Epstein-Barr Virus (EBV) SM Protein Induces and Recruits Cellular Sp110b To Stabilize mRNAs and Enhance EBV Lytic Gene Expression

Abstract

Promyelocytic leukemia protein (PML) nuclear bodies or nuclear domain 10s (ND10s) are multiprotein nuclear structures implicated in transcriptional and posttranscriptional gene regulation that are disrupted during replication of many DNA viruses. Interferon increases the size and number of PML nuclear bodies and stimulates transcription of several genes encoding PML nuclear body proteins. Moreover, some PML nuclear body proteins colocalize at sites of viral DNA synthesis and transcription. In this study, the relationship between lytic Epstein-Barr virus (EBV) replication and Sp110b, a PML nuclear body protein, was investigated. Sp110b is shown to physically and functionally interact with the EBV protein SM. SM is expressed early in the EBV replicative cycle and posttranscriptionally increases the level of target EBV lytic transcripts. SM bound to Sp110b via two distinct sites in Sp110b in an RNA-independent manner. SM also specifically induced expression of Sp110b during lytic EBV replication and in several cell types. Exogenous expression of Sp110b synergistically enhanced SM-mediated accumulation of intronless and lytic viral transcripts. This synergistic effect was shown to be promoter independent, posttranscriptional, and the result of increased stabilization of target transcripts. Finally, inhibiting Sp110b expression decreased accumulation of an SM-responsive lytic EBV transcript in EBV-infected cells. These findings imply that SM induces Sp110b expression, binds to Sp110b, and utilizes the recruited Sp110b protein to increase the stability of lytic EBV transcripts, indicating that Sp110b is a component of the cellular machinery that EBV utilizes to enhance lytic EBV replication.

The Epstein-Barr virus (EBV) nuclear protein SM is expressed early after entry of EBV into the lytic cycle of replication and is essential for EBV virion production (12, 19). SM increases the accumulation of several viral intronless lytic mRNA transcripts and specific cellular transcripts in both the nucleus and cytoplasm (38, 40, 42). SM also inhibits splicing and inhibits expression of the majority of cellular genes (38, 39). Thus, SM affects multiple cellular pathways involved in RNA processing and transport. In order to further elucidate its mechanism of action, we used the yeast two-hybrid assay to identify additional cellular proteins that interact with SM. During this investigation, we identified the promyelocytic leukemia protein (PML) nuclear body protein Sp110b, expressed from a splicing variant of Sp110 (4, 25), as a potential SM-interacting protein.

PML nuclear bodies (also known as nuclear domain 10 proteins [ND10s]) are dynamic nuclear structures composed of numerous proteins, some of which localize to the PML nuclear body under specific environmental conditions (for review, see reference 13, 30). Their function is still largely unknown, but many PML nuclear body proteins, including Sp110, have been implicated in regulation of gene transcription. The number of PML nuclear bodies in the nucleus increases in response to heat shock, viral infection, treatment with interferons, and heavy-metal exposure. RNA polymerase II transcripts localize near PML nuclear bodies, and the PML nuclear body is a preferred site for initial gene transcription in many viruses (13, 27, 30).

Several herpesviruses, adenoviruses, and papovaviruses begin their lytic replication and transcription at PML nuclear bodies (14). As replication progresses, PML nuclear body disruption occurs (13, 16, 26). During lytic EBV replication, several proteins are released sequentially from the PML nuclear body, whereas PML protein is either retained or recruited to the site of ongoing virus DNA replication (3). Because of the upregulation of PML nuclear body protein expression by interferons, and conversely, the disruption of PML nuclear bodies by infecting viruses, it has been suggested that PML nuclear body proteins may have antiviral functions. On the other hand, several herpesviruses, including herpes simplex virus, cytomegalovirus, and EBV, induce a number of interferon-stimulated genes by interferon-independent mechanisms, suggesting that specific interferon-stimulated genes may be utilized by incoming viruses to facilitate virus replication or other aspects of virion production (34, 35, 38).

Sp110 is a member of the Sp100/Sp140 family of nuclear body components expressed primarily in leukocytes and is induced by type I interferon treatment (4, 25). Sp110 is a 689-amino-acid protein which includes an Sp100-like domain, a putative nuclear localization sequence, a SAND domain, a plant homeobox domain, and a bromodomain in that order (Fig. 1). The Sp100-like domain is proposed to be involved in homodimerization and heterodimerization with other Sp100 family proteins (4, 41). The SAND domain, plant homeobox domain, and bromodomain are all common features of modular proteins involved in chromatin-mediated control of gene transcription (4, 6, 17). Sp110 also contains an LXXLL-type nuclear hormone receptor interaction motif (Fig. 1) and enhances expression of all-trans-retinoic acid-responsive reporters in cotransfection assays, suggesting that Sp110 may function as a nuclear hormone receptor coactivator (4).

FIG. 1.

Structural and functional domains of Sp110 and Sp110b. Common features of Sp110 and Sp110b include the Sp100-like domain located between amino acids 6 and 159, the putative nuclear localization signal (NLS) between amino acids 288 and 306, and the SAND domain between amino acids 452 and 532. In addition, alternative splicing adds a plant homeobox domain (PHD, amino acids 537 to 577) and a bromodomain (BD, amino acids 606 to 674) to Sp110. The LXXLL nuclear hormone interaction motif is shown with an arrow. Sp110 and Sp110b have predicted sizes of 689 and 539 amino acids, respectively.

Sp110b is a splice variant of Sp110 which terminates immediately after the SAND domain and therefore lacks both the plant homeobox domain and bromodomain, but a distinct function for Sp110b protein has not been established. However, a recent report suggests that Sp110b acts as a transcriptional repressor of retinoic acid receptor α (46). Further, it was shown that hepatitis B virus core protein interacts with Sp110b and relocalizes it from the nucleus to the cytoplasmic face of the endoplasmic reticulum, suggesting that hepatitis B virus core protein activates retinoic acid receptor α-mediated transcription by sequestration of Sp110b (46).

We have previously shown that expression of EBV SM induces several interferon-stimulated genes via induction of STAT1 (38). Since Sp110 was known to be an interferon-stimulated gene, it was possible that SM also induced Sp110b expression. The finding of a physical interaction between SM and Sp110b suggested that Sp110b might be an example of an interferon-stimulated gene whose cellular function is subverted by EBV to enhance EBV gene expression and EBV replication. We therefore investigated the functional consequences of an interaction between Sp110b and EBV SM.

MATERIALS AND METHODS

Plasmids.

SM cDNA cloned in the cytomegalovirus (CMV) promoter-driven vector pCDNA3 has been described previously (40). The cDNA cloned in reverse orientation in pCDNA3 (aSM) used as a control plasmid in some transfections has also been described previously (40). The bait plasmid for the yeast two-hybrid assay was constructed by insertion of SM cDNA lacking RXP repeat domains (SMΔRXP) (37) in frame with the DNA binding domain of GAL4 in the pAS2 vector (Clontech).

The cDNA library was derived from a human B lymphoblastoid cell line cloned in the pACT vector (Clontech) (kind gift of Erle Robertson). Hemagglutinin (HA)-tagged pCDNA3 was created by ligation of a synthetic hemagglutinin epitope tag into the polylinker of pCDNA3. Sp110b cDNA cloned in pCDNA3 has been described previously (4). An expression plasmid for HA-tagged full-length Sp110b was generated by excising and cloning Sp110b cDNA into HA-pCDNA3. The C-terminal 277 amino acids of Sp110b were excised from pACT clone 8, which was isolated by two-hybrid screening, and recloned in HA-pCDNA3. Plasmids N-ter c8 (clone 8 with amino acids 453 to 539 deleted) and C-ter c8 (clone8 with amino acids 304 to 453 deleted) were generated by PCR amplification and cloned in HA-pCDNA3. All constructs were sequenced prior to use in transfections. The reporter CMV-CAT has been described previously (40). The BMRF1 expression plasmid was generated by cloning the BMRF1 open reading frame in pCDNA3. CMV-RLΔi and SV40Δi were generated by excision of the intron from the corresponding intron-containing vectors CMVRL and SV40RL (Promega).

Yeast two-hybrid assay.

SMΔRXP cloned in pAS2 was used as the bait. Saccharomyces cerevisiae lines stably expressing SMΔRXP were established by transformation of S. cerevisiae strain Y190 and confirmed to express full-length bait protein by immunoblotting. The lymphoblastoid cell line (LCL) cDNA library cloned in pACT was used to transform S. cerevisiae cells expressing SMΔRXP, and 10⁷ transformants were screened. Potential interacting clones were identified by testing for histidine auxotrophy and lacZ expression as previously described (2). Analysis of one of these clones (clone 8) is reported herein.

Coimmunoprecipitation assays.

HeLa cells were transfected with HA-tagged clone 8 and Myc-tagged SM. At 48 h posttransfection, cells were harvested and resuspended in lysis buffer [0.1% Triton X-100, 1 mM dithiothreitol, and protease inhibitor cocktail (Sigma) in Tris-buffered saline], followed by sonication and removal of cellular debris by centrifugation. Complexes were precipitated with either anti-Myc (Upstate), anti-HA (Babco), or isotype-matched irrelevant monoclonal antibodies (Santa Cruz) at 4°C for 2 h. Immune complexes were harvested with protein G-agarose beads (Sigma) and washed extensively at 4°C with lysis buffer. The beads were suspended in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) buffer and analyzed by immunoblotting as previously described (7).

In vitro GST pulldown assays.

The glutathione S-transferase (GST) fusion protein SMΔRXP was expressed in Escherichia coli BL21(pLys) bacteria. Soluble protein fractions were generated by lysis in buffer containing 50 mM Tris (pH 8), 2 mM EDTA, 100 μg of lysozyme per ml, 0.1% Triton X-100, and bacterial protease inhibitor cocktail (Sigma). ³⁵S-labeled Sp110b and Sp110b peptides were produced in the rabbit reticulocyte lysate in vitro transcription-translation TNT system (Promega) by addition of ³⁵S-labeled methionine and cysteine (Amersham). Full-length Sp110b or peptides cloned in HA-pCDNA3 were used as templates. The ³⁵S-labeled Sp110b and Sp110b peptide lysates were precleared by incubation with glutathione-agarose beads (Sigma). Binding reactions were carried out in 100 μl of 0.1% NP-40-0.5 mM dithiothreitol-10% glycerol-1 mM phenylmethylsulfonyl fluoride in phosphate-buffered saline. The glutathione-agarose beads were washed five times with 1 ml of the same buffer. In some experiments, RNase A (100 μg/ml) was added to the last wash and incubated for 15 min at room temperature. In these experiments, the complexes bound to the glutathione-agarose beads were washed two more times. Proteins were eluted in SDS-PAGE protein loading buffer, electrophoresed, and analyzed by autoradiography.

Cell lines and transfections.

HeLa cells were grown at 37°C in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal calf serum and glutamine. AGS cells, a gastric epithelial cell line (ATCC CRL-1739), and AGS cells carrying an SM-knockout recombinant EBV (SMKO-AGS) were grown in F-12 nutrient mixture (Intvitrogen) supplemented with 10% fetal calf serum and glutamine. SMKO-AGS cells were generated by infecting AGS cells with supernatant from SMKO-293 cells which were induced to permit lytic EBV replication by transfection with Z and SM expression plasmids as previously described (19, 39). SMKO-293 (19), P3HR1, an EBV-positive Burkitt's lymphoma-derived cell line (36), and BJAB, an EBV-negative B lymphoma cell line (33), were grown in RPMI (Invitrogen) supplemented with 10% fetal calf serum and glutamine. Akata and EBV-negative Akata cells (kind gift of John Sixbey) have been described previously (43, 44). SM expression was induced in SM-BJAB cells by treatment with 4-hydroxytamoxifen as previously described (38). HeLa, AGS, SMKO-AGS and SMKO-293 cells were transfected with Lipofectamine Plus (Invitrogen) following the manufacturer's protocol. BJAB and P3HR1 cells were electroporated with a total of 20 μg of DNA as previously described (40), with equal amounts of SM, target gene, and Sp110b expression plasmids. When one of the plasmids was omitted, the total amount of DNA was kept constant by inclusion of empty vector DNA.

Reporter gene assays.

Cells were lysed in reporter lysis buffer (Promega) 48 h after transfection, and supernatants were stored at −80°C. Chloramphenicol acetyltransferase (CAT) assays were performed and quantitated by direct radiometry with a Packard InstantImager as previously described (40). The amounts of lysates were adjusted to keep the percent acetylation in the linear range for the CAT assay. Renilla luciferase assays were performed as per the manufacturer's protocol (Promega). Each data point was calculated from the mean of three independent transfections.

Run-on transcription assays.

Run-on transcription was performed with 25 × 10⁶ nuclei per reaction. Washed cells were lysed 48 h after transfection in 10 mM Tris (pH 8)-0.5% NP-40-3 mM MgCl₂. Nuclei were separated by centrifugation and stored at −80°C. One microgram of RLΔi cDNA was slot blotted onto a Zeta-Probe membrane (Bio-Rad). Bacteriophage λ DNA and 18S rRNA cDNA were used as negative and positive controls, respectively. In vitro run-on nuclear transcripts were generated with ³²P-labeled UTP, treated with DNase, and purified as previously described (40); 10⁷ cpm of each transcript was hybridized to immobilized cDNA for 40 h and washed at 65°C. Radioactivity bound to each DNA was quantitated with an Instant Imager (Packard Instruments).

RNA isolation.

We lysed 5 × 10⁵ or 1 × 10⁶ cells in 1 ml of RNA-Bee (Teltest, Friendswood, Tex.). The RNA was isolated by chloroform extraction of the aqueous phase, followed by isopropanol precipitation as per the manufacturer's instructions. The RNA was then resuspended in 5 mM EDTA, extracted with phenol and chloroform, and ethanol precipitated. For nuclear and cytoplasmic preparations, 30 h after transfection, the cells were harvested and lysed in NP-40 buffer as previously described (40). Nuclei were separated by centrifugation. RNA was then isolated from both the cytoplasmic and nuclear fractions as described above. Five micrograms per lane of each RNA was used for Northern blotting.

RNA half-life measurement.

BJAB cells were electroporated with 20 μg of plasmid DNA per 10⁷ cells; 10 μg of target, 3 μg of SM, and 7 μg of Sp110b plasmid were used in each transfection. Empty vector DNA was used in place of SM or Sp110b when either plasmid was omitted. At 40 h after transfection, cells were treated with 10 μg of actinomycin D per ml. Cells were harvested at various times after actinomycin D treatment, and RNA was isolated and analyzed by Northern blotting as described above.

Induction of lytic EBV replication.

Lytic EBV replication was induced in P3HR-1 cells by adding tetradecanoyl phorbol acetate (TPA) at 20 ng/ml; 2 mM sodium butyrate was used in addition to TPA in some experiments. Akata cells were induced with 1% rabbit anti-human immunoglobulin G (Dako) as previously described (45). RNAs were harvested for Northern blotting 48 h after induction.

Construction of Sp110 siRNA plasmid.

Ambion's pSilencer 2.0 U6 was used as the short interfering RNA (siRNA) expression vector. Sp110b siRNA constructs were designed and constructed following the manufacturer's protocol. Briefly, two complementary oligonucleotides derived from nucleotides 962 to 980 of the Sp110b sequence were synthesized, annealed, and ligated into pSilencer 2.0 U6. The sequences of the oligonucleotides were 5′-GATCCCGAAAGATGACTCAACTTGTTTCAAGAGAACAAGTTGAGTCATCTTTCTTTTTTGGAAA-3′ and 5′-AGCTTTTCCAAAAAAGAAAGATGACTCAACTTGTTCTCTTGAAACAAGTTGAGTCATCTTTCGG-3′.

RESULTS

SM interacts with Sp110b.

In order to identify cellular proteins that interacted with SM, a screen was performed with a yeast two-hybrid transcriptional activation assay. SMΔRXP was used to generate the bait plasmid because we found that SM containing the RXP repeats was susceptible to truncation in S. cerevisiae (data not shown). SM protein containing the RXP repeats is similarly unstable or poorly synthesized in bacteria (37). An EBV-transformed B LCL cDNA library was screened with SMΔRXP as the bait, and several interacting clones were isolated. The clone 8 cDNA insert was sequenced and found to encode the final 277 C-terminal amino acid residues (amino acids 312 to 539) of the human protein Sp110b fused in-frame with the Gal4 activation domain. As shown in Fig. 1, Sp110b is a shorter splice variant of the human protein Sp110, identical to Sp110 except for the terminal splice that results in an alternative final exon. Sp110b therefore contains an Sp100-like domain, a putative nuclear localization sequence, and SAND domain present in Sp110 but lacks the plant homeobox domain and bromodomain (Fig. 1).

In order to determine whether SM and the Sp110b peptide encoded by clone 8 interacted in vivo, coimmunoprecipitation experiments were performed. Plasmids that expressed Myc-tagged SM and a plasmid encoding HA-tagged Sp110b from amino acids 262 to 539 were cotransfected into HeLa cells, and immunoprecipitation from whole cell extracts was performed with anti-Myc or anti-HA monoclonal antibodies. Coimmunoprecipitation of SM and Sp110b was then evaluated by immunoblotting with anti-Myc or anti-HA monoclonal antibodies, respectively. As shown in Fig. 2, SM and the Sp110b peptide were coimmunoprecipitated, indicating that they could interact in vivo.

FIG. 2.

Coimmunoprecipitation of SM and Sp110b. Plasmids expressing Myc-tagged SM and HA-tagged Sp110b peptide (amino acids 262 to 539) were cotransfected into HeLa cells, and immunoprecipitation from whole-cell extracts was performed with anti-Myc, anti-HA, or irrelevant isotype-matched (control IP) monoclonal antibodies. Immunoprecipitates were immunoblotted with anti-HA antibodies (αHA WB) or anti-Myc antibodies (αMyc WB). The input lysate is shown at the left (input).

Sp110b contains two independent SM-binding regions.

GST pulldown experiments were employed to identify the region of Sp110b that interacts with SM. Several DNA subclones of Sp110b were created by restriction digestion of full-length Sp110b (Fig. 3A). Each subclone of the Sp110b gene, as well as full-length Sp110b, was transcribed and translated in vitro, and the resultant peptides were radioactively labeled with ³⁵S-labeled methionine and cysteine. They were then incubated with bacterially synthesized GST-SMΔRXP fusion protein immobilized on glutathione-Sepharose beads. As shown in Fig. 3B, two regions of Sp110b, both present in the interacting clone c8, were capable of independently mediating attachment to SM. One binding site was mapped to amino acids 453 to 539, in the SAND domain. The other SM-binding domain was mapped to amino acids 304 to 453, which are unique to Sp110b and are not homologous to Sp100 or Sp140.

FIG. 3.

Mapping of Sp110b interaction with SM in vitro. (A) A schematic representation of full-length Sp110b and Sp110b peptides that were used to test for interaction with SM in vitro. Clone 8 consists of amino acids 262 to 539 of Sp110b that were isolated in the yeast two-hybrid assay. C-ter clone 8 has amino acids 304 to 453 removed. N-ter clone8 has amino acids 453 to 539 removed from clone 8. ATG-Xmn contains amino acids 1 to 192 of Sp110b. delNLS clone 8 has amino acids 262 to 304 removed from clone 8. Sp110b peptides that interacted with SM in vitro are shown with a + at right, and brackets delineate two independent SM-interacting regions. (B) GST pulldown assays of Sp110b peptides with immobilized GST-SM. The Sp110b peptide used is shown above each panel. ³⁵S-labeled methionine and cysteine-labeled Sp110b and Sp110b peptides were incubated with control GST beads (lanes 2) or GST-SM (lanes 3 and 4). The input peptide (lanes 1) and the bound fractions (lanes 2, 3, and 4) were analyzed by SDS-PAGE and visualized by autoradiography. Beads were treated with RNase A after binding (lanes 4) as described in the text.

The first 192 amino acids of Sp110b (ATG-Xmn peptide), which contain the Sp100-like domain, did not bind to GST-SMΔRXP. The putative nuclear localization sequence region in Sp110b was also not required for binding to SM. Therefore, we concluded that the region between amino acids 262 and 539 of Sp110 and Sp110b contains two or more binding sites that interact directly and independently with SM. In order to determine whether either of the Sp110b-SM interactions was RNA dependent, after the wash steps, a fraction of the complexes bound to the glutathione-agarose beads was treated with RNase A and washed again. RNase treatment did not appreciably affect either interaction (Fig. 3B, lanes 4).

Sp110b enhances SM-mediated trans-activation of target genes.

In order to determine whether the physical interaction of Sp110b and possibly Sp110 with SM had a functional significance, we next performed assays of SM function in the presence and absence of exogenously expressed Sp110b and Sp110. We have previously reported that SM increases the expression of intronless viral gene and heterologous reporter gene mRNAs by increasing the accumulation of the target transcripts in both the nucleus and cytoplasm (37, 38, 40). We therefore evaluated the effect of cotransfecting Sp110 or Sp110b with SM on several reporter plasmids that expressed intronless mRNAs.

Initial experiments were performed in HeLa cells transfected with the reporter plasmid CMV-CAT (40). Transfection of SM plasmid increased CAT expression approximately fivefold over the level observed with transfection of empty vector. In these experiments, the amount of SM plasmid used (0.25 μg) was less than that required for maximal activation in order to allow detection of additional Sp110 or Sp110b effects. Coexpression of Sp110 or Sp110b with SM led to a further increase in CAT activity (approximately twofold over that observed with SM alone) which was modest but reproducible (Fig. 4A). To determine if the enhancing effects of Sp110 and Sp110b on activation by SM were gene dependent, a similar experiment was performed in HeLa cells transfected with an intronless Renilla luciferase reporter driven by a CMV promoter (RLΔi). In this experiment, cotransfection with SM and either Sp110 or Sp110b increased Renilla luciferase activity more than transfection of SM alone (Fig. 4B). The enhancing effect of Sp110 on SM-mediated activation of RLΔi was somewhat greater than that observed with CAT, perhaps because SM transactivates RLΔi less strongly than it does the CAT reporter (39). It should also be noted that with both reporters, Sp110 or Sp110b in the absence of SM had little effect on reporter gene activity, indicating that Sp110 and Sp110b were not acting at the transcriptional level to increase the amounts of target gene mRNAs.

FIG. 4.

Effect of Sp110 and Sp110b on SM transactivation of reporter genes. (A) CAT driven by the CMV immediate-early promoter (CMV-CAT) was used as the reporter gene in HeLa cells. Control plasmid (C) or SM plasmid was transfected with or without Sp110 or Sp110b plus the reporter plasmid. CAT assays were performed 48 h after transfection. Results are expressed as activation over the control. (B) Renilla luciferase driven by the CMV promoter (CMV-RLΔi) was used as the reporter gene in HeLa cells. Control plasmid (C) or SM plasmid was transfected with or without Sp110 or Sp110b plus the reporter plasmid. Renilla luciferase assays were performed 48 h after transfection. Results are expressed as relative luminescence units (RLU). (C) Renilla luciferase driven by the simian virus 40 early promoter (SV40-RLΔi) was used as the reporter, and transfections were performed with the same effector plasmids as in B. (D) CMV-RLΔi was used as the reporter plasmid and was transfected into BJAB cells with either control plasmid (C) or SM, with or without Sp110b, by electroporation. All results represent the average of three independent transfections. Empty vector DNA was added where necessary so that the total amount of DNA used for each transformation was equal.

Enhancing effect of Sp110b on SM is posttranscriptional.

If the synergistic effect of Sp110 were posttranscriptional, one would expect that it would be promoter independent, similar to gene activation by SM. We therefore replaced the CMV promoter in the RLΔi plasmid with a simian virus 40 promoter and examined the effect of Sp110 expression on SM-mediated activation. HeLa cells were transfected with the SV40-RLΔi plasmid and either empty vector or SM, with or without Sp110 or Sp110b. In this experiment, cotransfection of Sp110 or Sp110b with SM also led to a further increase in reporter activity over that achieved with transfection of SM alone (Fig. 4C). Therefore, the synergistic effect of Sp110 and Sp110b on posttranscriptional SM activation could be demonstrated with either of the two promoters tested.

In order to ask if the enhancing effect of Sp110 and Sp110b on SM transactivation was cell type dependent, similar experiments were performed in B cells. EBV-negative B lymphoma BJAB cells were transfected with CMV-RLΔi and either SM or SM plus either Sp110 or Sp110b, and luciferase assays were performed on cell lysates after 48 h. Again, Sp110b, in the presence of SM, increased Renilla luciferase expression over that observed with transfection of SM alone (Fig. 4D). Transfection of Sp110 led to toxicity in BJAB cells, with death of a large number of cells within 48 h. The reason for this cell type-dependent toxicity of Sp110 in BJAB cells, which was not observed in HeLa cells, remains to be explored. The plant homeobox domain and bromodomain of Sp110, which are absent in Sp110b, may confer some additional gene regulation properties on Sp110. It is also possible that Sp110 may function differently in BJAB and HeLa cells because of the presence of different cellular pathways in the two cell lines, particularly those involving cell type-specific transcription factors. Nevertheless, cotransfection of Sp110b synergistically enhanced expression of the SM target reporter above levels seen by transfection of SM alone regardless of the promoter, reporter gene, or cell type examined.

Since Sp110 has been reported to be a possible transcriptional coactivator, we wished to directly confirm that the effect of Sp110b observed in the reporter assays was posttranscriptional. Therefore, a nuclear run-on assay was performed on BJAB cells to compare the transcriptional initiation rate of the target gene in cells transfected or not transfected with Sp110b. Nuclei were prepared from cells transfected with CMV-RLΔi and SM with or without Sp110b and used in a run-on assay to measure the rate of Renilla luciferase transcript initiation. As shown in Fig. 5A, there was no increase in the rate of RLΔi transcript initiation when Sp110b was expressed. RNA was also isolated from an aliquot of the transfected cells from which the nuclei were harvested and analyzed by Northern blotting. As expected, SM expression resulted in increased accumulation of RLΔi mRNA, which was further increased by Sp110b expression (Fig. 5B). This indicates that SM and Sp110b interact so that coexpression synergistically increases accumulation of the target mRNA.

FIG. 5.

Effect of Sp110 and Sp110b on SM target gene expression. (A) Nuclear run-on assays were performed with nuclei from BJAB cells transfected with either Renilla luciferase (RL) and empty vector (C), Renilla luciferase plus SM (SM), or Renilla luciferase plus SM plus Sp110b (SM/Sp110b). Labeled nuclear transcripts were hybridized to immobilized RLΔi cDNA (left panel), 18S rRNA cDNA as a positive control (right panel), and bacteriophage λ DNA as a negative control (middle panel). (B) Northern blot of RNA from a fraction of the Renilla luciferase plus control, Renilla luciferase plus SM, or Renilla luciferase plus SM plus Sp110b-transfected BJAB cells used for nuclear run-on assays in A. (C) Northern blot analysis of BJAB cells transfected with the BMRF1 expression plasmid and either SM or SM plus Sp110b. The blot was stripped and reprobed with glyceraldehhyde-3-phosphate dehydrogenase (GAPDH) cDNA as a loading control (bottom panel). (D) Immunoblot analysis of BJAB cells transfected with either vector (C), Sp110b (Sp), SM or SM plus Sp110b (SM/Sp). The blot was probed with anti-SM antibody (upper panel), stripped, and reprobed with antiactin antibodies as a loading control (bottom panel).

Accumulation of the EBV early gene BMRF1 transcript is known to be enhanced by SM. In order to determine whether Sp110b exerted a similar cooperative effect with SM on BMRF1 expression, HeLa cells were transfected with a plasmid that expresses BMRF1, SM expression plasmid, and either empty vector or Sp110b. RNA was harvested from the transfected cells, and the effect of Sp110b on BMRF1 mRNA was measured by Northern blotting. As shown in Fig. 5C, Sp110b enhanced expression of BMRF1 mRNA, consistent with the effects seen on Renilla luciferase mRNA above.

Although Sp110b alone did not affect reporter gene expression, it was possible that Sp110b was enhancing SM expression and thereby indirectly increasing expression of the target gene. To examine this possibility, samples of BJAB cells transfected with SM with and without Sp110b were immunoblotted to measure the relative amounts of SM in the presence and absence of Sp110b. As shown in Fig. 5D, the amount of SM protein was not affected by overexpression of Sp110b.

Sp110b prolongs mRNA half-life in the presence of SM.

The above experiments suggested that Sp110 and Sp110b may increase SM target gene expression by enhancing the accumulation of target mRNAs. In order to show that Sp110b synergistically enhanced SM's ability to increase mRNA half-life, we measured the half-life of SM-responsive mRNAs in the presence and absence of Sp110b. BJAB cells were transfected with RLΔi and SM with either empty vector or Sp110b. Forty hours after transfection, the cells were treated with actinomycin D to inhibit further transcription, and RNA was harvested serially for 200 min after actinomycin D treatment. Northern blotting with a probe specific for the RLΔi mRNA revealed that whereas there was a significant decrease in the signal for the RLΔi mRNA in cells transfected with SM over 200 min, there was no detectable decay in the cells transfected with SM and Sp110b (Fig. 6A). A similar experiment was performed with BMRF1 used as the target gene, and RNA was harvested every 3 h for 15 h after actinomycin D addition. Again, Sp110b in the presence of SM increased the stability of BMRF1 mRNA over that observed with SM alone. Based on quantitation of the RNA levels at each time point, Sp110b increased the half-life of BMRF1 mRNA from 7.4 h to 15 h.

FIG. 6.

Effects of Sp110b on stability of SM target gene mRNAs. (A) Effect of Sp110b on stability of RLΔi mRNA. BJAB cells were transfected with CMV RLΔi and SM plasmids with either control plasmid or Sp110b plasmid; 40 h after transfection, actinomycin D was added. RNA was harvested at 0, 20, 40, 80, 120, and 200 min after actinomycin D addition and analyzed by Northern blotting for RL mRNA. (B) Effect of Sp110b on stability of BMRF1 mRNA. BJAB cells were transfected with BMRF1 and SM plasmids with or without Sp110b and analyzed as in A above. Harvest times in hours are shown below. (C) Effect of Sp110b on nuclear and cytoplasmic BMRF1 mRNAs in the presence of SM. BJAB cells were transfected with BMRF1 and SM with either control or Sp110b plasmid. RNA was prepared from separated nuclear (N) and cytoplasmic (C) fractions 30 h after transfection and analyzed by Northern blotting for BMRF1 mRNA.

Based on previous work by ourselves and others, SM is likely to increase both the nuclear accumulation and the export of its target mRNAs. mRNAs, particularly intronless mRNAs, are often more stable once exported from the nucleus (29). Thus, Sp110b could exert its synergistic effect with SM by either increasing nuclear stability or indirectly by facilitating SM-mediated export. Regardless of whether Sp110b affects nuclear mRNA export, an effect on nuclear stability should be manifested by increased nuclear accumulation of target mRNAs. Therefore, we compared the accumulation of an intronless SM target gene mRNA transcript in the nucleus and cytoplasm in the presence of Sp110b and SM versus that in the presence of SM alone.

BJAB cells were transfected with EBV BMRF1 and SM with or without Sp110b. Thirty hours after transfection, the cells were harvested, and a nuclear-cytoplasmic separation was performed after detergent lysis, as described previously (40). Detection of BMRF1 mRNA by Northern blotting revealed that BMRF1 mRNA levels were increased in both the nucleus and cytoplasm of cells transfected with both SM and Sp110b over the corresponding levels in cells transfected with SM alone (Fig. 6B). Therefore, Sp110b enhances SM's ability to stabilize viral mRNAs in the nucleus. The question of whether Sp110b also has effects on SM-mediated export remains open and was not addressed by this experiment. Immunofluorescence microscopy of cells transfected with SM and Sp110b did not reveal any obvious effects of coexpression on the nuclear staining pattern of either protein (data not shown).

SM induces Sp110b expression in B lymphocytes and epithelial cells.

Sp110 and its splicing isoforms are known to be interferon-stimulated genes (25). SM has been demonstrated to induce several type I interferon-stimulated genes in B cells inducibly expressing exogenous SM (38). The mechanism is likely to be mediated by SM induction of STAT1, which is a component of the regulatory pathway for interferon-stimulated gene expression. We reasoned that Sp110b might therefore also be induced by SM and examined levels of Sp110b mRNA in cells expressing or not expressing SM. An inducible SM-expressing BJAB cell line that we have described previously (38) was either chemically induced or mock induced, and RNA was harvested from the induced and uninduced cells. Northern blot analysis was performed to detect Sp110b and revealed that Sp110b was upregulated during SM expression (Fig. 7A). Interestingly, a single RNA species was detectable by Northern blotting. Although Sp110 and Sp110b cDNAs have both been cloned previously, it had not been established which isoforms were expressed in B cells. Therefore, reverse transcription-PCR was performed on the RNA with primers that would detect both species. The results, shown in Fig. 7B, demonstrate that Sp110b is the predominant form induced by SM expression in BJAB cells.

FIG. 7.

SM induces Sp110b expression. (A) RNA was harvested from BJAB cells that express SM upon tamoxifen treatment, after induction with tamoxifen (I), or after mock induction (C). Sp110b was measured by Northern blotting 48 h after induction. The blot was stripped and reprobed with glyceraldehhyde-3-phosphate dehydrogenase (GAPDH) as a loading control. (B) Reverse transcription-PCR was performed on RNA (R) and genomic DNA (D) from induced BJAB-SM cells with Sp110- or Sp110b-specific primers. Products were analyzed by gel electrophoresis and ethidium bromide staining. Molecular size markers are shown at the left (lane M). (C) AGS cells were transfected with SM (SM) or empty vector (C), and expression of Sp110b was examined by Northern blotting of the transfected cell RNA 48 h after transfection. The blot was stripped and reprobed with glyceraldehhyde-3-phosphate dehydrogenase (GAPDH) as a loading control.

Sp110 has been reported to be primarily expressed in leukocytes, although EBV infects both epithelial cells and lymphocytes. Recently, Sp110b has been shown to be expressed in a variety of cell types (46). We therefore wished to determine whether SM could also induce Sp110b expression in cells other than B lymphocytes. AGS cells, which are derived from an EBV-negative gastric carcinoma cell line and are capable of supporting EBV replication (5), were transfected with SM or empty vector, and expression of Sp110b was examined by Northern blotting of the transfected cell RNA 48 h after transfection.(Fig. 7C). Sp110b was specifically induced by SM, demonstrating that SM induction of Sp110b is not limited to B lymphocytes but can also occur in cells of epithelial origin.

Sp110b is induced during EBV replication in Burkitt's lymphoma cells and gastric epithelial cells.

In order to investigate whether Sp110b was induced during lytic EBV replication, we examined Sp110b mRNA levels after induction of EBV replication in several cell types which are permissive for EBV replication. Lytic EBV replication was induced in EBV-infected Burkitt's lymphoma P3HR-1 cells by treatment with TPA or TPA plus sodium butyrate. RNA was harvested from mock-induced and chemically induced cells 48 h after induction and analyzed for Sp110b expression by Northern blotting. Sp110b expression was increased in TPA-treated cells and further induced in TPA- and butyrate-treated cells (Fig. 8A). Induction of Sp110b during lytic EBV replication was also examined in Akata cells, in which lytic replication is efficiently induced by anti-immunoglobulin G treatment (45). Accordingly, Akata cells were either induced or mock induced to permit lytic EBV replication, and RNA was harvested and analyzed by Northern blotting for Sp110b expression. As shown in Fig. 8B, Sp110b was also induced during lytic replication in Akata cells. The availability of EBV-negative Akata cell lines in which episomal EBV has been spontaneously lost allowed us to determine whether Sp110b expression upon immunoglobulin G treatment was specifically associated with EBV replication. The results of immunoglobulin G treatment of EBV-negative Akata lines demonstrated that Sp110b is not induced in the absence of EBV replication (Fig. 8B).

FIG. 8.

Sp110b is induced during EBV replication. (A) P3HR1 cells were induced with TPA (T), or TPA and butyrate (TB) or mock induced (C). RNA was harvested 48 h after induction and analyzed by Northern blotting. (B) EBV-infected Akata cells and EBV-negative Akata cells (Akata-EBV) were treated with 1% anti-human immunoglobulin G (Ig) or mock induced (C). RNA was harvested and analyzed as in A above. (C) AGS cells infected with SM-deleted recombinant EBV (SM-KO AGS) were transfected with the Z expression plasmid (Z) or Z and SM expression plasmids (Z + SM). RNA was harvested 48 h after transfection and analyzed by Northern blotting with the Sp110b probe. (D) Northern blotting for Sp110b was performed in 293 cells carrying the SM-deleted recombinant (SMKO-293) after transfection with Z or Z and SM expression plasmids as in A. Blots were stripped and rehybridized with the glyceraldehhyde-3-phosphate dehydrogenase (GAPDH) probe as a loading control (bottom panels).

In order to determine whether EBV replication was associated with Sp110 induction in other cell types, we employed the recently described SM-deleted recombinant EBV (19). The SM-deleted EBV which was generated in 293 cells was passaged into AGS cells, and EBV infection was confirmed by observation of green fluorescent protein expression and detection of EBV antigen by immunofluorescence microscopy (data not shown). The SM-deleted virus is defective for EBV lytic replication unless complemented with SM (19, 39). Upon induction of lytic viral replication by transfection with BZLF1 and SM but not by BZLF1 alone, Sp110b mRNA levels were increased in both cell types (Fig. 8D and E). These data indicate that Sp110b is induced during lytic EBV replication in epithelial cells and specifically as a result of SM expression. In the epithelial and B-cell lines examined, a single species of Sp110 mRNA was detected, of a size consistent with Sp110b, indicating that this is the predominant species induced by EBV replication.

Inhibiting Sp110b expression diminishes the ability of SM to enhance EBV lytic mRNA expression.

Since SM induced Sp110b expression and bound to Sp110b protein, which enhances the ability of SM to increase viral mRNA accumulation, it appeared likely that Sp110b plays an important role in EBV replication. In order to further investigate the function of Sp110b during EBV lytic viral replication, an Sp110b siRNA plasmid (si110b) was constructed. The siRNA coding sequence was not homologous to any known sequence in the human genome other than Sp110. The sequence was cloned in the U6 promoter-driven vector pSilencer2.0 U6 (Ambion), and the commercially available pSilencer2.0 NC (Ambion) was used as a negative control. Transfection of si110b efficiently decreased Sp110b expression in HeLa cells transfected with Sp110b, whereas pSilencer2.0 NC had no effect (Fig. 9A).

FIG. 9.

Sp110b siRNA inhibits Sp110b and BMRF1 expression. (A) HeLa cells were transfected with empty vector or Sp110b and irrelevant siRNA (negative control, NC) or anti-Sp110b siRNA plasmid (si110b). RNA was harvested 48 h after transfection and analyzed by Northern blotting for Sp110b. (B) P3HR-1 cells were transfected with either Sp110b siRNA plasmid (si110b) or pSilencer2 NC and treated with TPA and butyrate (TB) or mock treated (C). RNA was harvested 48 h after transfection and analyzed for Sp110b expression by Northern blotting. (C) RNA from P3HR-1 cells harvested as in (B) above were analyzed for BMRF1 mRNA expression by Northern blotting. The blots shown in all three panels were stripped and reprobed for glyceraldehhyde-3-phosphate dehydrogenase (GAPDH) as a loading control (lower panels).

EBV-infected P3HR1 cells were then transfected with si110b or pSilencer2.0 NC, and lytic replication of EBV was induced by addition of TPA and butyrate. The cells were harvested 40 h after lytic induction, and Northern blot analysis was performed with probes specific for Sp110b and BMRF1 mRNAs. Both Sp110b and BMRF1 expression levels were decreased in cells transfected with the Sp110b siRNA compared to levels seen in cells transfected with the control siRNA plasmid (Fig. 9B). These data further confirm, in a biologically relevant manner, that the interaction between SM and Sp110b may be important for increased stabilization of viral mRNAs and therefore suggest that in an EBV-infected cell undergoing lytic viral replication, Sp110b enhances SM's posttranscriptional activation of viral genes and may be important for EBV gene expression and virion production.

DISCUSSION

In this study, we demonstrate that SM induces expression of the human protein Sp110b and that SM interacts directly with Sp110b in vitro and in vivo. In addition, two distinct sites of Sp110b that bind SM were mapped. Binding of Sp110b to SM was found to be resistant to RNase treatment and therefore likely to be due to a direct protein-protein interaction. We also found that SM specifically induces expression of Sp110b in the cell types that EBV infects. Exogenous expression of both Sp110b and SM synergistically enhanced SM-mediated accumulation of intronless reporter transcripts and lytic EBV transcripts that are stabilized by SM. This synergistic effect was shown to be posttranscriptional, resulting in increased amounts of target transcript in both the nucleus and cytoplasm. Finally, we found that Sp110b is induced during EBV replication and that inhibiting expression of Sp110b decreased expression of an early lytic viral transcript (BMRF1) that had previously been shown to be SM dependent. The findings of this study suggest that Sp110b is part of the cellular machinery which EBV utilizes to enhance expression of lytic viral transcripts that are otherwise poorly expressed (22, 29). The findings also imply that by inducing and binding Sp110b, SM increases the posttranscriptional stability of lytic EBV transcripts.

Several proteins in the PML nuclear body, including Sp110, are upregulated by interferon expression (4, 20, 28). Herpesviruses, including EBV, have been shown to cause redistribution or degradation of PML nuclear body proteins upon infection or during lytic reactivation (3, 8, 13, 16, 31). These findings have led to the suggestion that PML nuclear bodies play an antiviral role and that disruption of the PML nuclear body may be required for successful virus replication. PML protein has been reported to mediate some of the antiviral effects of interferon against herpes simplex virus (9). It has also been suggested that some components of the PML nuclear body may act as transcriptional repressors, and viral disruption of the PML nuclear body is necessary for progression of the lytic cycle. Consistent with such a model, several viral proteins, such as herpes simplex virus ICP0, cytomegalovirus IE1, and EBV BZLF1, have been implicated in specifically mediating PML nuclear body disruption (1, 16, 26). In the case of ICP0, the viral protein leads to proteasome-mediated degradation of both PML and Sp100 (10, 15). Although it has been argued that the effect of BZLF1 on PML nuclear bodies is due to overexpression of BZLF1, EBV lytic replication is nevertheless associated with loss of Sp100, hDaxx, NDP55, and PML itself, albeit more slowly, from the PML nuclear body (3).

An alternative but not mutually exclusive hypothesis has been proposed that components of the PML nuclear body facilitate viral transcription and DNA replication. The PML nuclear body has been shown to be an important site for immediate-early CMV transcription, and both herpes simplex virus and CMV replication compartments are associated with the periphery of ND10 (23, 32). Transfection of Kaposi's sarcoma-associated herpesvirus (human herpesvirus 8) genes results in the assembly of replication compartments associated with PML nuclear bodies (47). Although the PML nuclear body is disrupted upon EBV lytic replication, replicating EBV compartments in the nucleus nevertheless contain retained or recruited PML protein (3). Consistent with this second view of the PML nuclear body as a nexus for initial viral transcription is the finding that CMV pp71 may anchor viral DNA to Daxx in the PML nuclear body (24). Prior to this study, SM had been shown to bind nuclear shuttling proteins such as CRM1 (7) and REF/Aly (21), but other cellular SM-interacting proteins had not been identified. Identification of the nuclear protein Sp110b as an SM-interacting protein indicates that SM may bind proteins in the PML nuclear body in order to enhance expression of lytic viral transcripts.

Although Sp110b has not been studied extensively, previous work implicates the longer splicing isoform Sp110 as a transcriptional regulator (4). Sp110 contains a potential nuclear hormone-binding domain and acts as a coactivator of glucocorticoid and all-trans-retinoic acid-responsive promoters (4). Sp110b, although otherwise identical to Sp110, lacks the plant homeobox domain and bromodomain found in Sp110. A recent study indicates that Sp110b may act as a transcriptional corepressor of the retinoic acid receptor (46). In contrast, the effects of Sp110b on mRNA stability in the presence of SM are posttranscriptional. Furthermore, enhancement of transcript expression by Sp110b appears to be completely SM dependent, suggesting that Sp110b does not have intrinsic RNA binding ability. Rather, the data suggest that SM recruits Sp110b to mRNA and that Sp110b further enhances mRNA stability, possibly by protecting against nucleolytic degradation. In Hep-2 cells infected with an adenovirus vector encoding Sp110, Sp110 localized near the nuclear membrane, but in the presence of Sp140, Sp110 also localized to nuclear bodies. Therefore, Sp110 has been hypothesized to be a potential bridge between the nuclear body and the nuclear membrane (4). The possibility therefore exists that Sp110b may also have effects on nuclear mRNA export, and this aspect of its interaction with SM remains to be studied further.

Sp110 and Sp110b are induced by treatment with type I interferons, as are many genes encoding PML nuclear body proteins (18, 20, 28). It has generally been assumed that induction of interferon-stimulated genes during viral infection is a host cellular defense response. However, induction of interferon-stimulated genes during viral infection is frequently not merely a secondary consequence of interferon production by the host cell. Rather, several viruses appear to induce interferon-stimulated genes directly by distinct mechanisms mediated by one or more viral proteins. For example, CMV gB induces the expression of interferon-stimulated genes via IRF-3, and SM has been shown to induce the expression of several interferon-stimulated genes by upregulating expression of STAT1, particularly the STAT1β isoform (35, 38). These findings have led to the model that some interferon-stimulated genes may have functions that have been subverted by infecting viruses to enhance virus replication.

The interferon-stimulated gene product viperin inhibits human CMV replication when it is induced by treatment of cells with interferon prior to CMV infection. However, when induced by CMV infection, viperin is redistributed from the endoplasmic reticulum, where it normally localizes, to the Golgi apparatus and no longer inhibits CMV replication (11). Viperin was also found at sites of CMV replication, leading to the suggestion that CMV may utilize viperin to enhance its replication. The findings that SM induces Sp110b and that EBV lytic gene expression is enhanced by Sp110b support the model that induction of some interferon-stimulated genes and the association of replicating DNA viruses with PML nuclear bodies is an evolutionary adaptation that facilitates viral replication.