The Polypyrimidine Tract-binding Protein Is Required for Efficient Dengue Virus Propagation and Associates with the Viral Replication Machinery*

The polypyrimidine tract-binding protein (PTB) functions primarily as an IRES trans-acting factor in the propagation of hepatitis C virus and picornaviruses. PTB interacts with secondary structures within the 3′- and 5′-untranslated regions of these viral genomes to mediate efficient IRES-mediated viral translation. PTB has also been reported to bind to the untranslated region of the single-stranded RNA dengue virus (DENV), suggesting a similar function for PTB in flaviviruses. Indeed small interfering RNA-mediated PTB knockdown inhibited the production of infectious DENV, and this inhibition was specific to PTB knockdown and not due to a nonspecific anti-viral state. In fact, PTB depletion did not inhibit the production infectious yellow fever virus, another flavivirus. Nevertheless, whereas PTB knockdown led to a significant decrease in the accumulation of DENV viral RNAs, it did not impair translation. Moreover, PTB was shown to interact with the DENV nonstructural 4A protein, a known component of the viral replication complex, and with the DENV genome during infection. These data suggest that PTB interacts with the replication complex of DENV and is acting at the level of viral RNA replication.

The Flaviviridae family of viruses has single-stranded positive polarity RNA genomes, which are mRNAs coding viral proteins required for replication. The viral RNA-dependent RNA polymerase, NS5, in conjunction with other viral nonstructural (NS) proteins and host cellular proteins, copies complementary negative strand RNAs from the genomic RNA template, which in turn serve as templates for the synthesis of new positive strand RNAs (2). Replication, transcription, and assembly of virions require cis-acting elements within the 5Ј-and 3Ј-UTRs of the viral genomic RNA (3)(4)(5)(6). These elements interact with host cellular RNA-binding proteins including the polypyrimidine tract-binding protein (PTB). PTB, a heterogeneous nuclear ribonucleoprotein (7), is involved in multiple aspects of cellular mRNA metabolism, including the regulation of alternative splicing, RNA stability, and internal ribosomal entry (IRES)-mediated translation of viral and cellular mRNAs (8 -10).
The role of PTB in the propagation of the positive singlestranded RNA viruses has largely been studied with HCV and picornaviruses, for which PTB primarily functions as an IRES trans-acting factor (ITAF), activating viral translation initiation (11,12). Its role in RNA replication of these viruses is however still debated (13)(14)(15)(16). PTB has also been reported to bind to the UTR of the DENV-4 (17) and Japanese encephalitis virus (18) flaviviruses. Flaviviruses are not known to contain an IRES within their RNA genomes (19), and thus the functional relevance of PTB in flavivirus propagation remains to be demonstrated. This study reports that RNA interference-mediated PTB depletion abrogated DENV production and reduced viral RNA levels. The interactions of PTB with the DENV genome and nonstructural 4A (NS4A) protein, components of the viral replication machinery, were observed during infection. Depletion of PTB and its paralog neuronal PTB (nPTB) did not alter translation from the DENV RNA. The requirement of PTB for flavivirus replication in general was explored using PTB-and nPTB-depleted cells for the propagation of YFV, with no significant inhibition of propagation observed. These data indicate a critical role for PTB in DENV and suggest an effect on RNA accumulation.

EXPERIMENTAL PROCEDURES
Cell Lines and Viruses-The dengue virus 2 New Guinea C strain (DENV-2) and yellow fever virus 17D strain (YFV) used in this study were propagated in the C6/36 mosquito and Vero cells, respectively, as previously described (20). Baby hamster kidney  and Vero cells were used for the quantification of DENV and YFV by plaque assay. In brief, the cells were seeded in 24-well plates and infected the next day with virus. The cells were processed for plaque determination 3 and 6 days post-infection for YFV and DENV, respectively. All of the statistical analyses were carried out with GraphPad Prism 4 (GraphPad Software). DENV was used at a multiplicity of infection (MOI) of 1 or 10 as indicated. YFV was used at a MOI of 1. The cells were incubated with the virus for 1 h at 37°C with occasional rocking. After 1 h, the cells were rinsed, overlaid with complete medium, and incubated at various time points.
siRNA Transfection Procedures and Analyses of Gene Expression-The siRNAs P1, P9 (PTB-specific siRNA duplexes), N3 (nPTB-specific siRNA duplex), and C2 (nonspecific siRNA duplex) used have been previously reported (21)(22)(23) and were synthesized by Invitrogen. siRNA transfections were performed on Huh-7 cells as described (24), with 33 and 25 nM of each siRNA for the first and second transfections, respectively. 72 h post-transfection, the cells were plated at a density of 1 ϫ 10 5 (for DENV infection) or 3 ϫ 10 5 (for YFV infection) cells/35-mm well and a day later were infected with virus at an MOI of 1. At the indicated time points, the supernatant and cells were collected for plaque, Western, and quantitative real time RT-PCR analyses (see Fig. 2A). For Western blot analysis, the harvested cells lysates were resolved by SDS-PAGE and transferred onto a nitrocellulose membrane, and the appropriate proteins were detected by labeling using either a polyclonal (a gift of the Garcia-Blanco laboratory, Duke University Medical Center) or monoclonal (Zymed Laboratories Inc.) anti-PTB, anti-nPTB (Abcam) anti-fibrillarin (Cell Signaling Technology), anti-GAPDH (Cell Signaling Technology), or anti-tubulin (Santa Cruz Biotechnology) antibodies. The levels of specific proteins detected were quantified using Quantity One (Bio-Rad).
RT-PCR and Quantitative Real Time PCR-Total RNA was prepared from uninfected and infected cells using the TRIzol reagent (Invitrogen) according to the manufacturer's instructions. The RT reaction was performed at 42°C as described (25). All of the quantitative real time PCRs were performed on the iCycler iQ TM Multi-Color real time PCR detection system (Bio-Rad) with the following conditions: 40 cycles of 30 s denaturation at 95°C, 30 s annealing at 55°C, and 30 s extension at 72°C. Fluorescent detection of SYBR Green I (Power SYBR Green PCR Master mix; Applied Biosystems) was carried out at the extension phase. All of the cDNA standards used in quantitative real time PCR were identical in size and sequence to the targets and were generated by PCR using Taq polymerase (Promega Corporation) and 50 nM of each primer, with the following cycling conditions: a 95°C step for 3 min; 25 cycles of 95°C for 30 s, 55°C for 30 s, and 72°C for 1 min; and an elongation step at 72°C for 7 min. The quantification of viral and cellular nucleic acids in cell culture studies was normalized against the expression levels of actin. The sequences of the primers used are shown in supplemental Table S1. All of the statistical analyses were carried out with GraphPad Prism 4.
RNA Immunoprecipitation (RIP)-Isolation of PTB-associated RNAs under native conditions was performed by immunoprecipitation as described previously (26) using the anti-PTB monoclonal antibody from precleared lysates of Huh-7 cells uninfected or infected with DENV-2 at an MOI of 1. Briefly uninfected or infected cells were harvested and either left untreated or cross-linked with 0.6% formaldehyde and disrupted by sonication in RIP assay buffer (50 mM Tris-Cl, pH 7.5, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.05% SDS, 1 mM EDTA, and 150 mM NaCl), and the resultant lysates were precleared before immunoprecipitation. Immune complexes were precipitated with a slurry of 50% Ultralink protein A beads (Pierce) and washed five times with high stringency RIP assay buffer (50 mM Tris-Cl, pH 7.5, 1% Nonidet P-40, 1% sodium deoxycholate, 0.1% sodium dodecyl sulfate, 1 mM EDTA, 1 M NaCl, and 0.2 mM phenylmethylsulfonyl fluoride). The co-immunoprecipitated RNA was isolated by TRIzol, and subjected to RT-PCR for the detection of PTB-associated cellular and viral nucleic acids using the primers shown in supplemental  Table S2. The cycling parameters for the real time PCR were used for amplification, and the PCR products were visualized by gel electrophoresis. The primer sequences for GAS2L1 and SMARCA2 used have been previously reported (27). The RIP analyses were performed in three independent experiments.
RNA Constructs, in Vitro Transcription, and Transfection-The 5Ј and 3Ј-UTR regions used for the construction of the DENV-2 luciferase reporter construct (pD2-5FLuc3) were obtained from the DENV-2 strain NGC viral RNA. A 171-bp fragment of the 5Ј-UTR inclusive of the first 75 nucleotides of the capsid sequence and a 483-bp of the 3Ј-UTR with the last 41 nucleotides of the NS5 sequence were amplified using primers 5D2F and 5D2BamR and primers 3D2NotF and 3D2R, respectively (supplemental Table S3 for primer sequences). The amplified products were gel-purified (QIAquick gel extraction kit; Qiagen) and cloned into the pGemT-Easy vectors (Promega). The firefly luciferase gene was amplified from the pGL3-Basic vector (Promega) using the primers FlucBamF and Fluc-NotR and cloned into the pBluescript II-SK vector (Stratagene) at the BamHI and NotI sites (pBS-Fluc). The pD2-5Lu3c plasmid was generated by sequential cloning of the 5Ј-UTR fragment to the EcoRI and BamHI sites of the pBS-Fluc, followed by the 3Ј-UTR fragment at the NotI and blunt-ended SacII sites. The reporter plasmid was transcribed using a Megascript T7 Transcription kit (Ambion) according to the manufacturer's instructions, in the presence of an m 7 GpppA nucleotide (New England Biolabs). The reporter transcripts were transfected into Huh-7 cells using Lipofectamine 2000 (Invitrogen) as previously described (28), together with a transfection control Renilla luciferase, obtained from in vitro transcription of a linearized pRL vector (Promega). The analyses of both firefly and Renilla luciferase levels were performed using the dual luciferase assay kit (Promega) on a Tecan Infinite M200 luminometer.
Plasmid Constructs-For the construction of plasmid standards for quantitative real time PCR, extracted total cellular and DENV genomic RNA were reverse transcribed, used as tem-plates for PCR as described above, and the amplified DNA fragments were subcloned into the pGEMT vector (Promega). For the construction of the plasmid pcTAPNS4A expressing the full-length DENV-2 NS4A protein with a TAP tag at the C terminus, the full-length DENV-2 NGC strain cDNA (a gift from Andrew Davidson, School of Medical Sciences, Bristol, UK) was used as a template for PCR amplification of the required full-length DENV-2 NS4A gene. The TA cloningbased pGEM vector (Promega) was used for the subcloning of the full-length DENV-2 NS4A gene. Plasmid pcTAPNS4A were constructed by inserting the DENV-2 NS4A gene product from pGEM subcloning vector into pcTAP (Interplay Mammalian TAP system; Stratagene) by use of EcoRI and XhoI, respectively. All of the plasmids were confirmed by restriction and DNA sequence analyses.
Transfection and Generation of a Stable Cell Line Expressing DENV-2 NS4A-TAP Fusion Protein-DNA construct (pcTAPNS4A or pcTAP-control vector) were transfected into Huh-7 cells using Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. Cell monolayers were trypsinized 24 h after transfection and transferred into T25cm 2 flasks or 100-mm-diameter culture dishes. The clones were selected by growth in complete Dulbecco's modified Eagle's medium containing 500 g/ml of G418. The isolated clones were then cultured and expanded. Expression of the DENV-2 NS4A-TAP fusion protein was confirmed by immunoblotting and immunofluorescence assays at various time points.
Expression and Purification of Tagged DENV-2 NS4A-containing Complexes-DENV-2 NS4A fusion protein containing complexes were isolated using the InterPlay mammalian TAP system (29) according to the manufacturer's instructions (Stratagene). Briefly, two T175-cm 2 cell culture flasks of confluent Huh-7 cells stably expressing DENV-2 NS4A-TAP protein or control TAP protein were harvested. Following harvest, the cells were washed three times in phosphate-buffered saline (PBS) and lysed in lysis buffer by three successive rounds of freezing in dry ice and thawing. The cell extracts were centrifuged for 10 min at 16,000 ϫ g at 4°C, the supernatants were passed over streptavidin resin, and the eluted complexes were bound to calmodulin-coated beads. Following elution from the calmodulin column, the proteins were resuspended in gel-loading buffer. The samples were resolved on 10% SDS-PAGE gels, fixed in 50% (v/v) methanol and 10% (v/v) acetic acid, and stained using Coomassie Blue staining solution.
Mass Spectrometry-In-gel reduction, alkylation, and trypsin digestion were performed using standard procedures prior to subsequent analysis by mass spectrometry, as described previously (30). The mass spectral data were processed into peak lists (containing the precursor ion m/z and charge state and the m/z and intensity of the fragment ions) and searched against the SwissProt and NCBI nonredundant databases using Mascot software (Matrix Science). All of the results were manually verified.
Co-immunoprecipitation and Immunoblotting-Two T175cm 2 cell culture flasks containing confluent Huh-7 cells were infected with DENV-2 at an MOI of 10 and lysed at different time points post-infection. In brief, the cells were lysed in immunoprecipitation buffer (1% Nonidet P-40, 150 mM NaCl, 50 mM Tris-HCl, pH 7.5, 1 mM EDTA) with complete protease inhibitor mixture (Roche Applied Science) for 30 min at 4°C and then cleared by centrifugation for 15 min at 18,000 ϫ g at 4°C. For co-immunoprecipitation assays, protein A-or G-agarose (Invitrogen) was preincubated with an anti-DENV NS4A (mouse polyclonal; a gift from Dr Ralf Bartenschlager, University of Heidelberg, Heidelberg, Germany) or rabbit polyclonal anti-PTB antibody for 1 h at 4°C and then mixed with cell lysate for 4 h at 4°C on a rotator. Thereafter, the agarose beads were washed three times with cold immunoprecipitation buffer, and the bound complexes were eluted in gel loading buffer. The samples were analyzed by 10% SDS-PAGE gel. Resolved proteins were detected by either anti-DENV NS4A or anti-PTB as primary antibodies. Anti-DENV envelope and NS5 antibodies are included as controls for the experiments. Bound antibodies were visualized using alkaline phosphatase-conjugated goat anti-rabbit or mouse secondary antibodies and 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium (Chemicon).
Immunogold Labeling and Electron Microscopy Analysis-DENV-infected and mock infected Huh-7 cells (3 days postinfection) were processed for LR White resin embedding. In brief, the cells were first fixed with 4% paraformaldehyde and 0.1% of glutaldehyde in PBS, pH 7.4. The cells were then scraped and centrifuged at 3000 rpm for 10 min. The fixatives were removed by a series of washing with PBS, and the cell samples were processed for dehydration with increasing concentration of ethanol. The dehydrated cells were further infiltrated with LR White resin and embedded into BEEM capsules. The embedded cells are then cut, ultrathin sections of the infected cells were picked onto Formvar-coated grids, and immunolabeling was then carried out at room temperature. The sections were washed for 10 min with PBS. This was followed by three washes (5 min each) with 0.05 M glycine-PBS before incubating for 30 min in PBS with 5% bovine serum albumin. The sections were then washed in 0.1% bovine serum albumin in PBS and exposed to anti-PTB or anti-DENV NS4A primary antibodies at 1:100 dilution for 1 h. Six washes in 1% bovine serum albumin in PBS were done before exposing the sections for 1 h to the protein A colloidal gold (Agar Aids) (5 and 15 nm at 1:20 dilution for PTB and DENV NS4A proteins, respectively). The above steps were repeated with the second primary antibody for double labeling experiments. The sections are post-fixed in 2% glutaraldehyde in PBS for 5 min and washed (two times for 5 min each) in PBS and distilled water (four times for 5 min each). Finally, the sections were embedded in 2% uranyl acetate for 5 min and dried before viewing under the EM208 transmission electron microscope (Philips).

DENV Infection Does Not Induce PTB Cleavage or Increase in
Cytoplasmic Localization-PTB has been shown to be involved in the propagation of the positive strand RNA HCV and picornaviruses. Following infection with these viruses, PTB relocates to the cytoplasm and is proteolytically cleaved (13,14,24,32,33). Hence, we asked whether DENV infection would lead to similar effects. Huh-7 cells were infected at a MOI of 1, and the subcellular localization of DENV envelope (Env) and PTB was assessed 48 h post-infection (p.i.) by indirect immunofluorescence. PTB was localized predominantly in the nucleus of DENV-infected cells, which was indistinguishable from the pattern seen in the flanking uninfected Huh-7 cells (Fig. 1A). Similar PTB distribution profiles were also observed in Huh-7 cells infected with DENV at MOI of 10 (data not shown). Additionally, Western blot analysis of cell lysates from DENV-infected cells harvested at 1-6 days p.i. did not show the presence of cleaved PTB fragments, or a preferential turnover of any PTB isoforms (Fig. 1B). These data suggest that in contrast to picornavirus and HCV infection, DENV does not detectably alter either the integrity or cellular localization of PTB.
PTB Is Required for the Propagation of DENV-To examine the role of PTB in DENV propagation, PTB-specific siRNAs (P1 and P9) targeting different regions of PTB mRNAs (21,22) were used to reduce PTB levels in Huh-7 cells. Control cells were mock transfected (Mock), treated with a nonspecific siRNA (C2), or left untreated (DG) (Fig. 2B). PTB depletion did not significantly affect the levels of fibrillarin, a nucleolar protein involved in ribosome biogenesis (34) (data not shown). This is consistent with the observed continued growth of cells transiently depleted of PTB (21) and minimal changes in the proteome of PTB siRNA-treated cells (23). PTB-depleted and control cells were subsequently infected with DENV at an MOI of 1, and the samples were harvested according to the schedule in Fig. 2A. Cellular supernatants were titrated for infectious virus, whereas the DENV RNA was quantified by real time quantitative RT-PCR.
PTB knockdown resulted in a statistically significant and persistent inhibition of infectious DENV production (Fig. 2). Significant effects of PTB depletion on the production of infectious DENV were observed from 2.5 days p.i. onwards, with a maximum of 7-fold inhibition obtained (Fig. 2, C and D). There was a concomitant decrease in the levels of DENV RNA (Fig. 3). The extent of inhibition of DENV propagation paralleled the degree of PTB knockdown, as shown with infected P1-and P9-treated cells (Figs. 2 and 4A). The PTB levels of P9-treated cells were generally observed to be higher than that in P1-treated cells throughout the course of the study (Fig. 4A).
To test the possibility that inhibition of DENV propagation was caused by the activation of an antiviral cellular response caused by treatment with siRNAs, the levels of interferon-␤ mRNA were assessed by real time quantitative RT-PCR. Interferon-␤ transcript levels did not increase upon treatment with any of the siRNAs and only showed an increase with the progression of infection (Fig. 4B). The  combined data therefore indicated that RNA interference treatment did not lead to the activation of a nonspecific antiviral state and that the inhibition of DENV propagation was specific to PTB depletion.
PTB Associates with DENV RNA during Infection-A role for PTB in RNA synthesis of single strand RNA viruses, although still debated, has been proposed (13)(14)(15)(16). Because PTB was reported to bind to sequences of flaviviruses in vitro (17,18), we first sought to determine whether PTB associated with the DENV RNA in vivo during infection by RIP analyses. Uninfected or DENV-infected Huh-7 cells (MOI of 1) were harvested 2 days p.i., and the lysates, either untreated or crosslinked with formaldehyde, were immunoprecipitated in the presence or absence of anti-PTB monoclonal antibody. The input lysates and RIP eluants were then analyzed for the presence of endogenous mRNA and viral RNA ribonucleoprotein complexes by RT-PCR using primers specific to sequences of the DENV UTRs and PTB-associated and nonassociated mRNAs. Specific bands with the expected sizes were observed with primers to PTB-associated, but not to PTB nonassociated genes, in both uninfected and infected cells (Fig. 5). A higher band (ϳ230 bp), observed to be co-amplified in the RIP eluant of uninfected cells with the GAS2L1 primers, was attributed to the amplification of as yet unspliced PTB-associated GAS2L1 RNA. Additionally, bands specific to the DENV sequences were only detected in RIP-PCR samples from infected but not from uninfected lysates. These products were easily detected in the input samples. In contrast, DENV sequences were not amplified from RIP eluants of DENV-infected cell lysates that were immunoprecipitated with an anti-fibrillarin antibody (data not shown). These data strongly suggested the in vivo association of PTB with the DENV RNA during viral infection.

PTB Interacts with DENV NS4A and Co-localizes within the DENV Replication
Complex-The replication of flaviviruses involves the formation of a replication complex that includes NS proteins and the viral RNA template (2). Because DENV RNA was shown to associate with PTB during infection and was reduced with PTB depletion, we sought to determine whether PTB could form part of the DENV replication complex. A role for NS4A in the replication of flavivirus has been documented (35,36). Cellular proteins associated with DENV NS4A protein were purified from Huh-7 cells stably expressing NS4A-TAP fusion using tandem affinity purification (29). Both NS4A-TAP and NS4A were detected in perinuclear compartments and in the nucleus during DENV infection of Huh-7 (Fig. 6A). Additionally, ultrastructural analysis revealed induction of membrane rearrangements in Huh-7 cells stably transfected with pcTAPNS4A (data not shown), typical of flavivirus infection, indicating that the TAP sequences did not perturb the biological function of the viral protein.
DENV NS4A-containing complexes were first purified from Huh-7 cells stably transfected with pcTAPNS4A vector and visualized by Coomassie Blue staining (Fig. 6B). As controls, parallel purifications were performed with cells stably trans-  fected with the parental pcTAP vector. The eight most prominent bands were excised, subjected to tryptic digestion, and analyzed by mass spectrometry for identification. Interestingly, one of the proteins excised was identified as PTB.
The interaction of NS4A with cellular PTB was further examined by co-immunoprecipitation of cell lysates from DENVinfected Huh-7 cells with PTB-and DENV NS4A-specific antibodies (Fig. 6, C and D). Following immunoprecipitation with an anti-PTB antibody, the association of NS4A with cellular PTB protein was evaluated by immunoblotting using DENV NS4A-specific antibody. The anti-PTB serum co-immunoprecipitated NS4A (Fig. 6C). Equally the reverse immunoprecipitation revealed that anti-DENV NS4A antibody brought down PTB (Fig. 6D). PTB or DENV NS4A proteins were not detected in immunoblots of controls using protein A-linked agarose in the absence of the primary antibodies (control panels in Fig. 6, C and D). In contrast, PTB was not observed to be co-immunoprecipitated with either DENV envelope or NS5 protein from the infected cells (data not shown). Similar results were obtained with lysates treated with RNase A (200 g/ml) prior to immunoprecipitation, suggesting that the interaction between PTB and NS4A did not require a bridging RNA (data not shown).
Co-localization of DENV NS4A protein with PTB in DENV-infected cells was further confirmed by immunogold labeling of these proteins at the ultrastructural level. A close association of the DENV NS4A (15-nm gold particles) protein with PTB (5-nm gold particles) was observed within the typical DENV-induced replication complex (convoluted membranes) (Fig. 6E). Taken together, these data strongly suggested a close association of DENV NS4A with cellular PTB during infection.
PTB Depletion Does Not Inhibit DENV Translation-PTB primarily functions as an ITAF in IRES-mediated viral translation of HCV and picornaviruses. To assess whether PTB serves similar translational function in DENV propagation, Huh-7 cells treated with PTB-specific siRNA (P1) a nonspecific siRNA (C2) or mock transfected (Mock) were transfected with the DENV-2 luciferase reporter mRNA. Luciferase activity, which is an indication of translation from the reporter mRNA, was measured 2-30 h post-transfection (Fig. 7, A and C). PTB knockdown did not inhibit translation of the DENV mRNA compared with the C2 and mock transfected control cells.
It is of note that many mammalian cells express both PTB and its paralog, nPTB, which share a large degree of functional overlap and cross-regulation (23). Indeed an increase in the level of nPTB was observed upon PTB depletion (Fig. 7, A and   FIGURE 5. PTB interacts with the DENV-2 RNA genome. Huh-7 cells were either uninfected (control) or infected with DENV-2 and the lysates harvested for RIP 2 days p.i. Non-cross-linked or cross-linked lysates were immunoprecipitated in the presence or absence of ␣-PTB monoclonal antibody. The RIP eluants and the input samples were analyzed for the presence of endogenous mRNA and viral RNA ribonucleoprotein complexes by RT-PCR using primers specific to sequences of the DENV UTRs (5ЈD2-UTR and 3ЈD2-UTR), PTB-associated (GAS2L1 and ACT) and PTB nonassociated (SMARCA2) mRNAs. The expected sizes of the respective amplified products are indicated in parentheses. ϩRT and ϪRT represent RT reactions carried out in the presence and absence of reverse transcriptase. A 100-bp DNA ladder was used as molecular mass markers in the electrophoretic analyses. The viral protein was detected with anti-NS4A antibody (green, arrow). B, TAPtagged purification of NS4A complexes from Huh-7 cells. Cells stably transfected with the control pcTAP or pcTAP-NS4A plasmids were lysed, and the DENV NS4A protein complexes were sequentially purified using streptavidinbinding peptide-coated and calmodulin-binding peptide-coated beads. The samples were resolved on SDS-PAGE gels, and the bands were visualized by Coomassie Blue staining. The arrow indicates the band identified as PTB by mass spectrometry. M denotes the protein marker. C, immunoprecipitation of NS4A with anti-PTB antibody. DENV infected Huh-7 cells were lysed at different time points p.i. and immunoprecipitated with anti-PTB antibody and protein A-linked agarose. The presence of DENV NS4A protein was detected using anti-NS4A antibody. D, reciprocal immunoprecipitation of PTB with anti-DENV NS4A antibody. DENV infected Huh-7 cells were lysed at different time points p.i. and immunoprecipitated with anti-NS4A antibody and protein A-linked agarose. The presence of PTB protein was detected using anti-PTB antibody. For both immunoprecipitation studies, background control was included using protein A-linked agarose in the absence of the primary antibody. E, the cellular localization of NS4A and PTB by electron microscopy. Close association of DENV NS4A (15 nm of immunogold, arrows) and PTB proteins (5 nm of immunogold; arrowhead) were observed within the convoluted membranous (CM) replication complex of DENV-infected cells. ER denotes the endoplasmic reticulum. B). It was therefore possible that an inhibitory effect of PTB depletion on DENV translation was masked by the compensatory up-regulation of nPTB. To address this, nPTB-specific (N3) or PTB and nPTB-specific (P1/N3) siRNAs was used to reduce the levels of nPTB, or both PTB and nPTB, in Huh-7 cells (Fig. 7A). The depletion of nPTB, or the simultaneous depletion of PTB and nPTB did not show an inhibition in the translation of the DENV RNA (Fig. 7C), suggesting that PTB does not function at the translation stage in DENV replication cycle. The dual depletion of PTB and nPTB did, however, result in a more profound inhibition of DENV-2 propagation than observed with depletion of PTB alone (see next section and Fig.   7D). We conclude from this that increased expression of nPTB can partly rescue viral production in PTB-depleted cells, but neither protein appears to be required for translation of DENV-2 mRNAs.
PTB Is Not Required for Efficient YFV Propagation in Huh-7-To assess whether the requirement of PTB for virus production extended across the flaviviruses, the effect of PTB depletion was next tested on the propagation of YFV strain 17D. Huh-7 cells were either treated with P1 and P9 siRNAs for PTB depletion, with the control cells either mock transfected or treated with C2 siRNA (supplemental Fig. S1). The cells were subsequently infected with YFV at an MOI of 1. Because the kinetics and development of cytopathic events caused by YFV infection was more rapid than DENV in hepatoma cells (37), the growth kinetics of YFV were monitored for 24 h, with the samples harvested at 6, 12, and 24 h p.i. As in DENV infection, PTB protein cleavage was not observed with YFV infection. Interestingly however, unlike DENV, the production of infectious YFV was not significantly different between control and PTB-depleted cells (supplemental Fig. S1).
Given the aforementioned compensatory up-regulation of nPTB with PTB depletion, it was conceivable that a double knockdown of PTB and nPTB would lead to a decrease in YFV replication and perhaps even to a much more profound inhibition of DENV propagation. To address this, Huh-7 cells that were untreated (DG), mock transfected (Mock), or transfected with PTB or/and nPTB siRNAs were infected with either DENV-2 or YFV (MOI of 1), and the virus produced was assessed at 2.5 days and 12 h p.i., respectively. nPTB depletion did not show inhibition of both DENV and YFV propagation ( Fig. 7D and supplemental Fig. S1). Compared with the knockdown of PTB, simultaneous depletion of PTB and nPTB led to a larger inhibition of DENV propagation (Fig. 7D). Conversely, YFV propagation was not significantly different between control, PTB-depleted, and concomitant PTB-and nPTB-depleted cells (supplemental Fig. S1). These data therefore suggest that in contrast to DENV, PTB is not required for efficient YFV propagation.

DISCUSSION
PTB is required for efficient proliferation of poliovirus, encephalomyocarditis virus (24,33,38,39), and HCV (5,11,13,40). Here we show that PTB is also required for efficient DENV-2 propagation in human cells in culture. Unlike IREScontaining viruses, DENV infection did not induce detectable PTB cytoplasmic translocation or cleavage. Although predominantly found in the nucleus of cells, PTB shuttles between the nucleus and cytoplasm (41) and can therefore influence DENV propagation. Indeed a significant proportion of PTB has been reported to be present within the cytoplasm (27,42). Reciprocally, several DENV and flaviviral components, including NS3, NS4, and NS5, undergo a "nuclear experience" during the virus replication cycle (43)(44)(45)(46)(47). An interesting emerging view is that positive strand RNA virus replication may not be exclusive to the cytoplasm and that interactions with the nucleus may facilitate virus replication (48).
We observed an unexpected dichotomy in the requirement for PTB in DENV and YFV propagation. DENV and FIGURE 7. PTB and nPTB in DENV replication. A and B, nPTB protein is upregulated by PTB knockdown. Western blot analysis of proteins harvested from Huh-7 cells treated with mock transfection, control C2, PTB (P1), nPTB (N3) or PTB and nPTB (P1/N3) siRNAs (A) or with mock transfection or PTB (P9) siRNA (B). The cell lysates were collected at day 3 post-transfection. The blots were probed for nPTB (panels 1), PTB (panels 3), and GAPDH (panels 2 and 4). C, depletion of PTB and nPTB did not inhibit translation of DENV reporter RNA. Huh-7 cells treated with mock transfection, control C2, PTB (P1), nPTB (N3), or PTB and nPTB (P1/N3) siRNAs were transfected with the D2-5FLuc3 and RLuc control RNA transcripts, and the cell lysates were harvested for luciferase analyses at the indicated time points post-transfection D, propagation of DENV in Huh-7 cells which were untreated (DG), mock transfected (Mock), or transfected with PTB or/and nPTB siRNAs at 2.5 days p.i. The bars indicate the range of plaque-forming units/ml from three replicates. The asterisks indicate statistically significant differences (p Ͻ 0.001) by one-way analysis of variance and Tukey's post-test.
YFV belong to different sero-groups within the Flavivirus genus (2), with reported differences in the infection pattern and propagation kinetics in liver cells (37). One explanation for the differences in the replication kinetics, and perhaps the differential requirement for PTB between these flaviviruses, may lie within their UTRs. Although DENV and YFV show significant conservation in the predicted RNA structures within their 5Ј-UTRs, they differ significantly within their 3Ј-UTRs (19,49,50). PTB interacts with uncharacterized sequences in the 3Ј-UTRs of the positive genomic strand of DENV-4 and the negative strand of Japanese encephalitis virus RNAs (17,18).
RNA viruses are known to subvert cellular proteins to become part of their replication strategy (51). Current studies on the function of PTB in virus propagation have been focused on its role as an ITAF, likely acting as an RNA chaperone to stabilize the IRES structure to facilitate translation initiation (52,53). Flaviviruses have a type I 5Ј-m 7 GpN-cap structure and translate mainly via a cap-mediated process (2). Although DENV has been reported to be able to switch to a cap-independent translation when required, this was shown not to be mediated via an IRES structure (54). Therefore a similar role for PTB acting as an ITAF in DENV-2 and possibly in flavivirus propagation was unlikely. Indeed, we show here that PTB depletion did not alter the translation efficiency of a DENV mRNA reporter.
A role of PTB in viral RNA replication as opposed to translation, although limited, has been proposed (13,15,16,40,(55)(56)(57). The replication of the mouse hepatitis virus requires PTB, which induces RNA conformational changes when bound to the complementary negative strand RNA (15,55,56). PTB is required for the efficient replication of positive stranded RNA feline calicivirus in a temperature-dependent manner, consistent with a role for PTB as an RNA chaperone (58). Given our data showing interactions of PTB with the DENV genome and NS4A, its localization in the replication machinery, it is attractive to speculate that PTB could also have similar chaperone function in the replication of DENV RNAs.