Early Growth Response-1-dependent Apoptosis Is Mediated by p53*

The early growth response-1 (EGR-1) protein is an anti-proliferative signal for certain tumor cells and is required for apoptosis induced by stimuli that elevate intracellular Ca2+. We present evidence that EGR-1 transactivates the promoter of the p53 gene and up-regulates p53 RNA and protein levels. Inhibition of p53 function with dominant-negative p53 mutants abrogates EGR-1-dependent apoptosis. These findings establish a direct functional link between EGR-1 and the p53-mediated cell death pathway and suggest that mutant forms of p53 in tumor cells may provide resistance to the anti-proliferative effects of EGR-1.

The early growth response-1 (EGR-1) protein is an anti-proliferative signal for certain tumor cells and is required for apoptosis induced by stimuli that elevate intracellular Ca 2؉ . We present evidence that EGR-1 transactivates the promoter of the p53 gene and upregulates p53 RNA and protein levels. Inhibition of p53 function with dominant-negative p53 mutants abrogates EGR-1-dependent apoptosis. These findings establish a direct functional link between EGR-1 and the p53-mediated cell death pathway and suggest that mutant forms of p53 in tumor cells may provide resistance to the antiproliferative effects of EGR-1.
Apoptosis, or programmed cell death, a genetic process of coordinated deletion of selective cells, is essential for metazoan development and homeostasis (1)(2)(3)(4). The primordial forms of apoptosis in Caenorhabditis elegans and Drosophila have been recapitulated in mammalian cells, and striking similarities have been observed in the cell death programs of invertebrates and vertebrates (5)(6)(7). Apoptosis is characterized by cell membrane blebbing, chromatin condensation, changes in nuclear architecture, and oligonucleosome-length DNA fragmentation (8,9). The apoptotic pathways consist of an early component that includes molecular events that are specific for an inducer or a group of inducers and of downstream effector components that are common to diverse apoptotic signals (10). The common components include a basal cell death machinery composed of initiator, amplifier, and effector proteases belonging to the interleukin-1 converting enzyme subfamily or an interleukin-1 converting enzyme-related family (10 -14). Downstream targets of these proteases include the interleukin-1 converting enzyme subfamily proteases themselves; the nuclear enzymes poly(ADP-ribose) polymerase and DNA-dependent protein kinase, which are involved in DNA repair; the nuclear protein U1 ribonucleoprotein and nuclear lamins; and cytoplasmic components such as protein kinase C␦ and cytoskeleton components such as actin (cited in Ref. 12).
Intracellular calcium is a key second messenger implicated in the activation of the apoptotic program in diverse cell types and tissues (15)(16)(17)(18). Intracellular calcium levels become elevated after the activation of T lymphocytes by anti-CD3 or that of B lymphocytes by anti-IgM antibodies; after withdrawal of survival factors, such as testosterone in the prostate gland; or after exposure to certain exogenous stimuli, such as calcium ionophores or thapsigargin (TG), 1 a potent inhibitor of the Ca 2ϩ -dependent ATPase in the endoplasmic reticulum (19). Elevation of intracellular calcium causes induction of immediate-early genes that further trigger a cascade of downstream events leading ultimately to cell death. Although several immediate-early genes such as early growth response-1 (Egr-1; also referred to as zif268, NGF-IA, TIS8, and Krox-24), nur77, and par-4 have been functionally linked to apoptosis caused by intracellular calcium elevation (20 -24), the precise downstream events that are important for successful apoptosis via this pathway have not been delineated.
Egr-1 was first identified as an immediate-early gene induced by mitogenic stimulation and during membrane depolarization and seizure (25,26). Subsequently, EGR-1 was shown to be induced by diverse exogenous stimuli (27). EGR-1 is a nuclear protein that contains three zinc finger motifs of the C 2 H 2 subtype that bind to a GC-rich consensus DNA sequence TGCG(T/g)(G/A)GG(C/a/t)G(G/T) (where lowercase letters indicate bases of relatively lower binding affinity) or to a (TCC) n motif (28,29). Structure-function mapping studies suggest that the NH 2 terminus of EGR-1 confers transactivation function to the protein (27,30). The Egr-1 gene was localized to human chromosome 5q31.1, a region known to be often deleted from patients suffering from therapy-induced acute myeloid leukemia (31). Moreover, studies using diverse tumor cells suggest that endogenous levels of EGR-1 act to impede proliferation (32,33). Consistent with an anti-tumor role for EGR-1, apoptosis-inducing stimuli, such as TG and ionizing radiation, upregulate EGR-1 expression (20,34). Furthermore, our previous studies, which used antisense oligomers to block EGR-1 expression or a dominant-negative EGR-1 mutant to inhibit EGR-1 function, confirmed that EGR-1 is essential for apoptosis induced by TG or by ionizing radiation (20,34).
Because EGR-1 is induced very early in the apoptotic process (20,34), it is expected to mediate the activation of downstream genes that play crucial roles in growth control. EGR-1-binding sites (EBS) that conform to the GC-rich consensus sequence have been identified in the promoters of genes such as thymidine kinase, an enzyme integral to DNA biosynthesis; cell cycle regulators such as cyclin D1 and the retinoblastoma susceptibility gene Rb; and (TCC) n motifs have been identified in the promoter regions of genes encoding growth factors such as platelet-derived growth factor and basic fibroblast growth factor; growth factor receptors such as epidermal growth factor-receptor and the insulin-like growth factor-receptor; and protooncogenes c-Ki-ras and c-myc (cited in Refs. 20 and 27). However, none of these genes has been shown to be directly involved in apoptosis as a consequence of EGR-1 induction.
The tumor suppressor gene p53 is a central mediator of cell cycle growth arrest and apoptosis (reviewed in Refs. [35][36][37]. The p53 protein encodes a transcription factor that functions as a transcriptional activator or repressor, depending upon the promoter context (38 -41). The p53 protein can up-regulate the expression of a number of downstream genes, such as p21/waf1, insulin-like growth factor binding protein-3, bax, and fas/apo1, implicated in growth inhibition and apoptotic cell death (42)(43)(44)(45). These regulatory effects require binding of p53 protein to consensus binding sites in the promoter region of target genes (35,46,47). The p53-dependent downstream induction or activation of the Fas/Apo1 pathway leads to the activation of a cascade of downstream effector proteases resulting in apoptotic death (12,45). p53-dependent apoptotic pathways that are independent of the ability of this protein to function as a transcriptional activator have been also suggested (48,49). The induction of apoptosis requires p53 in most cellular experimental systems, but there are multiple examples of apoptosis in the absence of functional p53 protein, suggesting the presence of p53-independent pathway(s) for apoptosis (50 -53). The precise molecular components of the p53-independent pathway, however, are not known.
In a number of different types of cancers, p53 is the most commonly mutated tumor suppressor gene (54). Lack of p53 expression or function is associated with an increased risk of tumor development (55)(56)(57)(58). Most tumor types either contain no p53 protein owing to loss of the p53 alleles or of chromosome 17 where p53 is located or contain point mutations in the p53 gene that result in mutant protein products that are functionally inactive (54). Most of the commonly occurring mutations in p53 are located in the DNA-binding or transactivation domains of the protein and render p53 inactive in the transcription of downstream proapoptotic genes such as bax (39,40,54,59). Some mutant forms of p53 protein can oligomerize with wildtype p53 protein and abrogate its growth-inhibitory functions and thus act as dominant-negative inhibitors of wild-type p53 (46).
Wild-type but not mutant p53 protein can cause apoptosis in response to a wide range of exogenous stimuli such as adenovirus E1A protein, anti-cancer agents, or ionizing radiation (62)(63)(64)(65)(66)(67)(68). Moreover, hypoxia has been shown to induce apoptosis that is dependent on p53, and hypoxic conditions are suggested to provide a selective pressure for p53 mutations (69). Several immediate-early genes encoding transcription factors such as c-Fos, c-Jun, EGR-1, Nur77, and c-MYC have been implicated in apoptosis (20, 22-24, 70 -73), and c-Fos and c-MYC have been shown to require wild-type p53 for apoptosis (70,72,73). It is not known, however, whether the apoptotic action of the other immediate-early genes is dependent on wild-type p53 function. Studies directed at identifying whether p53 is required or not for the transduction of an apoptotic signal from immediate-early proapoptotic genes encoding transcription factors, which couple the early events in the plasma membrane to long term cellular phenotypic changes in the cell, should provide valuable insights into the downstream targets of the transcription factors and help identify a link with components of the cell death pathways. In the course of our studies aimed at understanding the mechanism of EGR-1-mediated apoptosis, we have identified p53 as a direct transcriptional target of EGR-1. Our data suggest that the proapoptotic action of EGR-1 is mediated via wild-type p53.

EXPERIMENTAL PROCEDURES
Cell Cultures and Plasmid Constructs-Human melanoma cells A375-C6 and transfected cell lines that expressed pCMV-mEGR1 or vector were cultured as described previously (20,34). Pools of approximately 200 transfected clones were maintained as cell lines. Expression constructs m143 and m175 that encode mutant p53 proteins (46) were kindly provided by Bert Vogelstein (Johns Hopkins University School of Medicine, Baltimore, MD). The constructs pCMV-⌬TA, pCMV-⌬RM, pCMV-⌬ZF, and pCMV-⌬TA/RM that encode mutant EGR-1 proteins have been previously described (27,74). Reporter construct p53(2.2ϩ1.6)-CAT, which contains about 2.2 kb of sequence upstream of the human p53 cap site and about 1.6 kb of sequence downstream of the p53 cap site placed downstream of the chloramphenicol acetyltransferase (CAT) cDNA (75), was kindly provided by David Reisman (University of South Carolina, Columbia, SC). Polymerase chain reaction was employed to synthesize a 0.8-kb p53 promoter fragment from p53(2.2ϩ1.6)-CAT DNA template. The reactions used a downstream primer (5Ј-TTTTTGGCTCTAGACTTTTGAGAA-3Ј) complementary to the region ϩ11 to Ϫ9 (underlined) and an upstream primer (5Ј-TTTT-TCTAGAGAAGCGCCTACGCTCCC-3Ј) corresponding to the region Ϫ758 to Ϫ774 (underlined) in the p53 promoter. The polymerase chain reaction product was subcloned in the XbaI site upstream of the CAT sequence in pGCAT-C vector and the construct, p53(0.8)-CAT, thus obtained was confirmed for orientation.
Western (Immunoblot) Analysis-Whole-cell protein extracts were prepared and analyzed by using Western blot analysis as described previously (20,34). EGR-1 (sc-110) and p53 (DO-1) antibodies were obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA), and the actin antibody was purchased from Sigma.
Transfections and CAT Assays-Transfections were performed by the calcium phosphate coprecipitation method, and stable transfectants were selected in 300 g/ml G418-sulfate as described previously (20). Pools of about 200 stably transfected clones were maintained as cell lines. CAT assays were performed by using thin layer chromatography as described previously (76,77).
Northern Analysis-Total RNA extraction and Northern blot analysis were performed as described previously (77). The cDNA probe for Egr-1 has been described previously (77). The cDNA for human p53 was purchased from the American Type Culture Collection (Rockville, MD).
Electrophoretic Mobility Shift Assay and Supershift Assay-Preparation of nuclear extracts from transfected cells and electrophoretic mobility shift assay were preformed as described previously (76). Two complementary primers, 5Ј-GCGCCTACGCTC-3Ј and 5Ј-GAGCGTAG-GCGC-3Ј, corresponding to the upper and lower strands of EBS-1 (Ϫ770 to Ϫ762) in the upstream region of the p53 promoter were synthesized at the Macromolecular Synthesis Facility, University of Kentucky. The primers were annealed, end-labeled with [␥-32 P]ATP and used as probe. The reaction mixture consisting of nuclear extract, 5 ϫ reaction buffer (1 M Tris, pH 7.5, 2.5 M potassium chloride, 0.5 M EDTA, 1 M dithiothreitol, glycerol, bovine serum albumin), and 1 M poly(dI-dC)⅐poly(dI-dC) was incubated on ice for 15 min. Probe was added, and the mixture was further incubated at room temperature for 20 min. The reaction mixture for the supershift assay consisted of all of the reagents described above for electrophoretic mobility shift assay plus 2 l of EGR-1 antibody (sc-110X from Santa Cruz Biotechnology, Inc.) or preimmune antibody. The reaction mixtures were preincubated for 20 min on ice, then incubated with the probe for 40 min, and subjected to electrophoresis on polyacrylamide gels to separate the bound complexes from the free probe.

EGR-1 Sensitizes Cells to
Apoptosis-To examine ectopically overexpressed EGR-1 as a cause or enhancer of apoptosis, human A375-C6 cells were stably transfected with pCMV-mEGR1, an expression construct for mouse EGR-1 (mEGR1), or with vector for a control. For each of the two constructs, we examined three transfected cell lines for EGR-1 expression by Western blot analysis. Data representative of the transfected cell lines shown in Fig. 1A confirmed overexpression of EGR-1 in the pCMV-mEGR1-transfected cell line as compared with the vector-transfected cell line. These transfected cell lines were then exposed to 250 nM or 1 M TG for various periods of time and examined for apoptosis by TUNEL. As seen in Fig. 1B, vector-transfected cells consistently showed a significantly lower number of apoptotic cells (p Ͻ 0.0001 by the Student's t test) than did the EGR-1-transfected cells at all the time points ranging from 6 to 36 h of exposure to TG. These findings suggested that ectopic mEGR1 provided enhanced susceptibility to TG-inducible cell death (Fig. 1B). Thus, consistent with a role for EGR-1 as a positive regulator of apoptosis, ectopic overexpression of EGR-1 sensitizes cells to apoptosis.
One logical mechanism by which EGR-1 induces apoptosis may involve the transcriptional up-regulation of downstream genes. To address this question, A375-C6 cells were stably transfected with pCMV-mEGR1 or the following mEGR1 deletion mutant constructs: pCMV-⌬TA that lacks the transactivation domain (amino acids 1-240); pCMV-⌬RM that lacks the repression module (amino acids 284 -330); pCMV-⌬ZF that lacks the first two zinc fingers (amino acids 331-374); or pCMV-⌬TA/RM that lacks both the transactivation domain and the repression module (amino acids 1-314) ( Fig. 2A and Ref. 74). Whole-cell protein extracts from each stably transfected cell line were first examined for expression of endogenous or ectopic EGR-1 or mutant proteins by Western blot analysis. As seen in Fig. 2B, EGR-1-transfected cells contained about 12-fold higher EGR-1 levels than did the vector-transfected cells, which contained only endogenous EGR-1. Moreover, the levels of mutant EGR-1 proteins (⌬TA, ⌬ZF, ⌬RM, or ⌬TA/RM) produced by the mutant constructs were similar to those produced by full-length EGR-1 (Fig. 2B). Next, the transfectants were transiently transfected with the EBS-containing reporter construct, EBS-CAT (78,79), and analyzed for CAT activity. As seen in Fig. 2C, the cells expressing pCMV-mEGR1 or pCMV-⌬RM showed relatively higher CAT activity than did the cells expressing the vector, pCMV-⌬TA, pCMV-⌬ZF, or pCMV-⌬TA/RM. When the transfectants were examined for TG-inducible apoptosis by TUNEL, those expressing pCMV-mEGR1 or pCMV-⌬RM showed a higher percentage of TUNELpositive cells than did transfectants that expressed vector, pCMV-⌬TA, pCMV-⌬ZF, or pCMV-⌬TA/RM (Fig. 2D). These findings suggested that constructs that lacked the transactivation domain or the DNA-binding zinc finger domain of EGR-1 failed to cause transcriptional activation or to sensitize the cells to TG-inducible apoptosis. Thus, both the DNA binding and transactivation functions of EGR-1 are essential for an enhanced apoptotic response to TG. These findings suggest that EGR-1 induces apoptosis by a mechanism involving transcriptional regulation of downstream target genes.
EGR-1 Induces p53 Expression-To identify candidate EGR-1-responsive genes essential for mediating apoptosis, a Gen-Bank data base search was conducted to identify putative EGR-1-binding sites in the promoter regions of genes known to be functionally involved in apoptosis. Interestingly, two putative EGR-1 consensus binding sites, EBS-1 and EBS-2 ( Fig.  3A), were identified in the promoter region of the p53 gene (75) (GenBank accession number X54156). The p53 gene has been also shown to be essential for apoptosis induced via activation of diverse intracellular signaling pathways (50 -53, 62-68). Because regulation of the p53 gene at the transcriptional level contributes significantly to the control of p53 protein levels and to phenotypic responses (80), the ability of EGR-1 to induce the p53 promoter was tested. Reporter constructs, p53(2.2ϩ1.6)-CAT, which contains about 2.2 kb of sequence upstream of the p53 cap site and about 1.6 kb of sequence downstream of the p53 cap site (Fig. 3A and Ref. 75); p53(0.8)-CAT, which contains the 0.8 kb upstream region of p53 (Fig. 3A); or the empty CAT vector, pGCAT-C, were examined in transient transfection and CAT assays for inducibility with the pCMV-mEGR1 construct. As seen in Fig. 3B (left panel), cells transfected with pCMV-mEGR1 showed a 4 -5-fold higher CAT activity from the p53 reporter construct p53(2.2ϩ1.6)-CAT than did the empty CAT reporter construct. Also, both of the p53 reporter constructs p53(2.2ϩ1.6)-CAT and p53(0.8)-CAT showed induction of CAT activity with pCMV-mEGR1 but not with the pCB6 ϩ empty vector (Fig. 3B, middle panel), suggesting that the 0.8-kb region upstream of the cap site in the p53 promoter was sufficient for transcriptional activation by EGR-1. To determine whether the transactivation and the DNA binding functions of EGR-1 were necessary for p53 induction, the cells were transiently cotransfected with p53(2.2ϩ1.6)-CAT and pCMV-mEGR1, pCMV-⌬TA, pCMV-⌬ZF, or vector. As seen in Fig. 3B (right  panel), pCMV-mEGR1, but not the vector, pCMV-⌬TA, or pCMV-⌬ZF, showed a 3-fold induction of CAT activity from the p53-promoter construct. Thus, EGR-1 mutants with deletion of the transactivation domain or DNA-binding domain showed minimal CAT activity, suggesting that these domains of EGR-1 are required for induction of the p53 promoter.
To ascertain that EGR-1 protein from the mEGR-1 overproducer cells actually bound to EBS-1 of the p53 promoter, electrophoretic mobility shift assays were performed. Nuclear protein extracts prepared from cells stably transfected with pCMV-mEGR-1 or vector were incubated with radiolabeled EBS-1. As shown in Fig. 3C (left panel), nuclear extracts from the cells transfected with vector produced two primary bound complexes: cI and cII. On the other hand, nuclear extracts from cells transfected with mEGR-1 produced three primary complexes; cI and cII were similar in intensity to those seen in extracts from vector-transfected cells, whereas the third com-FIG. 1. Ectopically expressed EGR-1 enhances the apoptotic response to TG. A375-C6 cells were stably transfected with vector or pCMV-mEGR1 that encodes full-length mouse EGR-1 protein, and transfected clones were pooled to obtain cell lines. A, the cell lines were examined for expression of EGR-1 protein or actin (for a control) by Western blot analysis. B, the cell lines were exposed for the indicated time periods to either 250 nM or 1 M TG, and apoptotic cells were quantified by TUNEL. A total of 400 cells were scored for TUNEL-positive cells in each experiment, and the data represent means of three experiments. The error bars represent standard deviation. plex, cIII, was predominant in extracts from mEGR-1 overproducers (Fig. 3C, left panel). These results suggested that complex cIII most likely contained EGR-1 bound to EBS-1. To ascertain whether complex cIII represented EGR-1, the nuclear extracts were preincubated either with a preimmune antibody or with EGR-1 antibody before incubation with the radiolabeled probe. These experiments showed that complex cIII was supershifted with the EGR-1 antibody but not with the preimmune antibody, suggesting that it contained EGR-1 (Fig. 3C,  right panel). Together with the results of CAT assays described above, these findings suggest that the EGR-1 protein can directly bind to and transcriptionally activate the p53 promoter.
To ascertain that the induction of p53 was not merely restricted to its promoter, we used Northern blot analysis to examine the ability of EGR-1 to induce the endogenous p53 gene at the RNA level (Fig. 3D) and Western blot analysis to examine the ability of EGR-1 to induce the endogenous p53 gene at the protein level (Fig. 3E). Three different pools of stable transfectants that overexpressed EGR-1 showed 4 -5fold higher levels of p53 RNA and protein than did transfectants containing the vector (Fig. 3, D and E). These findings suggest that ectopic EGR-1 induced expression of the p53 gene, causing increased production of p53 RNA and protein.
p53 Is Required for TG-inducible Apoptosis-Because A375-C6 cells contain wild-type p53 (34), we used these cells to examine the functional requirement of wild-type p53 protein in TG-inducible apoptosis. These studies used dominant-negative mutants of p53, pCMV-m143 (Val to Ala at amino acid 143), and pCMV-m175 (Arg to His at amino acid 175) that had been previously shown (46) to oligomerize with and inhibit the function of wild-type p53 protein. A375-C6 cells were transfected with expression constructs pCMV-m143 or pCMV-m175 or with vector for a control, and pools of stably transfected clones were maintained as cell lines. When whole-cell protein extracts from these transfected cell lines were subjected to Western blot analysis, the levels of p53 protein in cells transfected with the p53 mutant constructs were higher than the levels of endogenous wild-type p53 protein in cells transfected with vector ( Fig.  4A), indicating successful overexpression of the p53 mutants. Next, the stably transfected cells were tested for their ability to compete with endogenous wild-type p53 for transactivation of reporter constructs PG 13 -CAT (containing intact p53-binding sites) or MG 15 -CAT (containing mutations in the p53-binding site) (46). The stably transfected cells expressing m143, m175, or vector were transiently transfected with PG 13 -CAT or MG 15 -CAT. When cell extracts were tested for CAT activity, transfectants expressing mutant p53 proteins showed much lower CAT activity from PG 13 -CAT than did transfectants containing FIG. 2. Both DNA binding and transactivation functions of EGR-1 are essential for an enhanced apoptotic response. A, to identify the structural domains of EGR-1 required to enhance apoptosis, the following expression constructs were used: pCMV-mEGR1, pCMV-⌬TA, pCMV-⌬RM, pCMV-⌬ZF, and pCMV-⌬TA/RM (see text for details). B, A375-C6 cells were stably transfected with the above constructs or vector for a control, and pools of transfected clones were maintained as cell lines. Whole-cell protein extracts from each pool of transfectants were subjected to Western blot analysis with EGR-1 antibody and then with actin antibody. The intensity of EGR-1 or mutant EGR-1 protein signals was determined by densitometric scanning and was normalized with respect to the corresponding signal for actin. The relative protein levels for endogenous EGR-1 in vector-transfected cells, full-length EGR-1, and different mutants are shown. C, the stably transfected cell lines were then transiently transfected with the EBS-CAT reporter construct that contains three tandem EGR-1-binding sites, and CAT assays were performed. D, the stably transfected cells were left unexposed or exposed to either 250 nM or 1 M TG for 24 h and examined for apoptosis by TUNEL (see legend for Fig. 1B). ⌬TA, transcription activation-deficient; ⌬RM, repression module-deficient; ⌬ZF, finger 1-and 2-deficient; ⌬TA/RM, transcription activation-and repression module-deficient. the vector (Fig. 4B). As expected, all the three transfectants produced low level CAT activity from MG 15 -CAT (Fig. 4B). These results, in agreement with those obtained from other cell lines (46), suggest that mutant forms of p53 protein can inhibit the ability of endogenous wild-type p53 in the A375-C6 cells to cause transcriptional activation.  3. EGR-1 induces p53 expression. A, the p53 promoter constructs, p53(2.2ϩ1.6)-CAT and p53(0.8)-CAT, were used to test the ability of EGR-1 to transactivate the p53 promoter. Both constructs contain two putative EGR-1-binding sites; EBS-1 (5Ј-CGCCTACGC-3Ј) located from nucleotide positions Ϫ770 to Ϫ762 and EBS-2 ((TCC) n sequence) located from nucleotide positions Ϫ139 to Ϫ186. B, A375-C6 cells were transiently cotransfected with pCMV-mEGR1 and either p53(2.2ϩ1.6)-CAT or pGCAT-C (left panel); with vector or pCMV-mEGR1 and p53(2.2ϩ1.6)-CAT or p53(0.8)-CAT (middle panel); or with p53(2.2ϩ1.6)-CAT and vector, pCMV-mEGR1, pCMV-⌬TA (⌬TA), or pCMV-⌬ZF (⌬ZF, right panel). Protein extracts were prepared from the transfected cells and were examined for CAT activity. C, electrophoretic mobility shift assays were performed by using radiolabeled EBS-1 and nuclear extracts from cells stably expressing mEGR1 or vector (left panel). The supershift assay was performed either with no antibody or with 2 l of EGR-1 antibody or preimmune rabbit serum as indicated (right panel) to ascertain the presence of EGR-1 in the bound complex. The bound complexes cI, cII, cIII, and supershift with EGR-1 antibody are indicated. D, total RNA was prepared from stably transfected cell lines expressing pCMV-mEGR1 or vector and was subjected to Northern analysis by using p53 cDNA as a probe (left panel). The corresponding ethidium bromide-stained agarose gel shows loading of the RNAs (right panel). E, whole-cell protein extracts were prepared from the stably transfected cell lines, and 20 g of each protein extract was subjected to Western blot analysis for p53 protein by using antibody DO-1. Data are shown for p53 protein expression in three different mEGR1-transfected cell lines (mEGR1.L1, mEGR1.L2, and mEGR1.L3) and in one vector-transfected cell line.
We then determined the susceptibility of transfected cells expressing p53 mutant proteins or vector to TG-inducible apoptosis by exposing them to either 250 nM or 500 nM TG for 48 h. When these cells were assayed for apoptosis with TUNEL, the cell lines expressing mutant p53 proteins showed a significantly lower (p Ͻ 0.0001 by Student's t test) percentage of TUNEL-positive cells than did the cell lines expressing the vector (Fig. 4C). These experiments suggested that the p53 mutants, which interfere with the transactivation function of endogenous wild-type p53, inhibited the ability of TG to cause apoptosis in the transfectants. Thus, wild-type p53 function is essential for apoptosis induced by TG.
p53 Is a Downstream Mediator of EGR-1 Function-Our previous studies have shown that EGR-1 protein is induced by TG in A375-C6 cells (20). To determine whether p53 induction was temporally downstream of EGR-1-induction, we exposed A375-C6 cells to TG for various periods of time; whole-cell protein extracts prepared from the cells were then subjected to Western blot analysis for EGR-1 or p53 protein. As seen in Fig.  4D, EGR-1 protein levels increased at 1 h of exposure to TG, and peak levels were reached at 2 h; thereafter, the levels decreased sharply at 3 and 5 h. The levels of p53 protein were low until 1 h of exposure to TG and then began to increase at 2 h of exposure, and peak levels were reached at 3 h. These findings suggest that p53 induction is temporally downstream of EGR-1 induction by TG.
We reasoned that if wild-type p53 functions at a level downstream of EGR-1 in a linear pathway, the dominant-negative mutants of p53 should abrogate EGR-1-dependent apoptosis. To address this question, we stably cotransfected A375-C6 cells with pCMV-mEGR-1 plus vector, pCMV-mEGR-1 plus pCMV-m175, pCMV-m175 plus vector, or with vector alone. The stably transfected cell lines, thus obtained, were tested for their responsiveness to TG-inducible apoptosis by exposing them to 250 nM or 1 M TG for 24 h, and the apoptotic cells were scored with TUNEL. As seen in Fig. 4E, cell lines coexpressing mEGR1 and m175 showed a significantly lower (p Ͻ 0.0001 by Student's t test) percentage of apoptotic cells than did cell lines coexpressing mEGR1 and vector at each of the two different concentrations of TG, suggesting that mutant p53 proteins can abrogate the ability of EGR-1 to cause apoptosis. Thus, wildtype p53 function is essential for EGR-1-dependent apoptosis induced by TG, and p53 is a downstream effector of EGR-1 function.

FIG. 4. p53 is required for TG-inducible and EGR-1-dependent apoptosis.
A, whole-cell protein extracts from stably transfected cell lines expected to contain either vector or mutant p53 constructs (pCMV-m143 or pCMV-m175) were examined for p53 expression or actin as a loading control by Western blot analysis. B, to determine the effect of ectopically overexpressed mutant p53 proteins on the transactivation function of endogenous wild-type p53, mutant-p53 overexpressor cells or vector-transfected cells were transiently transfected with PG 13 -CAT (containing 13 intact p53 binding sites) or with MG 15 -CAT (in which the 15 p53-binding sites were mutated). Protein extracts from the transiently transfected cells were tested for CAT activity. C, A375-C6 cells that stably expressed vector or the mutant p53 proteins m143 or m175 were tested by TUNEL for susceptibility to TG-inducible apoptosis (see legend for Fig. 1B). D, A375-C6 cells were treated with TG (500 nM) for various periods of time; whole-cell protein extracts from the cells were then subjected to Western blot analysis for EGR-1, p53, or actin protein levels. The intensity of the signal in each lane was determined by densitometric scanning and the signals for EGR-1 or p53 were normalized with respect to the corresponding signal for actin. The relative normalized expression levels are shown as a function of time. E, A375-C6 cells were cotransfected with pCMV-mEGR1 plus vector, with pCMV-mEGR1 plus pCMV-m175, with pCMV-m175 plus vector, or with vector alone, and transfected cell lines stably expressing the plasmids were tested for apoptosis by TUNEL (see legend for Fig. 1B).

DISCUSSION
The p53 protein is required for apoptosis induced by diverse exogenous signals (33,(42)(43)(44)(45), and the findings of this study have expanded the range of p53-inducers to EGR-1 and intracellular Ca 2ϩ . The demonstration that wild-type p53 function is essential for apoptosis induced by intracellular Ca 2ϩ elevation is of broad biological significance because intracellular Ca 2ϩ elevation is central to apoptosis in a number of in vitro and in vivo model systems (15)(16)(17)(18)20). Consistent with the observations from our previous studies indicating that inhibition of EGR-1 expression or function abrogates apoptosis (20), the findings of the present study indicated that ectopic EGR-1 enhances the apoptotic action of TG. Because the proapoptotic action of EGR-1 requires intact transactivation and DNA binding functions, we examined genes that are central mediators of the apoptotic process for EGR-1-responsive sites in their promoter. This led to the identification of p53 as a novel downstream target of EGR-1. Two putative EGR-1 consensus binding sites, EBS-1 and EBS-2, were identified in the p53 promoter. EBS-1 conforms to the GC-rich consensus motif and EBS-2 conforms to the (TCC) n motif. We showed by electrophoretic mobility shift and supershift assays that EGR-1 binds to EBS-1. These findings showed direct interaction of EGR-1 with the p53 promoter element. Furthermore, we showed that EGR-1 can cause transcriptional up-regulation of the p53 promoter. The ability of EGR-1 to induce the p53 promoter required the transcriptional activation functions of EGR-1, as evident from the finding that full-length EGR-1 but not mutants that lacked the transactivation or DNA-binding domain up-regulated the p53 promoter. EGR-1 also caused induction of p53 RNA and protein. These findings are consistent with the previous observation that up-regulation of the p53 gene at the transcriptional level contributes to up-regulation of p53 protein levels (80). Most importantly, we showed that when transactivation function of p53 protein was blocked with two different dominant-negative mutants, the ability of EGR-1 to enhance TG-inducible apoptosis was also abrogated. These findings indicate that the transactivation function of p53 protein is required for an apoptotic response to TG and EGR-1. The proapoptotic gene bax is known to be transcriptionally induced by wild-type p53 but not by mutant p53 proteins (39,40,44), and A375-C6 transfectants that expressed dominant-negative p53 mutants showed down-regulation of endogenous bax as compared with vector-transfected cells that expressed only endogenous wild-type p53, 2 suggesting that bax induction by p53 may mediate the apoptotic action of TG and EGR-1. Together, these findings establish the p53 gene as a direct target of EGR-1 that is functionally important for apoptosis. Because EGR-1 has been implicated in both positive and negative regulation of growth, these findings can be extended in future studies to address the relevance of EGR-1-mediated p53 induction in other growth control processes.
EGR-1 enhances the apoptotic action of TG or ionizing radiation but blocks the growth arresting action of interleukin-1 (20,34,77). The mechanisms by which EGR-1 induces these distinct phenotypic effects are unclear. EGR-1-binding sites have been identified in the promoter regions of a number of genes, including those encoding growth factors, growth factor receptors, cell cycle regulators, and protooncogenes (27), but these genes have not been shown to be transcriptionally upregulated or to have biological relevance in the proapoptotic action of EGR-1. Toward the goal of determining the mechanism of EGR-1-dependent apoptosis, we used ectopic EGR-1 overexpression and determined its impact on downstream events. A375-C6 transfectants that stably expressed EGR-1 were confirmed by TUNEL studies to show increased susceptibility to apoptosis induced by TG as compared with control vector-transfected cells that contained only endogenous levels of EGR-1, confirming that ectopic overexpression of EGR-1 sensitizes cells to apoptosis. The successful selection of transfectants stably expressing high levels of EGR-1 protein suggested that EGR-1 was not sufficient on its own to cause apoptosis. Because EGR-1 protein contains several phosphorylation sites and phosphorylated forms of the protein have been shown to bind the EBS motif (27), we are investigating the possibility that exposure to TG causes post-translational phosphorylation of EGR-1 protein that results in activation of its transregulatory function leading to induction of downstream genes and apoptosis.
Because EGR-1 protein acts either as a transcriptional upregulator or suppressor of promoter constructs depending on the promoter context and cell background (27), we determined which of these two transregulatory functions was important for apoptosis in A375-C6 cells. Wild-type EGR-1 or the mutant that lacked the restriction module but contained the transactivation and DNA-binding domains showed transcriptional activation of the EBS-CAT reporter construct and an enhanced apoptotic response to TG. On the other hand, EGR-1 mutants that lacked the transactivation domain or DNA-binding domain did not transactivate the reporter construct and did not show an enhanced apoptotic response to TG. Because these cells contained comparable amounts of the ectopic proteins, the findings suggested that the DNA-binding and transactivation domains of EGR-1 are necessary for an enhanced apoptotic response to TG. In the light of recent reports indicating that EGR-1 can directly bind to intracellular proteins (81,82), it would be reasonable to speculate that EGR-1 may enhance apoptosis by sequestration of key growth/survival promoting proteins by protein-protein interactions. However, this possibility was ruled out by the observation that the proapoptotic functions of EGR-1 were absent in the DNA-binding and transactivation mutants, which together spanned the entire EGR-1 protein and thus included the domains shown to bind other proteins. Thus, the proapoptotic function of EGR-1 requires the DNA binding and transactivation functions of EGR-1. By contrast, the proapoptotic functions of another immediate-early gene product c-Fos are dependent on transcriptional repressor functions of the protein that cause negative regulation of survival genes in a p53-dependent manner (70). Thus, immediateearly genes that encode transcription factors can regulate apoptosis by distinct p53-dependent mechanisms.
The present study has identified p53 as a functionally relevant and direct target of EGR-1 in the apoptosis pathway. Because p53 is a central component of apoptotic pathways in diverse model systems (33,(42)(43)(44)(45), this observation establishes the importance of EGR-1 in apoptosis in our model system, i.e. A375-C6 cells exposed to TG. Because EGR-1 is an indispensable component of the apoptotic or growth arrest pathways in A375-C6 cells (20,77), i.e. loss of EGR-1 function is not compensated by other EGR-1 family members in these cells, our model system is ideal for studying EGR-1 function and downstream targets. However, because EGR-1 knock-out mice do not show any abnormalities related to growth, differentiation, or apoptosis (60, 61), EGR-1 function may be a dispensable component of apoptotic pathways in other model systems. It is possible that other members of the EGR-1 family may have compensated for the loss of EGR-1 function and produced the normal growth, differentiation, or apoptotic features in the knock-out mice. To ascertain that the importance of EGR-1 in apoptosis is not restricted to our model system, it is necessary to study EGR-1 function in a number of other apoptosis model systems wherein the results are not influenced by functional redundancy of other EGR-1 family members. EGR-1 has been shown to function as an anti-proliferative protein in a number of tumor cell types (32,33). Because the presence of functional p53 is required for successful transduction of the EGR-1-dependent apoptotic signal, tumor cells that contain mutant forms of p53 or that lack p53 can be expected to be unresponsive or resistant to the EGR-1-dependent apoptotic pathway. A number of reports have suggested that the process of tumor progression is linked to the absence of wild-type p53 tumorsuppressor function (54 -58). In such tumor cells, transduction of an apoptotic signal by EGR-1 is expected to be unsuccessful. On the other hand, because p53 protein is wild-type in normal cells, EGR-1 up-regulation may constitute the important cellular event directed at preventing malignant transformation or tumor progression by induction of apoptosis via p53 induction. Further studies on a broad panel of primary cell cultures and tumor models are required to test the validity of this hypothesis.