TopBP1 Mediates Mutant p53 Gain of Function through NF-Y and p63/p73

Abstract

Nearly half of human cancers harbor p53 mutations, which can promote cancerous growth, metastasis, and resistance to therapy. The gain of function of mutant p53 is partly mediated by its ability to form a complex with NF-Y or p63/p73. Here, we demonstrate that TopBP1 mediates these activities in cancer, and we provide both in vitro and in vivo evidence to support its role. We show that TopBP1 interacts with p53 hot spot mutants and NF-YA and promotes mutant p53 and p300 recruitment to NF-Y target gene promoters. TopBP1 also facilitates mutant p53 interaction with and inhibition of the transcriptional activities of p63/p73. Depletion of TopBP1 in mutant p53 cancer cells leads to downregulation of NF-Y target genes cyclin A and Cdk1 and upregulation of p63/p73 target genes such as Bax and Noxa. Mutant p53-mediated resistance to chemotherapeutic agents depends on TopBP1. The growth-promoting activity of mutant p53 in a xenograft model also requires TopBP1. Thus, TopBP1 mediates mutant p53 gain of function in cancer. Since TopBP1 is often overexpressed in cancer cells and is recruited to cooperate with mutant p53 for tumor progression, TopBP1/mutant p53 interaction may be a new therapeutic target in cancer.

INTRODUCTION

The tumor suppressor protein p53 generally functions through a specific DNA binding activity. Mutations of p53 are found in almost half of human cancers. Most of these mutations occur within the DNA-binding domain of p53, destroying its specific DNA binding activity. It is also well recognized that mutant p53 (mutp53) acquires new functions (gain of function) in promoting cancer cell proliferation, metastasis, genomic instability, and resistance to chemotherapy (33). The combined effects of both loss of tumor suppression and newly gained oncogenic properties may explain the high prevalence of mutp53 in human cancers.

There are several potential mechanisms for mutp53 gain of function in transcriptional regulation. mutp53 can interact with NF-Y, a heterotrimeric transcription factor that recognizes the CCAAT consensus motif and regulates many cell cycle-related genes such as cyclin A, cyclin B, Cdk1, Cdc25C, etc. (7). Through the interaction, mutp53 and p300 are recruited to NF-Y target gene promoters and are responsible for aberrant expression of the above-mentioned NF-Y target genes and consequently abnormal proliferation. mutp53 can form a complex with p63/p73 and block the DNA binding activities of p63 and p73 and therefore inactivate their proapoptotic functions (9, 30, 39). mutp53 was also reported to bind non-B DNA in a DNA structure-selective manner rather than a sequence-specific manner. This binding was proposed to be the basis for its interaction with the matrix attachment region resulting in inhibition of the transcription factor recruitment and transcriptional repression (12). The full scope of mutp53 in carcinogenesis remains to be explored. Understanding its mechanistic aspect would be imperative for us to devise badly needed therapeutic strategies targeting the mutp53 gain of function in cancer.

TopBP1 (topoisomerase IIβ binding protein) contains nine BRCA1 carboxyl-terminal (BRCT) motifs (35). TopBP1 appears to serve as a scaffold to modulate many processes of DNA metabolism, such as DNA damage checkpoint, replication, and transcription (10). The activation of checkpoint kinase 1 (Chk1) requires chromatin loading of ATR (ATM [ataxia-telangiectasia mutated]–Rad3-related kinase)/ATRIP (ATR-interacting protein) and Rad9-Hus1-Rad1 (9-1-1) clamp. The 9-1-1 complex binds and tethers TopBP1 to ATR/ATRIP (5). TopBP1 contains a conserved ATR-activating domain and activates ATR (23). Initially it was proposed that the 9-1-1 complex recruits TopBP1 to stalled replication forks (5). Yan and Michael later used Xenopus egg extracts and showed that TopBP1 binds to the stalled fork first. It then recruits the 9-1-1 complex. Their data suggest a role of replication stress sensor for TopBP1 (46, 47). Recruitment of TopBP1 to double-strand breaks or stalled replication forks was recently shown to be dependent on its interaction with 53BP1 (4) or MDC1 (43). The sensing step is followed by an interaction with a DNA helicase, BACH1, which might facilitate the unwinding of double-stranded DNA for an additional replication protein A (RPA) coating, and subsequent loading of ATR/ATRIP and the 9-1-1 complex (14). TopBP1 is also directly involved in DNA replication initiation. The loading of Cdc45 and DNA polymerases α and ε to replication origins requires TopBP1 (16, 42). This function is mediated by its association with Treslin/TICRR (TopBP1-interacting, checkpoint, and replication regulator) in a Cdk2-dependent manner (24, 36).

Besides a direct involvement in DNA replication, a role that is shared among all eukaryotes, TopBP1 also regulates transcription in metazoa. Through this regulation, TopBP1 controls cell cycle progression in an additional layer. TopBP1 is required to restrict the transcriptional activities of E2F1 and p53 during G₁/S transition (26–29). The repression of E2F1 proapoptotic activity by TopBP1 involves recruitment of Brg1/Brm chromatin-remodeling complex (28) and requires activation of phosphatidylinositol 3-kinase (PI3K)/Akt. Akt phosphorylates TopBP1 at Ser¹¹⁵⁹ and induces its oligomerization, which then induces TopBP1 to bind and repress E2F1, Miz1 (29), and an ePHD (extended plant homeodomain) protein SPBP (stromelysin 1 platelet-derived growth factor responsive element-binding protein) (37). Regulation of p53 by TopBP1 appears to be more direct. The carboxyl-terminal BRCT repeats of TopBP1 bind to p53 DNA-binding domain and inhibit p53 sequence-specific DNA binding activity (26). TopBP1 also binds Miz1 and represses its transactivation activity on the p21^Cip1 promoter (17). With the multilayered involvement of TopBP1 in cell cycle progression, it is not surprising that a knockout of TopBP1 in mice leads to early embryonic lethality due to a proliferation defect and increased apoptosis (21).

With a report that a missense variant (R309C) occurred at an elevated frequency in familial breast cancer cases compared to healthy controls from Finland (22), TopBP1 has been suggested as a breast cancer susceptibility gene. However, a recent German study shows that this R309C allele is a common polymorphic variant at least in Germany and that the R309C allele does not confer an increased risk for breast cancer (2). On the other hand, the expression of TopBP1 has been found to be upregulated in many breast cancer tissues, and its overexpression is associated with higher tumor grade and shorter patient survival time (26). Moreover, overexpression of TopBP1 in cultured cells to a level seen in human breast tumors inhibits many functions of p53, suggesting a potential oncogenic role for TopBP1 by inactivating p53 in cancer (26). In this report, we further show that TopBP1 binds to many common forms of mutant p53 in cancer cell lines and facilitates the interaction between mutant p53 and NF-Y or p63/p73. We also demonstrate an in vivo role for TopBP1 in mediating the mutant p53 gain of function regarding aberrant proliferation and increased chemotherapy resistance. Our results suggest that TopBP1/mutant p53 interaction may serve as a novel therapeutic target in cancer therapy.

MATERIALS AND METHODS

Cell culture and transfection.

C33A, H1299, and SKOV-3 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS), penicillin (50 IU/ml), and streptomycin (50 μg/ml). SKBr3 cells were grown in McCoy's medium supplemented with 10% FBS. OVCAR-3 cells were grown in RPMI 1640 medium supplemented with 10% FBS. All cells were grown in a humidified incubator at 37°C with 5% CO₂ and 95% air. C33A cells and H1299 cells were transfected with a standard calcium phosphate method or Gene Pulser Xcell electroporation system (Bio-Rad) according to the manufacturer's instruction. OVCAR-3 and SKBr3 cells were transfected using Lipofectamine 2000 (Invitrogen). After transfection, cells were incubated for 48 h before analysis. SKOV-3 cells were transfected with the Gene Pulser Xcell electroporation system (Bio-Rad).

Plasmid construction.

The TopBP1 and p53-related constructs have been described previously (26). pCMV-Neo-Bam-p53-WT (CMV stands for cytomegalovirus, Neo stands for neomycin, Bam stands for BamHI, and WT stands for wild type), -p53-V143A, -p53-R175H, -p53-R248W, -p53-R249S, and -p53-R273H mutants were obtained from Bert Vogelstein through Addgene (1). GFP-p53-R175H (GFP stands for green fluorescent protein) and GFP-p53-R273H were obtained by cloning the BamHI-digested fragments of pCMV-Neo-Bam-p53 R175H and pCMV-Neo-Bam-p53 R273H into pGFP-C1. FLAG-NF-YA, -NF-YB, and -NF-YC were provided by Keiko Funa (15). To construct GST-NF-YA (GST stands for glutathione S-transferase), -YB, and -YC, NF-YA, -YB, and -YC were amplified by PCR with the following primers. For NF-YA, the forward primer was 5′-CGCGGATCCATGGAGCAGTATACAGC-3′, and the reverse primer was 5′-GCGCTCGAGTTAGGACACTCGGATG-3′. For NF-YB, the forward primer was 5′-CGCGGATCCATGACAATGGATGGTGAC-3′, and the reverse primer was 5′-GCGCTCGAGTCATGAAAACTGAATTTG-3′. For NF-YC, the forward primer was 5′-CGCGGATCCATGTCCACAGAAGGAG-3′, and the reverse primer was 5′-GCGCTCGAGTCACTCGCCGGTCAC-3′. The PCR products were digested with BamHI/XhoI, cloned into pGEX6P1, and verified by sequencing. The shp53pLKO.1 puro (KO stands for knockout, and puro stands for puromycin) (from Bob Weinberg) (11) and scramble short hairpin RNA (shRNA) pLKO.1 (from David Sabatini) were obtained through Addgene. pLKO.1-shTopBP1 plasmids were obtained from the RNAi Consortium (Open Biosystems, RHS4533-NM_007027) and screened for knockdown of TopBP1 by transient transfection of HEK293T cells, followed by Western blotting. Two different pLKO.1-shTopBP1 constructs (siTopBP1-1 and siTopBP1-2) yielded the best knockdown and were chosen for subsequent experiments.

Establishment of stable cell lines.

To prepare the stable cell line expressing human papillomavirus (HPV) E6, C33A cells were transfected with a control pcDNA3 vector or pcDNA3-E6 (gift of Xinbin Chen) and then subjected to selection with neomycin (G418; 600 μg/ml) 48 h after transfection. To obtain the stable cell lines expressing shp53 or shTopBP1, C33A cells were infected with the lentivirus expressing scramble, p53, or two different TopBP1 shRNAs, and H1299 cells were infected with the lentivirus expressing scramble or two different TopBP1 shRNAs. The cells were grown for 48 h and then selected with puromycin (2 μg/ml). To establish the mutp53-expressing stable H1299 cell lines, the H1299 cells stably expressing scramble shRNA or TopBP1 shRNA (siTopBP1-1) were transfected with an empty vector pCMV or various mutp53 constructs, followed by a second selection with neomycin (G418; 600 μg/ml). Overexpression or knockdown was confirmed by Western blotting.

Immunoprecipitation, Western blot analysis, and immunofluorescence studies.

Cells were harvested 48 h after transfection in TNN buffer (27), and immunoprecipitation was carried out as described previously (27). The specific signals were detected with appropriate antibodies. The antibodies specific to p53 (FL-393), Bax (N-20), p300 (C-20), Cdc2 p34 (17), GST (B-14), Myc (A-14), p21 (C-19), and p63 (4A4) were purchased from Santa Cruz. The NF-YA and NF-YB antibodies were purchased from Rockland. The cyclin A antibody was purchased from Upstate Biotechnology (06-138). The β-actin and FLAG antibody were purchased from Sigma. p53 (Ab-1) was purchased from Calbiochem. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibody was purchased from Alexis. p73 (5B429) was purchased from Imgenex. GFP antibody was purchased from Clontech. Poly(ADP-ribose) polymerase (PARP) antibody and monoclonal and polyclonal TopBP1 antibodies were purchased from BD Transduction Laboratories and Bethyl Laboratories, respectively. A goat polyclonal TopBP1 antibody (L20) for immunoprecipitation was purchased from Santa Cruz. For immunofluorescence studies, H1299 cells were plated on collagen-coated coverslips in 6-well plates, transfected with HcRed1-TopBP1 and EGFP-p53-R175H (EGFP stands for enhanced GFP) or -R273H, and then incubated for 48 h before analysis. The cells were fixed in 3% paraformaldehyde for 20 min, and the nuclei were stained with Hoechst 33258 for visualization. Images were captured on a Zeiss fluorescence microscope (Axio Observer inverted microscope).

GST pulldown assay.

The GST fusion proteins were induced by 0.1 mM IPTG (isopropyl-β-d-thiogalactopyranoside) in Escherichia coli strain BL21 and purified. The GST portion of the GST-TopBP1 and GST-NFYA fusion proteins was excised by PreScission protease (Pharmacia). One microgram of purified GST or GST-NFYA, -NFYB, -NFYC, or -TopBP1 was incubated in NETN-A buffer (50 mM NaCl, 1 mM EDTA, 20 mM Tris, 0.5% NP-40) with 2 μg purified TopBP1 or NFYA, and rotated at 4°C for 3 h. GST-NFYA, -NFYB, -NFYC, or -TopBP1 was pulled down with glutathione-Sepharose, the beads were washed six times with NETN-B buffer (100 mM NaCl, 1 mM EDTA, 0.2 mM phenylmethylsulfonyl fluoride [PMSF]), and the proteins were then subjected to SDS-PAGE and analyzed by Western blotting with anti-TopBP1 or anti-NFYA antibody.

Apoptosis assay.

We used flow cytometry and caspase activity to assay cell apoptosis. For flow cytometry, the treated C33A cells were collected and stained with annexin V-phycoerythrin (PE) (Pharmingen) and 7-amino-actinomycin (7-AAD) (BD Bioscience). The annexin V-PE/7-AAD profile was analyzed by flow cytometry. At least 10,000 cells were counted, and annexin V-positive (annexin V⁺) 7-AAD-negative (7-AAD⁻) and annexin V⁺ 7-AAD⁺ cells were scored as apoptotic. To measure caspase activity, the treated C33A cells were harvested and then assayed for caspase 3/7 cleavage according to the manufacturer's instructions (Caspase-Glo 3/7; Promega). An aliquot of the harvested cells was lysed in SDS lysis buffer for Western blotting in both assays.

Luciferase assay.

The expression constructs (5 μg for pcDNA4-Myc-p63α and 5 μg for HA-p73α [HA stands for hemagglutinin]), a reporter promoter plasmid (1 μg pGL3-Bax promoter-Luc [Luc stands for luciferase]; gift of Moshe Oren [8]), and 1 μg of β-galactosidase plasmid were transfected into H1299 stable cell lines expressing scramble small interfering RNA (siRNA) or TopBP1 siRNA together with mutp53-R175H (mutp53 stands for mutant p53) or p53-R273H. Cells were harvested 2 days later. An aliquot of cells was lysed in SDS lysis buffer for Western blot analysis; the other cells were lysed in reporter lysis buffer (Promega) for luciferase activity, and β-galactosidase activity was measured (27). The luciferase activity was normalized against the β-galactosidase activity. All assays were carried out in triplicate.

Colony formation assay.

Ten micrograms of the indicated constructs (pCMV empty vector, pCMV-p53-WT, or different mutp53) were transfected into H1299 stable cell lines expressing either scramble siRNA or TopBP1 siRNA in three replicates. Two days later, the cells were selected by neomycin (G418; 600 μg/ml). The medium was changed every 3 days and supplemented with neomycin. At day 21 after selection, the cells were fixed with methanol and stained with crystal violet followed by washing with water. The colonies were counted.

Cell proliferation assay.

H1299 stable cell lines expressing the indicated constructs (5 × 10⁵ cells) were each plated in three replicates into 10-cm-diameter plates. An aliquot from each plate was counted with a Beckman Coulter Counter. One-quarter of the remaining cells were replated. The assay was repeated on days 2, 4, and 8, respectively. A portion of each cell line was harvested at day 8 in SDS lysis buffer for Western blotting.

Xenograft experiment.

Male nu/nu mice (4 weeks old; strain code 088) were purchased from Charles River Laboratories (Cambridge, MA). The animals were cared for and maintained in accordance with institutional guidelines. After 1 week of acclimation, animals were randomly distributed into the indicated 6 groups (5 mice per group). H1299 stable cells expressing siScramble or siTopBP1 together with pCMV, pCMV-p53-R175H, or pCMV-p53-R273H (1.5 × 10⁶ cells per site in 100 μl phosphate-buffered saline [PBS]) were injected subcutaneously into the right side of the flank. The mice were monitored twice a week. Tumor size was measured on the indicated day with a caliper and calculated based on the formula π/6 (length × depth × width). The animals were euthanized on day 34 after injection, and the tumors were harvested, weighed, and further processed for histological and biochemical analyses.

Immunohistochemical staining.

Sections of paraffin-embedded tissue (5 μm) were mounted on Bond-Rite slides from Richard-Allan Scientific. The slides were coated with chrome-alum gelatin and heated at 60° for 2 h. Paraffin was removed from the sections by three changes of xylenes and rehydrated through graded alcohols from absolute to 70% for 5 min each. For immunohistochemistry, the slides were rinsed with deionized water and high-temperature antigen retrieval was performed in pH 9 buffer (10 mM Tris, 1 mM EDTA) in a pressure cooker for 10 min. They were then allowed to cool for 20 min and transferred to Tris buffer (0.05 M Trizma base, 0.15 M NaCl, 0.1% Triton X-100, pH adjusted to 7.6 with HCl). Endogenous peroxidases were quenched with an aqueous solution of H₂O₂ for 5 min. Biotin was blocked by treating the tissue samples with streptavidin (Jackson Immuno Research, West Grove, PA) (10 μg/ml) in PBS for 15 min, rinsing with Tris buffer, and applying biotin (Sigma-Aldrich, St. Louis, MO) (200 μg/ml) in PBE buffer (PBS with 1% bovine serum albumin [BSA], 1 mM EDTA, and 0.15 mM NaN₃). Nonspecific binding was blocked with 3% goat serum in PBE buffer on the sections for 20 min. The primary antibodies were diluted in PBE buffer and applied to the sections for 1 h at room temperature. TopBP1 antibody (clone 33; BD Transduction Laboratories, San Jose, CA) was used at a dilution of 1:250 with pH 9 antigen retrieval. A pantropic p53 antibody (clone PAb421; Calbiochem) was used at a dilution of 1:40 with no antigen retrieval. Negative controls were incubated with 3% goat serum. The slides were rinsed with Tris buffer after each step. Immunodetection was performed with a biotinylated anti-mouse secondary antibody and streptavidin horseradish peroxidase (HRP) (Signet Pathology Systems, Dedham, MA) for 20 min each. The chromogen used was 3-3′diaminobenzidine (Scy Tek, Logan, UT). After 7 min, the slides were rinsed with water and lightly counterstained with Mayer's hematoxylin. The sections were dehydrated through graded alcohols (70% to absolute) and three xylene baths for 5 min each. The coverslips were mounted with Permount.

ChIP assay.

Chromatin immunoprecipitation (ChIP) assay was performed following the protocol described previously (40). C33A cells grown in 15-cm-diameter dishes were infected with adenovirus expressing TopBP1 or lentivirus expressing TopBP1 siRNA. Two days later, the cells were either left untreated or treated with adriamycin (5 μM) for 5 h and then cross-linked with formaldehyde. The cells were collected, and chromatin was extracted and sonicated. Portions (one tenth) of supernatants were saved for control input PCR. The other chromatin was precleared with protein G plus/protein A-agarose beads and then immunoprecipitated with 4 μg of each antibody (p53, p63, p73, p300, NF-YA, and NF-YB). The antibodies for p53 (FL-393), p63 (H-129), p73 (H-79), and p300 (C-20) were from Santa Cruz. The antibodies for NF-YA and -YB were from Rockland. The antibody-bound complexes were recovered on protein A/G beads. Immunoprecipitates were washed under stringent conditions, and cross-links of chromatin were reversed by incubating samples at 65°C overnight. The resulting DNA was purified and analyzed by PCR. The PCR conditions were as follows: (i) 4 min at 94°C; (ii) 30 to 40 cycles, with 1 cycle consisting of 30 s at 94°C, 40 s at 58°C, and 40 s at 72°C; and (iii) 7 min at 72°C. Quantitative PCR was performed in triplicate on an MX3005P thermal cycler (Stratagene) using SYBR green dye method with ROX dye added as a reference. The PCR primers for p21 and GAPDH promoters have been described previously (26, 34). The other PCR primers were as follows. For Cdc25C, the forward primer was 5′-GAATGGACATCACTAGTAAGGCGCG-3′, and the reverse primer was 5′-GCAGGCGTTGACCATTCAAACCTTC-3′. For Cdk1, the forward primer was 5′-GAACTGTGCCAATGCTGGGA-3′, and the reverse primer was 5′-GCAGTTTCAAACTCACCGCG-3′. For cyclin A2, the forward primer was 5′-GAGTCAGCCTTCGGACAGCC-3′, and the reverse primer was 5′-CCAGAGATGCAGCGAGCAGC-3′. For Bax, the forward primer was 5′-TCAGCACAGATTAGTTTCT-3′, and the reverse primer was 5′-GGGATTACAGGCATGAGCTA-3′. For Puma, the forward primer was 5′-GATTACAGGCATGCGCCACA-3′, and the reverse primer was 5′-ACCCACACTGATGATCACAC-3′.

For re-ChIP assay, cells and chromatin were treated as described above; chromatin was immunoprecipitated using 4 μg of TopBP1 mouse antibody or control healthy mouse IgG; antibody-chromatin complexes were eluted in 10 mM dithiothreitol and incubated at 37°C for 30 min. The supernatants were then diluted 20:1 in re-ChIP buffer (1% Triton X-100, 2 mM EDTA, 150 mM NaCl, 20 mM Tris) and nutated at 4°C overnight with 4 μg of antibodies (p53, NFYA, or control normal rabbit IgG). Blocked protein G/A-agarose beads were added for 2 h, and then the beads were washed and eluted, DNA was purified, and PCR was performed as described above.

RNA extraction and real-time reverse transcription-PCR (RT-PCR).

C33A cells were infected with lentivirus expressing siScramble or siTopBP1 or infected with adenovirus expressing TopBP1 or an empty vector adenovirus for 48 h. RNA was then extracted using TRIzol reagent (Invitrogen). Quantitative PCR was performed in triplicate on an MX3005P thermal cycler using SYBR green dye method to track the progress of the reactions with ROX dye added as a reference. GAPDH was run in parallel with test genes. The PCR conditions were a denaturation step (30 s at 95°C), annealing (1 min at 55°C), and 2 min at 72°C. Results were analyzed with MxPro 4.0 quantitative PCR (QPCR) software (Stratagene). The PCR primers for p21, Bax, Noxa, and GAPDH promoters have been described elsewhere (26). The other PCR primers were as follows. For Cdc25C, the forward primer was 5′-GTATCTGGGAGGACACATCCAGGG-3′, and the reverse primer was 5′-CAAGTTGGTAGCCTGTTGGTTTG-3′. For Cdk1, the forward primer was 5′-CCTTGCCAGAGCTTTTGGAATACC-3′, and the reverse primer was 5′-GACATGGGATGCTAGGCTTCCTGG-3′. For cyclin A2, the forward primer was 5′-AGCAGCCTGCAAACTGCAAAGTTG-3′, and the reverse primer was 5′-TGGTGGGTTGAGGAGAGAAACAC-3′.

Statistical analysis.

Statistical analyses were performed using JMP version 6.0 for Windows (SAS Institute, Cary, NC). Pearson correlation coefficient was calculated to evaluate correlations between the mRNA expression of TopBP1 and NF-Y targets or E2F1. An ordinal logistic regression with cumulative probability model was used to determine the association between tumor grade and TopBP1. The likelihood ratio test was used to test the model assumption. Time to disease progression is defined as the number of months or years between the date of the initial diagnosis and date of the first disease relapse/progression or the last follow-up date. The association of TopBP1 levels and overall survival or disease-free survival was analyzed using the Kaplan-Meier method and the Cox proportional hazards method. P values less than 0.05 were considered statistically significant.

RESULTS

TopBP1 interacts with mutp53 in vitro and in vivo.

Approximately 30% of the tumor-derived p53 mutations affect six “hot spot” residues at positions 175, 245, 248, 249, 273, and 282. We tested the interaction between TopBP1 and several mutant p53 proteins (mutp53s) with V143A, R175H, R248W, R249S, and R273H mutations. The wild-type p53 and mutp53 constructs were cooverexpressed with TopBP1 in a p53-null human lung cancer cell line H1299, and the lysates were immunoprecipitated with a p53 antibody. Consistent with our prior data (26), TopBP1 coimmunoprecipitated with wild-type p53 (Fig. 1A). TopBP1 was also present in mutp53 immunoprecipitates. These data indicate that TopBP1 interacts with mutp53 as well. Colocalization of GFP-p53(R175H) or (R273H) with HcRed-TopBP1 was seen when GFP-mutp53 and HcRed-TopBP1 were coexpressed in H1299 cells (Fig. 1B). The interaction under endogenous protein levels was examined in several cancer cell lines. Reciprocal coimmunoprecipitation was performed using both TopBP1 and p53 antibodies in several mutp53-harboring cancer cell lines: C33A (harboring R273C), OVCAR-3 (harboring R248Q), and SKBR3 (harboring R175H). Indeed, an interaction between endogenous TopBP1 and mutp53 could be detected in these cell lines (Fig. 1C).

Fig. 1.

TopBP1 binds mutant p53 proteins. (A) H1299 cells were transfected with an empty vector, TopBP1, TopBP1 with wild-type p53, or TopBP1 with various mutp53 proteins. The lysates were then immunoprecipitated with a p53 antibody (IP: p53), followed by immunoblotting (IB) as indicated. An aliquot of the cell lysates before immunoprecipitation was analyzed by Western blotting and probed with antibodies against TopBP1 or GAPDH (loading control). (B) TopBP1 can colocalize with mutp53. HcRed-TopBP1 was cotransfected with either GFP-p53-R175H or GFP-p53-R273H in H1299 cells. Forty-eight hours later, the cells were fixed, and the nuclei were stained with Hoechst 33258. The images were captured on a Zeiss fluorescence microscope. Bar, 10 μm. (C) TopBP1 binds mutp53(R273C), mutp53(R248Q) and mutp53(R175H) at endogenous levels. C33A cells (harboring mutp53R273C), OVCAR-3 cells (harboring mutp53R248Q), and SKBr3 cells (harboring mutp53R175H) were left untreated (no tx) or treated with adriamycin (Adr) at 5 μM for 5 h. The lysates were immunoprecipitated with a TopBP1 antibody (αTopBP1), a p53 antibody, or a control healthy mouse IgG as indicated, followed by the same procedures as described above for panel A. long exp., long exposure.

A role for TopBP1 to promote mutp53/NF-Y and mutp53/p63/p73 complex formation.

Studies have clearly shown that mutp53 does not merely lose normal activities of wild-type p53 but also acquires new functions (gain of function). These functions contribute to the maintenance and increase of resistance to chemotherapy (19, 38, 45). Although mutp53 loses its DNA binding activity, it can complex with NF-Y and recruit p300 coactivator to upregulate NF-Y target genes upon DNA damage. It has been shown that the induction of NF-Y target genes, such as cyclin A, cyclin B, and Cdk1 by DNA-damaging agents in mutp53-harboring cancer cells is dependent on mutp53 (7). Consistent with the published data, we also observed mutp53-dependent regulation of Cdk1 and cyclin A in C33A cells (Fig. 2A). To investigate whether the interaction between TopBP1 and mutp53 is involved in this NF-Y-mediated mechanism, we examined the effects of TopBP1 expression on the induction of NF-Y target genes upon cisplatin treatment in C33A cells. As shown in Fig. 2B, cisplatin treatment induced the expression of cyclin A and Cdk1. TopBP1 depletion inhibited their induction, whereas overexpression of TopBP1 increased their expression. These data suggest that TopBP1 might be required for mutp53/NF-Y actions in mutp53 gain of function. Another mechanism for gain of function of mutp53 is binding to p63/p73 and inhibiting the function of p63/p73 (25). We investigated the role of TopBP1 in the binding of mutp53 to NF-Y or p63/p73. C33A cells were either depleted of TopBP1 (Fig. 2C, top blots) or overexpressed with TopBP1 (Fig. 2D, top blots). Then the cells were treated with the DNA-damaging drug adriamycin. Consistent with the literature, the interaction between mutp53 and NF-Y or p63/p73 can be induced after adriamycin treatment. Importantly, depletion of TopBP1 inhibited these interactions, and TopBP1 overexpression enhanced these interactions. Altering TopBP1 expression did not affect the levels of NF-Y, p63/p73, and mutp53. These data support a role for TopBP1 to facilitate the formation of complexes between mutp53(R273C) and NF-Y or p63/p73 in C33A cells. Although p63 signal was not detected on the untreated/siGFP/p53 immunoprecipitation lane of Fig. 2C, top blots, the interaction between mutp53 and p63 at their endogenous levels under the unstressed condition has been reported in several cancer cell lines such as T47D and HaCaT cells (9, 39). We suspect that multiple stripping procedures prior to probing with p63 antibody in that particular experiment hampered our ability to observe their interaction. Indeed, repeated experiments did show an interaction between mutp53 and p63 in unstressed C33A cells (Fig. 2C and D, bottom blots). Depletion of TopBP1 also blocked their interaction. Another point worth explaining about the data in the top blot of Fig. 4D is the absence of further induction of TopBP1 by adriamycin treatment in the sample infected with an adenovirus overexpressing TopBP1 (AdTopBP1). We had observed in our previous paper (26) that when TopBP1 is overexpressed to a high level (e.g., through a recombinant adenoviral vector as in this case), its level sometimes cannot be further induced by adriamycin. We suspect that overexpressed TopBP1 might saturate its normal degradation system, thus dampening the effect of further stabilization by adriamycin. To clarify this issue, we decreased the multiplicity of infection during AdTopBP1 infection and did observe an induction of ectopically expressed TopBP1 by adriamycin (Fig. 2D, bottom right blots).

Fig. 2.

A role for TopBP1 in the formation of complexes between mutp53 and NY-F or p63/p73 in C33A cells. (A) C33A stable cell lines expressing either a control scrambled siRNA or p53 siRNA were left untreated or treated with adriamycin (Adr) at 5 μM for 5 h. The cells were harvested, and the cellular lysates were analyzed by Western blot or immunoblot (IB) analysis as indicated. (B) TopBP1 affects the expression of NF-Y target genes. (Left) C33A cells were treated with cisplatin (20 μM) at the indicated period of time (in hours) and then harvested for Western blot analysis. (Middle and right) TopBP1 in C33A cells was either knocked down by three different TopBP1 siRNAs (middle left blots, pSUPER-siScr and pSUPER-siTopBP1 [28]; middle right blots, lentiviral siScr and siTopBP1-1 and -2 as described in Materials and Methods) or overexpressed (right blots), and then cells were treated with cisplatin (20 μM) for 48 h. The cells were harvested and analyzed by Western blot analysis as indicated. (C) C33A cells were transfected with pSUPER-siTopBP1 or control siGFP. Two days later, the cells were left untreated or treated with adriamycin at 5 μM for 5 h. Immunoprecipitation was performed with a p53 antibody (IP: p53), or a control mouse IgG followed by immunoblotting as indicated. One tenth of cell lysates before immunoprecipitation was also analyzed by Western blotting (right blots). The absence of p63 signal on the untreated/siGFP/αp53 lane is probably due to multiple stripping procedures before probing with p63 antibody. A repeated experiment (bottom blots) detected this interaction and also showed the presence of TopBP1 in the mutp53 immunoprecipitate. (D) C33A cells were infected with an adenovirus expressing TopBP1 (AdTopBP1) or an empty-vector control virus (AdCMV) at a multiplicity of infection of 100 (top blots) or 25 (bottom blots). Forty-eight hours later, the same procedure was performed as described above for panel C.

Fig. 4.

TopBP1 interacts with p63 and p73 through mutp53 and prevents the occupancy of p63 and p73 on the promoters of their target genes. (A) Endogenous interaction between TopBP1 and p63 and p73. C33A cell lysates were immunoprecipitated with TopBP1, p63, or p73 antibody, followed by immunoblotting (IB) with indicated antibodies. The same amount of healthy mouse IgG was used as a control for immunoprecipitation (IP). (B) Interaction between TopBP1 and p63 and p73 is dependent on mutp53. C33A stable cell lines expressing either scramble siRNA or p53 siRNA were immunoprecipitated with TopBP1 antibody, followed by Western blot analysis as indicated. (C) H1299 stable cell lines expressing scramble siRNA and two different TopBP1 siRNAs were cotransfected with the expression vectors for p73α and mutp53(R175H), or mutp53(R273H), along with a p73 activity reporter plasmid (a Bax promoter-luciferase plasmid) and pCMV-βgal (βgal stands for β-galactosidase). Luciferase activity of transfected p73 was determined as fold induction relative to that of the empty vector control. Each sample was performed in triplicate. A portion of the cellular lysates was immunoblotted with the indicated antibody (bottom blots). Values that are significantly different (P < 0.005) for the p73 and p73+mutp53 groups are indicated by an asterisk. Values that are significantly different (P < 0.005) compared to the value for the corresponding siScr group are indicated by two asterisks (paired two-tailed t test). (D) The same procedures were performed as described above for panel C except a p63α expression construct was used for transfection to measure the effect of mutp53 and TopBP1 knockdown on the p63 transcriptional activity. *, P < 0.01 between p63 and p63+mutp53 groups; **, P < 0.05 compared with its corresponding siScr group (paired two-tailed t test). (E) ChIP assay was performed in C33A cells, which have been stably transduced with lentivirus expressing siScr or siTopBP1 (siTopBP1-1). Depletion of TopBP1 was confirmed by Western blot analysis (Fig. 3F and 5). The cells were left untreated or treated with adriamycin at 5 μM for 5 h and then processed for ChIP assay. IgG represents a control antibody using healthy mouse IgG. The eluted chromatin from protein A/G beads was subjected to PCR using the primers for the indicated p63- and p73-responsive promoters. The input represents 10% of the chromatin added to each immunoprecipitation reaction. (F) The ChIP experiments were repeated as described above for panel E, but the chromatin binding of p63 or p73 on the indicated promoters was quantified by real-time PCR instead. The data shown are the means ± SD of three experiments. The values represent relative fold abundance compared to the input chromatin. Input samples represent 8% of chromatin loaded to each immunoprecipitation reaction. *, P < 0.05; #, P < 0.005 compared with its corresponding siScr control (paired two-tailed t test).

TopBP1 interacts with NF-Y and recruits mutp53 to NF-Y target gene promoters.

We further tested the interaction between TopBP1 and NF-Y. As shown in Fig. 3A, endogenous TopBP1 coimmunoprecipitated with endogenous NF-YA and NF-YB in C33A cells. Since TopBP1 interacts with mutp53 and mutp53 can interact with NF-Y (7), the interaction between TopBP1 and NF-Y could be indirect and mediated by mutp53. To determine whether this interaction requires p53 and to gain additional support for the observation in Fig. 3A, we expressed Myc-tagged TopBP1 (Myc-TopBP1) with FLAG-tagged NF-YA (FLAG-NF-YA), -YB, or -YC in p53-null H1299 cells and immunoprecipitated the NF-Y complex with FLAG beads. Indeed, TopBP1 could be coimmunoprecipitated along with the NF-Y complex (Fig. 3B). Thus, TopBP1 interacts with NF-Y independently of mutp53. In this experiment, FLAG beads immunoprecipitated the entire NF-Y complex, because the expressed FLAG-NF-Y subunit formed a complex with other endogenous NF-Y subunits. This experiment did not distinguish which subunit interacted with TopBP1 and whether the interaction is direct. To investigate this, we prepared purified TopBP1 and GST-tagged NF-YA (GST-NF-YA), -YB, or -YC from E. coli and performed a GST-NF-Y pulldown assay (Fig. 3C, left blots). The results show that TopBP1 could directly interact with NF-YA. No interaction between TopBP1 and NF-YB was detected. There appeared to be a very weak interaction between TopBP1 and NF-YC, but the signal is much fainter compared with that from NF-YA (Fig. 3C, left blots). The interaction between TopBP1 and NF-YA was also confirmed by a reciprocal GST-TopBP1 pulldown assay (Fig. 3C, right blots).

Fig. 3.

TopBP1 interacts with NF-Y complex and promotes cooccupancy of mutp53 and p300 with NF-Y on NF-Y-responsive promoters during DNA damage. (A) Endogenous interaction between TopBP1 and NF-Y complex. The lysates from growing C33A cells were immunoprecipitated with two different TopBP1 antibodies (goat polyclonal and mouse monoclonal) or with the corresponding control IgG (goat or mouse) as indicated and then immunoblotted with NF-YA antibody (top blots) or NF-YB antibody (bottom blots). The NF-YA antibody detected both 42- and 35-kDa differentially spliced forms of NF-YA in this experiment. It appears that TopBP1 preferentially bound to 42- kDa NF-YA. exp., exposure. (B) Interaction between TopBP1 and NF-Y complex is independent of mutp53. Myc-TopBP1 was cotransfected with FLAG-NF-YA, -NF-YB, or -NF-YC in p53-null H1299 cells. Two days later, the cells were immunoprecipitated with anti-FLAG beads, followed by Western blot analysis. (C) TopBP1 interacts with NF-YA in vitro. (Left) Purified GST, GST-NF-YA, -NF-YB, or -NF-YC was incubated with purified TopBP1 protein, and GST-tagged proteins were pulled down by glutathione-Sepharose. (Right) Purified GST or GST-TopBP1 was incubated with purified NF-YA protein, followed by glutathione-Sepharose pulldown. The interacting TopBP1 and NF-YA were detected by immunoblotting with corresponding antibodies. GST-NF-Y, GST-TopBP1, and GST are shown by Ponceau S staining. (D) A role for TopBP1 in recruitment of mutp53 and p300 to NF-Y-responsive promoters during DNA damage. C33A cells were infected with an empty vector control adenovirus (AdCMV) or an adenovirus expressing TopBP1 (AdTopBP1) at a multiplicity of infection of 100 (left blots), or infected with lentivirus expressing scramble siRNA (siScr) or TopBP1 siRNA (siTopBP1-1) (right blots). Forty-eight hours later, the cells were treated with adriamycin at 5 μM for 5 h and then harvested for chromatin immunoprecipitation assay. The input represents 10% of chromatin added to each immunoprecipitation reaction. The immunoblots from the lysates are presented in panel F. (E) C33A cells were treated with adriamycin at 5 μM for 5 h. A re-ChIP assay was performed by using TopBP1 antibody in the first immunoprecipitation (IP), followed by p53 or NF-YA antibody in the second immunoprecipitation. (F) C33A cells infected with AdCMV, AdTopBP1, or lentivirus expressing scramble siRNA or TopBP1 siRNA as in panel D were harvested with SDS lysis buffer, and lysates were analyzed by Western blotting using the indicated antibodies. (G) The ChIP assay in C33A cells was performed as described above for panel D, and chromatin binding was quantified by real-time PCR. The data shown are the means plus standard deviations (SD) (error bars) from three experiments. The values represent relative fold abundance compared to the input chromatin. Input samples represent 8% of chromatin loaded for each immunoprecipitation reaction. Values that were significantly different from the value for the corresponding AdCMV or siScr control (paired two-tailed t test) are indicated as follows: ∗, P < 0.05; ∗∗, P < 0.01; #, P < 0.005.

According to the model proposed by Di Agostino and colleagues (7), mutp53 and p300 are recruited to NF-Y target gene promoters upon DNA damage and activate NF-Y target gene expression. The results in Fig. 1, 2, and 3A to C suggest that TopBP1 might promote the recruitment of mutp53 and p300 to NF-Y target gene promoters through its ability to bind both mutp53 and NF-YA. This hypothesis was tested by chromatin immunoprecipitation (ChIP) assay. C33A cells were either overexpressed with TopBP1 or depleted of TopBP1 expression and then treated with adriamycin. Altering the TopBP1 level did not affect the level of NF-Y, mutp53, or p300 (Fig. 3F). Overexpression or knockdown of TopBP1 also did not change the occupancy of NF-Y on the promoters of Cdc25C, Cdk1, and cyclin A (Fig. 3D). Nevertheless, overexpression of TopBP1 increased the promoter occupancy of mutp53 and p300 on the promoters, and depletion of TopBP1 inhibited their promoter occupancy (Fig. 3D and G). To further test the formation of complexes between TopBP1 and mutp53 or NF-YA on the promoters, a ChIP–re-ChIP assay was performed. Indeed, TopBP1/mutp53/NY-YA complex could be found on the NF-Y target gene promoters (Fig. 3E). Taken together, it can be concluded that TopBP1 interacts with NF-YA and recruits mutp53 and p300 to NF-Y target gene promoters.

TopBP1 facilitates the formation of complexes between mutp53 and p63/p73 and inhibits p63/p73 transcriptional activities and binding to target gene promoters.

The results in Fig. 2C and D suggest that TopBP1 might also facilitate the interaction between mutp53(R273C) and p63/p73. To further characterize the relationship between TopBP1, mutp53, and p63/p73, a reciprocal coimmunoprecipitation assay was performed between TopBP1 and p73 or p63 in C33A cells. As shown in Fig. 4A, endogenous TopBP1 and p73 or p63 coimmunoprecipitated. However, the interaction appears to be dependent on the presence of mutp53, as depletion of mutp53 significantly decreased the interaction between TopBP1 and p73 or p63 (Fig. 4B). Together with the data in Fig. 1 and 2C and D, the data suggest that TopBP1 interacts with mutp53 and facilitates mutp53 interaction with p63/p73. To further test this hypothesis, TopBP1 was depleted in H1299 cell lines which had been stably transduced with mutp53(R175H) or mutp53(R273H). We then investigated the role of TopBP1 on the activity of mutp53 in blocking p73 or p63 transcriptional activity. Using a Bax promoter-luciferase reporter assay, we demonstrated that the expression of mutp53(R175H or R273H) inhibited the transcriptional activity of p73 (Fig. 4C) or p63 (Fig. 4D), which reflected the gain of function of mutp53. Importantly, depletion of TopBP1 significantly blocked this activity. These results demonstrate a role for TopBP1 in the mutp53-mediated inhibition of p73 or p63 transcriptional activity. Finally, a ChIP assay of TopBP1-depleted C33A cells was performed. Altering TopBP1 did not affect p63 or p73 levels in C33A cells (Fig. 3F). However, the promoter occupancy of p63 and p73 on p21, Bax, and Puma promoters was enhanced when TopBP1 was depleted in C33A cells (Fig. 4E, left blots). The signals were further increased upon adriamycin treatment (Fig. 4E, right blots). As expected, occupancy of mutp53 was not observed on these promoters. The changes of chromatin binding were confirmed and quantified by real-time PCR (Fig. 4F). In conclusion, TopBP1 facilitates the formation of complexes between mutp53 and p63/p73 and prevents promoter occupancy of p63/p73, consequently inhibiting p63/p73 transcriptional activities.

TopBP1 regulates the expression of NF-Y and p63/p73 target genes in mutp53-bearing cells.

Based on the ChIP assay results, altering the TopBP1 level would affect the NF-Y and p63/p73 target gene expression in C33A cells. TopBP1 was overexpressed by infecting C33A cells with AdTopBP1 or depleted by infecting cells with two different lentiviruses expressing siTopBP1 (lenti-siTopBP1). Indeed, TopBP1 overexpression upregulated the RNA expression of cyclin A, Cdk1, and Cdc25C but downregulated p21, Bax, and Noxa expression (Fig. 5A, top graphs). In contrast, depletion of TopBP1 inhibited cyclin A, Cdk1, and Cdc25C expression but enhanced the expression of p21, Bax, and Noxa (Fig. 5A, bottom graphs). The changes in the protein levels were also observed by Western blot analysis (Fig. 5B).

Fig. 5.

A physiological role for TopBP1 in the control of endogenous NF-Y, p63, and p73 target genes and protein expression. (A) C33A cells were infected with adenovirus containing an empty vector or TopBP1 cDNA at a multiplicity of infection of 100 for 48 h (top graphs). For depletion of TopBP1 (bottom graphs), stable C33A cells infected with lentivirus expressing siScr or two different siTopBP1 were established as described in Materials and Methods. RNA was then extracted, and quantitative real-time RT-PCR analysis was performed using primers specific for the selected NFY, p63, and p73 target genes or GAPDH as indicated. Results were normalized to GAPDH levels and are expressed relative to the expression of the gene in AdCMV or siScr control. Aliquots of the cell lysates were analyzed by Western blotting (left two pairs of blots). Values that were significantly different from the value for the corresponding AdCMV or siScr control (paired two-tailed t test) are indicated as follows: *, P < 0.05; **, P < 0.01; #, P < 0.005. (B) C33A cells that have been infected with lentiviruses expressing siScr or two different siTopBP1 were left untreated or treated with adriamycin at 5 μM for 5 h. The cell lysates were then analyzed by Western blotting using the indicated antibodies. The two sets of blots are from two independent experiments.

TopBP1 facilitates the gain of function of mutp53 in repressing apoptosis during DNA damage.

Blocking p73 function by mutp53 has been shown to be responsible for chemotherapy resistance (19). Since TopBP1 facilitates mutp53 interaction with p63/p73 and downregulates their proapoptotic target genes, whether TopBP1 was required for mutp53 to confer this chemoresistance phenotype was investigated. We depleted or overexpressed TopBP1 in C33A cells and examined the effect on the sensitivity to chemotherapeutic agents such as cisplatin (Fig. 6A and D) and adriamycin (Fig. 6B and C). C33A cells were chosen for the experiments to avoid the effect of TopBP1 on E2F1, since these cells lack Brg1/Brm; therefore, TopBP1 does not regulate E2F1 in these cells (28). Indeed, depletion of TopBP1 sensitized C33A cells to cisplatin treatment, and TopBP1 overexpression inhibited the cisplatin chemosensitivity (Fig. 6A). To further determine whether this effect depended on mutp53, mutp53 was depleted in C33A cells by either HPV E6 (Fig. 6B) or p53 siRNA (Fig. 6C and D), and the sensitivity to adriamycin or cisplatin was tested. Depletion of mutp53 did sensitize C33A cells to chemotherapeutic drugs. The results also demonstrate that the chemosensitization by TopBP1 depletion in C33A cells is largely dependent on mutp53. We further tested the role of TopBP1 for mutant p53 function in another cell line. SKOV-3 cells are p53-null ovarian cancer cells. Reconstitution of a mutant p53(R248W) in these cells inhibited the sensitivity to cisplatin treatment (Fig. 6E and F). The effect of mutant p53 is largely mitigated upon depletion of TopBP1 (Fig. 6E) or enhanced by TopBP1 overexpression (Fig. 6F). Together, the results strongly suggest that TopBP1 facilitates the formation of complexes between mutant p53 and p63/p73 and participates in its gain of function to increase resistance to chemotherapy.

Fig. 6.

TopBP1 promotes resistance to DNA-damaging agents in a mutp53-dependent manner. (A) C33A cells were infected with Ad-siTopBP1 or control Ad-siGFP (left panels), or AdCMV or AdTopBP1 at a multiplicity of infection of 100 (right panels). Twenty-four hours later, the cells were either treated with cisplatin (20 μM) for 24 h or left untreated, and harvested for annexin V-PE/7-AAD staining and flow cytometry analysis. The data shown are the means plus standard errors (SE) (error bars) from three independent experiments. Aliquots of cell lysates were analyzed by Western blotting with TopBP1 and GAPDH antibodies (blots). (B) Stably transduced control pcDNA3 or pcDNA3-E6 C33A cells were infected with Ad-siGFP or Ad-siTopBP1 at a multiplicity of infection of 100. Forty-eight hours later, the cells were left untreated or treated with adriamycin at 5 μM for 5 h. The cells were then harvested for caspase 3/7 cleavage activity. The indicated values are fold induction (means plus SE) relative to that of control cells. Each sample was performed in triplicate. A portion of the cellular lysates was immunoblotted with the indicated antibody (right blots). The 89-kDa proteolytic product of PARP serves as an independent assay for apoptosis. The ratios of the signal intensities between 89-kDa and 116-kDa bands are shown below the blot. (C) The same assays as in panel B were performed in stably transduced siScr or sip53 C33A cells. The graph shows means plus SE from three independent experiments. The P values are based on a paired two-tailed t test. The ratios of the signal intensities between 89-kDa and 116-kDa bands on the PARP blot are shown below the blot. (D) C33A cells were infected with Ad-siGFP or Ad-siTopBP1 at a multiplicity of infection of 100 and lentivirus expressing scramble siRNA or p53 siRNA. Twenty-four hours later, the cells were either left untreated or treated with cisplatin at 20 μM for 24 h. The cells were then harvested and stained with annexin V-PE and 7-AAD and analyzed by flow cytometry. (Top) The graph represents mean plus SE from three experiments. (Bottom) Representative profiles of Annexin V-PE/7-AAD. The indicated numbers are the percentages of annexin-positive cells in each sample. (E and F) TopBP1 mediates the gain of function of mutant p53 in cisplatin resistance. SKOV-3 cells (p53-null cells) were transfected with R248W mutant p53 or an empty vector. The TopBP1 levels were either depleted by a lentivirus expressing TopBP1 siRNA (D) or overexpressed by cotransfection with a TopBP1 expression vector (E), and the cells were treated with cisplatin (2 μM) for 24 h. The cells were then harvested for caspase 3/7 cleavage activity. The indicated values are fold induction (means ± SE) relative to that of control cells. Each sample was performed in triplicate. The P values are based on a paired two-tailed t test. A portion of the cellular lysates was immunoblotted with the indicated antibody (right blots).

TopBP1 mediates the gain of function of mutp53 in promoting cellular proliferation.

mutp53 can aberrantly upregulate cell cycle-related genes such as Cdk1 through NF-Y (7). We also observed increased expression of Cdk1 in p53-null H1299 cells when they are expressed with mutp53(R175H) or mutp53(R273H), and the effect diminished when TopBP1 was knocked down (Fig. 4C, bottom blots). To further investigate whether TopBP1 plays a role in the growth-promoting activity of mutp53, first, a colony formation assay in H1299 cells was performed (Fig. 7A). As expected, wild-type p53 inhibited colony formation. An increase in colony formation was seen in several mutp53 proteins with the V143A, R175H, R248W, and R273H mutations. Importantly, depletion of TopBP1 inhibited the activity of mutp53. Second, stable H1299 cell lines expressing siTopBP1 with mutp53(R175H) or mutp53(R273H) were established, and their proliferation rates were measured. Expression of either R175H or R273H increased the growth rate of H1299 cells. This growth-promoting activity of mutp53 was inhibited by TopBP1 depletion (Fig. 7B). Last, H1299 xenografts were established in nude mice. Expression of mutp53(R175H) or mutp53(R273H) promoted the growth of H1299 xenografts. Remarkably, depletion of TopBP1 in these xenografts blocked the effect of mutp53 (Fig. 7C to E). Consistent with the observation in cell culture experiments (Fig. 4C, bottom blots, and 7B), mutp53 enhanced the expression of Cdk1 and cyclin A in the xenografts, and this effect was blocked by TopBP1 depletion (Fig. 7E). Taken together, TopBP1 mediates the in vivo growth-promoting function of mutp53.

Fig. 7.

TopBP1 depletion blocks the proliferation-stimulating activity of mutp53 in vivo. (A) H1299 stable cell lines expressing siScr or siTopBP1 (siTopBP1-1) were transfected with wild-type p53 (WT) or different mutp53 proteins as indicated. The colony formation assay was performed as described in Materials and Methods. Aliquots of the stable cell lysates were analyzed by immunoblotting with TopBP1 and GAPDH antibodies. The experiments were performed in triplicate and repeated twice. The values are means plus standard errors (SE) (error bars). Values that are significantly different (P < 0.05) from the value for the corresponding siScr group (two-tailed t test) are indicated by an asterisk. Representative pictures of colony formation assay are shown to the right of the graph. The numbers of colonies are shown in the right corner below each plate. (B) Proliferation rates of H1299 cell lines which have been stably transduced with siScr or siTopBP1 (siTopBP1-1) along with either an empty vector, mutp53(R175H), or mutp53(R273H). The cells were counted on days 0, 2, 4, 6, and 8 after plating. On day 8, a portion of the cells from each cell line were lysed and immunoblotted with the indicated antibody (right blots). Values that are significantly different (P < 0.001) from the value for the corresponding mutp53 siScr group or vector siScr group (t test, two-tailed) are indicated by an asterisk. (C to E) The stably transduced H1299 cell lines as indicated were used to establish xenografts in nude mice. Tumors were harvested on day 34. (C) (Left) Photograph of all tumors. Two mice were euthanized earlier (one on day 27 [#] and one on day 31 [##]) due to paralysis of the animals. (Right) Tumor weights on day 34. **, P < 0.005 compared with vector siScr group; *, P < 0.05 compared with the corresponding siScr group (t test, two-tailed). (D) Tumor volumes were measured, calculated, and compared between groups. *, P < 0.001 compared with the corresponding mutp53 siScr group or vector siScr group (t test, two-tailed). (E) (Left) A portion of tumor from the animals in each group was lysed and analyzed by immunoblotting with the indicated antibodies. (Right) Representative images of p53 and TopBP1 immunohistochemical staining of xenografts. Bar, 50 μm. (F) Correlations between TopBP1 and indicated gene expression levels were assessed in 58 mutp53-bearing breast cancer samples using the data set extracted from GEO (GSE4922) (20). Pearson correlation coefficients (r) and the corresponding P values are shown. (G) Ordinal logistic regression analysis of TopBP1 expression levels plotted against the histological grades of breast cancer (Elston grade 1, 2, or 3) in 58 mutp53-bearing breast cancer samples (20). The probabilities of each type of Elston grade are measured as the vertical distance between the curves, with the total across all Elston grade probabilities summing to 1. The bottom curve represents the probability for grade 1, and the top curve represents the probability for grades 1 and 2. The x axis represents the reported values of TopBP1 signal intensities on Affymetrix arrays. Probe set signal values were natural log transformed and scaled. These values were obtained directly from the GEO website.

TopBP1 is often overexpressed in human breast cancer, and its overexpression is associated with poor overall survival (26). Our prior study (26) suggests that its association with poor clinical outcome is partly due to an inhibitory activity toward wild-type p53 function by overexpressed TopBP1. The data presented here suggest that the effect of TopBP1 can be extended to mutp53-bearing cancers. Indeed, among 58 primary human breast cancer samples with p53 mutations (20), TopBP1 expression levels positively correlate with the levels of expression of several NF-Y target genes: cyclin A2, Cdk1, Cdc25C, and cyclin B1 and B2 (Fig. 7F). The correlation could be due to the possibility that TopBP1 and these genes are coregulated by other transcription factors such as E2F1. However, the statistical analysis indicates that the expression level of TopBP1 does not correlate with that of E2F1 among these mutp53-bearing breast tumors (Fig. 7F, right bottom graph). Thus, the correlation between TopBP1 and NF-Y target genes is not because of E2F1 in this data set. There was a statistically significant correlation between the expression levels of TopBP1 and the likelihood of the tumors being classified as high-grade tumors by pathologists in the same cohort of mutp53 breast cancer patients (Fig. 7G). Summing up our experimental results and clinical observations, we conclude that TopBP1 plays an important role in determining the levels of NF-Y targets in these human breast cancers.

DISCUSSION

The oncogenic properties of mutant p53 (mutp53) have contributed significantly to the progression and drug resistance of many human cancers. Our study identifies TopBP1 as a key player in promoting mutp53 gain of function. We also elucidate the molecular mechanism by which TopBP1 cooperates with mutp53 in the NF-Y and p63/p73-mediated processes (Fig. 8). First, TopBP1 binds to both NF-YA and mutp53 and recruits mutp53 and p300 to NF-Y target gene promoters to upregulate the expression of cyclin A, cyclin B, Cdk1, and Cdc25C, etc. In this manner, despite the fact that mutp53 with a mutation(s) in its DNA-binding domain loses binding activity to the p53 target gene promoters, TopBP1 indirectly confers new DNA binding properties by bridging mutp53 with another transcriptional factor (NF-Y). Through this mechanism, TopBP1 cooperates with mutp53 to promote cancer cell proliferation. Second, association between TopBP1 and mutp53 facilitates mutp53 binding to p63/p73. This formation of complexes inhibits the DNA binding activities of p63/p73, consequently blocking the expression of p21, Bax, Puma, Noxa, etc., and leads to resistance to chemotherapeutic agents in cancer cells. The mechanisms may be partially responsible for the high-grade phenotype of TopBP1-overexpressing breast tumors and the short survival time of the patients.

Fig. 8.

A model for the mechanism by which TopBP1 mediates the gain of function of mutant p53. TopBP1 recruits mutp53 and p300 to NF-Y target gene promoters and upregulates cell cycle-related genes. This mechanism may lead to an aberrant proliferation. TopBP1 also facilitates mutp53 interaction with p63/p73 and inhibits promoter occupancy of p63/p73. Through this action, TopBP1 may also facilitate the chemoresistance phenotype of mutp53 tumors.

Although TopBP1 can bind both wild-type p53 (26) and mutant p53, the cellular context and mechanism of action are quite distinct. Binding of TopBP1 to wild-type p53 occurs in healthy cells or cancer cells with wild-type p53 and inhibits p53 DNA binding activity (26). In contrast, TopBP1/mutant p53 interaction occurs in mutant p53-harboring cancer cells and “induces” mutant p53 to bind to NF-Y target gene promoters and upregulate their expression. In addition, while in healthy cells, TopBP1 does not interact with p63/p73, in mutant p53-bearing cancer cells, TopBP1/mutant p53 interaction leads to inhibition of p63/p73 function. Our findings may be very significant clinically in identifying the interaction between TopBP1 and mutp53 as a novel therapeutic target for mutp53-bearing cancers.

It is worth noting that wild-type p53 interacts with NF-Y, resulting in transcriptional repression of cell cycle-related genes (18). The CCAAT box is one of the most frequent elements on promoters, typically between −60 and −100 bp from the transcriptional start site. However, within the context itself, NF-Y is not a strong activator but rather functions as a promoter organizer that cooperates with the neighboring factors depending on cellular contexts. In fact, NF-Y is embedded in both positive and negative methyl histone marks, suggesting a function as an activator as well as a repressor (3). With the ability of TopBP1 to bind E2F1, p53, and NF-Y, TopBP1 might play a unique role to modulate transcriptional regulation of the cell cycle-related genes. It would be interesting to investigate whether TopBP1 is involved in NF-Y functions in healthy cells in the future. How the interaction between TopBP1 and mutp53 facilitates the formation of complexes with p63/p73 also warrants further investigation. We speculate that the association between TopBP1 and mutp53 might position mutp53 for a tighter association with p63/p73. Alternatively, TopBP1 may regulate other proteins, which then facilitate the formation of complexes between mutp53 and p63/p73.

Although aberrant cytoplasmic localization or undetectable expression of TopBP1 was reported in some breast cancer samples (13), analysis of TopBP1 expression by Western blotting, immunoprecipitation/Western blotting, and immunohistochemical staining in primary breast tissues by our group shows that the predominant pattern of aberrant TopBP1 expression in human breast cancer is its nuclear overexpression (26). Overexpression of TopBP1 is associated with high-grade tumors and shorter overall survival. The association between high TopBP1 RNA levels and high-grade tumor is also seen in three breast cancer microarray databases involving a total of 537 breast cancer samples via the Oncomine tool (26). There was no exclusively cytoplasmic staining in our case series, nor in feline and canine mammary tumors (31, 32). We have sequenced the C terminus (about 2.3 kDa) of TopBP1 cDNA from seven primary breast cancer samples and the complete TopBP1 cDNA from another two primary breast cancer tissues but did not find any mutation (data not shown). In the COSMIC database, there are four heterozygous simple missense mutations out of 485 solid tumors (two in ovarian cancer, one in pancreatic cancer, and one in renal cancer, but none in breast cancer). The significance of these rare heterozygous simple mutations remains to be determined. Therefore, despite the fact that we cannot rule out the possibility of a loss of TopBP1 function/expression in a very small fraction of breast cancers, its overexpression is much more prevalent in cancer. The analysis of four breast cancer databases (6, 20, 41, 44) clearly demonstrates that higher levels of TopBP1 are associated with an increased risk of relapse or death (Fig. 9). These data are consistent with our prior analysis when we examined TopBP1 proteins by Western blot analysis or immunohistochemical staining (26). The activities of TopBP1 in repressing E2F1 (27–29) and wild-type p53 functions (26), and its newly uncovered role in promoting mutp53 gain of function may be responsible for the association between high levels of TopBP1 and poor clinical outcomes in breast cancer.

Fig. 9.

Higher levels of TopBP1 in primary breast cancer tissues are associated with an increased risk of relapse or death. (A) We extracted the TopBP1 expression levels and clinical outcomes from four published data sets (6, 20, 41, 44) found with the Oncomine tool or extracted from GEO (GSE2034, GSE4922, and GSE7390). The Cox proportional hazards regression analysis was taken to evaluate the association of TopBP1 expression levels with survival. The hazard ratio shows the risk change over one unit of the difference of TopBP1 signal levels in each microarray data set. The values are relative within the same data set, and the absolute values cannot be compared between two different data sets. Therefore, the absolute values of hazard ratio depend on the scales of TopBP1 intensity signals in each data set and cannot be compared between different data sets. For example, unlike the studies of Desmedt et al. (6), Ivshina et al. (20), and Wang et al. (44), which all used Affymetrix U133A microarrays, the study of van de Vijver et al. (41) used the Agilent Technologies platform and therefore has a much different scale for TopBP1 values. Regardless, the significant P values indicate that there is a significant increase in the risk of progression or death when TopBP1 signals increase in the primary breast tumors. n represents the number of patients in each cohort. ChiSq, chi-square. (B) Kaplan-Meier analysis of breast cancer according to TopBP1 expression levels. The TopBP1 expression levels and clinical outcomes of two published data sets, the data sets of Wang et al. (44) (LN-, lymph node negative) and of van de Vijver et al. (41), were obtained using the Oncomine tool. Patients were ranked according to the levels of expression of TopBP1 in their breast tumors. We then divided the patients into either three groups (TopBP1 levels, high > intermediate > low; top graphs) or four groups (TopBP1 levels, high > intermediate2 > intermediate1 > low; bottom graphs). The numbers (N) of patients in each group are indicated. The numbers of patients in the groups were assigned so that the groups had almost equal numbers of patients. Kaplan-Meier curves were then derived from each group. The Wilcoxon test was used to evaluate the significance. The P values for the tests between all groups are indicated. The bracketed values are the P values for the differences between the high and low groups. Since it is impossible to equally divide the patients into three or four groups, we tested variations of groupings (e.g., n = 96, 95, and 95 versus 95, 96, and 95 versus 95, 95, and 96) and obtained similar results. A representative grouping is presented.