A Role for E1B-AP5 in ATR Signaling Pathways during Adenovirus Infection

Abstract

E1B-55K-associated protein 5 (E1B-AP5) is a cellular, heterogeneous nuclear ribonucleoprotein that is targeted by adenovirus (Ad) E1B-55K during infection. The function of E1B-AP5 during infection, however, remains largely unknown. Given the role of E1B-55K targets in the DNA damage response, we examined whether E1B-AP5 function was integral to these pathways. Here, we show a novel role for E1B-AP5 as a key regulator of ATR signaling pathways activated during Ad infection. E1B-AP5 is recruited to viral replication centers during infection, where it colocalizes with ATR-interacting protein (ATRIP) and the ATR substrate replication protein A 32 (RPA32). Indeed, E1B-AP5 associates with ATRIP and RPA complex component RPA70 in both uninfected and Ad-infected cells. Additionally, glutathione S-transferase pull-downs show that E1B-AP5 associates with RPA components RPA70 and RPA32 directly in vitro. E1B-AP5 is required for the ATR-dependent phosphorylation of RPA32 during infection and contributes to the Ad-induced phosphorylation of Smc1 and H2AX. In this regard, it is interesting that Ad5 and Ad12 differentially promote the phosphorylation of RPA32, Rad9, and Smc1 during infection such that Ad12 promotes a significant phosphorylation of RPA32 and Rad9, whereas Ad5 only weakly promotes RPA32 phosphorylation and does not induce Rad9 phosphorylation. These data suggest that Ad5 and Ad12 have evolved different strategies to regulate DNA damage signaling pathways during infection in order to promote viral replication. Taken together, our results define a role for E1B-AP5 in ATR signaling pathways activated during infection. This might have broader implications for the regulation of ATR activity during cellular DNA replication or in response to DNA damage.

The phosphoinositide 3-kinase-like kinase protein products of the ataxia telangiectasia (A-T) mutated gene (ATM) and the ATM-Rad3-related gene (ATR) function as key transducers of signals initiated in response to DNA damage (25, 29). ATM initiates pathways in response to agents, such as ionizing radiation, that cause double-strand breaks and controls apoptosis and G₁-S transition, intra-S, and G₂/M checkpoint responses through the selective, targeted phosphorylation of a number of key mediator and effector proteins such as MDC1, Chk2, BRCA1, and p53 (29). ATR, on the other hand, predominantly regulates pathways that control DNA replication during unperturbed S phase, at stalled replication forks initiated during S phase, or in response to genotoxic stress such as UV irradiation (26). ATR is recruited to replication sites or sites of damage through its cognate binding partner ATR-interacting protein (ATRIP) (10), which binds selectively to the replication protein A (RPA) complex, which in turn associates directly with single-stranded DNA (ssDNA) (39). Thus, ATR activation initiates pathways that prevent further origin firing, block cell cycle progression, stabilize stalled replication forks, and facilitate replication fork restart. ATR phosphorylates substrates such as Chk1, RPA32, Smc1, and Rad9 to effect its responses (25). The MRE11-Rad50-NBS1 (MRN) complex is recruited to sites of DNA damage, where it facilitates ATM and ATR activation and serves to transduce signals from ATM and ATR to downstream effector proteins (18). RPA is a heterotrimeric ssDNA binding protein complex that was first identified as being an essential cellular cofactor for simian virus 40 (SV40) DNA replication (35). RPA comprises three subunits, RPA70, RPA32, and RPA14, each possessing at least one oligonucleotide binding fold that is found in many ssDNA binding proteins (14). RPA binds numerous proteins involved in DNA metabolism; is required for chromosomal DNA replication, repair, and recombination; and is a substrate for both ATM and ATR kinases (14).

It is becoming increasingly apparent that in order to replicate their genomes efficiently in host cells, viruses employ numerous strategies to selectively activate and/or evade host cell genotoxic stress response pathways (21, 33). Thus, adenovirus (Ad) has evolved to bypass, or inactivate, host cell cycle checkpoints that would otherwise initiate cell cycle arrest or apoptotic programs in the infected cell (4, 21, 33). The large Ad E1B gene product, E1B-55K, functions in concert with the E4 gene product, E4orf6, to inactivate the DNA damage response, cell cycle arrest, and apoptotic signaling pathways during infection through the specific targeting of p53, the MRN component MRE11, and DNA ligase IV for 26S proteasome-dependent degradation by the cellular Cul5-elongin B/C-Rbx1 ubiquitin ligase complex (2, 6, 16, 27, 30). Through their ability to recruit the Cul5-containing E3 ubiquitin ligase complex, E1B-55K and E4orf6 also cooperate to promote late viral mRNA nuclear export (36). It has been demonstrated that as a consequence of E1B-55K/E4orf6-targeted MRE11 degradation, Ad5 avoids the activation of ATM-dependent signaling cascades that would otherwise recognize Ad linear double-stranded DNA ends as double-strand breaks and initiate DNA repair pathways resulting in Ad genome concatenation (30). Indeed, Ad5 E4 deletion viruses incapable of promoting the degradation of MRE11 activate ATM pathways and consequently promote Ad5 genome end-to-end concatemer formation through nonhomologous end joining (NHEJ) (30). Interestingly, the ATR substrates RPA32 and Chk1 are phosphorylated in response to Ad5 E4 deletion virus infection, suggesting that MRE11 degradation also prevents ATR activation during wild-type (wt) Ad5 infection (9). Other work suggests that Ad5 E4orf3 acts in concert with E1B-55K and E4orf6 to disrupt MRE11 function during infection by binding and recruiting MRE11 to promyelocytic leukemia-containing nuclear tracks and cytoplasmic aggresomes (1, 12). Ad avoids NHEJ-mediated genome concatemer formation by also promoting the E1B-55K/E4orf6-dependent degradation of DNA ligase IV, the protein required for the rejoining step in NHEJ (2). Given these few examples, it is apparent that Ad protein function is intimately involved in regulating specific DNA damage signaling pathways in order to facilitate viral replication.

The cellular protein Ad E1B-55K-associated protein 5 (E1B-AP5) was first identified as being an E1B-55K binding protein; E1B-55K binds E1B-AP5 in both Ad5-infected and Ad5-transformed cells (15). E1B-AP5 is a heterogeneous nuclear ribonucleoprotein (hnRNP) with significant sequence homology with hnRNP-U/scaffold attachment factor A (SAF-A) (15). Additionally, E1B-AP5 possesses a functional RGG box, binds both mRNA and ssDNA (19), and activates transcription through interactions with BRD7 (20). It was previously suggested that during viral infection, E1B-55K utilizes E1B-AP5 to promote the selective accumulation of viral transcripts in the cytoplasm while retaining cellular transcripts in the nucleus (15). E1B-55K also targets E1B-AP5 during cellular transformation, as the overexpression of E1B-AP5 reduces the E1A/E1B-mediated transformation of primary rat fibroblasts significantly (15). Recent work from our laboratory has shown that E1B-AP5 can function independently of E1B-55K and inhibit p53 transcriptional activity (3). Given the relationship between E1B-AP5 and p53 and the key role of E1B-55K in regulating the function of host DNA damage response/repair signaling pathways during infection, we resolved to determine whether E1B-AP5, akin to other E1B-55K binding proteins, functions specifically in DNA damage signaling pathways that are targeted during the Ad infectious life cycle.

In contrast to other E1B-55K binding proteins, we have shown that E1B-AP5 is not targeted for 26S proteasomal degradation during infection but rather that E1B-AP5 protein levels are elevated following Ad infection. Significantly, we demonstrate that E1B-AP5 is a key component of DNA damage signaling pathways that are initiated in response to Ad infection. Specifically, we have determined a novel role for E1B-AP5 as a key component of ATR signaling pathways that are activated during Ad infection. E1B-AP5 colocalizes with ATRIP and RPA at Ad5 and Ad12 replication centers and is required for the ATR-dependent phosphorylation of RPA32 and Smc1 during Ad12 infection. Consistent with its role in ATR signaling pathways, E1B-AP5 also participates in the ATR-dependent phosphorylation of RPA32 during Ad5 infection. Taken together, our studies identify E1B-AP5 as being a key regulator of ATR during Ad infection.

MATERIALS AND METHODS

Cells.

The human lung carcinoma cell line A549 was grown in HEPES-modified Dulbecco's modified Eagle's medium (DMEM) (Sigma-Aldrich) supplemented with 8% fetal calf serum (FCS) (Invitrogen) and 2 mM l-glutamine (Invitrogen). HeLa human cervical carcinoma cells and clonogenic HeLa cells expressing short hairpin RNAs targeting ATM (7) were grown in DMEM containing 2 mM glutamine supplemented with 8% FCS. All cells were kept at 37°C in a humidified 5% CO₂ atmosphere.

Viruses.

Adenoviridae used in this study were mastadenoviruses, wt Ad5 (subgroup C), wt Ad12 Huie (subgroup A), and Ad5 dl1520 (a large E1B-null adenovirus that has a termination codon at position 3 and an additional deletion from nucleotides 2496 to 3323, with a linker insertion generating a second stop codon at nucleotide 3336 introduced into its sequence). These viruses were described previously (32). Ad5 and Ad12 viruses were propagated on permissive human embryonic kidney 293 (HEK293) cells and human embryonic retinoblastoma 3 (HER3) cells, respectively, and titers were determined by plaque assay on HER911 and HER3 cells, respectively.

Plasmids.

Full-length E1B-AP5 cDNA (15) was cloned into pGEX for expression in BL21 (Stratagene). Plasmids p11d-RPA70, p11d-RPA32, and p3d-RPA14, competent for in vitro transcription/translation, were described previously (17).

Antibodies.

Anti-ATRIP, -Smc1, -Smc1-S966, and -RPA32-S4/8 antibodies were obtained from Bethyl Laboratories. Anti-ATM, -RPA70, and -RPA32 antibodies were obtained from Calbiochem; anti-MRE11 antibodies were obtained from Novus Biologicals; anti-ATR antibodies were obtained from Abcam; anti-phospho-Chk1 (S345) antibodies were obtained from Cell Signaling; anti-Chk1 antibodies were obtained from Santa Cruz; anti-SAF-B, -H2AX, and -γ-H2AX antibodies were obtained from Millipore; and anti-β-actin antibodies were obtained from Sigma. An anti-E1B-AP5 rabbit polyclonal antibody (960) was raised in house against full-length glutathione S-transferase (GST)-E1B-AP5. The monoclonal antibodies against Ad5 E1B-55K (2A6), Ad12 E1B-55K (XPH9), DNA binding protein (DBP) (B6-10), and p53 (DO-1) were all obtained as supernatant fluid from cultures of the appropriate hybridoma cell lines. The anti-Ad5 preterminal protein (pTP) pAb was a gift from Pieter van der Vliet. Secondary anti-mouse and anti-rabbit antibodies used for Western blotting were obtained from Dako. Secondary anti-mouse and anti-rabbit Alexa 488/546 antibodies used for immunofluorescence were obtained from Molecular Probes.

Ad infection and drug treatment.

Ad infection was carried out using subconfluent monolayers of cultured cells in Opti-MEM I reduced serum medium (Invitrogen). Cells were washed twice in phosphate-buffered saline (PBS) before the addition of the virus at the appropriate infectivity ratio. After 2 h at 37°C, medium containing virus was removed and replaced with DMEM supplemented with 8% FCS. Where appropriate, medium was supplemented with 5 mM caffeine (Sigma-Aldrich) to inhibit caffeine-sensitive kinases.

Immunoprecipitation.

Cells were washed twice in PBS, and whole-cell extracts were prepared in immunoprecipitation lysis buffer (150 mM NaCl, 1 mM EDTA, 50 mM Tris-HCl [pH 7.5], 1% Nonidet P-40). Lysates were homogenized and cleared by ultracentrifugation (40,000 × g for 30 min). Immunoprecipitation was carried out overnight at 4°C, typically with 5 μg anti-E1B-AP5 antiserum per 5 mg of cell lysate. Antigen-antibody complexes were isolated using protein G-Sepharose (Sigma-Aldrich). Immunoprecipitates were washed five times in immunoprecipitation lysis buffer, eluted in sample buffer (9 M urea, 50 mM Tris-HCl [pH 7.4], and 0.15 M β-mercaptoethanol-10% sodium dodecyl sulfate [SDS] [2:1, vol/vol] containing 0.1% [wt/vol] bromophenol blue), and separated by SDS-polyacrylamide gel electrophoresis (PAGE).

SDS-PAGE and Western blot analysis.

Cells were washed twice in PBS, and whole-cell extracts were prepared in lysis buffer containing 9 M urea, 50 mM Tris-HCl (pH 7.4), and 0.15 M β-mercaptoethanol. Samples were subsequently sonicated and cleared by centrifugation. Protein concentrations were determined by use of a Bradford assay (Bio-Rad). Fifty-microgram protein samples and immunoprecipitates were separated on 12% polyacrylamide gels run in the presence of 0.1 M Tris, 0.1 M Bicine, and 0.1% SDS. Separated proteins were electroblotted onto nitrocellulose filters (Gelman Sciences) and incubated with the appropriate antibodies. Antigens were visualized by enhanced chemiluminescence (ECL) (Amersham Pharmacia).

GST fusion protein preparation.

pGEX-E1B-AP5 was used to transform Escherichia coli BL21 competent cells (Stratagene). After the induction of GST protein expression with IPTG (isopropyl-β-d-thiogalactopyranoside) (Sigma-Aldrich), bacteria were collected by centrifugation (6,000 × g for 10 min) and resuspended in GST lysis buffer (2 mM EDTA, 1% Triton X-100 in PBS) at 4°C. Bacterial lysates were sonicated and centrifuged (15,000 × g for 30 min). Supernatants were incubated with glutathione-agarose (Sigma-Aldrich) for 1 h with rotation at 4°C. Beads were washed three times in GST lysis buffer and twice in GST wash buffer (2 mM EDTA in PBS). GST-E1B-AP5 protein was eluted by incubating beads in GST elution buffer (25 mM glutathione, 50 mM Tris-HCl [pH 8.0]). Purified protein was dialyzed against a solution containing 150 mM NaCl, 1 mM dithiothreitol, 10% (vol/vol) glycerol, and 50 mM Tris-HCl (pH 7.2); aliquoted; and stored at −80°C.

In vitro GST-protein binding assays.

[³⁵S]methionine-labeled RPA subunits were expressed individually in vitro using a TNT T7 coupled reticulocyte lysate system (Promega) according to the manufacturer's guidelines. Typically, 10 μl of ³⁵S-labeled proteins was incubated with 10 μg GST-E1B-AP5 on ice for 1 h. Glutathione-agarose (Sigma) was used to isolate protein complexes. Beads were washed three times in GST lysis buffer and twice in GST wash buffer. Proteins were eluted by incubation in GST elution buffer before the addition of SDS-PAGE sample buffer. After separation by SDS-PAGE, ³⁵S-labeled proteins were visualized by fluorography and autoradiography.

Immunofluorescence.

Cells were grown on glass slides and infected with virus at the appropriate infectivity ratio. To better visualize Ad replication centers, at the indicated time points, cells were washed with PBS, treated with preextraction buffer [10 mM piperazine-N,N′-bis(2-ethanesulfonic acid) (PIPES) (pH 6.8), 20 mM NaCl, 300 mM sucrose, 3 mM MgCl₂, 0.5% Triton X-100] for 5 min, and subsequently fixed in 4% (wt/vol) paraformaldehyde in PBS for 10 min. Alternatively, cells were initially fixed in 4% (wt/vol) paraformaldehyde in PBS for 10 min and then permeabilized in ice-cold acetone at −20°C for 20 min. See the text and figure legends for more information. Cells were then rehydrated in PBS and blocked in blocking buffer (10% FCS in PBS) for 1 h. After washing in PBS, cells were incubated with primary antibodies in 0.1% FCS in PBS for 2 h. Slides were washed three times in PBS before incubation with secondary antibodies for 1 h. Cells were washed three times in PBS and then mounted in Vectashield mounting medium (Vector Laboratories) containing 4′,6′-diamidino-2-phenylindole (DAPI). Cells were observed and images were acquired using a Zeiss LSM510-Meta laser scanning confocal microscope and processed with associated software. In instances where images of single cells are presented, these images are representative of most cells, unless otherwise stated.

Transfection of siRNAs.

Small interfering RNA (siRNA) oligonucleotides targeting ATR and E1B-AP5 were purchased from Ambion, and their sequences were as follows: 5′-GGAAUAUAAUACAGUUGUATT-3′ (ATR) and 5′-GCAGUGGAACCAGUACUAUTT-3′ (E1B-AP5). As a negative control, AllStars negative control siRNA was purchased from Qiagen. Cells were plated 24 h prior to transfection so as to be 30% confluent the next day. Oligofectamine reagent (Invitrogen) was used to deliver siRNA duplexes into cells, according to the manufacturer's instructions. Transfections were repeated 24 h later, and cells were infected 48 h after the second dose of siRNA.

RESULTS

E1B-AP5 protein levels increase during Ad infection.

As the other known E1B-55K binding proteins, p53, MRE11, and DNA ligase IV, are targeted for ubiquitin-dependent, proteasome-mediated degradation during Ad5 infection, we initially investigated the effect of Ad5 and Ad12 infection upon p53, MRE11, and E1B-AP5 protein levels in A549 cells. Consistent with previous reports, MRE11 and p53 protein levels were reduced dramatically following the infection of A549 cells with both Ad5 (Fig. 1Aii and iii) and Ad12 (Fig. 1Bii and iii). In contrast to p53 and MRE11, E1B-AP5 proteins levels were increased (Fig. 1Ai and Bi). E1B-AP5 protein levels mirrored the observed increases in levels of expression of Ad5- and Ad12-E1A (data not shown) and E1B-55K (Fig. 1Aiv and Biv). β-Actin was used to monitor loading efficiency in each instance (Fig. 1Av and 1Bv). To substantiate these findings, we also determined E1B-AP5 levels in Ad-infected cells by immunofluorescent confocal microscopy (Fig. 1C). In order to retain all of the nuclear E1B-AP5, cells were not preextracted prior to immunostaining in this instance. Consistent with the Western blot data, E1B-AP5 levels were elevated substantially in both Ad5- and Ad12-infected A549 cells (Fig. 1Ci and iv, respectively). Cells were costained with antibodies to Ad5 DBP and Ad12 E1B-55K in order to distinguish Ad-infected from noninfected cells (Fig. 1Cii and v, respectively). Collectively, these data indicate that E1B-AP5 is not targeted for proteasomal degradation during Ad infection and suggest that it might perform discrete functions during infection.

FIG. 1.

Effects of Ad infection on protein levels of E1B-55K binding proteins. A549 cells were infected with 10 PFU per cell of either wt Ad5 (A) or wt Ad12 (B). Cells were harvested at the appropriate times postinfection, and 50 μg protein samples was separated by SDS-PAGE and transferred onto nitrocellulose. Membranes were then Western blotted for E1B-AP5 (i), MRE11 (ii), p53 (iii), E1B-55K (iv), and β-actin (v) using the appropriate antibodies and visualized by enhanced chemiluminescence. (C) Immunofluorescent Zeiss LSM510-Meta confocal microscopic detection of E1B-AP5 protein expression in wt Ad5- and wt Ad12-infected A549 cells. Cells were infected with either 10 PFU/cell wt Ad5 or wt Ad12, subsequently fixed with 4% (wt/vol) paraformaldehyde in PBS, and permeabilized with acetone 24 h postinfection. E1B-AP5 (i and iv), Ad5 DBP (ii), and Ad12 E1B-55K (v) were then visualized using the appropriate antibodies. Nuclei are stained with DAPI and are shown in blue.

E1B-AP5 localizes to Ad replication centers during infection.

In order to determine possible functions for E1B-AP5 during infection, we first investigated, by confocal microscopy, the subcellular localization of the E1B-AP5 protein in both noninfected and Ad-infected A549 cells. In noninfected cells, E1B-AP5 was localized exclusively in the nucleus but was excluded from the nucleoli (Fig. 2Ai). E1B-AP5 had a distinctive granular staining pattern and colocalized to some extent with RPA32 (Fig. 2Aiii). The RPA complex binding protein ATRIP similarly displayed a granular staining pattern that overlapped partially with RPA32 (Fig. 2Aiv and vi). In Ad5-infected cells, we initially visualized replication centers by costaining with Ad pTP and DBP (Fig. 2Bi to iii). Consistent with previous studies (31), the RPA component RPA32 was redistributed to viral replication centers during Ad5 infection, where it colocalized to some extent with pTP (Fig. 2Biv to vi). Significantly, E1B-AP5 also relocalized to replication centers in Ad5-infected A549 cells, where it colocalized with RPA32 (Fig. 3Ai to iii). ATRIP was also redistributed to viral replication centers during Ad5 infection, where it colocalized with RPA32 (Fig. 3Aiv to vi) and, presumably, E1B-AP5. Consistent with the Ad5 studies, E1B-AP5 was also redistributed to Ad12 replication centers during infection, where it colocalized with RPA32 (Fig. 3Bi to iii); ATRIP was also redistributed to Ad12 replication centers during infection and, akin to the Ad5 studies, colocalized with RPA32 (Fig. 3Biv to vi).

FIG. 2.

(A) Localization of E1B-AP5, ATRIP, and RPA32 in mock-infected interphase A549 cells. Cells were grown on glass coverslips and then treated with preextraction buffer and fixed with 4% (wt/vol) paraformaldehyde as described in Materials and Methods. Antigens were detected using the appropriate antibodies. (B) Localization of pTP, DBP, and RPA32 at viral replication centers in Ad5-infected A549 cells. Twenty-four hours postinfection with 10 PFU/cell wt Ad5, cells were treated with preextraction buffer and then fixed with 4% (wt/vol) paraformaldehyde, whereupon antigens were detected using the appropriate reagents. Colocalization images were recorded using a Zeiss LSM510-Meta laser scanning confocal microscope. Nuclei are stained with DAPI and are shown in blue.

FIG. 3.

Colocalization of E1B-AP5 and RPA32, and ATRIP and RPA32, at Ad5 (A) and Ad12 (B) replication centers. Cells were infected with 10 PFU/cell of the appropriate virus. Ad5-infected and Ad12-infected cells were treated with preextraction buffer and then fixed with 4% (wt/vol) paraformaldehyde 24 h and 48 h postinfection, respectively. E1B-AP5, RPA32, and ATRIP localizations were visualized using the appropriate antibodies. Colocalization images were recorded using a Zeiss LSM510-Meta confocal microscope. Regions of substantial colocalization are shown as yellow in the merged image. Nuclei are stained with DAPI and are shown in blue.

E1B-AP5 recruitment to viral replication centers is not dependent upon E1B-55K.

Given the established relationship between E1B-AP5 and Ad E1B-55K, we examined the extent to which E1B-AP5 colocalized with E1B-55K in Ad-infected cells. Results from these experiments revealed that there was substantial E1B-AP5/E1B-55K colocalization at replication centers in both Ad5- and Ad12-infected A549 cells (Fig. 4Ai to iii and iv to vi, respectively). Interestingly, no E1B-AP5 staining was observed in Ad5 E1B-55K-containing cytoplasmic bodies (Fig. 4Ai to iii). In Ad12-infected cells, which were at a later stage of the infectious life cycle, a large proportion of the E1B-55K was found at the nuclear periphery and in cytoplasmic bodies and did not colocalize with E1B-AP5 (Fig. 4Avii to ix). To determine whether E1B-55K expression was necessary for E1B-AP5 recruitment to replication centers, we infected A549 cells with the Ad5 E1B-55K-deleted virus dl1520. E1B-AP5 retained the capacity to localize with DBP (Fig. 4Bi to iii) and RPA32 (Fig. 4Biv to vi) at replication centers in Ad5 dl1520-infected cells, demonstrating that E1B-55K is not required for the redistribution of E1B-AP5 that occurs during infection. ATRIP was also recruited to Ad5 replication sites independently of E1B-55K expression (Fig. 4Bvii to ix). Taken together, these experiments are important in establishing that E1B-AP5 is redistributed specifically to Ad replication centers independently of E1B-55K, where it colocalizes with E1B-55K, DBP, RPA32, and ATRIP, and moreover suggest that E1B-AP5 might have specific functions during Ad replication.

FIG. 4.

(A) Colocalization of E1B-AP5 with E1B-55K at Ad5 and Ad12 replication centers. Cells were infected with 10 PFU/cell of the appropriate virus. Infected cells were treated with preextraction buffer and then fixed with 4% (wt/vol) paraformaldehyde 24 h postinfection. E1B-AP5 and Ad E1B-55K were visualized using the appropriate antibodies. (B) Recruitment of E1B-AP5, RPA32, and ATRIP to Ad5 replication centers is independent of E1B-55K. Cells were infected with 10 PFU/cell of Ad5 dl1520 and treated with preextraction buffer and then fixed with 4% (wt/vol) paraformaldehyde 24 h postinfection. E1B-AP5, Ad5 DBP, RPA32, and ATRIP were visualized using the appropriate antibodies. Colocalization images were recorded using a Zeiss LSM510-Meta confocal microscope. Regions of substantial colocalization are shown as yellow in the merged image. Nuclei are stained with DAPI and are shown in blue.

E1B-AP5 binds ATRIP, RPA, and DBP in vivo.

As we determined that E1B-AP5 colocalizes with ATRIP, RPA32, and DBP at Ad replication centers during infection, we next investigated whether we could coimmunoprecipitate E1B-AP5 with these proteins from uninfected or Ad5-infected cells. Thus, at appropriate times following mock or Ad5 infection, E1B-AP5 complexes were immunoprecipitated, and Western blot analysis was used to determine whether ATRIP, RPA, or DBP was found to be associated with E1B-AP5 in vivo. Significantly, both ATRIP and RPA component RPA70 were found in E1B-AP5 complexes both prior to and following Ad5 infection (Fig. 5Ai); infection did not alter the ability of these proteins to associate with E1B-AP5. The validity of these protein-protein interactions was strengthened by the observations that immunoprecipitation of the hnRNP SAF-B did not coprecipitate RPA70 or ATRIP and that anti-E1B-AP5 antibodies were unable to coprecipitate ATM (Fig. 5Aii). Consistent with the microscopy studies, we also determined that E1B-AP5 associated with DBP during Ad5 infection (Fig. 5B). GST pull-down analysis revealed that E1B-AP5 also bound directly to the RPA components RPA32 and RPA70; E1B-AP5 did not bind RPA14 directly (Fig. 5C). Taken together, these data show that E1B-AP5 associates with components of DNA damage checkpoint signaling pathways that bind cellular ssDNA and also binds to a viral protein that interacts with viral ssDNA.

FIG. 5.

(A and B) Association of E1B-AP5 with ATRIP, RPA70, and DBP in uninfected and Ad5-infected A549 cells. Asynchronously growing, mock-infected, and Ad5-infected (10 PFU/cell) A549 cells were harvested 24 h postinfection and subjected to immunoprecipitation with either anti-IgG, -SAF-B, or -E1B-AP5 antibodies. Immunoprecipitates were collected on protein G-Sepharose beads, eluted with the appropriate sample buffer, separated by SDS-PAGE, and transferred onto nitrocellulose. ATM, ATRIP, RPA70, and DBP were visualized by Western blotting with the appropriate antibodies. (C) Association of E1B-AP5 with the RPA components RPA32 and RPA70 in vitro. Ten micrograms of GST-E1B-AP5 fusion protein was incubated with 20 μl of l-α-[³⁵S]methionine-labeled RPA70, RPA32, or RPA14. Bound proteins were precipitated using glutathione-Sepharose and selectively eluted with glutathione (see Materials and Methods). Proteins were separated by SDS-PAGE and subjected to fluorography (Amersham Pharmacia). Gels were dried and then subjected to autoradiography.

Rad9 and Rad17 localize to Ad replication centers during infection.

As we determined that ATRIP and RPA32 redistributed to sites of Ad replication during infection, we next investigated whether other ATR-ATRIP and RPA binding proteins displayed similar localization patterns in noninfected and Ad-infected cells. Interestingly ATR-ATRIP and RPA binding proteins Rad9 and Rad17 both displayed a granular, nuclear staining pattern in noninfected cells (Fig. 6Ai to iii and iv to vi, respectively) and relocalized to Ad5 (Fig. 6Bi to iii and Ci to iii) and Ad12 (Fig. 6Biv to vi and 6Civ to vi) replication centers during infection. Akin to E1B-AP5, ATRIP, and RPA32, Rad9 and Rad17 were recruited to Ad5 replication centers independently of Ad5 E1B-55K (Fig. 6Di to vi). Taken together with the immunofluorescent images presented in Fig. 3, these data suggest that E1B-AP5, RPA, ATR-ATRIP, Rad9-Rad1-Hus1, and Rad17-RFC complexes are all recruited to Ad replication sites, suggesting that they might function in a common pathway during infection.

FIG. 6.

Colocalization of Rad9, Rad17, and RPA32 at Ad replication centers. A549 cells were either mock infected (A) or infected with 10 PFU/cell of wt Ad5 (Bi to iii and Ci to iii), wt Ad12 (Biv to vi and Civ to vi), or Ad5 dl1520 (Di to vi). Ad5- and Ad12-infected cells were treated with preextraction buffer and then fixed with 4% (wt/vol) paraformaldehyde at 24 h and 48 h postinfection, respectively. Rad9, Rad17, and RPA32 were visualized using the appropriate antibodies. Colocalization images were recorded using a Zeiss LSM510-Meta confocal microscope. Regions of substantial colocalization are shown in yellow in the merged image. Nuclei are stained with DAPI and are shown in blue.

Ad5 and Ad12 differentially regulate the phosphorylation of known ATR kinase substrates during infection.

Given that ATR/ATRIP and ATR kinase substrates RPA32 and Rad9 are recruited to viral replication centers during Ad5 and Ad12 infection, we wished to determine whether ATR kinase substrates are phosphorylated in response to Ad infection. To do this, we infected A549 cells with either wt Ad5 or wt Ad12 for various times and determined the phosphorylation statuses of a number of ATR substrates by Western blot analysis using phosphate-specific antibodies (Fig. 7A). Ad12 infection promoted the hyperphosphorylation of RPA32, while Ad5 did not (Fig. 7Ai). Phosphospecific antibodies raised against amino acids S4 and S8 of RPA32 confirmed that these residues were specifically phosphorylated in response to wt Ad12 infection (Fig. 7Aii). In contrast, even though Ad5 infection progresses more quickly than Ad12 infection, Ad5 promoted only weak phosphorylation of S4 and S8 of RPA32 (Fig. 7Aii). Rad9 was also phosphorylated in response to Ad12 infection but not Ad5 infection (Fig. 7Aiii). Intriguingly, ATR substrate Smc1 was phosphorylated at S966 in response to both Ad5 and Ad12 infection (Fig. 7Aiv). In contrast to the marked phosphorylation of Smc1 observed during Ad infection, neither Ad5 nor Ad12 promoted the phosphorylation of S345 of Chk1 (Fig. 7Avi). Both viruses, however, promoted the phosphorylation of histone variant H2AX to similar extents (Fig. 7Aviii), suggesting that both Ad5 and Ad12 induce damage response pathways during infection. Significantly, however, Ad5 and Ad12 differentially regulate the phosphorylation of a number of known ATR substrates, suggesting a selective requirement for these proteins during Ad5 and Ad12 infection.

FIG. 7.

(A) Ad5 and Ad12 differentially regulate the phosphorylation of RPA32, Rad9, and Smc1 during infection. A549 cells were infected with 10 PFU/cell of either wt Ad5 or wt Ad12. Cells were harvested at the appropriate times postinfection, and 50 μg protein samples was separated by SDS-PAGE. After electrophoretic transfer onto nitrocellulose, membranes were Western blotted for RPA32 (i), RPA32 S4/8 (ii), Rad9 (iii), Smc1-S966 (iv), Smc1 (v), Chk1-S345 (vi), Chk1 (vii), γ-H2AX (viii), and H2AX (ix) using the appropriate antibodies. (B) E1B-AP5 is required for Ad12-induced phosphorylation of RPA32. A549 cells were initially treated with either nonsilencing (non-sil.) siRNA or siRNA oligonucleotides specific for the E1B-AP5 gene. Cells were subsequently infected with either Ad5 or Ad12 (at 10 PFU/cell), and whole-cell lysates were prepared at the appropriate times postinfection. After SDS-PAGE and transfer onto nitrocellulose, membranes were probed for E1B-AP5 (i), RPA32 (ii), RPA32 S4/8 (iii), Smc1-S966 (iv), Smc1 (v), γ-H2AX (vi), and H2AX (vii) with the appropriate antibodies. Antigens were visualized by ECL.

Ad12-induced phosphorylation of RPA32 is dependent upon E1B-AP5.

As RPA32 is a major substrate for phosphorylation during infection (Fig. 7A) and E1B-AP5 colocalizes with RPA32 at Ad replication centers (Fig. 3), we wished to determine whether E1B-AP5 was required for RPA32 phosphorylation during Ad12 infection. We initially treated A549 cells with either nonsilencing siRNAs or siRNAs directed against E1B-AP5 (see Materials and Methods). Following siRNA-mediated knockdown, we infected cells with either Ad5 or Ad12 and determined the protein phosphorylation status by Western blot analysis. E1B-AP5 protein expression was reduced significantly following treatment with E1B-AP5-specific siRNAs (Fig. 7Bi). Intriguingly, Ad12-induced RPA32 phosphorylation was dependent upon E1B-AP5 expression, as levels of both the hyperphosphorylated form of RPA32 (Fig. 7Bii) and the S4/S8-phosphorylated form of RPA32 (Fig. 7Biii) were reduced considerably following E1B-AP5 knockdown. Indeed, the modest phosphorylation of S4/S8 RPA32 observed during Ad5 infection was also dependent upon E1B-AP5 expression (Fig. 7Biii). Interestingly, the Ad12-induced phosphorylation of Smc1-S966 was also reduced following E1B-AP5 knockdown (Fig. 7Biv) though not as dramatically as that observed with RPA32. Ad5-induced phosphorylation of Smc1-S966 was not dependent upon E1B-AP5 (Fig. 7Biv). E1B-AP5 knockdown had only a limited effect upon the Ad5- and Ad12-induced phosphorylation of H2AX (Fig. 7Bvi). Taken together, these data suggest that E1B-AP5 selectively participates in kinase signaling pathways initiated following infection and, more particularly, is integral to the function of kinases that target RPA32 during Ad infection.

Ad12-induced phosphorylation of RPA32 is independent of ATM.

As RPA32, Smc1, and H2AX are all substrates for phosphorylation during Ad5 and Ad12 infection, we wished to determine the kinase(s) responsible. To investigate a requirement for ATM in Ad-induced phosphorylation, we compared the abilities of Ad5 and Ad12 to promote phosphorylation in HeLa cells and clonally derived HeLa cells, where endogenous ATM expression is ablated through the constitutive expression of specific short hairpin RNAs (HeLa shATM).

The ablation of ATM expression had only moderate effects on the abilities of Ad5 (Fig. 8A) and Ad12 (Fig. 8B) to promote the phosphorylation of RPA32 S4/S8 and Smc1-S966. Thus, low levels of RPA32 S4/S8 phosphorylation were seen in both HeLa cells and HeLa shATM cells after Ad5 infection (Fig. 8Aiii); RPA32 S4/S8 phosphorylation was reduced modestly in HeLa shATM cells relative to that in HeLa cells. Ad5 also retained the ability to promote Smc1-S966 phosphorylation in HeLa shATM cells, although levels of Smc1 phosphorylation were lower, by approximately half, than those seen after Ad5 infection of HeLa cells (Fig. 8Aiv). Interestingly, the ability of Ad5 to promote H2AX phosphorylation was most severely affected by the ablation of ATM expression. Thus, H2AX phosphorylation levels in HeLa shATM cells were reduced to 30% of the levels seen following Ad5 infection of HeLa cells (Fig. 8Avi). Akin to the studies with Ad5, the ablation of ATM had only limited effects upon the ability of Ad12 to promote the hyperphosphorylation of RPA32 and the phosphorylation of RPA32 S4/S8 and Smc1-S966. Thus, significant RPA32 hyperphosphorylation and RPA32 S4/S8 phosphorylation were seen in both HeLa and HeLa shATM cells following Ad12 infection (Fig. 8Bii and iii). The ability of Ad12 to phosphorylate Smc1-S966 was enhanced to some extent in Ad12-infected HeLa shATM cells relative to HeLa cells (Fig. 8Biv). Like the Ad5 studies, the ability of Ad12 to promote H2AX phosphorylation was most severely affected by the ablation of ATM expression. Thus, H2AX phosphorylation levels in HeLa shATM cell were reduced to around 20% of the levels seen following Ad12 infection of HeLa cells (Fig. 8Bvi). Taken together, these experiments demonstrate that ATM has only a minor role in promoting the phosphorylation of RPA32 following Ad5 and Ad12 infection but has a greater role in promoting the phosphorylation of H2AX.

FIG. 8.

Role of ATM and other caffeine-sensitive kinases in Ad5 (A)- and Ad12 (B)-induced phosphorylation events. HeLa cells and HeLa shATM cells were infected with either Ad5 or Ad12 (at 10 PFU/cell), following which they were incubated in the absence or presence of 5 mM caffeine. Cells were harvested at the appropriate times postinfection, and 50 μg protein samples was separated by SDS-PAGE. After electrophoretic transfer onto nitrocellulose, membranes were Western blotted for ATM (i), RPA32 (ii), RPA32 S4/8 (iii), Smc1-S966 (iv), Smc1 (v), γ-H2AX (vi), and H2AX (vii) using the appropriate antibodies and visualized by ECL.

In order to assess the contribution of ATR and potentially other phosphoinositide 3-kinase-like kinases in phosphorylation events during the Ad life cycle, we treated a subset of HeLa cells and HeLa shATM cells with the phosphoinositide 3-kinase-like kinase inhibitor caffeine after infection. Significantly, the treatment of cells with caffeine had a pronounced effect upon the abilities of Ad5 and Ad12 to promote phosphorylation in both HeLa and HeLa shATM cells (Fig. 8A and B). Thus, caffeine inhibited the ability of Ad5 to promote the phosphorylation of RPA S4/S8 (Fig. 8Aiii) and Smc1-S966 (Fig. 8Aiv). Moreover, caffeine treatment almost completely negated the ability of Ad5 to promote H2AX phosphorylation (Fig. 8Avi). Crucially, caffeine had similar effects on the ability of Ad12 to promote the phosphorylation of RPA32 (Fig. 8Bii and iii), Smc1-S966 (Fig. 8Biv), and H2AX (Fig. 8Bvi). It is apparent from these studies that caffeine-sensitive protein kinases play a major role in directing the phosphorylation of RPA32, Smc1, and H2AX in Ad5- and Ad12-infected cells.

Ad12-induced phosphorylation of RPA32 is dependent upon ATR.

Given the role of caffeine-sensitive kinases in the Ad-dependent phosphorylation of a number of proteins during Ad5 and Ad12 infection and the observation that a number of these proteins are known substrates for ATR kinase, we next addressed the absolute requirement for ATR in the phosphorylation of these proteins. In order to do this, we compared the abilities of Ad5 and Ad12 to promote RPA32, Smc1, and H2AX phosphorylation in HeLa cells and HeLa shATM cells that had been treated prior to infection with either nonsilencing siRNA or siRNAs specific for ATR. Key to the interpretation of our results, our analyses revealed that ATR protein expression was almost completely ablated following the treatment of HeLa and HeLa shATM cells with ATR-specific siRNAs (Fig. 9Ai and Bi). Significantly, the knockdown of ATR gene expression precluded the ability of Ad12 to promote RPA32 hyperphosphorylation (Fig. 9Bii). Indeed, the ability of Ad12 to promote RPA32 S4/S8 phosphorylation was also reduced considerably following ATR knockdown (Fig. 9Biii). In line with above-described experiments (Fig. 7), the ability of Ad12 to promote RPA32 phosphorylation was, in this context, independent of ATM (Fig. 9Bii and iii). The ability of Ad5 to promote low levels of RPA32 S4/S8 phosphorylation was also dependent upon ATR expression (Fig. 9Aiii). Consistent with the notion that ATR phosphorylates a number of its known substrates during Ad infection, the abilities of Ad5 and Ad12 to promote Smc1-S966 phosphorylation, despite the Ad-induced Smc1-S966 phosphorylation not being particularly marked in this instance, were also largely dependent upon ATR expression (Fig. 9Aiv and Biv). Interestingly, the abilities of Ad5 and Ad12 to promote H2AX phosphorylation were somewhat reduced following ATR knockdown (Fig. 9Avi and Bvi), although ATM and other unidentified kinases also play a role in the phosphorylation of H2AX during infection (Fig. 9Avi and Bvi). In summary, these data suggest that during infection, both Ad5 and Ad12 differentially utilize ATR kinase activity to promote the phosphorylation of a number of ATR-regulated effectors.

FIG. 9.

Role of ATR in Ad5 (A)- and Ad12 (B)-induced phosphorylation events. HeLa cells and HeLa shATM cells were initially treated with either nonsilencing (non-sil.) siRNA or siRNA oligonucleotides specific for the ATR gene. Cells were subsequently infected with either Ad5 or Ad12 (at 10 PFU/cell). Cells were harvested at the appropriate times postinfection, and 50 μg protein samples was separated by SDS-PAGE. After electrophoretic transfer onto nitrocellulose, membranes were Western blotted for ATR (i), RPA32 (ii), RPA32 S4/8 (iii), Smc1-S966 (iv), Smc1 (v), γ-H2AX (vi), and H2AX (vii) using the appropriate antibodies and visualized by ECL.

DISCUSSION

Role for E1B-AP5 during infection.

We present evidence here to demonstrate that E1B-AP5 is recruited to Ad5 and Ad12 replication centers during infection (Fig. 3). Consistent with reports demonstrating the recruitment of ATR and RPA32 to Ad5 and Ad-associated virus replication centers (1, 9, 31), we demonstrate that ATRIP and RPA32 are also recruited to Ad5 and Ad12 replication centers during infection of A549 cells (Fig. 2 to 4). Significantly, we have shown that E1B-AP5 colocalizes with both RPA and ATRIP at Ad replication centers and is found to be associated with these proteins in vivo (Fig. 3 and 5). The ATR-ATRIP binding partners Rad9 and Rad17 are also recruited to Ad replication centers (Fig. 6). Presumably, the recruitment of ATR/ATRIP/RPA complexes to both Ad5 and Ad12 replication centers reflects the cellular response to the production of viral ssDNA intermediates formed during Ad DNA replication. Although the functions of E1B-AP5, ATR, RPA, and other ATR-binding proteins in Ad replication are not known, it is conceivable, given the known role of the RPA complex in binding ssDNA during cellular DNA replication and in response to DNA damage (5, 14) and its known ability to promote large T-antigen-dependent SV40 DNA replication (13, 35), that E1B-AP5-RPA complexes could aid Ad-DBP in its role during Ad DNA elongation, where it binds to displaced Ad ssDNA and enhances the processing abilities of Ad polymerase (11). In this regard, it will be of interest to determine, by chromatin immunoprecipitation and confocal microscopy, whether E1B-AP5 and/or RPA is associated with Ad ssDNA during infection, especially as data presented here indicate that E1B-AP5 is associated with DBP in Ad-infected cells (Fig. 5C).

Role for E1B-AP5 in the activation of ATR-dependent signaling pathways during infection.

We present compelling evidence to demonstrate that E1B-AP5 function is intimately linked to ATR activity during infection (Fig. 7 to 9). Taken together, these data indicate that the Ad E1B-55K/E4orf6-mediated degradation of MRE11 is not sufficient to inhibit ATR kinase activation during infection. Using siRNAs specific for E1B-AP5 and ATR, we have shown that E1B-AP5 regulates, to different degrees, the ATR-dependent phosphorylation of RPA32 during infection (Fig. 7 to 9). These observations suggest that E1B-AP5 functions directly to regulate ATR activation. Given E1B-AP5's ability to bind ATRIP, RPA, and DBP, it is possible that E1B-AP5 recruits functional ATR complexes to Ad DNA, where it promotes the selective phosphorylation of ATR substrates. As ATRIP, the Rad17-RFC complex, and the 9-1-1 complex have all been shown to be required for the activation of ATR signaling (10, 38), the observation that ATRIP, Rad17, and Rad9 are all recruited to replication centers (Fig. 3 and 6) supports our data suggesting a role for ATR activation during Ad infection. Interestingly, however, data presented here indicate that Ad5 and Ad12 employ different strategies to differentially regulate ATR kinase activity during infection. Specifically, Ad12 promotes an appreciable ATR-dependent phosphorylation of RPA32 and Smc1, while Ad5 promotes relatively weak ATR-dependent phosphorylation of RPA32 and Smc1 (Fig. 7 to 9). At present, it is unclear why Ad5 and Ad12 have evolved divergent mechanisms to regulate ATR, but data presented here suggest that Ad5 and Ad12 utilize distinct components of the ATR pathway to promote replication. Given that Ad5 E4orf3 and Ad12 E4orf3 differentially regulate MRN during infection (31), it is possible that Ad5 and Ad12 could similarly mediate the differential regulation of ATR. The most striking difference between Ad5 and Ad12 in this regard is their respective abilities to promote RPA32 phosphorylation; Ad12 promotes appreciable RPA32 phosphorylation, whereas Ad5 promotes only very weak RPA32 phosphorylation, which is detected only by using a phosphospecific RPA32 S4/8 antibody and is not detected with the RPA32 antibody that detects both hypo- and hyperphosphorylated RPA32 (Fig. 7). These observations are entirely consistent with the previous report by Carson and colleagues showing that wt Ad5 does not promote RPA32 hyperphosphorylation (9).

In considering the cellular function of RPA as an ssDNA binding protein, it was previously shown that RPA is phosphorylated during S phase by cyclin-dependent kinases in order to promote cellular DNA replication (24). Following DNA damage, however, RPA32 is phosphorylated at additional sites by a number of different kinases including ATM, ATR, and DNA-PK. RPA phosphorylation after DNA damage disrupts its association with cellular DNA replication centers, inhibiting DNA replication (24). At present, it is not clear why Ad12 promotes ATR-dependent RPA32 hyperphosphorylation during infection while Ad5 does not. Given that RPA hyperphosphorylation inhibits cellular DNA replication and that the DNA-PK-mediated hyperphosphorylation of RPA32 is required for SV40 DNA replication (8, 13, 35), it is possible that Ad12 needs to restrict cellular DNA replication for efficient viral DNA replication, whereas Ad5 does not. Interestingly, previous studies investigating the DNA damage response caused by Ad infection noted that RPA32 was hyperphosphorylated and that Chk1 was phosphorylated at S345 in response to infection with Ad5 E4 deletion viruses but not with wt Ad5; infection with Ad5 E4 deletion viruses promoted viral genome NHEJ-mediated concatenation (9). Consistent with those reports, we were unable to demonstrate Chk1 S345 phosphorylation during wt Ad5 or wt Ad12 infection (Fig. 7A). Given the E4 deletion studies, it would be interesting to determine whether E1B-AP5 is required for Ad5 E4 deletion virus induction of RPA32 or Chk1 phosphorylation and, moreover, whether Ad12 E4 deletion viruses similarly promote Chk1 S345 phosphorylation.

Role for E1B-AP5 as a component of ATR pathways during DNA damage.

From the work presented here and elsewhere, it is becoming increasingly apparent that hnRNPs, in addition to their roles in RNA metabolism, transcriptional regulation, and chromatin remodeling, can also function in DNA damage response/repair pathways. For instance, hnRNP K protein levels are stabilized through the inhibition of Mdm2-dependent proteasome-mediated degradation in an ATM- or ATR-dependent manner following treatment with ionizing or UV irradiation, respectively (23). In this regard, hnRNP K acts as a cofactor for p53 and promotes p53-dependent transcription in response to DNA damage (23). hnRNP A18 gene expression is enhanced in response to UV irradiation, where it promotes the stability of a number of stress- and UV-induced mRNA transcripts, such as RPA32, and enhances their translation; cells lacking A18 are more sensitive to UV irradiation (37). Consistent with those studies, we demonstrate here that E1B-AP5 is induced in response to Ad infection (Fig. 1) and subsequently participates selectively in ATR signaling pathways during infection. Other viruses also regulate ATR during infection to promote viral replication. For instance, HIV-1 viral protein R promotes the ATR-dependent phosphorylation of Chk1 in order to induce G₂ arrest in infected cells (28), while human cytomegalovirus promotes Chk1 phosphorylation and its association with ATRIP at replication centers (22). In contrast, herpes simplex virus type 1 sequesters virally induced hyperphosphorylated RPA32 from viral replication centers, while herpes simplex virus type 1 immediate-early protein ICP0 disrupts the ATR-ATRIP association in infected cells (34). Given these findings, it will be of interest to determine whether E1B-AP5 associates with, or is sequestered from, other viral replication centers by other viral proteins in order to regulate ATR-dependent signaling pathways during infection. Moreover, it will be extremely interesting to establish whether E1B-AP5 participates in RPA and ATR signaling pathways that control cellular DNA replication in noninfected cells and in response to DNA damage. Furthermore, it will be important to determine whether other hnRNPs participate in the Ad-mediated regulation of DNA damage response/repair pathways during infection. These areas of investigation will be a major focus of our laboratory in the future.