Adenovirus 5 E1A-Mediated Suppression of p53 via FUBP1

ABSTRACT

Far-upstream element (FUSE) binding protein 1 (FUBP1) was originally identified as a regulator of the oncogene c-Myc via binding to the FUSE within the c-Myc promoter and activating the expression of the gene. Recent studies have identified FUBP1 as a regulator of transcription, translation, and splicing via its DNA and RNA binding activities. Here we report the identification of FUBP1 as a novel binding partner of E1A. FUBP1 binds directly to E1A via the N terminus (residues 1 to 82) and conserved region 3 (residues 139 to 204) of adenovirus 5 E1A. The depletion of FUBP1 via short interfering RNAs (siRNA) reduces virus growth and drives the upregulation of the cellular stress response by activating the expression of p53-regulated genes. During infection, FUBP1 is relocalized within the nucleus, and it is recruited to viral promoters together with E1A while at the same time being lost from the FUSE upstream of the c-Myc promoter. The depletion of FUBP1 affects viral and cellular gene expression. Importantly, in FUBP1-depleted cells, p53-responsive genes are upregulated, p53 occupancy on target promoters is enhanced, and histone H3 lysine 9 is hyperacetylated. This is likely due to the loss of the FUBP1-mediated suppression of p53 DNA binding. We also observed that E1A stabilizes the FUBP1-p53 complex, preventing p53 promoter binding. Together, our results identify, for the first time, FUBP1 as a novel E1A binding protein that participates in aspects of viral replication and is involved in the E1A-mediated suppression of p53 function.

IMPORTANCE Viral infection triggers innate cellular defense mechanisms that have evolved to block virus replication. To overcome this, viruses have counterevolved mechanisms that ensure that cellular defenses are either disarmed or not activated to guarantee successful replication. One of the key regulators of cellular stress is the tumor suppressor p53 that responds to a variety of cellular stress stimuli and safeguards the integrity of the genome. During infection, many viruses target the p53 pathway in order to deactivate it. Here we report that human adenovirus 5 coopts the cellular protein FUBP1 to prevent the activation of the p53 stress response pathway that would block viral replication. This finding adds to our understanding of p53 deactivation by adenovirus and highlights its importance in infection and innate immunity.

INTRODUCTION

Human adenovirus (HAdV) is an obligate intracellular parasite targeting, predominantly, terminally differentiated epithelial cells of the eye, airway, and gut (1). The viral genome cannot replicate efficiently in these cells due to the lack of cellular factors required for viral genome copying (2, 3). To overcome this limitation, the virus expresses early proteins that reprogram the infected cell and drive it into S phase; of these proteins, the early 1A (E1A) protein is critical (3, 4). E1A is the first gene transcribed after infection, and its primary function is to induce the cells to enter S phase, deregulate cellular gene expression to favor viral replication, and activate the expression of viral transcription units (2). The E1A protein itself possesses no intrinsic DNA binding activity; instead, it modulates the function of cellular proteins via a large variety of interactions that alter protein function (2). Despite a lack of DNA binding activity, E1A is found on a large number of cellular promoters (5, 6). The recruitment of E1A to these cellular promoters occurs via interactions with promoter-bound factors, such as E2F-DP complexes that regulate S-phase-specific genes (7) or C-terminal binding protein 1 (CtBP1) that plays a role in E1A-mediated transformation (8). E1A is also recruited to all viral promoters via DNA-bound cellular transcriptional regulators, and it is critical for the efficient activation of many of them (9, 10). Although the induction of S phase is essential for viral replication, it also triggers cellular defense mechanisms that lead to cell cycle arrest or apoptosis (11), a highly undesirable outcome for the virus. DNA tumor viruses, such as HAdV, have therefore developed multiple strategies to prevent either cell cycle arrest or apoptosis caused by the activation of the DNA damage response pathway. For HAdV, one of the primary mechanisms that prevent this undesirable outcome targets the p53 tumor suppressor protein. The viral E1B-55k protein together with E4-orf6 forms a complex with p53 that recruits the Cullin-containing ubiquitin ligase assembly that drives p53 ubiquitination and subsequent degradation via the proteasome (12). The viral E4-orf3 protein is able to suppress the activation of p53-regulated target genes via the induction of heterochromatin at p53-regulated promoter sites, preventing the expression of p53-induced genes (13). E1A also influences p53 target gene expression by associating with the Nek9 kinase located on p53-regulated promoters and driving their transcriptional silencing via an unknown mechanism (14). In addition to a direct effect on the expression of p53-regulated genes via interactions with Nek9, E1A is able to modulate p53 function indirectly by altering the activity of cofactors required for p53 transactivation, such as p300/CBP and related enzymes, which are involved in p53-mediated transactivation. p300 also directly acetylates p53 (15–17), which is essential for p53-mediated transactivation (16).

Although a number of previous studies have shown that the status of p53 in a cell can have an effect on HAdV replication (18–21), more-recent studies have shown no effect of p53 on viral replication and attribute the previously reported differences to unknown variables present in cells of different origins (22). Together, these findings paint a picture whereby the suppression of p53 function is of the utmost importance for the virus and highlight the complex nature by which HAdV is able to prevent the expression of p53-regulated genes in infected cells.

In the present study, we have identified a novel E1A binding protein, far-upstream element (FUSE) binding protein 1 (FUBP1), via affinity purification of E1A complexes followed by mass spectrometry (MS) identification. E1A was found to bind to FUBP1 via the N terminus and conserved region 3 (CR), and this interaction was found to be direct. The depletion of FUBP1 via short interfering RNA (siRNA) knockdown led to a mild growth defect of the virus. Interestingly, we found perturbations in FUBP1 nuclear localization during viral infection. FUBP1 knockdown resulted in higher E1A occupancy on viral promoters and a higher degree of acetylation of histone H3 at lysine 9 (K9) associated with these promoters. Interestingly, the depletion of FUBP1 led to the activation of p53-regulated genes during infection, including CDKN1A (p21), GADD45A, and PIG3. The upregulated expression of these genes was associated with higher p53 occupancy at the p53 binding sites within promoters and resulted in the hyperacetylation of K9 of H3. Unexpectedly, we found E1A to associate with p53 during infection via FUBP1, since the depletion of FUBP1 reduced the amount of E1A interacting with p53. Together, our results identify, for the first time, FUBP1 as a novel E1A binding protein that appears to be coopted by the virus for the suppression of p53 target genes during infection.

RESULTS

FUBP1 binds to the N terminus and CR3 of E1A.

Our initial identification of FUBP1 was done via affinity purification of cellular proteins associating with the region of E1A encoded by exon 2 of the gene (consisting of amino acids 186 to 289 of E1A289R), followed by MS-based identification, as we have recently done for other E1A binding proteins (9, 23, 24). E1A was found to associate with endogenous FUBP1 during the course of normal infection (Fig. 1A); we also observed that FUBP1 was able to interact with E1A243R and E1A289R (Fig. 1B). To map the interaction region, we used green fluorescent protein (GFP) fusions of E1A fragments (Fig. 1) and performed coimmunoprecipitation experiments on cotransfected cells. FUBP1 was found to interact with the N terminus and CR3 of E1A. This result was somewhat unexpected, as the region used in the original affinity purification consisted of the region encoded by exon 2, and our construct included residues 187 to 289 of E1A (25). However, our CR3 construct includes residues 139 to 204 (25), which includes a portion of the region encoded by exon 2, potentially explaining this seemingly paradoxical result. Together, these results demonstrate that E1A interacts with FUBP1 via the N-terminal region (residues 1 to 82) and via CR3.

FIG 1.

E1A interacts with FUBP1 during viral infection. (A) HT1080 cells were mock infected or infected with HAdV5 dl309 for 24 h. Cells were subsequently lysed and immunoprecipitated for E1A using the M73 monoclonal antibody. Immunoprecipitates were resolved by SDS-PAGE and blotted for FUBP1 and E1A using anti-FUBP1 and M73 antibodies, respectively. (B) HT1080 cells were transfected with HA-tagged FUBP1 and the indicated E1A constructs expressing either E1A243R or E1A289R or GFP fusions of E1A fragments. Immunoprecipitations were carried out with either M73 antibody for E1A243R and E1A289R or anti-GFP antibody for the GFP fusions. Immunoprecipitates were resolved by SDS-PAGE and blotted for E1A or HA. (C) E. coli-expressed and purified 6×His-E1A289R was mixed with either GST-FUBP1 or GST, also bacterially expressed and purified, and glutathione-agarose beads; incubated; washed; and resolved by SDS-PAGE. Pulldown mixtures were blotted for E1A using M73 monoclonal antibody. Inputs for GST and 6×His-E1A289R (1 μg of each protein, except for GST-FUBP1, which we used by volume, as described in Materials and Methods) were resolved by SDS-PAGE and stained with Coomassie, while the input for GST-FUBP1 was resolved by SDS-PAGE and Western blotted for FUBP1 using anti-FUBP1 antibody.

To determine whether E1A and FUBP1 interacted directly, we performed glutathione S-transferase (GST) pulldown experiments with a GST fusion of FUBP1 purified from Escherichia coli, together with 6×His-tagged E1A289R, also purified from bacterial cells (Fig. 1C). The expression level of the GST fusion of FUBP1 was extremely low, and we could never purify large enough quantities of the protein to be visible on a Coomassie-stained gel; we nevertheless obtained quantities readily detectable by Western blotting (Fig. 1C). In spite of this difficulty, we used purified FUBP1 to perform pulldown experiments with E1A289R. E1A289R was readily pulled down by purified GST-FUBP1, whereas the negative control, GST alone, was not able to interact with E1A. These results suggest that E1A interacts directly with FUBP1.

FUBP1 affects virus growth and is relocalized during infection.

We wanted to determine the effects of the depletion of FUBP1 on virus growth. To assess this, IMR-90 cells were depleted of FUBP1 by using siRNA (Fig. 2A). After depletion, cells were infected with dl309 expressing wild-type (wt) E1A, and virus titers were determined 48, 72, and 96 h after infection (Fig. 2A). The depletion of FUBP1 had a negative effect on virus growth, with a reduction of growth of about 2-fold at the 96-h time point, and a lesser effect at earlier time points.

FIG 2.

Depletion of FUBP1 reduces HAdV growth. (A) IMR-90 cells were depleted of FUBP1 by siRNA. After depletion, cells were infected with HAdV5 dl309 at an MOI of 10. Virus was harvested, and the titer was determined on 293 cells at the indicated time points. Error bars represent standard deviations (n = 3). (B) Infected IMR-90 cells were stained for E1A and FUBP1 using M73 and anti-FUBP1 antibodies or for DBP and FUBP1 using anti-DBP and anti-FUBP1 antibodies 24 h after infection. Secondary Alexa 488 and Alexa 594 antibodies were used to visualize E1A, DBP, and FUBP1. DAPI was used as a nuclear counterstain. Images were acquired by using a Zeiss LSM700 confocal laser scanning microscope using a 63× objective lens.

E1A was previously shown to affect the subcellular localizations of a variety of cellular proteins (9, 23, 26); we therefore wanted to see if FUBP1 is also affected in a similar manner. To determine this, IMR-90 cells were infected with dl309, fixed, and then stained for FUBP1, E1A, or DNA binding protein (DBP), 24 h after infection (Fig. 2B). In uninfected cells, FUBP1 showed a diffuse nuclear distribution consistent with reports of this protein's subcellular localization. In infected cells, FUBP1 relocalized within the nucleus and showed a distinct punctate staining pattern. FUBP1 staining in infected cells showed no distinct colocalization with E1A or viral replication centers (Fig. 2B). Together, these results show a modest effect of FUBP1 on viral replication and demonstrate that during infection, FUBP1 is relocalized within the nucleus of infected cells.

Effects of FUBP1 depletion on viral genome replication and gene expression.

Since we observed that the depletion of FUBP1 had a modest effect on virus growth, similar to what we have seen for other recently identified E1A binding proteins (9, 23), we wanted to determine how depletion affects viral genome replication and viral gene expression (Fig. 3). The depletion of FUBP1 in IMR-90 cells had a similarly modest effect on viral genome replication (Fig. 3A). In FUBP1-depleted cells, we observed fewer genomes than in cells transfected with control siRNA (siControl). The observed difference was small but reproducible. Similarly, we determined how the depletion of FUBP1 affects viral gene expression (Fig. 3B). In depleted cells, viral genes were expressed at a slightly higher level than in cells treated with a negative-control siRNA, with most genes being upregulated somewhere between 50 and 100%. E2A expression was affected the most 48 h after infection, when the gene was upregulated over 3-fold in depleted versus control-treated cells. We also observed a high variability in hexon mRNA levels at 24 h, and this is likely due to low levels of expression of this gene at this time after infection (27). The observed results were not due to variations in our normalization control (glyceraldehyde-3-phosphate dehydrogenase gene [GAPDH]), as raw quantification cycle (C_q) values from our samples showed little fluctuation due to either the knockdown of FUBP1 or infection (Fig. 3C) over the 48-h course of our experiment and were similarly unchanged for another housekeeping gene, the glucose 6-phosphate dehydrogenase gene (G6PD) (Fig. 3C). Overall, these results show a modest effect of FUBP1 depletion on viral genome replication and viral gene expression.

FIG 3.

Effects of FUBP1 depletion on viral genome replication and gene expression. (A) IMR-90 cells were transfected with siRNA as indicated and infected 24 h later with HAdV5 dl309 at an MOI of 10 for the indicated time points at which cells were harvested, and viral DNA was extracted and quantified by qPCR with a Bio-Rad CFX96 instrument. The pXC1 plasmid was used to generate the standard curve, and E1B-specific primers were used. Asterisks represent statistically significant differences between siControl and siFUBP1 conditions, with a P value of ≤0.025. Error bars represent standard deviations (n = 3). (B) Cells were treated as described above for panel A except that total cellular RNA was extracted by using TRIzol reagents. RNA was converted to cDNA by using Vilo master mix and then used for qPCR using the Bio-Rad CFX96 instrument. Data are presented as fold changes versus siControl-transfected cells, using the Pfaffl method for data analysis. Asterisks represent statistically significant differences, with a P value of ≤0.05. Error bars represent standard deviations (n = 3). (C) The same cells as in panel B were analyzed for levels of G6PD and GAPDH mRNAs after FUBP1 knockdown and/or infection with HAdV dl309. Raw C_q values are shown. Error bars represent standard deviations (n = 3).

FUBP1 localizes to viral promoters, affecting E1A occupancy and H3 acetylation.

We have observed that the depletion of FUBP1 affects viral gene expression during infection (Fig. 3). We therefore wanted to determine whether FUBP1 binds to viral promoters during infection together with E1A. We performed chromatin immunoprecipitation (ChIP) on infected cells using either E1A antibody or FUBP1 antibody; anti-rat rabbit IgG was used as a negative-control IgG (Fig. 4). E1A was observed on all viral promoters, as we observed previously (9). FUBP1 was also found to associate with the same promoters as E1A.

FIG 4.

FUBP1 associates with viral promoters. IMR-90 cells were infected with HAdV5 dl309 for 24 h at an MOI of 10 and fixed, and chromatin was immunoprecipitated by using a cocktail of M73 and M58 for E1A, anti-FUBP1 antibody, or rabbit anti-rat antibody, as a negative control. Following immunoprecipitation and DNA purification, the samples were analyzed by using the Bio-Rad QX200 droplet digital PCR system and plotted as a percentage of the input. MLP, major late promoter. Error bars represent standard deviations (n = 4).

Since we observed FUBP1 recruitment to viral promoters during infection, and FUBP1 depletion affected viral gene expression, we wanted to determine the effect that this had on E1A occupancy on these promoters and how this affected H3 K9 acetylation, a marker of active gene transcription (28). ChIP analysis showed enhanced E1A occupancy on all viral promoters following the siRNA-mediated depletion of FUBP1 (Fig. 5A). Similarly, we observed an enhanced acetylation of K9 of H3 after FUBP1 was depleted (Fig. 5B). Changes in the acetylation of K9 in H3 observed at viral promoters were not due to changes in overall viral genome chromatinization due to FUBP1 depletion, as these remained similar at all viral promoters analyzed (Fig. 5C). There was an overall difference in E1A occupancy at various viral promoters between the data shown in Fig. 4 and 5, and it is unclear why we observed this. It is likely that the transfection of siRNA, even a negative-control one (as in Fig. 5), affects the cells in ways that may reduce viral gene expression and, hence, E1A levels.

FIG 5.

FUBP1 depletion enhances E1A occupancy on viral promoters and increases histone acetylation. (A) IMR-90 cells were infected with HAdV5 dl309 for 24 h at an MOI of 30 and fixed, and chromatin was immunoprecipitated by using a cocktail of M73 and M58 for E1A or rabbit anti-rat antibody as a negative control. Following immunoprecipitation and DNA purification, the samples were analyzed by using the Bio-Rad QX200 droplet digital PCR system and plotted as a percentage of the input. Asterisks represent statistically significant differences, with P values of ≤0.025. Error bars represent standard deviations (n = 2). (B) IMR-90 cells were treated as described above for panel A, and chromatin was immunoprecipitated by using anti-histone H3 K9 antibody or rabbit anti-rat antibody as a negative control. Following immunoprecipitation and DNA purification, the samples were analyzed by using the Bio-Rad QX200 droplet digital PCR system and plotted as a percentage of the input. Asterisks represent statistically significant differences, with P values of ≤0.025. Error bars represent standard deviations (n = 2). (C) Chromatin was immunoprecipitated from IMR-90 cells treated in the same way as described above for panel A, using total histone H3 antibody or negative-control IgG. Following immunoprecipitation and DNA purification, the samples were analyzed by using the Bio-Rad QX200 droplet digital PCR system and plotted as a percentage of the input. Asterisks represent statistically significant differences, with P values of ≤0.05. Error bars represent standard deviations (n = 2).

FUBP1 is depleted from the c-Myc FUSE during infection, affecting c-Myc expression.

FUBP1 affects the expression of the c-Myc oncogene and localizes to the far-upstream element (29). Therefore, we wanted to determine how infection affects the recruitment of FUBP1 to the FUSE of the c-Myc promoter. ChIPs were performed, and the occupancy of E1A and FUBP1 was assessed. E1A was not found at the FUSE, but FUBP1 was (Fig. 6A). Interestingly, during infection, there was a >50% loss of FUBP1 at the FUSE, suggesting that this may affect c-Myc expression. To determine the effect that this had on c-Myc expression, we investigated its expression during viral infection and during infection of cells with FUBP1 depleted (Fig. 6B). In cells transfected with control siRNA, c-Myc mRNA levels dropped consistently throughout the infection and were lowest 48 h after infection, while in cells in which FUBP1 was depleted, c-Myc mRNA levels also dropped initially but then stabilized and remained only slightly lower than those in uninfected cells. This was unsurprising in light of our ChIP experiments that showed a loss of FUBP1 from the FUSE, which would be associated with reduced c-Myc expression. However, we did not expect that c-Myc mRNA levels would be higher after FUBP1 depletion, although it should be noted that late in infection, changes to mRNA may not directly correlate with protein levels due to a general inhibition of cellular translation by the virus.

FIG 6.

FUBP1 FUSE occupancy is reduced after infection, and c-Myc mRNA levels are decreased. (A) IMR-90 cells were either mock infected or infected with HAdV5 dl309 at an MOI of 10 for 24 h and then immunoprecipitated for E1A or FUBP1 or with negative-control IgG. Occupancy was analyzed at the FUSE of the c-Myc promoter using the Bio-Rad QX200 droplet digital PCR system and plotted as a percentage of the input. Error bars represent standard deviations (n = 4). (B) IMR-90 cells were transfected with siControl or siRNA targeting FUBP1. Twenty-four hours later, the cells were infected with HAdV5 dl309 or mock infected, and at the indicated time points, total cellular RNA was extracted by using the TRIzol reagent, reverse transcribed to cDNA by using Vilo master mix, and analyzed for mRNA levels by using the Bio-Rad QX200 digital-droplet PCR system. Data are represented as percentages of GAPDH mRNA levels. Asterisks represent statistically significant differences, with P values of ≤0.01. Error bars represent standard deviations (n = 4).

FUBP1 affects expression of p53-regulated genes, inhibits p53 promoter occupancy, and reduces p53-regulated promoter acetylation during infection.

FUBP1 was also implicated in the regulation of p53 target gene expression (30), and higher c-Myc levels lead to increased p53 activity (31). We therefore wanted to investigate whether the depletion of FUBP1 led to changes in p53-regulated genes (Fig. 7A). The expression levels of three p53-regulated genes were examined: p21, GADD45A, and PIG3. Under control knockdown conditions, the expression levels of p21 and PIG3 increased slightly during infection, while the expression level of GADD45A was slightly reduced. Interestingly, the depletion of FUBP1 caused an increase in p53-regulated gene expression, which was most pronounced with p21 and GADD45A compared to the nondepleted cells. Surprisingly, levels of p21 mRNA were very high in the cells and reached nearly 50% of the GAPDH mRNA levels 48 h after infection in FUBP1-depleted cells, versus only about 25% in nondepleted cells. Together, these results show that there is a greater induction of p53-regulated genes in FUBP1-depleted cells, which may contribute to the reduced virus growth.

FIG 7.

FUBP1 depletion induces expression of p53-regulated genes via enhanced p53 promoter occupancy and histone acetylation. (A) IMR-90 cells were transfected with siControl or siRNA targeting FUBP1. Twenty-four hours later, the cells were infected with HAdV5 dl309 at an MOI of 30 or mock infected, and at the indicated time points, total cellular RNA was extracted by using the TRIzol reagent, reverse transcribed to cDNA by using Vilo master mix, and analyzed for mRNA levels by using the Bio-Rad QX200 droplet digital PCR system. Data are represented as percentages of GAPDH mRNA levels. Asterisks represent statistically significant differences, with P values of ≤0.05. Error bars represent standard deviations (n = 4). (B) IMR-90 cells were transfected with siControl or siRNA targeting FUBP1. Twenty-four hours later, cells were infected with HAdV5 dl1520 at an MOI of 30 for 24 h and then immunoprecipitated for p53 or the IgG negative control. Occupancy was analyzed at the indicated promoters by using the Bio-Rad QX200 droplet digital PCR system and is plotted as a percentage of the input. Asterisks represent statistically significant differences, with P values of ≤0.05. Error bars represent standard deviations (n = 2). (C) IMR-90 cells were treated as described above for panel B, except that immunoprecipitation was carried out for acetylated histone H3 at lysine 9 (AcH3). Asterisks represent statistically significant differences, with P values of ≤0.025. Error bars represent standard deviations (n = 2).

The observation that p53-regulated genes are expressed at higher levels in cells in which FUBP1 was depleted suggested that FUBP1 is a negative regulator of p53 function that HAdV may be utilizing in order to suppress p53 activity. FUBP1 was previously shown to be a negative regulator of p53 DNA binding (32). Therefore, we wanted to determine whether the enhanced p53 target gene expression after FUBP1 knockdown and infection was due to enhanced p53 promoter occupancy and transactivation. To determine this, we performed ChIP analysis after infection of cells with HAdV dl1520 to ensure that functional p53 was present. This virus carries a mutation in the E1B-55k gene and results in no E1B-55k protein being expressed but is otherwise similar to dl309 (33). In cells depleted for FUBP1, p53 was consistently observed at higher levels at p53-regulated promoters (Fig. 7B); this directly correlated with higher levels of H3 K9 acetylation observed at these promoters (Fig. 7C) and was in agreement with our observations of higher expression levels of these genes after infection (Fig. 7A).

E1A stabilizes the p53-FUBP1 interaction.

Previous reports (32) indicated that FUBP1 binds to p53, negatively affecting its DNA binding activity and resulting in the inhibition of p53-mediated transactivation. Our results show an upregulated expression of p53-regulated genes after FUBP1 depletion, likely due to enhanced p53 promoter occupancy. We therefore hypothesized that E1A may be using FUBP1 as means of inactivating this potent tumor suppressor. To test this hypothesis, we investigated whether E1A affects the previously reported interaction between FUBP1 and p53 (32). The presence of E1A enhanced the interaction between FUBP1 and p53 (Fig. 8A); interestingly, E1A was also present in this complex. To determine whether FUBP1 was required for the recruitment of E1A to p53, we performed further coimmunoprecipitation assays following FUBP1 depletion. In this case, we used HAdV dl1520, which lacks a functional E1B-55k protein and is unable to degrade p53. The depletion of FUBP1 via siRNA (siFUBP1) resulted in a substantial reduction in the amount of E1A found to associate with p53 (Fig. 8B) after infection with dl1520, suggesting that E1A binds to p53 via FUBP1.

FIG 8.

E1A enhances the p53-FUBP1 interaction and is recruited to p53 via FUBP1. (A) HT1080 cells were cotransfected with plasmids expressing FUBP1 and p53, with or without a plasmid expressing E1A. Six hours after transfection, actinomycin D was added to a final concentration of 10 nM, and the cells were incubated for a further 16 h. Immunoprecipitations were carried out for p53 or FUBP1 as indicated. Immunoprecipitated E1A, p53, or FUBP1 was detected by using the respective antibodies. Inputs are shown. (B) HT1080 cells were transfected with siRNA targeting FUBP1 or a negative-control siRNA for 24 h. Cells were then infected with HAdV dl1520 at an MOI of 30 and transfected with a plasmid expressing wt p53 for 6 h. Actinomycin D was added at a final concentration of 10 nM for 16 h prior to immunoprecipitation for p53. Associated E1A was detected by using a cocktail of M58 and M73 antibodies. Inputs are shown.

DISCUSSION

The present study describes the identification of FUBP1 as a novel E1A binding protein that interacts with the N terminus and CR3 of E1A. GST pulldown assays show that the interaction is direct. The depletion of FUBP1 has a modest negative effect on virus growth and an enhancing effect on early viral gene expression while also having a modest negative effect on viral genome replication. Interestingly, we observed a substantial reorganization of the nuclear distribution of FUBP1 during viral infection. We also examined the effects of infection on the localization of FUBP1 to the FUSE within the c-Myc promoter. During infection, FUBP1 was lost from the FUSE, and c-Myc mRNA levels were reduced. Interestingly, when FUBP1 was depleted via siRNA, c-Myc mRNA levels were reduced to a lesser extent than without depletion. It is possible that the higher levels of c-Myc mRNA led to higher c-Myc protein levels (at least early in infection), thus driving p53 function, consistent with the known effects of c-Myc on p53 activity (31). It is also possible that p53 activity was upregulated, in FUBP1-depleted cells, via an as-yet-unknown mechanism. Ultimately, however, our results suggest that FUBP1 negatively regulates p53 during infection in cooperation with E1A (Fig. 9). Importantly, E1A was found to stabilize the p53-FUBP1 interaction (Fig. 8), and it was found in the complex with FUBP1 and p53 in a FUBP1-dependent manner. This resulted in reduced p53 promoter occupancy and transactivation (Fig. 7).

FIG 9.

Diagram depicting a model showing the effects of E1A and FUBP1 on p53 promoter occupancy and transactivation. Our data suggest that during infection or expression of E1A, E1A is targeted to FUBP1-p53 complexes and stabilizes them, preventing p53 promoter binding. When FUBP1 is depleted, p53 is no longer restricted from binding to target promoters and activates their gene expression. DBD, DNA binding domain.

The binding of FUBP1 to E1A was found to occur via the N terminus and CR3 of E1A (Fig. 1), yet our initial MS purification used the region of E1A encoded by exon 2 of the gene. The construct used for initial purification and identification was a myc-tagged fragment of E1A encompassing residues 187 to 289, while our CR3 construct used in the GFP coimmunoprecipitation assays in Fig. 2 encodes amino acids 139 to 204 of E1A, with a substantial overlap of our C-terminal construct. The overlap of 17 amino acids suggests that FUBP1 may be binding to CR3 at the C terminus of the domain, but binding to this region alone is weak, as we did not detect a substantial band in our coimmunoprecipitation assays using the GFP-exon 2 fusion construct. Two possible explanations provide reasons why the GFP-exon 2 construct did not immunoprecipitate FUBP1. First, it is possible that this region alone binds to FUBP1 weakly, whereas CR3 as a whole binds strongly; since MS is very sensitive, it was able to detect even a weak interaction not detectable by Western blotting. Second, we used a myc-tagged construct in our affinity purification, whereas in the coimmunoprecipitation experiments, a GFP fusion was used instead; the large GFP could hinder interactions that may otherwise be more stable with a smaller tag, such as the myc epitope. Unfortunately, we were unable to identify a single deletion mutant of E1A that lost the interaction with FUBP1 despite an exhaustive search with deletion mutants covering all of E1A289R. This is likely due to the two binding regions that have relatively high binding affinities for FUBP1 (Fig. 1). Interestingly, the binding of proteins by E1A via multiple domains seems to be a common mode of action for E1A. We previously observed this for CtBP (8), pCAF (34), and p300/CBP (25); others have been reported as well, such as pRb (35). Multiple interaction domains that can bind to a target independently offer advantages beyond simply stabilizing or strengthening an interaction. For example, if one site is occupied by a bound protein, binding to a different target can still occur via the second site. It also provides greater flexibility, particularly for a small protein, such as E1A, that binds multiple targets, where the blocking of a binding site can occur due to a crowded interaction region, as exemplified by the N terminus and CR3. Finally, this allows for novel interactions to be orchestrated by E1A, whereby a protein binding at the N terminus can be forced to artificially interact with one bound to CR3, for example, for the purpose of inducing novel posttranslational modifications.

The depletion of FUBP1 via siRNA had a negative impact on virus growth, with about a 2-fold reduction in viral titers after 96 h (Fig. 2). Unexpectedly, early viral gene expression was enhanced modestly by FUBP1 depletion (Fig. 3). These two observations seem at odds with one another, yet they are similar to what was previously observed with DNA replication-related element binding factor (DREF) (9), and they suggest that FUBP1 may have other functions during viral infection outside viral gene expression that are more influential on viral replication. Interestingly, the depletion of FUBP1 resulted in enhanced E1A occupancy at viral promoters as well as an enhanced acetylation of histone H3 at K9, in agreement with the expression results, since K9 acetylation is associated with active transcription (28). It is possible that the depletion of FUBP1 affects the duration of viral early or late phases, as was previously reported for the dl1520 mutant virus (36). An extended early phase with prolonged early protein expression could lead to an abridged late phase, reducing both the expression of late proteins and the replication of viral genomes. The latter scenario is more likely, as the depletion of FUBP1 did not affect late protein levels (data not shown) but modestly reduced viral genome replication to a similar degree as what was observed for viral growth.

Our results show that the depletion of FUBP1 via siRNA leads to higher levels of c-Myc mRNA (Fig. 6) and higher levels of expression of several p53-responsive genes (Fig. 7), one of which, p21, is also a direct target for regulation by FUBP1 (37). Although FUBP1 was previously shown to be important for the induction of c-Myc expression, recent studies have shown that it plays a more complex role (38). Importantly, FUBP1-null fibroblasts showed highly dysregulated levels of c-Myc expression, indicative of both positive and negative roles for FUBP1 in c-Myc regulation (38). The upregulated expression of p21, GADD45A, and PIG3 could hinder viral growth and explain the overall reduced virus yields observed after FUBP1 depletion (Fig. 2), despite higher viral gene expression levels, which already may be at a saturation level. The suppression of p53 activity is a common feature of many viruses, and HAdV has multiple mechanisms that target this critical tumor suppressor protein. FUBP1 was also shown to antagonize the activity of p53 (30) and was shown to play a role in hepatitis C virus replication by interfering with p53 function via inhibition of DNA binding to promoter regions (32). Indeed, our results show a similar mode of action of E1A. The presence of E1A enhanced the interaction between FUBP1 and p53 (Fig. 8), while the depletion of FUBP1 reduced the ability of E1A to interact with p53 and resulted in higher p53 occupancy on p53-regulated promoters. This higher occupancy was correlated with the hyperacetylation of H3 at K9 of these promoters and, ultimately, higher gene expression levels (Fig. 7). This demonstrates the importance of the inhibition of the p53 pathway during infection and demonstrates its importance in the regulation of cell growth and innate immunity (39, 40), as exemplified by the myriad of ways in which HAdV aims to deregulate this pathway. The degradation of p53 via E1B-55k–E4-orf6 is perhaps the major route of p53 inactivation by HAdV (12, 41), yet it is clearly not the only one (13, 14, 42). The reason for these additional mechanisms may stem from inefficiencies in p53 degradation (for example, we were readily able to detect the p53 protein during infection with dl309 24 h after infection) and the need to build up sufficient levels of E1B-55k and E4-orf6 for efficient degradation to happen. It is unlikely that the levels of these two proteins would be high early in infection to expediently degrade p53, a critical time for the virus and a period at which levels of E1A are relatively high. Whether the higher levels of c-Myc expression in FUBP1-depleted cells contribute to the induction of p53 is unclear; however, in normal cells, deregulated expression of c-Myc alone can drive p53-dependent and -independent apoptosis (43), which in itself is an undesirable outcome for the virus.

Interestingly, our results, particularly the effects of infection on the expression of p21 and PIG3, present a different picture than what was previously reported for infected cells (13). Whereas we observed an upregulation of the expression of p21 and PIG3, that previous study observed a significant reduction in the levels of mRNAs produced from these genes. It is unclear why these differences are observed. The most likely explanation is that they are due to the different cell types used: whereas we used a fibroblast cell line (IMR-90), the previous study used primary small epithelial airway cells. Importantly, we have consistently observed upregulated p21 and PIG3 expression in both IMR-90 and WI-38 cells (unpublished results). Nevertheless, these different results suggest an unappreciated complexity of the regulation of the p53 pathway by viral proteins that merits further investigation.

The present study has identified, for the first time, the cellular protein FUBP1 as a novel binding target of HAdV5 E1A. FUBP1 binds directly to the N terminus and CR3 of E1A, the two conserved transactivation domains of the protein, and it is recruited to viral promoters during infection, affecting viral gene expression. Our results show that the depletion of FUBP1 in infected cell is deleterious to virus growth, possibly by allowing higher levels of activation of p21 and other p53-responsive genes that interfere with viral replication. Importantly, E1A was found to stabilize the p53-FUBP1 interaction, while the depletion of FUBP1 enhanced p53 promoter occupancy and the acetylation of H3 at activating K9, indicative of transcriptional activity (28). FUBP1 has other known functions besides regulating p53 activity, and since it was previously shown to affect viral mRNA metabolism (44, 45), it would be of interest to investigate whether it plays any such role in HAdV mRNA regulation or translation. Overall, the myriad of pathways that FUBP1 regulates (46) hints at its function as a cellular hub protein that regulates multiple key processes and is yet another key protein that E1A targets (2) in order to suppress p53, via a mechanism summarized in Fig. 9. Importantly, our study highlights a mechanism by which p53 is inactivated via an oncoprotein and further supports the notion that FUBP1 is a negative regulator of p53 function and DNA binding.

MATERIALS AND METHODS

Antibodies.

Mouse monoclonal anti-E1A M73 and M58 antibodies were previously described (47) and were grown in-house and used as the hybridoma supernatant. Mouse monoclonal anti-72-kDa E2 DNA binding protein (DBP) antibody was previously described (48) and was used at a dilution of 1:400 for Western blotting. Anti-FUBP1, anti-adenovirus type 5, anti-histone H3, and anti-acetyl K9 H3 antibodies were purchased from Abcam (catalog numbers ab181111, ab6982, ab180727, and ab10812, respectively) and were used according to the manufacturer's specifications. Rat anti-hemagglutinin (HA) antibody (Roche), clone 3F10, was used at a dilution of 1:5,000 for Western blotting. Anti-p53 mouse monoclonal antibody clone 1C12 was obtained from Cell Signaling Technology (catalog number 2524). Secondary antibodies were purchased from Jackson ImmunoResearch.

Cell and virus culture.

IMR-90 (ATCC CCL-186) and HT1080 (ATCC CCL-121) cells were grown in Dulbecco's modified Eagle's medium (DMEM) (HyClone) supplemented with 10% fetal bovine serum (Seradigm), streptomycin, and penicillin (HyClone). All virus infections were carried out in serum-free medium for 1 h, after which saved complete medium was added, without the removal of the infection medium.

Chromatin immunoprecipitation.

Chromatin immunoprecipitation (ChIP) was carried out essentially as previously described (25). HT1080 cells were infected with the indicated adenoviruses at a multiplicity of infection (MOI) of 10 or 30 and harvested 24 h after infection for ChIP analysis.

PCRs were carried out for HAdV5 early and major late promoters using EvaGreen master mix for droplet digital PCR (ddPCR) (Bio-Rad) according to the manufacturer's directions; 3% of total ChIP DNA was used as a template with a Bio-Rad QX200 droplet digital PCR instrument (Bio-Rad). The annealing temperature used was between 55°C and 65°C, depending on the primer set, and 40 cycles were run. Primers for viral promoters were described previously (9).

Immunofluorescence.

IMR-90 cells were plated at a low density (∼40,000 cells per chamber) onto chamber slides (Nalgene Nunc) and subsequently infected as described above. Twenty-four hours after infection, cells were fixed in 4% formaldehyde, blocked in blocking buffer (1% normal goat serum, 1% bovine serum albumin [BSA], and 0.2% Tween 20 in phosphate-buffered saline [PBS]), and stained with specific primary antibodies. M73 was used neat (hybridoma supernatant), E2 DBP antibody was used at a 1:100 dilution (hybridoma supernatant), FUBP1 antibody was used at a dilution of 1:300, and Alexa Fluor 488 and 594 secondary antibodies (Jackson ImmunoResearch) were used at a dilution of 1:600. After staining and extensive washing, slides were mounted by using Prolong Gold with 4′,6-diamidino-2-phenylindole (DAPI) (Invitrogen) and imaged by using a Zeiss LSM700 confocal laser scanning microscope. Images were analyzed by using the Zeiss ZEN software package.

Immunoprecipitation.

Transfected HT1080 cells were lysed in NP-40 lysis buffer (0.5% NP-40, 50 mM Tris [pH 7.8], 150 mM NaCl) supplemented with a protease inhibitor cocktail (Sigma). The cell lysate containing 1 mg of total protein was used for immunoprecipitation (IP) with the monoclonal M73 anti-E1A antibody. For immunoprecipitations of p53, FUBP1, and E1A, HT1080 cells were cotransfected with plasmids for the expression of all three proteins for 6 h, actinomycin D was added to a final concentration of 10 nM to stabilize p53 and activate the transcription of p53-regulated genes (49), and 18 h later, cells were lysed and immunoprecipitated for p53 or FUBP1 using their respective antibodies. For immunoprecipitations of p53 and E1A from infected cells, HT1080 cells were transfected with siRNA for FUBP1 or control siRNA for 24 h, infected with HAdV dl1520 and transfected with the p53 plasmid for 6 h, and then treated with 10 nM actinomycin D to ensure p53 stability for 16 h prior to lysis and immunoprecipitation of p53 using anti-p53 antibody 1C12.

PCR primers.

Primers used were GAGGCTATTCTGCCCATTTG and CCTCCTCGTCGCAGTAGAAA for c-Myc and CTCTTTTGGAGGTGGTGGAG and CCCACACATGATTTGTTTGC for the FUSE. Primers not listed were previously described (9, 14, 23, 27, 50).

Plasmids.

The expression plasmid for pcDNA3.1-E1A was described previously (26), and it expresses all E1A isoforms. The expression plasmid pGEX-6P-1-FUBP1 was generated by cloning FUBP1 in frame with the GST protein. pCGNM-HA-FUBP1 was a generous gift from Hye-Jung Chung and David Levens and was previously described (51). The plasmid for the expression of p53, pcDNA3-p53WT, was obtained from Addgene (catalog number 69003).

Protein purification and GST pulldown assay.

The glutathione S-transferase fusion with FUBP1 was made by subcloning the cDNA into pGEX-6P1 (GE Healthcare Life Sciences) in frame with the N-terminal GST tag. His-tagged E1A289R was made by subcloning the entire E1A289R cDNA into the pET42 vector (Novagen) in frame with a C-terminal 6×His tag. Proteins were expressed in Escherichia coli strain BL21(DE3) and purified on their respective resins, according to the manufacturer's specifications. The GST pulldown assay was carried out as previously described (25). GST-FUBP1 was expressed at low levels in bacterial cells or was insoluble; we therefore used a 50-μl volume in GST pulldown assays instead of 1 μg while maintaining other proteins at the 1-μg level.

Real-time gene expression analysis.

HT1080 or IMR-90 cells were infected with dl309 (52) at an MOI of 10. Total RNA was extracted by using the TRIzol reagent (Sigma) at the indicated time points, according to the manufacturer's instructions. A total of 1.25 μg of total RNA was used in the reverse transcriptase reaction by using SuperScript Vilo reverse transcriptase (Invitrogen) according to the manufacturer's guidelines, using random hexanucleotides for priming. The cDNA was subsequently used for real-time expression analysis using the Bio-Rad CFX96 real-time thermocycler. Analysis of expression data was carried out by using the Pfaffl method (53), and values were normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA levels and compared between siControl- and siFUBP1-transfected cells. Total E1A was detected as previously described (27).

siRNA knockdown.

siRNA knockdown was carried out as previously described (25). Briefly, IMR-90 cells were transfected with FUBP1-specific Silencer siRNA (catalog number s16967; Life Technologies) by using SilentFect reagent (Bio-Rad) according to the manufacturer's specifications, using a 10 nM final siRNA concentration. Silencer Select negative-control siRNA 1 (Life Technologies) was used as the negative siRNA control.

Statistical analysis.

Statistical analyses were performed as previously described (54). Student's independent-sample t test was conducted for ChIP and quantitative PCR (qPCR) assays. P values were one tailed, and values of ≤0.05 were considered statistically significant. n in the figure legends indicates the number of biological replicates used for statistical analysis.

Transfections.

Cells were plated in 10-cm plates at a density of 2.0 × 10⁶ cells/plate 24 h prior to transfection. Transfection mixtures were prepared by mixing 1 ml of serum-free DMEM, 10 μg of total plasmid DNA, and 20 μl of a 1-mg/ml solution of linear 25-kDa polyethylenimine reagent from Polysciences (catalog number 23966-2). This mixture was vortexed for 10 s and incubated at room temperature for 20 min. The complexes were then added to the cells, and the cells were incubated for 24 to 48 h.

Viral genome quantification.

HT1080 cells depleted for FUBP1 or treated with the control siRNA were lysed in lysis buffer (50 mM Tris [pH 8.1], 10 mM EDTA, and 1% SDS) on ice for 10 min. Lysates were sonicated briefly in a Covaris M220 focused ultrasonicator to break up cellular chromatin and subjected to digestion using proteinase K (NEB), according to the manufacturer's specifications. Following digestion, viral DNA was purified by using an EZ-10 gel extraction kit (Bio Basic). PCRs were carried out by using SYBR Select master mix for CFX (Applied Biosystems) according to manufacturer's directions by using 2% of total purified DNA as the template, using a CFX96 real-time PCR instrument (Bio-Rad). A standard curve for absolute quantification was generated by serially diluting the pXC1 plasmid containing the left end of the HAdV5 genome starting with a concentration of 1.0 × 10⁷ copies per reaction down to 1.0 copy per reaction. The primers used were the same as those used for expression analysis of the E1B region, the annealing temperature used was 60°C, and 40 cycles were run.

Viruses.

Viruses used in this study included the HAdV5 dl309 mutant (52) expressing wt E1A but deleted for much of the E3 region, which was generously donated by Joe Mymryk. The HAdV5 dl1520 mutant, carrying a mutation in E1B-55k, was generously donated by Patrick Hearing and was previously described (33). All viruses were amplified in low-passage-number 293 cells, and titers were also determined on these 293 cells prior to performing assays. All infections were carried out in serum-free medium for 1 h at an MOI of 10 unless otherwise specified in the figure legends.

Virus growth assay.

IMR-90 cells depleted for FUBP1 via siRNA or transfected with control siRNA were infected with HAdV5 dl309 at an MOI of 10 in serum-free medium. Virus was adsorbed for 1 h at 37°C under 5% CO₂. Virus titers were determined 24, 48, and 72 h after infection by plaque assays performed on 293 cells by serial dilution.