Wild-type p53 and p73 negatively regulate expression of proliferation related genes.

When normal cells come under stress, the wild-type (WT) p53 level increases resulting in the regulation of gene expression responsible for growth arrest or apoptosis. Here we show that elevated levels of WT p53 or its homologue, p73, inhibit expression of a number of cell cycle regulatory and growth promoting genes. Our analysis also identified a group of genes whose expression is differentially regulated by WT p53 and p73. We have infected p53-null H1299 human lung carcinoma cells with recombinant adenoviruses expressing WT p53, p73 or beta-galactosidase, and have undertaken microarray hybridization analyses to identify genes whose expression profile is altered by p53 or p73. Quantitative real-time PCR verified the repression of E2F-5, centromere protein A and E, minichromosome maintenance proteins (MCM)-2, -3, -5, -6 and -7 and human CDC25B after p53 expression. 5-Fluorouracil treatment of colon carcinoma HCT116 cells expressing WT p53 results in a reduction of the cyclin B2 protein level suggesting that DNA damage may indeed cause repression of these genes. Transient transcriptional assays verified that WT p53 repressed promoters of a number of these genes. Interestingly, a gain-of-function p53 mutant instead upregulated a number of these promoters in transient transfection. Using promoter deletion mutants of MCM-7 we have found that WT p53-mediated repression needs a minimal promoter that contains a single E2F site and surrounding sequences. However, a single E2F site cannot be significantly repressed by WT p53. Many of the genes identified are also repressed by p21. Thus, our work shows that WT p53 and p73 repress a number of growth-related genes and that in many instances this repression may be through the induction of p21.

Differential modulation of cellular and viral promoters by p73 and p53.

p73 has been shown to transcriptionally activate genes positively responsive to wild-type p53. In order to undertake a comparative study of functions of p53 and p73 we have cloned the cDNA of p73 from MCF-7 cells. Adenovirus onco-protein E1A inhibits the transactivation by p73; a deletion mutant of E1A incapable of interacting with p300 and CREB-binding protein (CBP) fails to disrupt the transactivation. Furthermore, CBP increases the transactivation mediated by p73 suggesting that CBP may function as a co-activator and E1A inhibits p73-mediated transactivation by sequestering p300 or CBP. We show that p73 can transcriptionally inhibit a number of cellular and viral promoters. However, wild-type p53, p73 alpha and p73 beta differ in their ability to inhibit transcriptional activity of different promoters. While wild-type p53 inhibits the promoters of the human cytomegalovirus (CMV) immediate-early gene, the long terminal repeat of human immunodeficiency virus type 1 (HIV LTR), human cyclin A (cyc A) gene, and insulin-like growth factor receptor I (IGF-I-R), p73 alpha only inhibits the HIV LTR and cyc A promoters significantly; and p73 beta inhibits the CMV, HIV LTR and cyc A promoters. A mutant of p73 alpha having amino acid substitutions at positions 268 and 300 on the presumptive DNA-binding domain fails to transactivate the p21 promoter but represses the CMV and the HIV LTR promoter quite efficiently showing that the mechanisms of transactivation and repression by p73 are different. Interestingly, p73 alpha transactivates the IGF-I-R promoter, which is inhibited by wild-type p53; p73 beta has no significant effect on this promoter. This is a unique situation where p73 alpha differs from p73 beta as well as p53.

The human oncoprotein MDM2 uses distinct strategies to inhibit transcriptional activation mediated by the wild-type p53 and its tumor-derived mutants.

Human MDM2 (hMDM2) inhibits transcriptional activation mediated by wild-type p53 and its tumor-derived mutants. We present evidence to show that hMDM2 interacts with the tumor-derived mutants of p53 and inhibits transcriptional activation of the human c-myc promoter mediated by the tumor-derived mutants of p53 through two domains. These two domains of hMDM2 are able to function independent of each other. Interaction with either of the domains is sufficient for inhibition of mutant p53-mediated transactivation. One of these domains is the same as the wild-type p53 interaction domain of hMDM2, whereas a second domain is situated within amino acid 190 and 276 residues and is specific for mutant p53. hMDM2 does not inhibit transcriptional activation mediated by the transcriptional activator VP16, suggesting that the inhibition is not mediated by inactivation of a general transcription factor. The transactivation and the oligomerization domains of mutant p53 are dispensable for its interaction with hMDM2. Thus, both hMDM2 and p53 recognize each other through unique domains. These observations suggest that forms of hMDM2 incapable of interacting with the wild-type p53, and are often expressed in transformed cells, would inhibit mutant p53-mediated transactivation and antagonize the tumorigenic function of mutant p53. This inhibitory function of hMDM2 may account for infrequent co-occurrence of p53 mutation and hMDM2 overexpression in cancer cells. Our results also suggest distinct mechanisms for wild-type and mutant p53-mediated transcriptional activation.

The human oncoprotein MDM2 arrests the cell cycle: elimination of its cell-cycle-inhibitory function induces tumorigenesis.

The human oncoprotein MDM2 (hMDM2) overexpresses in various human tumors. If amplified, the mdm2 gene can enhance the tumorigenic potential of murine cells. Here, we present evidence to show that the full-length human or mouse MDM2 expressed from their respective cDNA can inhibit the G0/G1-S phase transition of NIH 3T3 and normal human diploid cells. The protein harbors more than one cell-cycle-inhibitory domain that does not overlap with the p53-interaction domain. Deletion mutants of hMDM2 that lack the cell-cycle-inhibitory domains can be stably expressed in NIH 3T3 cells, enhancing their tumorigenic potential. The tumorigenic domain of hMDM2 overlaps with the p53-interaction domain. Some tumor-derived cells, such as Saos-2, H1299 or U-2OS, are relatively insensitive to the growth-inhibitory effects of hMDM2. These observations suggest that hMDM2 overexpression in response to oncogenic stimuli would induce growth arrest in normal cells. Elimination or inactivation of the hMDM2-induced G0/G1 arrest may contribute to one of the steps of tumorigenesis.

Transcriptional activation of the human epidermal growth factor receptor promoter by human p53.

The human epidermal growth factor receptor (EGFR) promoter is activated by both wild-type and tumor-derived mutant p53. In this communication, we demonstrate that EGFR promoter sequence requirements for transactivation by wild-type and mutant p53 are different. Transient-expression assays with EGFR promoter deletions identified a wild-type human p53 response element, 5'-AGCTAGACGTCCGGGCAGCCCCCGGCG -3', from positions --265 to --239. Electrophoretic mobility shift analysis and DNase I footprinting assays indicated that wild-type p53 binds sequence specifically to the response element. Using circularly permuted DNA fragments containing the p53-binding site, we show that wild-type p53 binding induces DNA bending at this site. We further show that the EGFR promoter is also activated by tumor-derived p53 mutants p53-143A, p53-175H, p53-248W, p53-273H, and p53-281G. However, the transactivation by mutant p53 does not require the wild-type p53-binding site. The minimal EGFR promoter from positions --104 to --20 which does not contain the wild-type p53-binding site is transactivated by the p53 mutants but not by the wild-type protein, showing a difference in the mechanism of transactivation by wild-type and mutant p53. Transactivation of the EGFR promoter by p53 may represent a novel mechanism of cell growth regulation.

The MDR1 downstream promoter contains sequence-specific binding sites for wild-type p53.

We have examined the interaction of the wild-type p53 protein with the downstream promoter of the human multidrug resistance gene-1 (MDR1). Our findings indicate that wild-type p53 inhibits reporter activity driven by the MDR1 downstream promoter (base pairs -189 to +133 relative to the major transcriptional initiation site) in a dose-dependent manner in cotransfection assays in the BHK and the Saos-2 cell lines. A 123 base-pair segment of DNA (-119 to +4 relative to the major transcriptional initiation site), a 193 base-pair segment (-189 to +4), and a 135 base-pair segment (-2 to +133) have been isolated from the MDR1 downstream promoter which, like the full promoter, are negatively controlled by wild-type p53. In addition, we show sequence-specific binding of wild-type p53 protein to the MDR1 downstream promoter. These in vitro results suggest that the presence of wild-type p53 negatively affects expression of the MDR1 gene product, p-glycoprotein, at the transcriptional level.

Wild-type human p53 transactivates the human proliferating cell nuclear antigen promoter.

The wild-type p53 protein is a transcriptional activator implicated in the control of cellular growth-related gene expression. Here, using a number of different cell lines and transient-transfection-transcription assays, we demonstrate that at low levels, wild-type p53 transactivates the human proliferating cell nuclear antigen (PCNA) promoter. When expressed at a similar level, the tumor-derived p53 mutants did not transactivate the PCNA promoter. We identified a p53-binding site on the human PCNA promoter with which p53 interacts sequence specifically. When placed on a heterologous synthetic promoter, the binding site functions as a wild-type p53 response element in either orientation. Deletion of the p53-binding site renders the PCNA promoter p53 nonresponsive, showing that wild-type p53 transactivates the PCNA promoter by binding to the site. At a higher concentration, wild-type p53 inhibits the PCNA promoter but p53 mutants activate. Transactivation by p53 mutants does not require the p53-binding site. These observations suggest that moderate elevation of the cellular wild-type p53 level induces PCNA production to help in DNA repair.

N-terminal 130 amino acids of MDM2 are sufficient to inhibit p53-mediated transcriptional activation.

The human oncoprotein MDM2 binds with the tumor suppressor p53 and inhibits p53-directed transactivation. In this report we show that deletion of 336 amino acids from the C-terminus of human MDM2 does not decrease its efficiency to bind p53 in vivo and inhibit p53-directed transactivation. Even further deletion of MDM2 from the C-terminus up to amino acid 131 does not reduce its ability to inhibit p53-mediated transactivation. Since deletion up to amino acid 131 also deletes many antigenic sites of MDM2 and the truncated protein cannot be immunoprecipitated by the antibodies available to us, two internal deletions were made to define the p53-interaction domain. Internal deletion of four amino acids beginning at 110 residue (amino acids 110 to 113) did not reduce p53-binding or inhibition of p53-directed transactivation whereas internal deletion of amino acids 60 to 65 reduces but does not abolish these activities. Sequential deletion of amino acids from the N-terminus leads to sequential destruction of p53-binding and inhibition of transactivation capability of MDM2. Fourteen amino acids can be deleted from this end without any reduction of these activities. Deletion of 28 N-terminal amino acids residues drastically reduces, but does not abolish the p53-binding ability of the protein, as well as inhibition of p53-directed transactivation. Deletion of 58 amino acids from the N-terminus of the oncoprotein abolishes its ability to bind p53 in vivo and to inhibit p53-directed transactivation. These results locate the p53-binding domain of MDM2 within amino acids 14 to 154 and inhibition of transactivation domains of MDM2 within amino acid residues 14 to 130 suggesting possible p53-independent biological functions of the 491 amino acid long oncoprotein.

Cell-type-specific induction of the UL9 gene of HSV-1 by cell signaling pathway.

The origin-binding protein, encoded by the UL9 gene of herpes simplex virus-1 (HSV-1), has the properties of an initiator of DNA replication. In this communication, we report that the UL9 promoter contains a cAMP-response element (CRE). Transient expression analyses show that dibutyryl cyclic AMP, known to elevate intracellular cAMP level, can induce the UL9 promoter in a rat pheochromocytoma cell line (PC12) but not in a non-neuronal human cell line (HeLa). Interestingly, a transcription factor that increases expression of a neuropeptide gene by interacting with CRE can also activate the UL9 promoter independent of cell type. Thus, our data suggest that extracellular stimuli, capable of interacting with the signaling pathway in neuronal cells, can activate UL9 gene expression, and different proteins may regulate UL9 expression in different cell types.

Wild-type human p53 activates the human epidermal growth factor receptor promoter.

We show that wild-type human p53 transactivates the human epidermal growth factor receptor (EGFR) promoter in vivo in a dose-dependent manner, implicating p53 in promotion of cell proliferation. This activation is sensitive to the expression of cellular oncoprotein MDM2 and human papillomavirus type 18 (HPV-18) E6 protein. The p53 response element is localized within -15 and -569 of the promoter. The EGFR promoter does not have a TATA box, and has low activity in Saos-2 cells in the absence of p53. Results from our in vivo transient transfection assays suggest that p53-binding sites, without any other known promoter element, can act as bidirectional promoters in the presence of wild-type p53. Gel retardation analyses suggest that p53 may serve to nucleate TBP on a promoter. We propose that p53 successfully nucleates the transcription complex, possibly via direct interaction with TFIID, and activates the EGFR promoter.

The tumor suppressor p53 and the oncoprotein simian virus 40 T antigen bind to overlapping domains on the MDM2 protein.

The oncogene mdm2 has been found to be amplified in human sarcomas, and the gene product binds to the tumor suppressor p53. In this report, we describe the dissection of the MDM2-binding domain on p53 as well as the p53-binding domain on MDM2. We also demonstrate that the oncoprotein simian virus 40 T antigen binds to the product of cellular oncogene mdm2. We have constructed several N- and C-terminal deletion mutants of p53 and MDM2, expressed them in vitro, and assayed their in vitro association capability. The N-terminal boundary of the p53-binding domain on MDM2 is between amino acids 1 and 58, while the C-terminal boundary is between amino acids 221 and 155. T antigen binds to an overlapping domain on the MDM2 protein. On the other hand, the MDM2-binding domain of p53 is defined by amino acids 1 and 159 at the N terminus. At the C terminus, binding is progressively reduced as amino acids 327 to 145 are deleted. We determined the effect of human MDM2 on the transactivation ability of wild-type human p53 in the Saos-2 osteosarcoma cell line, which does not have any endogenous p53. Human MDM2 inhibited the ability of human p53 to transactivate the promoter with p53-binding sites. Thus, human MDM2 protein, like the murine protein, can inactivate the transactivation ability of human p53. Interestingly, both the transactivation domain and the MDM2-binding domain of p53 are situated near the N terminus. We further show that deletion of the N-terminal 58 amino acids of MDM2, which eliminates p53 binding, also abolishes the capability of inactivating p53-mediated transactivation. This finding suggests a correlation of in vitro p53-MDM2 binding with MDM2's ability in vivo to interfere with p53-mediated transactivation.

p53 and SV40 T antigen bind to the same region overlapping the conserved domain of the TATA-binding protein.

In this report we demonstrate that the cloned human TATA-binding protein (TBP) interacts with T antigen. TBP co-immunoprecipitates with T antigen when incubated with the T antigen-specific monoclonal antibody PAb419, and Protein-A agarose. Gel retention analysis with a radiolabeled TATA box-containing probe showed that the complex of TBP and T antigen can bind to the TATA box. Recently, p53 has also been shown to interact with TBP. Using TBP deletion mutants and co-immunoprecipitation experiments with p53 or T antigen, we show that both p53 and T antigen bind to the same region, amino acids 203-275, within the conserved C-terminal domain of TBP. Binding of p53 and T antigen to the same domain on TBP may lead to competition between the two proteins for transcriptional function.

Analysis of the promoter sequences of the UL9 gene of herpes simplex virus type 1.

The origin-binding protein, encoded by the UL9 gene of herpes simplex virus-1 (HSV-1), has the properties of an initiator of DNA replication. Since onset of lytic replication of the virus can be critically controlled by availability of the initiator protein, we have analyzed the sequences upstream of the UL9 gene in a monkey kidney cell line (Vero). In this communication, we report the presence of three UL9 transcripts and multiple transcription initiation sites of the UL9 gene in HSV-1-infected cells. The presence of multiple transcription initiation sites for the UL9 gene is consistent with the absence of a TATA element 5'- to the UL9 ORF. A 395-bp upstream sequence encompassing all UL9 transcription initiation sites was active in expressing a reporter gene and was induced by viral immediate-early proteins. This result is consistent with the earlier observation that immediate-early proteins of HSV-1 induce viral early genes.

Analysis of the herpes simplex virus type 1 OriS sequence: mapping of functional domains.

The herpes simplex virus type 1 (HSV-1) OriS region resides within a 90-bp sequence that contains two binding sites for the origin-binding protein (OBP), designated sites I and II. A third presumptive OBP-binding site (III) within OriS has strong sequence similarity to sites I and II, but no sequence-specific OBP binding has yet been demonstrated at this site. We have generated mutations in sites I, II, and III and determined their replication efficiencies in a transient in vivo assay in the presence of a helper virus. Mutations in any one of the sites reduced DNA replication significantly. To study the role of OriS sequence elements in site I and the presumptive site III in DNA replication, we have also generated a series of mutations that span from site I across the presumptive binding site III. These mutants were tested for their ability to replicate and for the ability to bind OBP by using gel shift analyses. The results indicate that mutations across site I drastically reduce DNA replication. Triple-base-pair substitution mutations that fall within the crucial OBP-binding domain, 5'-YGYTCGCACT-3' (where Y represents C or T), show a reduced level of OBP binding and DNA replication. Substitution mutations in site I that are outside this crucial binding sequence show a more detrimental effect on DNA replication than on OBP binding. This suggests that these sequences are required for initiation of DNA replication but are not critical for OBP binding. Mutations across the presumptive OBP-binding site III also resulted in a loss in efficiency of DNA replication. These mutations influenced OBP binding to OriS in gel shift assays, even though the mutated sequences are not contained within known OBP-binding sites. Replacement of the wild-type site III with a perfect OBP-binding site I results in a drastic reduction of DNA replication. Thus, our DNA replication assays and in vitro DNA-binding studies suggest that the binding of the origin sequence by OBP is not the only determining factor for initiation of DNA replication in vivo.

A 269-amino-acid segment with a pseudo-leucine zipper and a helix-turn-helix motif codes for the sequence-specific DNA-binding domain of herpes simplex virus type 1 origin-binding protein.

The UL9 gene of herpes simplex virus (HSV) codes for a DNA-binding protein (OBP) that interacts sequence specifically with the origin of replication. This protein is essential for HSV DNA replication in cultured cells. The UL9 gene was cloned into a plasmid vector downstream of the SP6 RNA polymerase promoter. By using in vitro transcription and translation systems, a full-length OBP was synthesized. This synthetic protein is recognized by an antiserum generated against the C-terminal decapeptide of OBP and is functionally active in binding to OriS sequence specifically. The in vitro-synthesized protein has sequence specificity for binding similar to that found for the in vivo-generated OBP. A total of 14 in-frame deletion and insertion mutants of the UL9 gene were generated and expressed in vitro. Using these deletion mutants, we determined that the 269-amino-acid stretch defined by amino acids 564 to 832 localizes the OriS-specific DNA-binding domain. The N-terminal boundary is between amino acids 565 and 596, while the C terminus lies between amino acids 833 and 805. This segment contains a helix-turn-helix moiety and a pseudo-leucine zipper, neither of which alone can support DNA binding. The other leucine zipper from amino acids 150 to 173 is not required for the in vitro sequence-specific DNA-binding activity of OBP.

Preferential binding of simian virus 40 T-antigen dimers to origin region I.

The sequence components that direct high-affinity binding of simian virus 40 (SV40) T antigen to SV40 origin region I are composed of two recognition pentanucleotides separated by a spacer. This region has binding sites for two T-antigen monomeric units. We extended the tripartite region I sequence by one and two sets of spacers and pentanucleotides and also shortened the region by one pentanucleotide. Our T-antigen-binding studies with these constructs show that the protein has a strong preference for binding to an even rather than an odd number of pentanucleotides separated by spacer sequences. Gel retardation assays reveal that the size of the complex formed between the 17-base-pair region I sequence and T antigen did not increase when the sequence was extended with one spacer-pentanucleotide sequence but did increase with two such units. DNase I footprinting and fragment assay experiments indicate that the protein did not protect a pentanucleotide that was not paired with another pentanucleotide. The unpaired pentanucleotide resumed its binding activity when it was paired with a spacer and another pentanucleotide sequence. We propose that T antigen binds to region I as a preformed dimer.

Analysis of Ori-S sequence of HSV-1: identification of one functional DNA binding domain.

Using gel retardation assays, we have detected an Ori-S binding activity in the nuclear extract of HSV-1 infected Vero cells. The sequence-specific DNA binding activity seems to be identical to that described by Elias et al. (Proc. Natl. Acad. Sci. USA 83: 6322-6326, 1986). This activity fails to retard a mutant origin DNA that has a 5 bp deletion in the reported protein binding site along with an A to T substitution at a position 16 base-pairs away from the site. This mutant also failed to replicate in a transient replication assay, thus correlating binding of the factor on the origin to replication efficiency. Using crude nuclear extracts as the source of the factor and with the help of footprint and gel retardation analyses, we confirmed that protection is only observed on the preferred site of binding on and near the left arm of the Ori-S palindrome. In order to analyze the sequence specificity of the binding we have generated a set of binding site mutants. Competition experiments with these mutant origins indicate that the sequence 5'-TTCGCACTT-3' is crucial for binding.

The p53 complex from monkey cells modulates the biochemical activities of simian virus 40 large T antigen.

We have compared the ATPase, DNA-binding, and helicase activities of free simian virus 40 (SV40) large T antigen (To) and T antigen complexed with cellular p53 (T+p53). Each activity is essential for productive viral infection. The T+p53 and To fractions were prepared by sequential immunosorption of infected monkey cells with monoclonal antibodies specific for p53 and T antigen. The immune-complexed T fractions were then assayed in parallel. For ATP hydrolysis, the Vmax for T+p53 was 143 nmol of ADP per min per mg of protein, or 18-fold greater than for To. ATP had no effect on the stability of the T+p53 complex. The T+p53 complex was significantly more active than To in hydrolyzing dATP, dGTP, GTP, and UTP. Of the nucleotide substrates tested, the greatest relative increase (T+p53/To) was in hydrolyzing dGTP and GTP. In DNase footprinting assays performed under replication conditions, the T+p53 complex protected regions I, II, and III of origin DNA while equivalent amounts of To protected only regions I and II. Region III is known to contribute to the efficiency of DNA replication and contains the SP1-binding sites of the early viral promoter. The T+p53 fraction was also a more efficient helicase than To, especially with a GC-rich primer and template. Thus, the T+p53 complex has enhanced ATPase, GTPase, DNA-binding, and helicase activities. These findings imply that complex formation between cellular monkey p53 and SV40 T antigen modulates a number of essential activities of T in SV40 productive infection.

Only one of the origin binding forms of SV40 T antigen has helicase activity.

SV40 T antigen exists in monomeric and multimeric forms. We have separated the individual components by glycerol gradient centrifugation. Helicase activity is found to be associated with monomeric forms only. Dimers and other multimeric forms have no discernable helicase activity. However, results obtained from DNA binding experiments carried out with separated forms of T antigen indicate that both monomers and dimers bind to region I and region II of SV40 origin of replication. Possibly monomeric T antigen unwinds DNA at the replication fork while both monomeric and dimeric forms are utilized for positioning of T antigen at the origin of replication.

ATP enhances the binding of simian virus 40 large T antigen to the origin of replication.

Simian virus 40 large T antigen initiates DNA replication by binding to the origin of replication. We examined the binding of T antigen to origin regions I, II, and III under conditions designed for efficient in vitro replication functions. We found that 4 mM ATP enhanced the binding of T antigen to regions I and II of the origin DNA by 4- to 20-fold. DNase-footprinting and fragment assays showed that ATP extended the DNase protection domain of T antigen bound to region II by 5 to 10 base pairs at both ends of the core origin of replication. This alteration suggests a change in the conformation of T antigen, bound DNA, or both.