![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Department of Genetics, Franklin College of Arts and Sciences, University of Georgia, Athens, Georgia,1 Molecular Biology Institute and Department of Microbiology, Immunology and Molecular Genetics, University of California, Los Angeles, California,2 Department of Microbiology and Immunology, University of Michigan Medical School, Ann Arbor, Michigan3
Received 13 May 2004/ Accepted 20 July 2004
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Mouse adenovirus type 1 (MAV-1) specifically targets endothelial cells and cells of the monocyte-macrophage lineage (34, 52), whereas epithelial cells are the main cell type for human adenovirus infection (31). Unlike human adenovirus, MAV-1 only produces a single 200-amino-acid E1A protein (4, 53). Although MAV-1 E1A is not an essential gene for viral replication in cell culture (61), it is a virulence gene in both outbred and inbred mice (52, 55), emphasizing the importance of MAV-1 E1A for viral pathogenesis in vivo. Since human adenovirus E1A itself does not bind to DNA (51), direct and/or indirect association with cellular proteins is essential for E1A to carry out its multiple functions during viral infection.
The function of MAV-1 E1A in viral pathogenesis is undoubtedly related to binding to cellular proteins. MAV-1 E1A contains all three conserved regions (CR1, CR2, and CR3) (4, 6) found in the human adenovirus 289R E1A protein (42). However, the C terminus of MAV-1 E1A, lacking the conserved motif PXDLS, is completely different from that of human adenovirus E1A (3), and thus it is unlikely to bind to a mouse homologue of CtBP. Like human adenovirus E1A, MAV-1 E1A interacts with pRb and p107, mainly through the CR1 domain (53). MAV-1 E1A is also similar to human adenovirus E1A in that it plays a major role in antagonizing the antiviral effects of interferons (35). However, all three conserved regions (CR1, CR2, and CR3) are required for MAV-1 resistance to the interferon responses, whereas only CR1 of human adenovirus E1A is needed for such resistance (1).
Because of these differences between MAV-1 and human adenovirus E1A and to broaden our understanding of the roles of E1A in viral pathogenesis, we looked for cellular proteins that interact with MAV-1 E1A. We used an unbiased experimental approach, glutathione S-transferase (GST) pulldown coupled with mass spectrometry analysis. In the work presented here, we identified mouse Sur2 as a protein that interacts with MAV-1 E1A.
Human Sur2, a subunit of Mediator, binds to CR3 of human adenovirus E1A (13). Mediator complexes function as molecular bridges to link the transcriptional regulators with RNA polymerase II to regulate transcription (11). The interaction between Sur2 and human adenovirus E1A-CR3 is required for binding Mediator and for transcriptional activation by human adenovirus E1A (13, 56), suggesting that recruitment of the Mediator complex transactivates the transcription of other viral early genes. Even though human Sur2 interacts with human adenovirus E1A in virus-infected cells (57), the importance of Sur2 in adenovirus replication has not been directly tested because Sur2–/– human cells have not been available. In the work reported here, we demonstrated that mouse Sur2 is a critical factor for MAV-1 replication in mouse embryonic fibroblasts (MEFs). Our findings indicate that Sur2 is important for efficient viral replication, providing a new molecular basis to investigate viral pathogenesis in mice.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Sur2–/– and Sur2+/+ MEFs were obtained as follows. Sur2 knockout mice were generated by injecting Sur2+/– embryonic stem (ES) cell clones (56) into C57BL/6 blastocysts. Sur2 heterozygous mice (129SvEV x C57BL/6) were intercrossed to generate Sur2–/– embryos. Generation and characterization of Sur2 knockout mice will be described in greater detail elsewhere (J. Stevens and A. Berk, unpublished data). Sur2+/+ and Sur2–/– MEFs were generated from individual embryonic day 9.5 (E9.5) littermate embryos. Embryos were dissociated by trypsin digestion and plated in single wells of a 24-well plate containing ES growth medium (15% fetal bovine serum [HyClone], 2 mM L-glutamine [Invitrogen], 50 µg of penicillin-streptomycin [Invitrogen], 0.1 mM nonessential amino acids [Cellgro], 10–4 M ß-mercaptoethanol, 1,000 U of leukemia inhibitory factor [Chemicon] per ml). MEFs were passaged by a 3T3-like protocol until the cells escaped replicative senescence. They were then maintained in Dulbecco's modified Eagle's medium containing 10% fetal calf serum.
Wild-type MAV-1 was the standard MAV-1 stock originally obtained from S. Larson (5). pmE109 is an MAV-1 E1A null mutant virus; the dlE105, dlE102, and dlE106 viruses are MAV-1 E1A CR1, CR2, and CR3 deletion mutants, respectively (53). Mouse gammaherpesvirus 68 was obtained from Jason Weinberg, University of Michigan.
Plasmid constructions. A plasmid, pAS2E1A, containing the full-length MAV-1 E1A cDNA fragment from plasmid Z112.F (4) was generated through several intermediate vectors with an adaptor (5'-GACATGCTCATGAGCATGTCGC-3'). pAS2E1A has the full-length E1A cDNA fragment plus an additional 66 nucleotides (5'-GAATTCATGG CTTACCCATA CGATGTTCCA GATTACGCTA GCTTGGGTGG TCATATGGCC ATGAGC-3') at the 5' end of the E1A cDNA sequence. The E1A cDNA and linker were gel purified from EcoRI-digested pAS2E1A and ligated to EcoRI-digested vector pGEX-4T-1 (Amersham Biosciences) to generate plasmid pGST-mE1A. The correct open reading frame of pGST-mE1A was confirmed by DNA sequencing.
Plasmids pGST-wtE1A, pGST-Nter1, pGST-Nter2, pGST-Cter1, pGST-Cter2, and pGST-Cter3 were made by PCR with primers containing a 5' EcoRI site and a 3' SalI site to amplify MAV-1 E1A fragments encompassing full-length protein or residues 1 to 45, 1 to 113, 157 to 200, 125 to 200, and 90 to 200, respectively, from plasmid pCME1A (53) as a template (Fig. 1). The same primers as used in pGST-wtE1A cloning were used to construct pGST-CR1
(deletion of amino acids 36 to 77 of E1A), pGST-CR2 (deletion of amino acids 112 to 128) and pGST-CR3 (deletion of amino acids 136 to 153) by PCR with pCMV-CR1, pCMV-CR2, and pCMV-CR3 as templates, respectively (53) (Fig. 1). The reaction mix contained the 5' primer (1 ng/µl), 3' primer (1 ng/µl), plasmid template (1 ng/µl), 1.25 units of Pfu polymerase (Stratagene), 1x Pfu reaction buffer, and 200 nM each dATP, dCTP, dGTP, and dTTP in 100 µl. After 30 cycles (melting at 94°C for 30 s, annealing at 58°C for 30 s, and extension at 72°C for 45 s), the PCRs were subjected to an additional 7-min incubation at 72°C (GeneAmplifer 9600, Applied Biosystems). Each PCR yielded a single product that was subsequently purified with a PCR purification kit (Qiagen), digested with EcoRI and SalI, gel purified, and ligated to pGEX-4T-1 that had been digested with EcoRI and SalI. pGST-E3gp11K contains the full-length MAV-1 E3 gp11K gene (8).
|
Pellets were thawed on ice and resuspended (for 200 ml of original liquid culture) in 10 ml of ice-cold STE buffer (10 mM Tris-HCl, 1 mM EDTA, 150 mM NaCl) with 100 µl of lysozyme solution (30 mg/ml), 1 µl of 100 mM phenylmethylsulfonyl fluoride, and 2 µl of protease inhibitor cocktail (Sigma). Samples were incubated on ice for 15 min before addition of 100 µl of 1 M dithiothreitol and 1.4 ml of 10% Sarkosyl. Samples were sonicated for a total time of 30 s and then centrifuged at 15,000 x g for 30 min to pellet debris. Supernatants were transferred to a fresh 50-ml tube, and 4 ml of 10% Triton X-100 was added. The samples were diluted with STE buffer to a final 20-ml volume and incubated at room temperature for 30 min. The GST fusion proteins were mixed gently overnight at 4°C with a 1-ml bed of prepared glutathione-Sepharose 4B (Pharmacia Biotech) in phosphate-buffered saline. The beads were washed four times with 25 ml of ice-cold phosphate-buffered saline. The GST fusion proteins were eluted with three successive 1-ml volumes of elution buffer (50 mM Tris-HCl [pH 8.0], 10 mM reduced glutathione, 0.1% Sarkosyl). The eluates were pooled, and the reduced glutathione and Sarkosyl were removed by overnight dialysis against phosphate-buffered saline (pH 7.4). The purified GST fusion proteins were stored at –70°C until use.
Large-scale GST pulldown assays. Equal amounts of purified GST fusion proteins or GST proteins were mixed with 100 µl of glutathione beads and incubated for 2 h at 4°C. Nuclear extraction of 109 3T6 cells or 5 x 108 MBMECs was performed according to the method of Dignam et al. (22); 15 ml of the 3T6 or MBMEC nuclear extracts was preabsorbed against 266 µl of glutathione beads and then with 500 µl of GST-glutathione beads (loaded with 250 µg of GST) at 4°C for 4 h in GST binding buffer (125 mM NaCl, 50 mM Tris-HCl [pH 7.4], 0.1% NP-40). Equal aliquots of the preabsorbed nuclear extracts were then added to glutathione beads bound to 25 µg of GST or GST-mE1A and rocked overnight at 4°C. The beads were washed twice with GST binding buffer and twice with GST wash buffer (250 mM NaCl, 50 mM Tris-HCl [pH 7.4], 0.1% NP-40). The beads were washed once again with GST binding buffer just before they were eluted with 30 µl of 2x sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer (37) and boiled for 15 min. The bound proteins were separated by SDS-10% PAGE and either Coomassie stained for mass spectrometry analysis or transferred to polyvinylidene difluoride membranes for immunoblotting.
Mass spectrometry analysis. Bands of interest were cut out of the Coomassie-stained protein gels, digested in-gel with trypsin, and used for tandem mass spectrometry-mass spectrometry analysis with a matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) machine (Micromass Company) by the Michigan Proteomic Consortium at the University of Michigan. Two search engines, MS-FIT (http://prospector.ucsf.edu) and Mascot (http://www.matrixscience.com), were used to search databases to identify the proteins.
Antibodies. Monoclonal antibody against human Sur2 (BD Pharmingen) cross-reacts with mouse Sur2 and was used at a 1:1,000 dilution in Western blot assays. Anti-p107 (sc-318) and anti-p130 (sc-317) polyclonal antibodies were from Santa Cruz Biotechnology. Anti-murine Rb and anti-MAV-1 E1A (AKO7-147) rabbit polyclonal antibodies were described (53). AKO7-147 and normal rabbit serum were purified by DEAE Affi-Gel blue (Bio-Rad) affinity chromatography according to the manufacturer's instructions. Normal rabbit serum was used as a control where indicated. Mouse monoclonal antibodies against MAV-1 E1A (monoclonal antibody10B10) and MAV-1 E3gp11K (monoclonal antibody 11H9) were generated at the University of Georgia Monoclonal Facility with purified GST-mE1A and GST-gp11K fusion proteins as the antigen, respectively. Both monoclonal antibody 10B10 and monoclonal antibody 11H9 were purified with an Affi-Gel protein A MAPS II kit (Bio-Rad) according to the manufacturer's instructions; a 1:500 dilution of monoclonal antibody 10B10 or 1:1,000 dilution of monoclonal antibody 11H9 was used in Western blots. Monoclonal antibody against ß-actin was from Sigma (A5441) and used at a 1:5,000 dilution in Western blots.
Western blot assays. Samples were dissolved in SDS-PAGE sample buffer and resolved on SDS-polyacrylamide gels. The proteins were electrotransferred onto polyvinylidene difluoride membranes at 14 V for 16 h at 4°C, and the membranes were blocked by incubation at room temperature for 1 h in TBS-Tween (150 mM NaCl, 10 mM Tris [pH 7.4], 0.1% Tween 20) containing 5% nonfat dry milk. Anti-rabbit (1:12,000 from Amersham Biosciences) or anti-mouse (1:12,000 from Amersham Biosciences) immunoglobulin G-horseradish peroxidase-conjugated antibody was used as the secondary antibody. Immunoblots were developed with SuperSignal West Pico chemiluminescent substrate (Pierce Biotechnology, Inc.).
Immunoprecipitations. 3T6 cells or MBMECs were lysed in E1A lysis buffer (250 mM NaCl, 50 mM Tris [pH 7.4], 0.1% NP-40) containing protease inhibitor cocktail (1:50) (Sigma). The lysates were diluted in E1A binding buffer (125 mM NaCl, 50 mM Tris [pH 7.4], 0.1% NP-40) and then preabsorbed simultaneously against rabbit normal immunoglobulin G (IgG) and protein A-agarose beads (Pharmacia Biotech) by rocking at 4°C for 2 h. The lysates were then incubated with polyclonal antibody against MAV-1 E1A (AKO7-147) (53) or rabbit normal IgG, followed by incubation with 30 µl of protein A-agarose (Oncogene Research Products). The agarose beads were pelleted, washed twice with E1A binding buffer and twice with E1A lysis buffer, and once again with E1A binding buffer. The immunoprecipitated proteins were eluted by boiling in SDS-PAGE sample buffer and subjected to SDS-PAGE and immunoblotting.
Viral growth curves. Sur2+/+ and Sur2–/– MEFs were infected with wild-type MAV-1 at a multiplicity of infection (MOI) of 0.05, 0.1, 1, or 5 or with mouse gammaherpesvirus 68 at an MOI of 0.01. MAV-1 plaque assays were carried out on 3T6 cells as described previously (18). Briefly, cells were harvested at various times post infection by scraping in their medium. The cell suspensions were subjected to three cycles of freezing and thawing, and the cell debris was spun out of the supernatant. Tenfold serial dilutions of supernatants were plated in triplicate on 3T6 cells, and plaques were counted at day 9 after plating. For the mouse gammaherpesvirus 68 plaque assay, only supernatants were harvested, and 10-fold serial dilutions of supernatants were plated in duplicate on NIH 3T3 cells as described (47).
Southern blots. Sur2+/+ and Sur2–/– MEFs were infected at an MOI of 0.05, 0.1, 1, or 5 and harvested at various times post infection by scraping the cells off the plates. Viral DNA was isolated by the method of Hirt (28). Equal amounts of DNA samples were digested with HindIII and RNase A and electrophoresed on a 0.7% agarose gel. After staining with ethidium bromide and photography, the gel was soaked in 0.8 M NaCl-0.4 M NaOH for 30 min, rinsed with water, and then soaked in 1.5 M NaCl-0.5 M Tris (pH 7.5) for 30 min. The DNA was transferred to a positively charged nylon membrane (Boehringer) by capillary transfer. The DNA was UV cross-linked to the membrane by incubation for 12 s in a Fisher UV cross-linker.
The membrane was prehybridized in 10 ml of PerfectHyb solution (Sigma) at 65°C for 2 h. Plasmids pMBA, pMBB, and pMBC contain MAV-1 genome sequence from 31.7 to 64.2 map units, 64.2 to 94.5 map units, and 6.5 to 31.7 map units, respectively. They were pooled, digested with HinfI, and labeled with Klenow DNA polymerase with [
-32P]ATP. The free label was removed by passage over a G25 Sepharose spin column, and 106 to 107 cpm of the probe was added to 10 ml of PerfectHyb solution to hybridize at 65°C for 5 to 18 h. The membranes were washed twice with 2x SSC (1x SSC is 0.15 M NaCl plus 0.015 M sodium citrate)-0.1% SDS at room temperature for 15 min, twice with 2x SSC-0.1% SDS at 65°C for 30 min, and once with 0.5x SSC-0.1% SDS at 65°C for 15 min. A phosphorimager was used to quantify the signals.
RNase protection assays.
Sur2+/+ and Sur2–/– MEFs were infected at an MOI of 0.05, 0.1, 1, or 5. Total RNA was extracted with TRI Reagent (Molecular Research Center, Inc.) following the manufacturer's instructions. An equimolar pool of linearized plasmids were used as templates to make an [
-32P]UTP-labeled multiplex RNase protection assay probe set by T7 or T3 polymerase transcription. prK+7, a genomic E1A plasmid containing MAV-1 genome sequence nucleotides 1 to 820 in a pBluescript(+) vector, was digested with BamHI at nucleotide 360 in the MAV-1 sequence (GenBank accession no. NC_000942). The full-length probe was 489 nucleotides (MAV-1 nucleotides 360 to 820 plus 29 nucleotides of vector), and the length protected from RNase digestion after hybridization to an E1A mRNA was 392 nucleotides.
A genomic hexon plasmid, pHEX, was constructed by ligation of MAV-1 genome sequence nucleotides 16432 to 16769 with vector pBS2SK–. pHEX was digested at nucleotide 16432 in the viral sequence at a BamHI site. The full-length probe was 395 nucleotides (MAV-1 nucleotides 16432 to 16769 plus 58 nucleotides of vector) and the protected size was 337 nucleotides. The protected length of L32 was 80 nucleotides and it was used as an internal loading control (29). pZU14 is an E3gp11K cDNA clone used to make an E3 probe (7). The full-length probe is 714 nucleotides, and the protected E3 signal is 645 nucleotides. pH61 (MAV-1 E2A clone) (61) was digested with SmaI, and the full-length probe is 470 nucleotides. The protected E2A signal is about 440 nucleotides. pZ571 is an E4 cDNA clone (36) and was digested with BamHI. The full-length probe is 767 nucleotides, and the protected E4 signal is about 700 nucleotides. The linearized plasmid mouse ß-actin was from the Maxiscript in vitro transcription kit (Ambion Inc.). The full-length probe of ß-actin is 276 nucleotides with T3 transcriptase, and the protected length is 245 nucleotides.
The RNase protection assays were carried out as described by Hobbs et al. (29). Briefly, 7.5 µg of total RNA was hybridized with the combined MAV-1 probe set and either the L32 or ß-actin probe overnight at 56°C. Saccharomyces cerevisiae tRNA was used as a negative control in RNase protection assays. After digestion with RNase A and T1, samples were ethanol precipitated and electrophoresed on 5% polyacrylamide-8 M urea gels. After drying, protected mRNA signals were visualized with a phosphorimager, and quantitation was performed by normalizing the mRNA species of interest to L32 or ß-actin signals.
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Differences in posttranslational modifications between proteins expressed in bacterial versus mammalian cells might affect the interactions of MAV-1 E1A with cellular proteins. We previously showed that MAV-1 E1A binds to mouse pRb and p107 proteins in an in vitro mixing experiment (53). We reasoned that if the GST-mE1A fusion protein is capable of binding to these known interacting proteins, GST pulldown assays with the GST-mE1A fusion protein would be a valid experimental approach to screen for cellular proteins that interact with MAV-1 E1A. Since the pRb and p107 used in those experiments were in vitro-translated proteins, we first further verified that MAV-1 E1A interacts with endogenous Rb family proteins in virus-infected cells.
MBMECs were infected with wild-type MAV-1 at an MOI of 5 and harvested at 40 h postinfection. Mock and pmE109 (E1A null mutant) infections were used as negative controls. Coimmunoprecipitation experiments were carried out with antibodies to the Rb family proteins to show the protein-protein interactions. Western blots of the immunoprecipitates were probed with antibody against MAV-1 E1A. As expected, MAV-1 E1A signals were only detected in wild-type MAV-1-infected samples that had been immunoprecipitated with specific antibodies against p130, p107, or pRb (Fig. 2A, lanes 5, 8, and 11) and not with control normal rabbit serum (lane 2), demonstrating that MAV-1 E1A binds to Rb family proteins in infected cells. Coimmunoprecipitation experiments performed by immunoprecipitation with anti-MAV-1 E1A antibodies (rabbit polyclonal and mouse monoclonal antibodies) and detection on Western blots with antibodies against pRb, p107, and p130 confirmed the MAV-1 E1A-pRb, E1A-p107, and E1A-p130 interactions (data not shown). The interaction between MAV-1 E1A and p130 had not been demonstrated previously.
|
Sur2 interaction with MAV-1 E1A CR3 domain. Nuclear extracts from two cell lines, MBMECs and 3T6 cells, were used to screen for potential additional interacting proteins in GST pulldown assays. 3T6 cells or MBMECs were used in a single GST pulldown assay and visualized in a Coomassie-stained gel (Fig. 3A; 3T6 cell data not shown). The GST fusion proteins alone were used as a negative control. The purified GST-Cter1 protein (Fig. 1A) was used to screen for proteins that specifically bind to the unique C terminus of MAV-1 E1A. The protein bands that were only present in the GST-mE1A sample (Fig. 3A, lane 6) and not in the GST sample (Fig. 3A, lane 5) were considered potential specific MAV-1 E1A binding proteins. Several specific bands were cut out of the gels, and the identification of proteins was carried out by tandem mass spectrometry-mass spectrometry analysis. One band, indicated by the arrowhead at approximately 150 kDa, was identified as mouse Sur2. We identified mouse Sur2 as a GST-mE1A-interacting protein from both MBMEC (Fig. 3A) and 3T6 (data not shown) cells. Several additional bands were identified by mass spectrometry-mass spectrometry analysis. The characterization of these as potential MAV-1 E1A-interacting proteins will be described elsewhere following verification with specific antibodies (L. Fang and K. R. Spindler, unpublished data).
|
We investigated whether MAV-1 E1A binds to mouse Sur2 in virus-infected cells (Fig. 3C). Whole-cell lysates from either MAV-1-infected or mock-infected cells were subjected to immunoprecipitation with rabbit polyclonal anti-E1A antibody. The anti-E1A immunoprecipitates were analyzed for the presence of mouse Sur2 with anti-Sur2 monoclonal antibody in Western blots. Mouse Sur2 was only detected from the MAV-1-infected sample that was immunoprecipitated with anti-E1A antibody (Fig. 3C). The results demonstrate that the interaction between MAV-1 E1A and mouse Sur2 occurred in MAV-1-infected cells.
CR3 of human adenovirus E1A is necessary and sufficient for binding to Sur2 (13, 56, 57). Despite the presence of the CR3 domain in MAV-1 E1A, there is no significant similarity to human adenovirus E1A outside of the conserved regions (6), making it intriguing to investigate which regions of MAV-1 E1A are important for binding to mouse Sur2. We used two different experimental approaches to map the region required for MAV-1 E1A interaction with mouse Sur2. First, GST pulldown assays were used. In addition to GST-mE1A, containing a 22-amino-acid linker (Fig. 1A), we constructed another full-length MAV-1 E1A directly downstream of GST, GST-wild-type E1A (Fig. 1A). A series of GST fusion proteins were also prepared that contained either truncated or deleted forms of MAV-1 E1A, as shown in Fig. 1A. GST pulldown assays were carried out with these constructs, and the bound proteins were subjected to Western blotting with anti-Sur2 antibody (Fig. 4A). The same membrane was stripped and reprobed with anti-GST antibody to verify equivalent loading of GST fusion proteins (Fig. 4B). The mouse Sur2 protein was detected in samples with GST-E1A constructs containing the CR3 of E1A (Fig. 4A, lanes 2 to 5 and 9 and 10), and absent in samples with GST-E1A constructs lacking CR3 (Fig. 4A, lanes 6 to 8 and 11). Mouse Sur2 did not bind the negative controls GST (Fig. 4A, lane 1) or GST-gp11K (Fig. 4A, lane 12). The results indicate that the CR3 domain is required for the protein-protein interaction.
|
|
Sur2 is important for MAV-1 replication. We investigated the functional role of mouse Sur2 in MAV-1 infection with Sur2+/+ and Sur2–/– MEFs. Cytopathic effects after MAV-1 infection were quite different between Sur2–/– and Sur2+/+ MEFs. Wild-type MEFs started to show dramatic cytopathic effect at an MOI of 1 or 5 at 3 days postinfection, with only a slight cytopathic effect at an MOI of 0.05 (data not shown). At 7 days postinfection, the infected wild-type MEFs showed cytopathic effect at all multiplicities, whereas Sur2–/– MEFs showed little or no cytopathic effect except at the higher MOIs of 1 and 5 (Fig. 6). The results showed that there was a delay in the appearance of cytopathic effect in Sur2–/– MEFs upon MAV-1 infection compared to wild-type cells at equivalent MOIs.
|
|
The plaque assays measured the final outcome of viral infection. Every step in the infectious process, such as viral early gene expression, viral genome DNA replication, late gene expression, DNA packaging, viral assembly, and secretion of mature virions, could affect the final viral yield. We examined some of these steps to determine the potential roles of Sur2 in MAV-1 infection.
We tested whether viral DNA replication differed in Sur2–/– and Sur2+/+ MEFs by analyzing MAV-1 genome DNA in Southern blots (Fig. 8). At MOIs of 1 and 5, MAV-1 DNA was detected at 24 h postinfection in both cell types. The kinetics of viral DNA accumulation during the course of viral infection were shown by the increasing intensity of viral DNA bands in virus-infected Sur2+/+ MEFs. In contrast, less DNA accumulation was observed in Sur2–/– MEFs. This indicates that there was no defect in the time of onset of viral DNA synthesis at MOIs of 1 and 5, but the accumulation of viral DNA was substantially decreased in Sur2–/– cells. We did not detect any MAV-1 DNA from Sur2–/– MEFs at an MOI of 0.05, whereas in Sur2+/+ MEFs it increased to a level comparable to that seen in Sur2+/+ MEFs at MOIs of 1 and 5 (Fig. 8). The defect in MAV-1 DNA replication in Sur2–/– MEFs, taken together with overall lower virus yield in the same conditions (Fig. 7A), suggested that mouse Sur2 is required for some step at or prior to DNA replication.
|
|
To confirm that the RNase protection assay results for mRNA levels were reflected in protein levels, we analyzed the expression of two MAV-1 early proteins, E1A and E3gp11K, and compared them to that of ß-actin (Fig. 10). As expected, the levels of MAV-1 E1A and E3gp11K protein accumulation were higher in Sur2+/+ MEFs than in Sur2–/– MEFs when the cells were infected at an MOI of 1 (Fig. 10). The decreased protein recovery (and loading; see the figure legend) of E3 gp11K, E1A, and ß-actin seen at 5 and 7 days postinfection in Sur2+/+ MEFs (Fig. 10A, lane 3 and 4) was likely due to the cytopathic effect at this time in infection (Fig. 6). Quantitation (Fig. 10B) shows that E1A and E3gp11K protein levels were relatively high even when the cytopathic effect was apparent in the Sur2+/+ MEFs. At an MOI of 0.1, E1A and E3gp11K proteins were not detected by Western blots in Sur2–/– MEFs even at 7 days postinfection (data not shown), supporting the hypothesis that mouse Sur2 is required for MAV-1 replication at a low input multiplicity. In contrast, both early viral proteins were readily detected in Sur2+/+ MEFs at 3 days postinfection (data not shown). The correlation of the Western blot data with the RNase protection assay results strongly argues that mouse Sur2 plays a critical role in an early phase of MAV-1 replication.
|
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The polyclonal antibody against MAV-1 E1A (AKO7-147) can recognize the E1A deletion mutant proteins, including the CR3-deleted E1A, in immunoprecipitations (61). Therefore, the absence of the mouse Sur2 signal from the CR3 deletion mutant-infected sample immunoprecipitated with anti-E1A antibody was not due to a reduced efficiency of viral infection or failure of antibody recognition. However, we also noticed that less mouse Sur2 coimmunoprecipitated with MAV-1 E1A from dlE105 and dlE102 virus-infected samples than wild-type MAV-1-infected samples (Fig. 5, lanes 6 and 8). It is possible that the deleted forms of MAV-1 E1A protein might have conformational changes resulting in reduced binding affinity to mouse Sur2. A formal possibility that we believe is less likely is that CR1 and CR2 have direct but small effects on the Sur2-E1A interaction in viral infection.
With a variety of single-amino-acid human adenovirus E1A mutations in GST pulldown assays, an intact zinc finger structure (four cysteine residues that bind a single Zn2+ ion) in human adenovirus E1A CR3 was shown to be important for the Sur2-E1A CR3 interaction (13). MAV-1 E1A CR3 has the same four cysteine residues, and it is likely that this conserved zinc finger structure in MAV-1 E1A is important for binding to mouse Sur2. One question that has not been addressed for human adenovirus or MAV-1 is which regions of Sur2 protein are required for the Sur2-E1A interaction. Human Sur2 protein has been reported to interact with three transcription factors to date, Elk-1 (13, 56), ESX (2), and C/EBPß (40). Elk-1 is activated by extracellularly regulated kinase (ERK) in the mitogen-activated protein kinase signal transduction pathway and binds serum response elements of targeted genes. The region of Sur2 that binds to Elk-1 is unknown. ESX, an Ets factor, is an epithelial cell-specific transcription factor, and residues 352 to 625 of Sur2 are important for binding to ESX (2). It has been suggested that E1A competes with C/EBPß to bind to Sur2 (40). Mapping the Sur2 region required for binding to E1A will allow us to postulate whether E1A interferes with the binding of Sur2 to these transcription factors. In turn, it will help us understand the physiological functions of Sur2 and address why adenovirus E1A targets Sur2.
The conservation of adenovirus E1A binding to Sur2 in human adenovirus and MAV-1 indicates its importance for adenovirus replication. However, due to adenovirus species specificity and the availability of only mouse Sur2 knockout cells, Sur2 function in adenoviral replication had not been directly tested prior to this study. No Sur2 protein was detected in Sur2–/– MEFs by Western blots (data not shown). Expressing human Sur2 in Sur2–/– mouse stem cells rescues their defective transcription activation function, showing that Sur2 is the only missing factor in the knockout cells (56). The data in Fig. 6 to 10 suggest that the MAV-1 replication defect is due at least in part to a defect in viral early gene transcription.
The importance of mouse Sur2 protein for viral replication is specific for MAV-1, because there was no growth defect for mouse gammaherpesvirus 68 in Sur2–/– MEFs (Fig. 7) (13, 40, 56). The transcription activities of many transcription factors are not affected by human Sur2 (2, 13, 56), indicating the narrow target range of Sur2. This raises the interesting question of why the adenovirus E1A protein conserves the ability to bind to such a selective Mediator subunit.
The molecular mechanisms of Sur2 function in MAV-1 infection have not been fully investigated. It has been proposed that the primary mechanism by which E1A recruits the Mediator complex to transactivate the transcription of viral early genes is that E1A stimulates and also stabilizes the assembly of a transcription preinitiation complex on promoter DNA through stable Sur2-E1A CR3 interaction (17, 57). However, the ability of MAV-1 to replicate in Sur2–/– MEFs suggests that there is a Sur2-independent virus replication pathway. This raises many possibilities. First, it is possible that the Mediator complex or transcription preinitiation complex is recruited to viral gene promoters with decreased efficiency in the absence of Sur2. It is possible that E1A binds other components of the Mediator complex, and it will be interesting to see whether E1A can coimmunoprecipitate other Mediator components in Sur2–/– MEFs.
Second, other regions of human adenovirus E1A protein, including CR1 and CR2, also have very strong transcriptional activity, primarily through binding of p300/CBP (10, 58) and Rb family proteins (23), respectively. Binding of Rb family members and binding of p300/CBP are two independent mechanisms utilized by human adenovirus E1A protein to manipulate the cell cycle from G1 to S phase (51). The MAV-1 E1A protein contains all three conserved regions, and interactions between MAV-1 E1A and Rb family members are shown in this work. Another possibility is that E4 gene products also play a role in MAV-1 replication. The human adenovirus E4 ORF6/7 protein has been shown to induce E2F DNA binding to the viral E2A promoter and thus functionally compensate for the loss of E1A in human adenovirus infection (44, 49). MAV-1 E4 ORFd (36) has 17% identity and 43% similarity to human adenovirus E4 ORF6/7 protein (L. Fang and K. R. Spindler, unpublished data). It is reasonable to speculate that MAV-1 uses one or more of these mechanisms to replicate if Sur2 is absent in the infected host cells.
The CR3 deletion mutant virus (dlE106) replicates to a level comparable to wild-type virus in 3T6 cells at an MOI of 5 (61), but its CR3-deleted E1A protein does not bind to mouse Sur2 (Fig. 5), indicating that the MAV-1 E1A-mouse Sur2 interaction itself is not absolutely essential for MAV-1 replication. Moreover, the fact that the human adenovirus E1A null mutant virus (dl312) can replicate at high MOIs clearly demonstrates that E1A is not an essential gene for human adenovirus replication (26, 32, 43, 50). MAV-1 E1A is dispensable for MAV-1 replication in cell culture, demonstrated by growth of the MAV-1 E1A null mutant (pmE109) at MOIs as low as 1 (61). This indicates the existence of an E1A-independent viral replication pathway and may also explain why MAV-1 is able to replicate in the absence of the mouse Sur2-E1A CR3 interaction in Sur2–/– MEFs. We are currently testing the viral replication of E1A null and CR3 deletion mutant viruses in Sur2–/– MEFs. The data in the work presented here clearly showed that mouse Sur2 is a critical factor for efficient MAV-1 replication.
We note that there are differences between MAV-1 infection in cell culture and in mice. The E1A null mutant pmE109 is less virulent than wild-type MAV-1 in mice (52). The 50% lethal dose of pmE109 is 2 and 4 logs higher than that of wild-type MAV-1 in inbred SJL/J and outbred NIHS mice, respectively (55). Both pmE109 and CR3 deletion (dlE106) mutant viruses replicated to significantly lower levels in several inbred stains of mice than did wild-type MAV-1 (L. Fang and K. R. Spindler, manuscript in preparation). This suggests that mouse Sur2-E1A binding may be more important for MAV-1 replication in mice than cells in culture.
A multiplicity-dependent growth phenotype has been observed for many viruses, including human adenoviruses (26, 32, 43, 50), herpes simplex virus type 1 (16, 20, 24), cytomegalovirus (14, 45), and African swine fever virus (41). dl312, an E1A null mutant of human adenovirus, shows the multiplicity-dependent phenotype (26, 32, 43, 50). At low MOIs, there is a delay and reduced expression of viral early genes, but at higher MOIs, the expression of viral early genes can be enhanced (43). The UL82-deficient mutant of human cytomegalovirus displays a multiplicity-dependent phenotype in cell culture (14). The defect of severely restricted viral replication of this UL82-deficient mutant virus at low input MOIs can be rescued by higher input MOIs. Very similarly, the severe replication defect of ICP0 null mutant herpes simplex virus type 1 in Vero or BHK cells at low MOIs can be overcome at higher multiplicities (16, 20, 24). A thymidine kinase gene deletion mutant of African swine fever virus shows a growth defect on swine macrophages at low MOIs, but the defect is not apparent at high MOIs (41).
It is interesting that adenovirus E1A, herpes simplex virus type 1 ICP0, and human cytomegalovirus UL82 all function in the early phase of viral infection and in gene activation (9, 14, 15, 19, 30, 33, 39), and mutations in these genes all show multiplicity-dependent effects. E1A is the first transcribed gene in adenovirus infection (43, 51), and it transactivates the transcription of the other viral early genes, including E4 (9, 33). However, E4 gene expression is seen in the absence of E1A at high MOIs (46), suggesting that there is an E1A-independent pathway to activate the transcription of E4 genes. Therefore, it is possible that E4 gene products play a role in this multiplicity-dependent effect in MAV-1-infected Sur2–/– MEFs. There might be a threshold in Sur2–/– MEFs for MAV-1 replication. At a relatively low MOI, there might not be enough accumulation of E1A or E4 gene products. However, at a higher MOI, due to the increase in copy number of viral promoters in cells, the viral E1A and/or E4 gene products might accumulate to high enough levels to support MAV-1 replication. In any case, MAV-1 does not replicate as well in Sur2–/– MEFs as in Sur2+/+ MEFs, indicating that Sur2 is important for efficient MAV-1 replication.
|
ACKNOWLEDGMENTS |
|---|
This work was supported by NIH grants R01 AI023762 to K.R.S. and CA25235 to A.J.B.
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
