the WAF1 gene encoding a cell cycle inhibitor but are no longer able to activate promoters containing p53REs from the BAX or IG-FBP3 pro-apoptotic genes (29 -31). Interestingly, we demonstrate that both the human and mouse Fas p53RES are activated by the discriminatory mutants. This is the first report on p53REs from pro-apoptotic genes that are still activated by p53 discriminatory mutants. This activation does not depend on adjacent gene sequences or on the structure of the natural promoter, since it is effective in a synthetic construction containing only the sequence corresponding to the p53RE. In the same DNA context, WAF1 p53RE is activated by the discriminatory mutants, whereas BAX p53RE is not. Since activation of p53REs by the p53 discriminatory mutants is an intrinsic property of these elements, this activation should be effective on the endogenous Fas gene. Indeed, we observed that the Ala-143 p53 discriminatory mutant was able to activate Fas expression in two different cellular models (21,48). The Fas gene is therefore activated by p53 mutants able to induce growth arrest but unable to induce cell death. Thus, the death FIG. 5. Mouse and human FAS p53RE are activated by p53 discriminatory mutants. A, reporter gene plasmids containing the region of the murine Fas p53RE (pPymFA), the bax promoter (pbaxluc), and the WAF1 promoter (pWWP-luc) were used to co-transfect the H1299 cells with pCMV-hump53 plasmids expressing the wt or the indicated p53 mutant (mut, His-175 p53 mutant; wt, wild type p53; and dis, discriminatory Pro-175 p53 mutant). p53-dependent fold activation was calculated as indicated under "Materials and Methods." B, pE1B-hWAF1, pE1B-mFas, pE1B-hFAS, and pE1B-hBAX plasmids, containing p53REs derived from the indicated genes and cloned upstream the E1B minimal promoter in pGL3-E1bTATA, were used to co-transfect H1299 cells with or without pCMV-hump53 plasmids expressing wt p53, His-175 p53 mutant, or Pro-175 or Ala-143 discriminatory mutants. Luciferase assay was performed after 24 h of incubation at 37°C for Pro-175 mutant, and 6 h at 37°C and 18 h at 32°C for the ts Ala-143 discriminatory mutant. Results are presented as percent of activation relative to wt p53 used under the same experimental conditions. His-175 mutant never activates more than twice the corresponding p53RE (data not shown).

FIG. 4. p53 activates Fas transcription in the M1-LTR13 cells.
M1-LTR13 cells (expressing a ts p53) were cultured at 38°C and shifted at 10 6 ϫ 0.7 cells/ml to 32.5°C to activate p53. M1-S6 cells (expressing no p53) were cultured at 38°C and shifted at 10 6 ϫ 0.7 cells/ml to 32.5°C in presence of 100 units/ml of IL-6. Actinomycin D (Act-D) was added at 20 g/ml 3 h after temperature shift. Total RNAs were extracted at the indicated incubation times at 32.5°C, and Fas mRNA was revealed using the Northern blot technique. RNA amounts deposited in each lane have been equilibrated by measuring rRNAs intensity (data not shown). Hence, the observed decrease of the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) level in a time-dependent manner in the M1-LTR13 cells, treated or not with actinomycin D, is probably correlated with an apoptotic fate of these p53-expressing cells.
To demonstrate the efficiency of actinomycin D at the concentration used in this study, we have verified that addition of the drug prior to temperature shift completely inhibits induction of Fas mRNA in M1-S6 cells treated with IL-6 or in M1-LTR13 cells (data not shown).
receptor Fas may have a yet unknown function in cell cycle regulation. Alternatively, up-regulation of Fas by p53 may not be sufficient to induce apoptosis. The second hypothesis is supported by the death receptor function of Fas; activation of the Fas receptor requires the interaction with its specific ligand (FasL). In this context, a recent study has identified another death receptor, DR5 (KILLER), as a p53 target that is upregulated during genotoxic stress-induced apoptosis (49,50). In light of our results, it would be of interest to test whether DR5 presents the same regulation by the discriminatory mutants.
Fas in the Transcription-dependent p53 Response, a Model-Since particular p53 mutants can discriminate between the variable p53REs, it is tempting to speculate that wt p53 is also able to do so. Several lines of evidence suggest that this could indeed be the case. First, as for discriminatory mutants, cell cycle arrest and apoptosis activities of wt p53 can be separated as follows: low expression of wt p53 induces cell cycle arrest, and high expression preferentially results in apoptosis (51). Second, activation of p53 target genes in normal thymocytes following irradiation was shown to have different kinetics for the different target genes as follows: WAF1 is already fully activated 2 h after irradiation, activation of Fas begins, while bax is hardly induced at this time (26). Third, it has been shown that wt p53 binds in vitro more efficiently to WAF1 p53RE than to p53RE of the BAX gene (31). Finally, activation of different p53REs under the same synthetic promoter context by a moderate level of wt p53 (from the SV40 early promoter) leads to a gradual response of the elements; the most efficiently activated are h-WAF1 p53RE and m-Fas p53RE, whereas h-BAX p53RE is significantly less activated. 4 The precise mechanism underlying this wt p53 discriminatory effect is not known. As indicated by in vitro binding experiments, it could reflect a difference in the affinity of wt p53 for the different p53REs due to an intrinsic property of the p53 protein or to specific post-translational modifications (52). A recent report has suggested that different genotoxic treatments may cause different phosphorylations of p53 which, in turn, may result in different levels of activation of p53 target genes (40). A hierarchy of target gene activation may be proposed, in which the WAF1 p53RE can be regarded as a strong element, the BAX p53RE as a weak element, and the Fas p53RE as an intermediate element. Accordingly, we suggest a possible model of p53 transcription-dependent response to DNA damage, in which the response is variable, depending on the extent of p53 activation: low activation would lead to induction of WAF1 and cell cycle arrest to allow DNA repair; an intermediate activation would lead additionally to induction of Fas to sensitize the unrepaired cells to additional pro-apoptotic signals; and a strong p53 activation would lead to induction of all p53 target genes, including the pro-apoptotic gene bax, to bring about apoptosis in cells which are irreversibly altered. Nevertheless, it should be noted that specific mechanisms resulting from a particular cellular environment could interfere with this proposed basic scenario.