|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 279, Issue 31, 33001-33011, July 30, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The Evolutionarily Conserved Kaposi's Sarcoma-associated Herpesvirus ORF57 Protein Interacts with REF Protein and Acts as an RNA Export Factor*
From the Division of Virology, Institute of Biomedical and Life Sciences, University of Glasgow, Church Street, Glasgow, G11 5JR, Scotland, United Kingdom
Received for publication, December 1, 2003 , and in revised form, May 13, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
ORF57 (MTA) one of the earliest Kaposi's sarcoma-associated herpesvirus (KSHV) regulatory proteins to be expressed is essential for virus lytic replication. A counterpart is present in every herpesvirus sequenced, indicating the importance of this signature viral protein and those examined act post-transcriptionally, affecting RNA splicing and transport. In KSHV-infected cells, ORF57 protein was present in a complex with REF (Aly) and TAP (NXF1), factors involved in cellular mRNA export. The ORF57 N-terminal region interacts with REF, whereas both N- and C-terminal domains of REF interact with ORF57. The ORF57-REF interaction was direct, whereas TAP appeared to be recruited via REF. In somatic cells, ectopically expressed ORF57 protein was shown to function as a CRM1-independent nuclear mRNA export factor, promoting export of mRNAs that are poor substrates for splicing. The
-herpesvirus ORF57 protein, and its 
-1 herpesvirus ICP27 counterpart both export RNA through pathways involving REF and TAP proteins, although divergence of these herpesvirus subfamilies occurred some 180-210 million years ago. The TAP-mediated cellular mRNA export pathway is CRM1-independent. However, human immunodeficiency virus type 1 Rev protein-mediated RNA export, which is CRM1-dependent, was considerably inhibited by ORF57, suggesting that Rev and ORF57 compete for a common export component. These data strengthen arguments that TAP and CRM1 pathways converge in accessing similar components of the nuclear pore complex. We propose that ORF57-mediated RNA export may use different export factors to accommodate the KSHV-infected host cell environments, for example, in B-cells or endothelial cells and during the different phases of lytic virus replication. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Kaposi's sarcoma (KS)1-associated herpesvirus (KSHV, also known as human herpesvirus 8), the most recently identified human herpesvirus (1), is associated with KS, primary effusion lymphoma, and multicentric Castleman's disease (reviewed in Refs. 2 and 3). KSHV, in the same
-herpesvirus subfamily as Epstein-Barr virus (EBV) and herpesvirus saimiri, is capable of establishing both lytic and latent replication cycles (reviewed in Refs. 4 and 5). In KS, the virus localizes to tumor progenitor endothelial cells (6), most of which are latently infected (7). In cell culture, KSHV replication is generally studied using B-cell lines, such as BCBL-1, generated from primary effusion lymphoma material (8). Most BCBL-1 cells are latently infected, although there is some spontaneous virus reactivation. Addition of the phorbol ester 12-O-tetradecanoylphorbol-13-acetate (TPA) to these cells efficiently induces the lytic cycle and produces virions (8).
ORF50 (also known as RTA) and ORF57 (also known as MTA) proteins are the earliest KSHV regulatory genes to be expressed in the lytic cycle (9, 10) and essential for this phase of virus replication (11). A counterpart to ORF57 is present in every herpesvirus sequenced, indicating the importance of this signature viral protein. These protein homologues act post-transcriptionally affecting RNA splicing and transport, and they include EBV MTA (12, 13), herpesvirus saimiri ORF57 (14), and herpes simplex virus type 1 (HSV-1) ICP27 (15). ICP27 protein, studied in some detail, acts at transcriptional (16) and post-transcriptional levels, influencing pre-mRNA processing (17) and promoting nuclear export of viral RNAs (18, 19). ICP27 has been shown to bind RNA targets that are widely distributed throughout the HSV-1 genome (20). Like ICP27, KSHV ORF57 protein predominantly localizes to the nucleus (21-23) and can shuttle between the nucleus and cytoplasm (23). A post-transcriptional mode of ORF57 action was suggested as its expression increased the accumulation of cytoplasmic mRNA (21, 22).
The vast majority of HSV-1 genes are unspliced and ICP27 protein inhibits pre-mRNA splicing (24-26). By contrast, although transcript mapping is not yet complete, the KSHV genome contains several spliced genes (reviewed in Refs. 4 and 27), including the ORF57 gene itself (23). Furthermore, because expression of ORF57 gave only a slight decrease in reporter gene activity from intron-containing reporter constructs (21, 22), in virus-infected cells ORF57 may not profoundly inhibit pre-mRNA splicing. It appears that the ORF57 herpesvirus homologues may have shared and separate features; for example, EBV MTA protein only partially complemented ICP27 when inserted into an ICP27-null virus (28).
Several nuclear replicating viruses, including herpesviruses and human immunodeficiency virus type 1 (HIV-1), encode proteins that promote nuclear export of viral mRNAs and that also may inhibit export of cellular RNAs (reviewed in Ref. 29). ICP27 protein stimulates the export of intronless HSV-1 mRNAs in Xenopus laevis oocytes via REF (also known as Aly) and TAP (also known as NXF1) proteins that are involved in cellular mRNA export (19). In HSV-1-infected cells, ICP27 protein interacts directly with REF and is present in a complex with TAP (18, 19). HIV-1 Rev protein promotes the nuclear export of unspliced and partially spliced viral mRNAs (30-32) via CRM1, a member of the karyopherin (or importin/exportin) family (33). A Rev region containing a leucine-rich nuclear export signal (NES) directly interacts with CRM1 that, in turn, interacts with components of the nuclear pore complex (NPC) to promote RNA export with Ran GTPase as a cofactor (reviewed in Ref. 34). CRM1 function is specifically inhibited by leptomycin B (LMB) that also inhibits Rev-mediated HIV-1 RNA export (33, 35). Export of ribonucleoprotein (RNP) complexes containing cellular mRNAs is mediated by nuclear transport receptors that bind their cargoes and interact directly with components of the NPC (NPC reviewed in Ref. 36).
REF belongs to a conserved superfamily of RNA-binding proteins that mediate the association between mRNPs and nuclear transport receptors (37, 38). Members of this superfamily are distinguished by conserved motifs at their N and C termini (Fig. 2A), between which a RNP-type RNA-binding domain (RBD) (39) and variable regions are located (39, 40). In the mouse, REF is encoded by at least three different genes (39-41) that differ at multiple positions in the variable regions and there are multiple spliced variants (40). Murine REF1-II is generated by alternative splicing of REF1-I (42) and lacks the N-terminal variable region present in REF1-I and REF2-I (Fig. 2A), whereas murine REF2-I and REF2-II differ by a single amino acid insertion in REF2-I (40). REF1-II and REF2-I proteins exhibit 95% identity in their RBD (40) but differ at multiple positions within the C-vr region. REF recruits the export receptor TAP protein to cellular mRNPs (38, 40) forming part of the exon junction complex recruited to mRNAs during splicing (37, 43). Independent of splicing, REF can also associate with intronless mRNAs, influenced by both mRNA sequence and length (44).
|
In metazoan cells, the mRNP nuclear transport receptor consists of a large subunit (TAP) and a small subunit (p15). TAP, which itself facilitates the nuclear export of cellular mRNA (45, 46), heterodimerizes with p15 that further stimulates mRNP export (47, 48). TAP proteins represent a class of nuclear transport receptors that, via an essential C-terminal domain, directly interact with nucleoporin components of the NPC (49-51). The interaction of TAP with cellular RNAs is likely mediated by adaptor proteins such as REF, through protein-protein interactions (38, 40, 44). However, recent studies of Drosophila melanogaster cells and Caenorhabditis elegans indicate that while REF and other exon junction complex components contribute to mRNA export (52, 53), they appear to be dispensable, with additional adaptors bridging the interaction between TAP and cellular mRNAs (53). TAP is also required for the export of unspliced simian type D retroviral RNAs, and binds directly to the constitutive transport element RNA sequence (54, 55). TAP-mediated cellular mRNA export and constitutive transport element-mediated RNA export do not require any karyopherin or Ran GTPase and are not inhibited by LMB (54, 56).
We show that in KSHV-infected BCBL-1 cells, ORF57 protein was present in a complex with both REF and TAP proteins. The N-terminal region of ORF57 interacts with REF, whereas the N- and C-terminal domains of REF interact with ORF57. The interaction of ORF57 and REF was direct, with TAP being recruited via REF. In somatic cells, ORF57 was shown directly to function as a nuclear mRNA export factor using an assay that allows quantification of ORF57-mediated stimulation of RNA nuclear export in cultured cells. We show that overexpression of ORF57 bypasses nuclear retention and stimulates the nuclear export of unspliced reporter mRNAs that are otherwise exported inefficiently. The ORF57-mediated RNA export was CRM1-independent. These data extend what is known of this herpesvirus family of mRNA export factors, and we suggest that ORF57 protein utilizes more than one nuclear export route, perhaps in different cell types and at different stages of lytic viral infection.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
Plasmids—Plasmid pGST-ORF57 expressing an N-terminal glutathione S-transferase (GST) fusion of ORF57 amino acids (aa) 1-455 full-length (FL) protein driven by a tac promoter was constructed by cloning ORF57 cDNA into pGEX-5X-3 (Amersham Biosciences). ORF57 sequences were excised from pKS4 (kind gift of Dr. L. Bello) containing ORF57 FL cDNA cloned into the EcoRI site of pEGFP-C1 (Clontech). Plasmids pGBKT7-ORF57 FL and pGBKT7-ORF57 small (containing the ORF57 gene 2nd exon, aa 181-455) were used for in vitro transcription/translation of the ORF57 protein. ORF57 FL and 2nd exon DNA sequences were excised as XhoI and PstI fragments from pKS4 and pKS1 (23), respectively, and cloned into SalI and PstI sites of pGBKT7 (Clontech), resulting in loss of both XhoI and SalI sites from pGBKT7-ORF57 FL and small. Plasmids pcDNA4-gORF57 (FL) and pcDNA4-cORF57 (FL) contain full-length genomic and cDNA sequences, respectively, driven by a human cytomegalovirus (HCMV) promoter encoding the ORF57 (aa 1-455) protein. ORF57 genomic and cDNA sequences were PCR amplified with primers containing BamHI and XhoI sites using pKS3 (23) and pKS4 as templates, respectively, and cloned into pcDNA4/HisMax B (Invitrogen). Deletion mutants of ORF57 FL were generated by PCR from pKS4 using primers with BamHI and XhoI sites and cloned into pcDNA4/HisMax C. They encoded ORF57 aa 17-455, 1-215, 181-328, 329-455, and 387-455; identities were confirmed by DNA sequencing of both strands. Plasmids GFP-gORF57 FL (pKS3) and GFP-ORF57 small (PKS1) were as described (23).
Plasmid pCMV-HA-RevM10-ORF57 was generated by cloning ORF57 cDNA into XhoI-SalI sites of pCMV-RevM10 (48) to express ORF57 as a C-terminal fusion with the HIV-1 RevM10 mutant. Plasmids pGST-REF2-I and its deletion mutants expressing GST fusions of REF2-I, as described previously (19), were kindly provided by Dr. S. A. Wilson. Plasmid pCMV128 containing the chloramphenicol acetyltransferase (CAT) coding region within an intron sequence was as described (57, 58). Recombinant plasmids pGEXCS-REF1-II, pRSETB-REF1-II (and their mutants) expressing GST, and histidine (His) fusions, respectively, of REF1-II (40, 44), pCMV-HA-Rev, and pCMV-HA-RevM10 expressing HIV-1 Rev and export deficient mutant RevM10 proteins (48) were kindly provided by Dr. E. Izaurralde. Plasmids pCMV-HARev and pCMV-HA-RevM10 (48) contain an N-terminal hemagglutinin (HA) tag for detection of Rev, RevM10, and RevM10-ORF57 fusion proteins. Plasmids pCH110 (Amersham Biosciences) encoding 
-galactosidase, pEGFP-C1 (Clontech), and pcDNA3.1 (Invitrogen) were used in transfections.
Cell Culture and Antibodies—BCBL-1 cells from a KSHV-positive, EBV-negative primary effusion lymphoma (8) were grown as described (59), and the KSHV-negative and EBV-negative B-cell lymphoma cell line BJAB was grown in RPMI 1640 medium supplemented with 10% fetal calf serum and 2 mM L-glutamine. Human embryonic kidney 293 epithelial cells were grown in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal calf serum, 2 mM L-glutamine, 1% (v/v) non-essential amino acids, 100 units/ml penicillin, and 0.01% streptomycin.
A rabbit antibody against a synthetic peptide corresponding to KSHV ORF57 aa 182-195 (DGESPRFDDSIIPR), kindly provided by Dr. G. Hayward and termed anti-ORF57 (GH) Ab, was used for Western blotting at 1:2,500 dilution. Rabbit polyclonal antibodies to murine REF (KJ70, raised against the RBD of recombinant REF1-II) (44) and to TAP (KJ60) (60), kindly provided by Dr. E. Izaurralde, were used for Western blotting at 1:2,000 dilution. Western immunoblot analysis was performed with anti-GFP mouse mAb (Clontech, 1:1,000 dilution) and anti-HA mouse mAb (Sigma, 1:1,000) with the ECL detection system (Amersham Biosciences).
Chemical Induction of the KSHV Lytic Cycle and Preparation of Cell Extracts—BCBL-1 cells (0.2 x 106 cells/ml) were treated with TPA (20 ng/ml) (8) or left untreated for 72 h. Soluble protein extracts were prepared essentially as described (61) in 800 µl of lysis buffer (50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 1 mM EDTA, 0.1% Nonidet P-40, 0.5 mM phenylmethylsulfonyl fluoride) containing protease inhibitor mixture (Roche Diagnostics) and passed five times through a 26-gauge needle. As appropriate, cell extracts were treated with 10 units of RNase (ONETM Ribonuclease, Promega) at 37 °C for 15 min.
Recombinant Protein Expression in Escherichia coli—Expression of GST-REF (44) and GST-TAP (60) proteins was performed as described. His-tagged REF2-I was purified as described (19). GST-ORF57 fusion protein was prepared as described for GST-heterogeneous nuclear RNP K (61) with modifications: E. coli BL21 cells (500 ml) were induced with 1.0 mM isopropyl--D-thiogalactopyranoside for 16 h at 28 °C and resuspended in 5 ml of modified lysis buffer (20 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1 mM EDTA, 0.1% Nonidet P-40, 10% (w/v) glycerol, 5 mM -mercaptoethanol) with 100 µg/ml lysozyme and sonicated on ice. Triton X-100 was added to 2% (v/v), extracts were kept at 4 °C for 30 min, and GST-ORF57 fusion protein was purified using glutathione-Sepharose 4B beads.
In Vitro Pull-down Assays—Pull-down assays (19, 61) were performed using TPA-treated and untreated BCBL-1 extracts (
200 µg of protein) or [35S]methionine-labeled ORF57 (FL, small, or deletion mutants) protein (10 µl) synthesized in vitro using the TNT T7 Quick-coupled transcription/translation system (Promega) according to the manufacturer's instructions plus 5 µg of GST-REF2-I fusion protein or GST alone bound to 20 µl of glutathione-agarose beads. In some experiments, 5 µg of GST-ORF57, GST-REF1-II fusion proteins, or RNase were added. Bound proteins were eluted using 20 µl of elution buffer (50 mM Tris-HCl, pH 8.0, 20 mM glutathione), resolved by SDS-PAGE, and visualized by autoradiography or Western blotting (61) with the ECL detection system (Amersham Biosciences).
Immunoprecipitation Assays—Using in vitro synthesized radiolabeled proteins, 10 µl of 35S-labeled ORF57 and purified recombinant GST-TAP or GST proteins were incubated for 1 h at 4 °C with 5 µl of rabbit anti-TAP antibody or rabbit preimmune serum Ab (purified IgG) in 200 µl of in vitro immunoprecipitation buffer (50 mM Tris-HCl, pH 8.0, 100 mM NaCl, 1 mM EDTA, 2 mM EGTA, 0.1% Nonidet P-40, 1% bovine serum albumin, 0.5 mM phenylmethylsulfonyl fluoride) containing protease inhibitor mixture, as appropriate 5 µl of RNase was added. Then, 75 µl of a 50% slurry in immunoprecipitation buffer of protein A-protein G (50:50 ratio) Sepharose beads (Amersham Biosciences) was added and the mixture was incubated further for 1 h at 4 °C. The beads were washed with cold immunoprecipitation buffer, bound proteins were eluted, resolved by SDS-PAGE, and visualized by autoradiography (61).
For immunoprecipitations with BCBL-1 cell extracts, 200 µg of protein were pre-cleared with 5 µl of rabbit preimmune serum Ab (purified IgG) and 100 µl of a 50% slurry of protein A-Sepharose beads for 1 h at 4 °C. Then, 6 µl of anti-ORF57 (GH) Ab or control rabbit preimmune serum Ab (purified IgG) was added in 300 µl of modified buffer E (10 mM Tris-HCl, pH 8.0, 100 mM NaCl, 5% glycerol, 1 mM EDTA, 2 mM EGTA, 0.1% Nonidet P-40, 0.5% sodium deoxycholate, 0.5 mM phenylmethylsulfonyl fluoride) with the protease inhibitor mixture (19). Immunoprecipitated proteins were resolved by SDS-PAGE and visualized by Western blotting (61).
DNA Transfections, RNA Export Assays, and RNase Protection Assay—DNA transfections and CAT export assays were performed essentially as described (48, 62) with the following modifications. Subconfluent 293 cells (6 x 105 cells/well) were transfected using Polyfect (Qiagen) according to the manufacturer's instructions. As appropriate, the DNA mixture for transfection included 0.25 µg of pCMV128 reporter plasmid and 1.25 µg of GFP-gORF57 (FL or small), pCMVRevM10-ORF57, pCMV-RevM10, and 0.5 µg of pCMV-Rev. Transfection efficiency was determined by including 0.5 µg of pCH110 plasmid DNA encoding -galactosidase. The total amount of DNA transfected was brought to 2 µg with appropriate amounts of plasmid DNAs without inserts. The cells were harvested 48 h after transfection and CAT activity was measured as described previously (48, 62). -Galactosidase activity was measured using the -galactosidase reporter gene system (Promega) following the manufacturer's instructions. To compensate for any differences in transfection efficiencies between experiments, CAT activities were normalized with respect to the -galactosidase internal control (48, 62). Standard deviations of the CAT activity values from three separate experiments are shown as error bars. As appropriate, at 12 h post-transfection, medium containing 5 nM LMB (Sigma) or an equivalent amount of ethanol used to solubilize the LMB was added, cells were harvested after a further 12 h, and -galactosidase and CAT activities were measured.
RNase protection assays were performed as described (77). Briefly, total RNA was isolated from transfected cells with TRIzol reagent (Invitrogen). To isolate cytoplasmic RNA, cells were lysed in 130 mM NaCl, 5 mM KCl, 25 mM Tris, 0.2% Nonidet P-40, 0.1% sodium deoxycholate, 0.002% dextran sulfate. Following centrifugation, the supernatant was extracted twice with phenol-chloroform, and the RNA was precipitated. Antisense riboprobes radiolabeled with [32P]CTP encompassing nucleotides 291-411 and 1-100 of the -galactosidase and CAT open reading frames, respectively, were generated by in vitro transcription. Following DNase digestion, the probes were purified on denaturing 6% acrylamide, 7 M urea gels, eluted, precipitated, and resuspended in 100% formamide at 3.0 x 105 cpm/µl.
The riboprobes (1 µl each) were mixed with 10 µg of cellular RNA and 40 µg of yeast total RNA in a final volume of 50 µl of hybridization buffer consisting of 40 mM Pipes, pH 6.4, 400 mM NaCl, 1 mM EDTA, and 80% formamide. The RNA mixtures were denatured by incubation at 85 °C for 5 min. Hybridizations were carried out overnight at 50 °C, then 300 µl of RNase digestion buffer (10 mM Tris, pH 8.0, 5 mM EDTA, 300 mM NaCl) was added to each sample. RNase digestions were for 45 min at 37 °C with 2 units of RNase T1 (Roche Diagnostics) and 12 µg of RNase A per sample. Digestions were stopped by adding proteinase K (50 µg per sample) and SDS to a final concentration of 0.5%. Samples were further incubated for 15 min at 37 °C, extracted with phenol-chloroform, precipitated, resuspended in 80% formamide, and analyzed on a 6% acrylamide, 7 M urea denaturing gel, followed by autoradiography.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
ORF57 Protein Interacts with the Cellular RNA Export Factor REF—As some herpesvirus homologues of ORF57 interact with REF protein isoforms, we examined for a direct interaction between ORF57 and REF protein using a pull-down assay. Glutathione beads containing purified recombinant GSTREF2-I fusion protein or GST alone (see Fig. 2B, lanes 2 and 9) were incubated with in vitro synthesized 35S-labeled ORF57 FL (aa 1-455) or ORF57 small (aa 181-455) (Fig. 1A, lanes 1 and 2, respectively), and pull-down assays were performed. Bound proteins, eluted from resin, were separated by SDS-PAGE and visualized by autoradiography. GST-REF2-I pulled down both ORF57 FL and ORF57 small, whereas GST alone did not (Fig. 1A, compare lanes 5 and 6 with lanes 3 and 4). Expression of ORF57 following TPA treatment was examined; Western blotting of BCBL-1 cell extracts using anti-ORF57 (GH) Ab showed a readily detectable
50-kDa ORF57 band at 24 h post-induction in TPA-treated cell extracts (data not shown), and at 72 h a faster migrating ORF57 processed product (45-kDa) was also present (Fig. 1B, lane 2); these protein bands were absent in untreated BCBL-1 cell extracts (Fig. 1B, lane 1).
|
To examine ORF57 interaction with REF, the pull-down assay was performed with BCBL-1 cell extracts and GSTREF2-I fusion protein, followed by Western blotting using anti-ORF57 Ab. ORF57 protein was present in input TPA-treated cell extracts but absent in untreated BCBL-1 cell extracts (Fig. 1C, compare lane 2 with lane 1). GST-REF2-I pulled down a
50-kDa ORF57 band from TPA-treated cell extracts but not from untreated extracts (Fig. 1C, lane 3 upper band, compare with lane 4), and the GST controls were negative (Fig. 1C, lanes 5 and 6). The band below ORF57 (Fig. 1C, lanes 3 and 4) is a GST-REF2-I species that cross-reacts with anti-ORF57 Ab (Fig. 1C, lanes 3 and 4, lower band marked with asterisk); it was found with TPA-treated cell extracts, and with untreated cell extracts with no detectable ORF57 (Fig. 1C, compare input lane 1 with lane 2). Co-immunoprecipitation was used to examine for an interaction between ORF57 and REF in infected BCBL-1 cell extracts. Immunoprecipitates generated with anti-ORF57 Ab were fractionated by SDS-PAGE, transferred to nitrocellulose membranes, and the resulting Western blots were probed using anti-REF Ab. REF was co-immunoprecipitated with ORF57 from TPA-treated cell extracts but not from untreated extracts (Fig. 1D, compare lanes 3 and 4). REF was not detected in immunoprecipitates generated using anti-ORF57 Ab with extracts from KSHV-negative control BJAB cells (Fig. 1D, lane 5) nor in immunoprecipitates generated using a nonspecific control preimmune serum Ab with extracts of TPA-treated and untreated BCBL-1 cells (Fig. 1D, lanes 6 and 7) indicating specificity of the immunoprecipitation assay. The presence of ORF57 could not be demonstrated by Western blotting of immunoprecipitates obtained with anti-ORF57 and anti-REF Ab, as the IgG heavy chain of the available rabbit Abs is of similar size to the ORF57 protein, and gave a strong masking signal on Western blots using rabbit anti-ORF57 Ab (data not shown). The IgG heavy chain that masks the ORF57 protein band can be seen in Fig. 3A, lanes 2-4.
|
Mapping the Regions of REF and ORF57 Proteins Required for Their Interaction—To map the REF protein (see Fig. 2A) regions required for interaction with ORF57, GST-REF2-I FL (aa 1-218) and its deletion mutants were visualized by Coomassie Blue staining (Fig. 2B) and then equal amounts of each GST-REF2-I fusion protein were used with 10 µl of in vitro synthesized 35S-labeled ORF57 FL (Fig. 2C, lane 1) in pull-down assays. Fusion proteins GST-REF2-I FL (Fig. 2C, lane 9) and mutants that contained REF aa 1-74 (Fig. 2C, lane 8), aa 1-152 (Fig. 2C, lane 6), aa 1-198 (Fig. 2C, lane 5), aa 74-218 (Fig. 2C, lane 3), and aa 153-218 (Fig. 2C, lane 4) bound labeled ORF57, whereas REF aa 75-152 did not (Fig. 2C, lane 7). No ORF57 binding was detected with GST alone (Fig. 2C, lane 2). RNase treatment (Fig. 2D, lanes 5 and 7) did not abolish the interaction of in vitro synthesized 35S-labeled ORF57 FL with either REF2-I FL (aa 1-218) or REF2-I (aa 1-198) making it unlikely that these interactions are tethered by RNA. No ORF57 binding was detected with GST alone (Fig. 2D, lanes 2 and 3). Thus REF regions at the N and C termini, located between aa 1-73 and 153-218, interact with ORF57 with the C-terminal region showing stronger binding, whereas the central RBD showed no interaction.
REF protein isoforms REF2-I and REF1-II exhibit 95% identity in their RBD (40) but differ at multiple positions within the C-vr region, and REF1-II protein lacks the N-vr region present in REF2-I (see Fig. 2A). To examine the specificity of ORF57 binding to a spliced isoform of REFs, an equivalent amount of 35S-labeled ORF57 FL protein (Fig. 2E, lane 1) was used in pull-down assays with GST fusion proteins of REF1-II FL (aa 1-163) or REF1-II fragments aa 14-102 and 103-163. Similar to the interaction with REF2-I (Fig. 2C), 35S-labeled ORF57 bound to full-length REF1-II (Fig. 2E, lane 3) and C-terminal aa 103-163 (Fig. 2E, lane 4), whereas ORF57 showed no interaction with the REF1-II RBD domain (aa 14-102) (Fig. 2E, lane 2). No ORF57 binding was detected with GST alone (Fig. 2E, lane 5).
The ORF57 protein regions required for interaction with REF were mapped by pull-down assays using equal amounts of input GST-REF2-I FL fusion protein, with 35S-labeled ORF57 FL (aa 1-455) or ORF57 fragments containing aa 17-455, 1-215, 181-328, 329-455, and 387-455. The 35S-labeled ORF57 deletion mutant proteins (Fig. 2F) were translated in vitro (Fig. 2, G and H), and equally adjusted input levels of these were used in binding assays. GST-REF2-I bound ORF57 (FL) and deletion mutants containing ORF57 aa 17-455, 1-215, and 181-328 (Fig. 2I, lanes 1, 2, 4, and 5), whereas C-terminal fragments aa 329-455 and 387-455 showed no detectable binding (Fig. 2I, lanes 3 and 6). Thus, ORF57 aa 181-215 was the minimum region required for interaction with REF. However, ORF57 sequences closer to the N terminus (aa 17-180) also appear to contribute to binding as a stronger REF interaction was observed with ORF57 aa 1-215 (Fig. 2I, compare lanes 4 and 5). This difference in avidity also was demonstrated by the relatively weak interaction observed between REF2-I and ORF57 small (lacking the N-terminal 180 aa) compared with that with ORF57 FL (see Fig. 1A, compare lanes 7 and 8). In this experiment, input ORF57 small and ORF57 FL did not differ substantially in their relative radiolabeling (see Fig. 1A, compare lanes 1 and 2).
ORF57 Is Present in a Complex That Includes REF and TAP Proteins—Because REF proteins interact with the export receptor TAP/NXF1 that targets mRNPs to the NPC (38, 40, 44), we tested for the presence of TAP protein in the complex with ORF57. We performed Western blot analysis with anti-TAP Ab on immunoprecipitation samples obtained using anti-ORF57 Ab (shown in Fig. 1D). An ORF57-TAP interaction was identified with TPA-treated cell extracts but not with untreated extracts (Fig. 3A, compare lanes 3 and 2), and RNase treatment had no effect on the interaction (Fig. 3A, lane 4). To confirm the presence of TAP in a complex with ORF57, extracts of BCBL-1 cells were also examined using the GST-ORF57 fusion protein in a pull-down assay. GST-ORF57 bound TAP present in both TPA-treated and untreated cell extracts (Fig. 3B, lanes 2 and 3), whereas GST alone did not (Fig. 3B, lanes 4 and 5). Thus TAP was present in a complex with ORF57 and REF proteins. As expected from Figs. 1 and 2, REF protein from both TPA-treated and untreated BCBL-1 cell extracts bound to recombinant GST-ORF57 (Fig. 3C, compare lanes 3 and 4), whereas GST alone did not bind REF from treated cell extracts (Fig. 3C, lane 2).
To investigate whether TAP binds directly or indirectly to ORF57, the ability of 35S-labeled ORF57 FL synthesized in vitro (Fig. 3D, lane 6) to bind GST-TAP or GST alone bound to glutathione beads, in the presence or absence of purified REF protein, was examined. A weak interaction between ORF57 and GST-TAP was observed (Fig. 3D, lane 5), indicating that TAP might directly bind ORF57. However, incubation of beads with 5 µg of recombinant His-REF2-I prior to the pull-down considerably enhanced the ORF57-TAP interaction (Fig. 3D, lane 2), RNase treatment did not disrupt complex formation (Fig. 3D, lane 3). As expected, recombinant His-REF2-I alone pulled down labeled ORF57 (Fig. 3D, lane 1), whereas GST alone beads incubated with purified His-REF1-II did not bind ORF57 (Fig. 3D, lane 4). Thus the ORF57 and TAP interaction in vitro is strengthened by the addition of REF protein.
To examine further the ORF57 and TAP interaction by in vitro immunoprecipitation, equivalent amounts of purified GST-TAP or GST proteins were mixed with 35S-labeled ORF57 FL and incubated with 5 µl of anti-TAP, anti-ORF57 (GH) Ab, or preimmune serum Ab and immunoprecipitations were performed. ORF57 protein was immunoprecipitated with anti-TAP Ab following preincubation with purified recombinant GST-TAP (Fig. 3E, lane 4) or as a positive control with anti-ORF57 (GH) Ab (Fig. 3E, lane 5) but not following preincubation with purified GST (Fig. 3E, lane 3). RNase added at the time of preincubation had no negative effect on binding (Fig. 3E, lane 2) and preimmune serum Ab did not bring down ORF57 following preincubation with GST-TAP (Fig. 3E, lane 6). These data confirmed that ORF57 was capable of forming a complex with TAP protein directly, albeit at low efficiency.
Ectopic Expression of ORF57 Mediates the Nuclear Export of Inefficiently Spliced Pre-mRNA—To investigate the role of ORF57 in mRNA export, a sensitive assay that examines TAP-mediated stimulation of RNA nuclear export in cultured cells with the pCMV128 reporter plasmid was used (48). This plasmid is related to pDM138 but has a CMV promoter instead of an SV40 promoter, and harbors the CAT coding sequence inserted into an intron, which is not efficiently spliced out (see Fig. 4F) (57, 58). Cells transfected with pCMV128 reporter plasmid alone yield only trace levels of CAT enzyme activity as the unspliced pre-mRNA is retained in the nucleus (57, 58). However, co-transfection of pCMV128 reporter with the recombinant plasmids expressing proteins that bypass nuclear retention, promotes nuclear export of the inefficiently spliced pre-mRNA, and leads to an increase in CAT activity. ORF57 protein was tested using this assay. Plasmid GFP-gORF57 FL DNA, which allows ORF57 expression to be detected with anti-GFP Ab, was co-transfected into human 293 cells with pCMV128 and -galactosidase reporter DNAs. Expression of ORF57 FL resulted in a 6-fold increase in CAT enzyme activity in comparison to the value obtained with pEGFP-C1 empty vector (Fig. 4A, compare bars 1 and 2). Figs. 1A and 2I showed that although ORF57 aa 181-215 was the minimum region required for interaction with REF, sequences closer to the N terminus (aa 17-180) also contributed to REF binding. Thus we examined the contribution of the ORF57 N terminus (aa 1-180) using an N-terminal deletion mutant that expresses ORF57 aa 181-455. With GFP-ORF57 small although CAT activity increased by 4-fold (Fig. 4A, bar 3), it did not reach the levels obtained by ORF57 (FL) (Fig. 4A, bar 2), demonstrating that the N-terminal domain not only contributes to REF protein binding but also is required for ORF57-mediated mRNA export. The expression of GFP-ORF57 and GFP alone from the transfected plasmids was confirmed by Western blotting with anti-GFP Ab (Fig. 4B).
|
Stimulation of RNA export by a test protein could reflect its capability to bind the target RNA as well as its export activity. To determine the role of ORF57 in mRNA nuclear export when directly targeted to mRNAs, ORF57 was fused to the C terminus of the HIV-1 Rev mutant, RevM10 protein. RevM10 is defective in RNA export because of point mutations in its NES (63) but RevM10 protein does bind a functional HIV-1 Rev response element (RRE), a structured cis-acting RNA target (see Ref. 63, reviewed in Ref. 64) used to target Rev fusion proteins to RRE-bearing RNAs (47, 63). Reporter plasmid pCMV128, used in this assay, harbors an RRE inserted downstream of the CAT coding sequence within the inefficiently spliced intron (see Fig. 4F), thus pCMV128 pre-mRNA contains an RRE to which Rev or export deficient RevM10 mutant protein binds directly (57, 58). A RevM10-ORF57 fusion protein, tethered to pCMV128 cat mRNA via the RRE, would promote the nuclear export of inefficiently spliced RNAs, bypassing nuclear retention and rescuing the export deficiency of RevM10. We observed that expression of RevM10-ORF57 stimulated CAT activity by 8.5-fold as compared with RevM10 alone (Fig. 4C, compare bars 1 and 3). Co-expression of untethered ORF57 FL and RevM10-ORF57 fusion together stimulated CAT activity by some 12-fold (Fig. 4C, bar 5). Co-expression of the pEGFP-C1 vector did not affect export by RevM10-ORF57 (Fig. 4C, bar 4) and RevM10 had little effect on export activity of GFP-gORF57 FL (compare Fig. 4, C, bar 2, with A, bar 2). Western blotting with Ab directed against an HA tag present in RevM10 showed that RevM10 (Fig. 4D, compare lanes 6 and 7) and RevM10-ORF57 protein (Fig. 4D, compare lanes 1, 2, and 3) levels were unaffected by addition of pEGFP-C1 empty vector or pGFP-gORF57.
To confirm that the stimulation of CAT protein expression observed in the assays shown in Fig. 4, A and C, reflects an increased export of the unspliced cat mRNA encoded by pCMV128, we performed an RNase protection assay with cytoplasmic (Fig. 4E, lanes 5-8) and total RNA (Fig. 4E, lanes 1-4) fractions derived from transfected cells. Unspliced cat mRNA (Fig. 4E, upper panel) was detected at low levels in the cytoplasm of cells co-transfected with pCMV128 and the control vectors encoding either GFP or RevM10 alone (Fig. 4E, lanes 5 and 7). In contrast, the cat mRNA was exported to the cytoplasm of cells co-expressing ORF57 fused either to GFP or RevM10 (Fig. 4E, lanes 6 and 8). Whereas the -galactosidase mRNA (Fig. 4E, lower panel, lanes 1-8) from control pCH110 plasmid was detected at similar levels in the cytoplasm of cells co-transfected with pCH110 and the vectors encoding: either GFP or RevM10 alone (Fig. 4E, lanes 5 and 7) or ORF57 fused either to GFP or RevM10 (Fig. 4E, lanes 6 and 8).
ORF57-mediated RNA Export Is CRM1-independent but Shares Components with the Rev Export Pathway—Rev-mediated mRNA export requires its NES to bind the export receptor CRM1 (33, 65), an interaction specifically inhibited by the addition of LMB (35). To determine whether ORF57-mediated nuclear export of mRNAs was dependent on the CRM1 export factor, pCMV128 DNA and pRevM10-ORF57 or pRev/RevM10 expression vectors were co-transfected into 293 cells. At 12 h post-transfection, 5 nM LMB, or an equivalent amount of ethanol used to solubilize LMB, was added and cells were harvested after a further 12 h. With RevM10-ORF57 fusion protein the maximum CAT activity (Fig. 5A, bar 3) showed only a slight decrease in the presence of LMB (Fig. 5A, compare bars 3 and 4), similar to the decrease in export activity of RevM10-TAP (47). While as expected, the maximum cat export ability of the positive control Rev protein (Fig. 5B, bar 3) was inhibited by 80% (Fig. 5B, compare bars 3 and 4), consistent with previous reports (35, 66). A similar decrease in CAT activity to RevM10-ORF57 fusion protein was also seen with the negative control RevM10 (Fig. 5, A and B, compare bars 1 and 2), which is inactive in RNA export. Western blotting with anti-HA Ab showed that the amounts of RevM10-ORF57 fusion (Fig. 5C, lanes 1 and 2) and Rev (Fig. 5C, lanes 3 and 4) proteins expressed were unaffected by addition of LMB (Fig. 5C, lanes 2 and 4).
|
Co-expression of ORF57 (from pGFP-gORF57 FL) with Rev reduced CAT activity to 58% of the level obtained in the absence of ORF57 (Fig. 5B, bar 6), whereas the empty vector pEGFP-C1 had no effect (Fig. 5B, bar 5). Western blotting with anti-HA Ab revealed that Rev protein levels were not altered by the co-expression of ORF57 (Fig. 5C, compare lanes 5 and 6). Thus, co-expression of ORF57 with Rev impairs Rev-mediated RNA export, and the inhibition was observed under conditions in which ORF57 itself stimulated RNA export.
-Galactosidase protein and mRNA levels (data not shown) were unaffected by co-expression of ORF57 with Rev, indicating no overall inhibition of mRNA export. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
In extracts of BCBL-1 cells in which the KSHV lytic cycle was induced by TPA treatment, ORF57 protein forms a complex that includes REF and TAP proteins. The data indicate that ORF57 and REF proteins interact directly, and suggest that TAP is recruited to the complex via its known interaction with REF, as found with the Drosophila UAP56 protein that is essential for mRNA export (67, 68). The N-terminal region of ORF57 was shown to interact with REF as does a similar region of the HSV-1 ICP27 protein (18, 19). The N- and C-terminal domains of REF2-I, which also interact with EBV MTA (69), were required for ORF57 interaction and, in contrast to ICP27 (19), the REF RBD was not required. REF1-II lacks an N-terminal variable region (Fig. 2A) and interacts with ORF57 via its C-terminal domain. Thus REF binding to ORF57 requires one or another of the N- or C-terminal conserved variable regions, known to bind both TAP, RNA, and potentiate REF multimerization (44, 70). However, it is possible that differences in secondary structure and folding of REF recombinants may have occurred as only portions of the full-length REF protein were expressed. It may be that in vivo, with the full-length REF proteins these weaker interactions do not occur and only one region of the correctly folded protein interacts with ORF57. In BCBL-1 cells treated with TPA, ORF57 protein was localized to the nucleus and exhibited a punctate pattern that co-localized with REF protein (data not shown).
The -2 herpesvirus KSHV ORF57 protein and its -1 herpesvirus HSV-1 ICP27 counterpart both export RNA through pathways involving REF and TAP proteins. These cellular proteins are evolutionarily conserved (71). Divergence of the three ancestral branches leading to the -, -, and -herpesvirus subfamilies has been estimated as occurring 180-210 million years ago (72), and prior to the mammalian radiation of 60-80 million years ago. The functional adaptation of the ancestral herpesvirus regulatory protein to interact with ancestral REF and TAP proteins is therefore likely to have preceded divergence of the herpesvirus subfamilies.
The ability of ORF57 protein to export unspliced mRNA was assayed using a construct with suboptimal splice sites flanking the cat reporter gene. Used extensively to study HIV-1 Rev (57, 58), this assay was used to demonstrate that TAP protein increased CAT protein activity thereby inducing the cytoplasmic accumulation of intron-containing RNAs (48, 62). This assay was also used to demonstrate that EBV MTA protein promotes RNA export (73). Expression of ORF57 significantly increased CAT protein activity, bypassing nuclear retention of inefficiently spliced cat mRNA and promoting its export to the cytoplasm. ORF57 small (lacking aa 1-180) stimulated CAT activity, albeit to a reduced level in comparison to ORF57 FL, thus both N- and C-terminal regions of the protein contributed to RNA export.
A RevM10-ORF57 fusion protein was used to tether ORF57 to the RRE-containing cat reporter RNA. RevM10 has been shown to be non-functional and its defective NES function can be complemented by fusion to proteins such as TAP (47, 48), and others containing leucine-rich NESs (74, 75). Using the RevM10-ORF57 fusion protein, enhanced CAT activity was observed as compared with the levels with untethered ORF57 or RevM10 alone. Thus expression of ORF57 as a fusion with RevM10 was able to rescue the export deficient function of RevM10. The higher CAT activity found with RevM10-ORF57 as compared with ORF57 alone is likely because of direct binding of the fusion protein to cat mRNA. Interestingly, co-transfection of plasmids expressing RevM10-ORF57 and ORF57 proteins stimulated CAT activity more than RevM10-ORF57 alone. Augmentation of the export activity of RRE-tethered RevM10-ORF57 could occur by RRE-independent binding of ORF57 to cat mRNA or via ORF57 recruitment by RevM10-ORF57 as ORF57 protein is capable of self-interaction.2 The export activity of ORF57 alone was comparable with that obtained with GFP-TAP alone (48, 62) and activity of RevM10-ORF57 was comparable with that obtained with RevM10-TAP alone; both TAP export activities were dramatically increased by the addition of exogenous p15, a TAP co-factor (48). This suggests the presence of a similarly acting, as yet unidentified, ORF57 co-factor, either viral or cellular, that modulates its export ability during KSHV lytic replication.
LMB that binds to and inactivates CRM1 (33) caused a modest decrease in ORF57 stimulation of CAT activity, similar to the decrease in export activity of TAP (47), arguing that, unlike with HIV-1 Rev, ORF57-mediated RNA export is CRM1-independent. The TAP export pathway is CRM1-independent (54, 56). Taken together with the protein interaction data, it appears that in KSHV, ORF57 promotes viral mRNA export via REF recruiting TAP as for export of cellular mRNAs and as suggested for ICP27 (18, 19). However, co-expression of untethered ORF57 with HIV-1 Rev tethered to cat mRNA reduced the CAT activity, implying a competition between ORF57 and Rev for common component(s) that become limiting, resulting in reduced RNA export. Furthermore, this reduction in cat export activity occurs using an assay in which ORF57 itself promotes RNA export. These data suggest that ORF57- and Rev-mediated export pathways share a common component(s). Intriguingly, recent observations suggest that TAP- and CRM1-mediated export pathways converge (76) as CRM1 competes strongly with TAP for binding to nucleoporin Nup 214, and prevents export by the TAP transport domain (76). Competition for a nucleoporin(s) is a possible explanation for the reduction by ORF57 of Rev-mediated export activity. A further example of export pathways converging, involving structurally distinct groups of proteins, is demonstrated by the competition for NPC binding shown by the NPC-binding domain of TAP with the karyopherins (importin -like receptors CRM1 and exportin-t) involved in protein transport (50, 77). The data presented can be interpreted to strengthen arguments that TAP and CRM1 pathways converge in accessing similar components of the nuclear pore complex.
In Xenopus, nuclear export of viral RNAs mediated by ICP27 and of ICP27 protein itself is insensitive to LMB (19), and ICP27 protein export in HSV-1-infected cells is blocked by a dominant negative TAP mutant (18). ICP27 contains a CRM1-independent NES required for efficient protein export (18) and EBV MTA contains a novel CRM1-independent NES (69, 73). However, ICP27 appears to mediate the export of some HSV-1 RNAs via a CRM1-dependent pathway (78). And, although EBV MTA binds REF, an interaction crucial for its CRM-1-independent RNA export activity (69), it also has been reported to interact with CRM1, an association important for its export function (79). These herpesvirus proteins involved in RNA export may exit the nucleus by more than one pathway and use different export components throughout the viral replication cycle. This idea is supported by the observation that an HSV-1 ICP27 virus mutant (d3-4) that fails to bind REF (19) is only partially defective in growth compared with wild type HSV-1 (80, 81). Our data extends what is known of this herpesvirus family of export factors. We suggest that ORF57-mediated RNA export may use different export factors (such as both TAP and CRM1) to accommodate the KSHV-infected host cell environments, for example, in B-cells (primary effusion lymphoma) or endothelial (KS) cells, and perhaps during the different phases of lytic virus replication. Given evidence for the commonality and functional redundancy of components involved in cellular RNA export (71), it will be most important to determine whether ORF57 recognizes export factors other than REF.
|
FOOTNOTES |
|---|
* This work was supported in part by Medical Research Council Grant G9826324 (to J. B. C.), Association for International Cancer Research Grant 01/242 and Cancer Research UK Grant C7934 (to D. J. B.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Recipient of an UK Commonwealth Postgraduate Scholarship.
To whom correspondence should be addressed: Institute of Virology, University of Glasgow, Church St., Glasgow, G11 5JR, Scotland, UK. Tel.: 44-141-330-4027; Fax: 44-141-337-2236; E-mail: b.clements{at}vir.gla.ac.uk.
1 The abbreviations used are: KS, Kaposi's sarcoma; KSHV, Kaposi's sarcoma-associated herpesvirus; TPA, 12-O-tetradecanoylphorbol-13-acetate; HSV-1, herpes simplex virus type 1; EBV, Epstein-Barr virus; HIV-1, human immunodeficiency virus type 1; RBD, RNA-binding domain; LMB, leptomycin B; NPC, nuclear pore complex; GST, glutathione S-transferase; FL, full-length; HCMV, human cytomegalovirus; CAT, chloramphenicol acetyltransferase; HA, hemagglutinin; RRE, Rev response element; Pipes, piperazine-N,N'-bis(2-ethanesulfonic acid); ORF, open reading frame; NES, nuclear export signal; RNP, ribonucleoprotein; Ab, antibody.
2 P. Malik and J. B. Clements, unpublished results.
|
ACKNOWLEDGMENTS |
|---|
We thank Dr. Tom Hope for the kind gift of plasmid pCMV128, Dr. J. McLauchlan for valuable comments on the manuscript, and Dr. M. Koffa for helpful discussions.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
- Chang, Y., Cesarman, E., Pessin, M. S., Lee, F., Culpepper, J., Knowles, D. M., and Moore, P. S. (1994) Science 266, 1865-1869
[Abstract/Free Full Text] - Schulz, T. F. (1998) J. Gen. Virol. 79, 1573-1591[Medline] [Order article via Infotrieve]
- Boshoff, C., and Weiss, R. A. (2001) Philos. Trans. R. Soc. Lond. B Biol. Sci. 356, 517-534[Medline] [Order article via Infotrieve]
- Dourmishev, L. A., Dourmishev, A. L., Palmeri, D., Schwartz, R. A., and Lukac, D. M. (2003) Microbiol. Mol. Biol. Rev. 67, 175-212
[Abstract/Free Full Text] - Sarid, R., Olsen, S. J., and Moore, P. S. (1999) Adv. Virus Res. <