p53 Suppresses the Activation of the Bcl-2 Promoter by the Brn-3a POU Family Transcription Factor*

The Brn-3a POU family transcription factor has been shown to strongly activate expression of the Bcl-2proto-oncogene and thereby protect neuronal cells from programmed cell death (apoptosis). This activation of the Bcl-2 promoter by Brn-3a is strongly inhibited by the p53 anti-oncogene protein. This inhibitory effect of p53 on Brn-3a-mediated transactivation is observed with nonoverlapping gene fragments containing either the Bcl-2 p1 or p2 promoters but is not observed with other Brn-3a-activated promoters such as in the gene encoding α-internexin or with an isolated Brn-3a binding site from the Bcl-2 promoter linked to a heterologous promoter. In contrast, p53 mutants, which are incapable of binding to DNA, do not affect Brn-3a-mediated activation of the Bcl-2 p1 and p2 promoters. Moreover, Brn-3a and p53 have been shown to bind to adjacent sites in the p2 promoter and to directly interact with one another, bothin vitro and in vivo, with this interaction being mediated by the POU domain of Brn-3a and the DNA binding domain of p53. The significance of these effects is discussed in terms of the antagonistic effects of Bcl-2 and p53 on the rate of apoptosis and the overexpression of Brn-3a in specific tumor cell types.

The rate of apoptosis (programmed cell death) is controlled by the balance between proteins that activate processes resulting in such death and other proteins that act to inhibit these processes. Thus, for example, the Bcl-2 proto-oncogene was originally identified on the basis of its activation by chromosomal translocation in non-Hodgkin B cell lymphomas (1) and was subsequently shown to protect a wide variety of different cell types from programmed cell death or apoptosis (for review see Ref. 2). Conversely, the p53 anti-oncogene protein, as well as inhibiting cellular proliferation, can also stimulate programmed cell death (for review see Ref. 3).
It is clear therefore that Bcl-2 and p53 represent proteins with opposite effects on the rate of apoptosis. Moreover, it appears that Bcl-2 can specifically inhibit p53-dependent apoptosis. Thus, although overexpression of p53 can induce apoptosis in different cell types, this is prevented by overexpression of Bcl-2 (4,5). In this regard, it is of interest that high levels of p53 are associated with low levels of Bcl-2 and vice versa, both during normal rat development (6) and in different types of tumors (7)(8)(9).
These findings suggest therefore that as well as being func-tionally antagonistic to one another, p53 and Bcl-2 may also be interlinked in terms of the processes regulating their expression. This possibility is supported by the finding that artificial overexpression of p53 in a murine leukemia cell line results in reduced Bcl-2 expression (10). Similarly, Bcl-2 levels are elevated in several tissues of knockout mice lacking functional p53 (10,11). Evidently, these results raised the possibility that the p53 transcription factor may have an inhibitory effect on Bcl-2 gene transcription. In agreement with this it has been shown that p53 can inhibit the Bcl-2 gene promoter acting via a negative regulatory element in the 5Ј-untranslated region of the Bcl-2 gene (12). Although this effect is evidently of interest, it is a relatively small one, with p53 producing a reduction in basal promoter activity between 1.5-and 2-fold via this region (12). Hence, it remains unclear whether p53 has a significant effect on Bcl-2 gene transcription in particular situations. We have recently characterized the effect of the Brn-3a POU family transcription factor on the expression of the Bcl-2 gene. Thus, the activity of the Bcl-2 promoter is increased approximately 30-fold by co-transfection with Brn-3a (13). Similarly, neuronal cell lines stably transfected with Brn-3a show approximately 15-fold overexpression of Bcl-2 protein derived from the endogenous Bcl-2 gene packaged within its normal chromatin structure. Moreover, such overexpression of Brn-3a and the corresponding enhancement of Bcl-2 expression can protect the primary neuronal cells from apoptotic stimuli such as withdrawal of neurotrophic factors (13).
In view of the significant effect of Brn-3a on Bcl-2 gene expression in neuronal cells, we have therefore investigated the effect of p53 on activation of the Bcl-2 promoter by Brn-3a.

MATERIALS AND METHODS
Plasmid Constructs-The Brn-3a and Brn-3b expression vectors contain full-length cDNA clones for these factors under the control of the Moloney murine leukemia virus promoter (14,15). Promoter-reporter constructs containing the promoters of Bcl-2 (13,16) and ␣-internexin (17,18) linked to a luciferase or chloramphenicol acetyltransferase gene have been described previously. An octamer-related motif from the human papilloma virus promoter, which binds Brn-3a (14), and a binding site for Brn-3a from the Bcl-2 promoter (13) were cloned upstream of the herpes simplex virus thymidine kinase promoter in the vector pBL CAT2 (19). p53 expression vectors contain cDNAs encoding wild type p53 or the R175H and C⌬30 mutants under the control of the cytomegalovirus immediate early promoter. The R175P and R273H mutants were a kind gift of M. Oren and K. Vousden. Plasmids used to produce in vitro transcribed and translated p53 and truncated proteins contain the appropriate region of p53 cloned into the pT7 vectors (20).
Transient Transfection-The ND7 cell line obtained by the immortalization of primary sensory neurons from dorsal root ganglia (21) was routinely cultured in L15 medium containing 10% fetal calf serum. Transient transfection was carried out according to the method of Gorman (22). Routinely, 1 ϫ 10 6 cells were transfected with 10 g of the reporter plasmid and 10 g of the expression vectors together with 2 g of pCMV␤ plasmid containing the Escherichia coli lacZ gene under the control of the constitutive cytomegalovirus immediate early promoter and cells that were harvested 48 h later. The efficiency of transfection of each sample was determined with a chemiluminescent assay for ␤-galactosidase activity using a commercial kit (Galactolight Plus, Tropics), and the values were subsequently used to equalize the values obtained from luciferase and CAT 1 assays. In all cases, the value obtained in transfections of the reporter plasmid with an expression vector lacking any insert was set at 100%, and all other values were compared with it.
Luciferase and Chloramphenicol Acetyltransferase Assay-Assays of luciferase activity were carried out using a commercially available kit (Promega) and a Turner luminometer, whereas assays of chloramphenicol acetyltransferase activity were carried out according to the method of Gorman (22).
Protein-Protein Interaction-The protein-protein interaction assay was performed according to the method described by Baniahmad et al., (23). Briefly, Brn-3a or Brn-3b⅐GST fusion proteins linked to glutathione-Sepharose beads were prepared and stored in NENT buffer (100 mM NaCl, 1 mM EDTA, 20 mM Tris (pH 8), 0.5 Nonidet P-40, 0.5% milk powder). Prior to use, approximately 1-2 g of the fusion proteins were incubated in 20% milk powder in NENT buffer for 15 min at room temperature. The beads were washed in 1 ml of NENT buffer and then in 1 ml of transcription wash buffer (20 mM HEPES (pH 7.9), 60 mM NaCl, 1 mM dithiothreitol, 6 mM MgCl 2 , 8.2% glycerol, 0.1 mM EDTA). Following SDS-polyacrylamide gel electrophoresis analysis and densitometry, the volumes of in vitro translated proteins were adjusted so that relatively equal amounts of each protein were used. The in vitro translated p53 proteins or equivalent amounts of the luciferase control proteins were then incubated with the beads in 100 l of transcription buffer for 1 h at room temperature. The beads were washed (five times with 1 ml of NENT buffer), and the proteins were solubilized in SDSloading buffer, heated to 100°C for 5 min, and resolved on an SDS-12% polyacrylamide gel. Following electrophoresis, the gel was dried and exposed to radiographic film or a PhosphorImager screen. The amounts of protein retained following the interaction studies were assessed by comparing the intensity of the bands resulting after the protocol with that resulting when equivalent amounts of input protein were run on a similar gel.
Immunoprecipitation-Immunoprecipitation was carried out to assess the interaction between Brn-3 proteins and p53 in vivo. Protein extracts were from rat tissues such as brain, ovary, and kidney (negative control). Tissues were homogenized in extraction buffer containing 50 mM Tris-HCl (pH 8.0) 170 mM NaCl, 0.5% Nonidet P-40, 50 mM NaF, and 10 g of the protease inhibitors leupeptin, aprotinin, and pepstatin per milliliter plus 1 mM phenylmethylsulfonyl fluoride. The tissue homogenate was centrifuged at 14,000 ϫ g for 10 min to pellet debris. The supernatant was precleared by incubatiing it with 25 l of protein A-protein G-agarose slurry for 30 to 60 min at 4°C. After centrifugation the supernatant was incubated overnight at 4°C with one of the following: 10 l of the anti-p53 antibody, 10 l of antibody to the Bad protein (which does not interact with the Brn-3 proteins), or no antibody. The immunocomplexes were then collected by incubation with 30 l of the protein A/protein G/agarose slurry for 30 min. The agarose beads were washed five times with buffer containing 10 mM NaCl, 1 mM EDTA, 20 mM Tris (pH 8), and 0.5% Nonidet P-40 and then boiled in 1ϫ SDS-sample buffer (2% SDS, 10% glycerol, 62 mM Tris-HCl (pH 6.8), 1% ␤-mercaptoethanol) and loaded onto an SDS 12%-polyacrylamide gel. A Western blot was produced, and this was probed with the antibodies to the Brn-3 proteins (1:1000 dilution).
Electrophoretic Mobility Shift Asssay-The electrophoretic mobility shift assay was carried out as described by Budhram-Mahadeo et al. (24). Briefly, 3 l of the in vitro translated p53 protein or luciferase (control) protein was added as indicated to 2 l of 10ϫ electrophoretic mobility shift assay buffer (10 mM HEPES (pH 7.9), 60 mM KCl, 4% Ficoll, 1 mM dithiothreitol, 1 mM EDTA) containing 2 g of poly(dI-dC) to prevent nonspecific interactions, along with specific or nonspecific competitor oligonucleotides, and kept at room temperature for 5 min. One nanogram of 5Ј-end-labeled oligonucleotide probe (labeled with T4 kinase and purified on a Sephadex G-25 column) was then added, and this combination was mixed briefly, spun in a microcentrifuge for 5 s, and then incubated on ice for 45 min to 1 h. The DNA-protein complexes were resolved from free DNA by gel electrophoresis on a 7% polyacrylamide gel run in 0.5ϫ Tris borate/EDTA for 2-2.5 h at 4°C. The oligonucleotide containing a consensus p53 binding site had the sequence 5Ј-CGAGAGACATGCCCAGGCATGCC-3Ј. The oligonucleotide containing the putative p53 binding motifs from the Bcl-2 promoter had the sequence 5Ј-TTACAAAAAGGAAACTTGACAGAGGATCATGCTG-TACTTAAAAAAGAGCT-3Ј. Homologies to the consensus p53 binding site PuPuPuCA(TA)TGPyPyPy (25) are underlined.

RESULTS
We have previously demonstrated that Brn-3a is able to strongly activate the full-length Bcl-2 promoter (13). We therefore tested the effect of co-transfecting a p53 expression vector on the activation of the Bcl-2 promoter by Brn-3a. As illustrated in Fig. 1, Brn-3a was able to strongly transactivate the full-length Bcl-2 promoter in accordance with our previous results (13). However, this transactivation was completely abolished by inclusion of a p53 expression vector in the cotransfections. Indeed, the inclusion of p53 in the transfections containing Brn-3a resulted in a reduction in promoter activity below that observed in the absence of Brn-3a. This effect was not observed with the internal control plasmid in which the cytomegalovirus immediate early promoter drives ␤-galactosidase expression and was therefore not due to a nonspecific effect of p53 on all promoter activity in the transfected cells. Hence, p53 can specifically abolish activation of the Bcl-2 promoter by Brn-3a as well as dramatically reduce promoter activity in ND7 cells, which are derived from sensory neurons and express endogenous Brn-3a (17, 26). To determine the specificity of this effect of p53 on activation of the Bcl-2 promoter, we examined the effect of p53 on activation of other promoters that we have previously shown to be responsive to Brn-3a. In particular, we have shown that a test promoter containing the herpes simplex virus thymidine kinase promoter, with an added octamer motif that binds Brn-3a, is responsive to activation by Brn-3a (14,17). We therefore tested the effect of p53 on the activation of this promoter by Brn-3a. In these experiments (Fig. 2) over-expression of p53 had no effect on the ability of Brn-3a to activate this test promoter. Indeed, p53 appeared to have a strong activating effect when transfected with the promoter in the absence of Brn-3a and had no effect on activation by Brn-3a.
In previous experiments, we have shown that the artificial test promoter containing an octamer motif can be activated by the isolated POU domain of Brn-3a, whereas activation of the Bcl-2 promoter requires an N-terminal domain of Brn-3a in addition to the POU domain (13,14,27). We therefore wished to test the effect of p53 on another promoter activated by Brn-3a, in which activation is also dependent upon the presence of the N terminus. We therefore tested the effect of p53 on the activation of the ␣-internexin promoter, in which activation requires the N terminus of Brn-3a (17). However, in these experiments no effect of p53 on the ability of Brn-3a to activate the ␣-internexin promoter was observed (Fig. 3). Hence, the inhibitory effect of p53 on stimulation of the Bcl-2 promoter by Brn-3a was specific to this promoter and was not observed with other Brn-3a-responsive promoters even when, as in the case of the ␣-internexin promoter, such activation required the Nterminal activation domain of Brn-3a.
In view of the specificity of this effect for the Bcl-2 promoter, we wished to characterize it further. Thus, the Bcl-2 promoter construct used in the experiments shown in Fig. 1 contains both of the two independent promoters, p1 and p2, that drive expression of the Bcl-2 gene (Fig. 4). We have previously demonstrated that Brn-3a is able to transactivate both the p1 and the p2 promoter (13). We therefore wished to investigate the effect of p53 on the activation of each of these promoters by Brn-3a. As shown in Fig. 5, p53 was able to abolish the activation of the p1 Bcl-2 promoter by Brn-3a, although in this case the activity of the promoter was not reduced below its basal level. Stronger inhibition of activation by Brn-3a, with promoter activity being reduced below the basal level, was observed for a fragment of the p2 promoter from Ϫ740 to Ϫ170 relative to the transcription start site (Fig. 6) and for a shorter fragment containing the region of the p2 promoter from Ϫ740 to Ϫ355 (data not shown), which does not contain the element  (Ϫ279 to Ϫ85) that mediates the previously described weak inhibitory effect of p53 on the basal activity of the Bcl-2 promoter (12) (see Fig. 4).
The region of the Bcl-2 promoter from Ϫ740 to Ϫ355, which mediates the inhibitory effect of p53 on Brn-3a transactivation, contains a site at Ϫ598 to Ϫ587 that has been shown to bind Brn-3a in DNA mobility shift assays (13) and that can confer a response to Brn-3a on a heterologous promoter (27). We therefore tested whether p53 was able to interfere with the activation of a heterologous thymidine kinase promoter containing this added Brn-3a binding site. In these experiments, however, no inhibition of promoter activation by Brn-3a was observed when p53 was included in the co-transfection experiments (Fig.  7). Thus, as in the case of an octamer motif linked to the heterologous promoter, p53 was unable to repress a heterologous promoter containing the isolated Brn-3a binding site from the Bcl-2 promoter in the absence of any other region of the promoter.
This finding suggested that p53 may need to bind to the Bcl-2 promoter between Ϫ740 and Ϫ355 to mediate its inhibitory effect on Brn-3a-mediated transactivation and that activation of a heterologous promoter by the isolated Brn-3a binding site is therefore not inhibited by p53. To investigate this possibility further, we tested the effect of p53 mutants that are unable to bind to DNA. As illustrated in Fig. 8, p53 mutants in which one or other of the arginines at amino acids 175 or 273 have been changed to histidine and which are unable to bind to DNA (28,29) were also unable to interfere with Brn-3a-mediated activation of the Bcl-2 promoter.
In contrast, the ability to repress Brn-3a-mediated activation was observed with another mutation at position 175 in which a proline residue was substituted for the arginine. This mutation results in a loss of the ability to induce apoptosis in fibroblasts but does not impair the ability to transactivate target promoters and retains the ability to bind to DNA (28,30). Similarly, the ability of p53 to inhibit Brn-3a-mediated activation was unaffected by the deletion of a regulatory domain that is contained in the last 30 C-terminal amino acids of the p53 protein.
These findings suggest, therefore, that the effect of p53 on Brn-3a-mediated activation of the Bcl-2 promoter may be dependent on the binding of p53 to the promoter. This dependence would explain why p53-mediated repression is ineffective on Brn-3a-activated promoters lacking such a p53 binding site or containing only an isolated Brn-3a binding site from the Bcl-2 promoter.
We therefore examined the sequence of the p2 Bcl-2 promoter from Ϫ740 to Ϫ355 for potential p53 binding sites. Between Ϫ558 and Ϫ535, this region contains two motifs that closely resemble the PuPuPuCA(TA)TGPyPyPy consensus binding site for p53 (25). These motifs are located close to the Brn-3a binding site at Ϫ598 to Ϫ581 in the p2 promoter (Fig.  4). In DNA mobility shift assays, this sequence was able to compete for p53 binding with the consensus binding p53 binding site, although an unrelated oligonucleotide was unable to do so (Fig. 9). Hence, the region of the p2 promoter that is necessary for the inhibitory effect of p53 on Brn-3a-mediated activation can also bind p53.
Although these experiments confirm the likely importance of p53 Suppresses Bcl-2 Promoter Activation DNA binding by p53 in its inhibitory effect, they do not indicate the manner in which the inhibitory effect of p53 is mediated following DNA binding. We therefore wished to determine whether Brn-3a and p53 can also interact directly with one another via a protein-protein interaction. To do this, we carried out pull-down assays using Brn-3a protein linked to glutathione S-transferase (GST) and in vitro transcribed and translated p53 protein. These experiments showed, that full-length p53 was readily able to interact with Brn-3a (Fig. 10A). This interaction could also be observed when p53 was linked to GST and mixed with in vitro transcribed and translated Brn-3a protein, but it was not observed when either p53 or Brn-3a alone was mixed with GST itself (data not shown).
Moreover, by using fragments of p53 in this assay we were able to demonstrate that this interaction could be observed with a p53 fragment containing the first 292 amino acids of the protein but not with a fragment containing only the N-terminal 106 amino acids, further confirming the specificity of this effect. Similarly, no interaction was observed with a fragment containing amino acids 202-393, even though such an interaction was observed with a fragment containing amino acids 44 -393. Hence, the interaction with Brn-3a maps to the DNA binding domain of p53 (Fig. 10A). Similarly, interaction with p53 could be observed with a GST fusion protein containing only the POU domain of Brn-3a (Fig. 10B), indicating that this region of Brn-3a is required for the interaction. A similar interaction between Brn-3a and p53 could also be demonstrated using a yeast two-hybrid assay (data not shown).
These data indicate therefore that interaction with Brn-3a and p53 mediated by the DNA binding domains of the two proteins can readily be observed in in vitro assays. To demonstrate that this interaction also occurs in vivo in intact cells, we immunoprecipitated extracts of different tissues with an antibody to p53 and then probed the resulting immunoprecipitates with an antibody to Brn-3a in a Western blotting assay. As illustrated, in Fig. 11, when extracts from rat brain were utilized in this assay, Brn-3a was readily detected in the immunoprecipitate using p53 antibody, indicating that the two factors interact. Indeed, the amount of Brn-3a detected was greater than that observed when an antibody to the estrogen receptor was used, indicating that the p53/Brn-3a interaction is comparable or stronger than that which we previously observed between Brn-3a and the estrogen receptor (24).
This interaction between p53 and Brn-3a was a specific one because, as shown in Fig. 11, no Brn-3a was observed in the immunoprecipitate in the absence of a primary antibody or when immunoprecipitation was carried out with an antibody to the Bad protein, which shows no interaction with Brn-3a (24). Similarly, no Brn-3a was observed in the immunoprecipitate with anti-p53 antibody when extracts from the rat kidney, which does not express Brn-3a, were used (data not shown).
Hence, the ability of p53 to inhibit activation of the Bcl-2 promoter by Brn-3a is paralleled by a direct interaction between these two proteins, which occurs both in vitro and within intact mammalian cells in vivo.

DISCUSSION
In this report we have shown for the first time that the p53 anti-oncogene protein is able to strongly inhibit the activation of the Bcl-2 promoter by Brn-3a. This effect is much stronger than the previously reported weak repression of basal Bcl-2 promoter activity via a negative element located in the 5Јuntranslated region of the Bcl-2 gene (12). Thus, in that study, the effect of p53 on the Bcl-2 promoter was relatively weak, producing an approximately 1.5-2-fold reduction in promoter activity on transfection with wild type p53 compared with the promoter activity observed with an inactive p53 mutant (12). In contrast, in our study, p53 was able to reduce the promoter activity in the presence of Brn-3a by approximately 50-fold compared with the activity observed in the presence of Brn-3a alone (see for example Fig. 1). Moreover, this effect is specific to the Bcl-2 promoter because it was not observed with the ␣-internexin promoter. This is of particular importance because this promoter, like the Bcl-2 promoter, requires the N-terminal activation domain of Brn-3a for its activation (13,17,27). Similarly, repression of Brn-3a-mediated activation by p53 was not observed with test promoters containing either a cloned octamer binding site for Brn-3a or a binding site for Brn-3a from the Bcl-2 promoter.
This finding indicates therefore that some feature of the architecture of the Bcl-2 promoter, apart from a Brn-3a binding site, is required for p53 to be able to repress this promoter.
Interestingly, mutants of p53, which are unable to bind to DNA, were not able to repress the stimulation of the Bcl-2 promoter by Brn-3a. It is therefore possible that the Bcl-2 promoter contains binding sites for p53 through which the inhibitory effect is mediated. Indeed, we have observed that p53 does bind to a region of the p2 promoter adjacent to the Brn-3a binding site. This element is present within the p2 promoter fragment in which activation by Brn-3a is inhibited by p53 but is absent from the construct containing only the Brn-3a binding site, which is not repressed by p53, indicating that it is likely to be important for the effect we observe.
In contrast, the smallest p2 promoter fragment (Ϫ740 to Ϫ355) in which Brn-3a-mediated activation is repressed by p53 does not contain the negative regulatory element (Ϫ279 to Ϫ85) that has previously been shown to mediate a weak effect on p53 on the basal activity of the Bcl-2 promoter (12). The inhibitory effect of p53 on Brn-3a-mediated transactivation can certainly occur in the absence of this element, and it is therefore unlikely that the inhibitory effect of p53 on Brn-3a-mediated activation of the full promoter involves this previously defined element.
Although the p53 binding site in the p2 promoter that we have identified is likely to be involved in the effects we observe, it cannot be the only element able to mediate the effect of p53 on Brn-3a-mediated activation, because this effect was also observed with the Bcl-2 p1 promoter construct, which lacks this element. It is noteworthy, however, that the full-length Bcl-2 promoter and p2 constructs that contain the p53 binding site demonstrated lower activity in the presence of Brn-3a and p53 than that observed in the absence of any added regulatory proteins. This effect was not observed for the p1 promoter, which lacks this motif. Clearly, further detailed studies of the Bcl-2 p1 and p2 promoters will be required to fully characterize the sequences involved in the p53 effect on Brn-3a-mediated activation.
It is already clear however that p53 can produce a strong inhibition of the stimulatory effect of Brn-3a on the p1 and p2 Bcl-2 promoters and that, at least in the case of the p2 promoter, these factors bind to adjacent sites in the promoter. Moreover, such an inhibitory effect of p53 on Brn-3a is paralleled by a direct interaction between Brn-3a and p53, which we have demonstrated both in vitro and in vivo in mammalian cells. It is likely therefore that, following DNA binding, a protein-protein interaction occurs between Brn-3a and p53 that prevents Brn-3a from stimulating the Bcl-2 promoter.
Although this interaction may directly inhibit the activity of Brn-3a, it is also possible that it prevents Brn-3a from interacting with another factor that is required for activation of the Bcl-2 promoter. Thus, we have shown that activation of the Bcl-2 promoter by Brn-3a occurs only upon transfection into neuronal cells and not in non-neuronal cell types (27). Moreover, we have shown that the stimulatory effect of Brn-3a on the Bcl-2 promoter can be inhibited by including in the transfections the isolated N terminus of Brn-3a unlinked to any DNA binding domain that would remove from the DNA any Brn-3a-interacting factor necessary for transcriptional activation (27). These experiments suggest therefore that Brn-3a may interact with a neuron-specific factor to mediate its effect upon the Bcl-2 promoter, thereby raising the possibility that interaction of Brn-3a with p53 prevents its interaction with another activating factor.
Whatever the precise mechanism of the effect of p53 on Brn-3a-mediated activation of the Bcl-2 promoter, its effect is likely to be of importance in terms of the regulation of the Bcl-2 gene in specific cells types in which Brn-3a is expressed. Thus, it may, at least in part, account for the opposite expression patterns of p53 and Bcl-2 both during normal rat development (6) and in different tumors (7,9,31) as well as the increased levels of Bcl-2 that are observed in p53 knockout mice (10,11). Thus, in the absence of p53 Brn-3a would be able to strongly stimulate the Bcl-2 promoter, whereas in the presence of p53 this effect would not be observed.
These effects are likely to be of particular importance in tumorigenesis because Brn-3a has been shown to be overexpressed in a number of different tumor cells, notably in cervical tumors (32), aggressive neuroendocrine tumors (33), and neuroblastomas (34,35). Moreover, Brn-3a has been shown to cooperate with the myc oncogene in transforming primary cells with such an oncogenic effect being dependent upon the presence of the same N-terminal domain that is involved in stimulating Bcl-2 gene expression (15).
Hence, Brn-3a may exert an oncogenic effect by stimulating the activation of Bcl-2, with such an effect being opposed by the p53 protein. This effect is of particular interest in the case of neuroblastomas because such tumors express high levels of Bcl-2 similar to those found in B cell lymphomas, but unlike lymphomas they do not exhibit any structural rearrangement of the Bcl-2 gene that could account for its overexpression. This finding suggests that specific transcription factors are likely to stimulate transcription of the unrearranged Bcl-2 gene in these cells (36). Moreover, in different patients, reduced Bcl-2 expression in neuroblastoma cells correlates with enhanced apoptosis and good prognosis/spontaneous regression of the tumor (31,37), and p53 and Bcl-2 exhibit a reciprocal expression pattern in these tumors (8).
It is therefore of interest that studies with a number of p53 mutants have indicated that mutants that do not bind to DNA cannot induce apoptosis (for review see Refs. 38 and 39), paralleling the need for DNA binding for p53 to inhibit the effect of Brn-3a on the Bcl-2 promoter. Similarly, such mutant studies have suggested that the ability to suppress tumor growth (28) and induce apoptosis (30,40) is more closely correlated with the ability of p53 to repress some promoters rather than with its ability to activate other promoters. Hence, the ability to interfere with Brn-3a-mediated activation of the Bcl-2 promoter may represent one of the means by which transcriptional inhibition by p53 acts to promote apoptosis and inhibit tumorigenesis.
In summary, our data demonstrate for the first time the ability of p53 to strongly repress the significant activation of the Bcl-2 promoter by Brn-3a. This effect is likely to play a key role in determining the level of Bcl-2 gene expression and hence of apoptosis, particularly in neuronal cells, as well as accounting for, at least in part, the previously reported reciprocal expression pattern of p53 and Bcl-2 in specific neuronal and tumor cell types.