Regulation of the Hepatitis B virus replication and gene expression by the multi-functional protein TARDBP

Hepatitis B virus (HBV) infects the liver and is a key risk factor for hepatocellular carcinoma. Identification of host factors that support viral replication is important to understand mechanisms of viral replication and to develop new therapeutic strategies. We identified TARDBP as a host factor that regulates HBV. Silencing or knocking out the protein in HBV infected cells severely impaired the production of viral replicative intermediates, mRNAs, proteins, and virions, whereas ectopic expression of TARDBP rescued production of these products. Mechanistically, we found that the protein binds to the HBV core promoter, as shown by chromatin precipitation as well as mutagenesis and protein-DNA interaction assays. Using LC-MS/MS analysis, we also found that TARDBP binds to a number of other proteins known to support the HBV life cycle, including NPM1, PARP1, Hsp90, HNRNPC, SFPQ, PTBP1, HNRNPK, and PUF60. Interestingly, given its key role as a regulator of RNA splicing, we found that TARDBP has an inhibitory role on pregenomic RNA splicing, which might help the virus to export its non-canonical RNAs from the nucleus without being subjected to unwanted splicing, even though mRNA nuclear export is normally closely tied to RNA splicing. Taken together, our results demonstrate that TARDBP is involved in multiple steps of HBV replication via binding to both HBV DNA and RNA. The protein’s broad interactome suggests that TARDBP may function as part of a RNA-binding scaffold involved in HBV replication and that the interaction between these proteins might be a target for development of anti-HBV drugs.

Replication of the virus occurs by reverse transcription of pregenomic RNA 7 . Four viral promoters, Core, Pre S1, Pre S2, and X, and two enhancers, enhancer I and enhancer II, control the transcription of HBV 8 . In addition, a vast number of host cellular transcription factors, including nuclear receptors, regulate HBV transcription by acting on the HBV promoters and enhancers 9 .
To further understand the molecular mechanisms behind the control of HBV replication and pathogenesis, we identified novel host factors that could play a role in the viral lifecycle using primary human hepatocytes derived from human hepatocyte transplanted chimeric mice, an infection system that we recently established 10 . One of the candidates we identified was a gene that codes for the trans-active response DNA binding protein (TARDBP). The protein was initially identified by Ignatius et al. as a factor that binds to TAR DNA of HIV 11 . TARDBP belongs to the family of heterogeneous nuclear ribonucleoproteins (hnRNPs) that serve multiple roles in the generation and processing of RNA, including transcription, splicing, transport, and mRNA stability 12,13 . TARDBP contains two RNA recognition motifs (RRMs), through which it binds to UG/TG repeats in RNA/ DNA, respectively, and a C-terminal glycine-rich domain that is considered important for protein-protein interactions 14 . In disease, TARDBP is known to be frequently mutated in sporadic and familial amyotrophic lateral sclerosis (ALS), as well as in patients with frontotemporal lobar degeneration (FTLD), providing evidence of a direct link between TARDBP abnormalities and neurodegeneration 15 . The role of this protein in HBV pathogenesis remains unreported.
In this study, we demonstrate that TARDBP is a novel host factor that can enhance HBV gene expression via both transcriptional and post-transcriptional mechanisms. We show that silencing of the protein results in inhibition of HBV gene expression. We further provide evidence that TARDBP activates transcription from the core promoter of HBV and that TARDBP forms complexes with other proteins known to support the lifecycle of HBV. Moreover, we report that the protein has an inhibitory effect on splicing of pregenomic (pg) HBV RNA, which may help the virus export noncanonical unspliced RNAs from the nucleus despite the normally close association between splicing and nuclear export.

Silencing of TARDBP inhibits HBV infection and gene expression in cell cultures. To examine
whether TARDBP plays a role in HBV gene expression, we performed gene silencing assays using the human hepatocyte cell line HepG2 with stably expressed NTCP (HepG2-hNTCP) as the HBV infection model 16 . The cells were transduced with lentiviruses containing shRNA targeting TARDBP or scramble control shRNA and selected in puromycin. Reduction in TARDBP expression was confirmed by western blotting and qPCR to be >70% (Fig. 1a). The effect of TARDBP silencing on HBV replication was examined by HBV infection of the cells for 12 days. Downregulation of TARDBP resulted in decreased levels of HBV secreted virions (Fig. 1b). Concordantly, knockdown of TARDBP also inhibited secretion of HBs and HBe proteins in the supernatant (Fig. 1c), production of the core protein (Fig. 1d), as demonstrated by CLEIA and western blotting, respectively, and production of HBV replicative intermediates as shown by both real-time PCR and Southern blotting (Fig. 1e). In addition, the 3.5-kb mRNA, which consists of precore and pregenomic (pg) mRNAs together with total HBV mRNAs were also remarkably reduced in TARDBP depleted cells (Fig. 1f). To further elucidate the role of TARDBP in HBV replication, we used T23 cells, a HepG2 cell line that is stably transfected with the 1.4x genome-length HBV expressing plasmid pTRE-HB-wt, and thus can constitutively produce HBV DNA in the supernatant 17  TARDBP stimulates the activity of the HBV core promoter. TARDBP is an RNA/DNA binding protein with predominantly nuclear localization; however, the protein has a nuclear export signal (NES) and a nuclear localization signal (NLS) through which it is believed to shuttle between the nucleus and the cytoplasm 18 . To study its role in HBV, we first investigated the localization of the protein in the stably transduced HBV cell line T23. As shown by immunofluorescence, TARDBP is primarily localized in the nucleus (Fig. 3a), hence its functions are expected to be restricted to the nucleus in our experimental setting. TARDBP was originally identified as a transcription factor that could repress the transcription of HIV-1 11 ; moreover, the protein has been shown to bind to DNA sequences in the promoter regions of a number of genes and regulate their expression 19,20 . We hypothesized that it could interact with one or more of the HBV promoters in order to regulate HBV replication. To this end, chromatin was prepared from the T23 cells and subjected to quantitative chromatin immunoprecipitation (ChIP) using a TARDBP antibody. The precipitated DNA was then subjected to real-time qPCR using primers specific for each of the four HBV promoters (Core, Sp1, Sp2 and X). PCR analysis showed that a DNA sequence corresponding to the core promoter was enriched in the TARDBP-antibody precipitate compared to the control, whereas the amount of the other three promoters was unchanged (Fig. 3b). The existence of a band corresponding to enriched TARDBP protein and the core promoter in the anti-TARDBP precipitate was validated by western blotting and PCR, respectively (Fig. 3c). This finding implied that TARDBP could regulate HBV transcription by interacting with the core promoter. To further characterize the relevance of the TARDBP-core promoter interaction, we employed a luciferase reporter assay. The pGL3-Cp plasmid is a luciferase plasmid whose expression is driven from the core promoter (Fig. 3d). The pGL3-Cp was co-transfected with an empty vector or with increasing concentrations of a TARDBP expressing plasmid in NTCP-HepG2 cells. The transcriptional activity was examined in these cells. As shown, we observed a dose dependent enhancement of core promoter activity by TARDBP (Fig. 3e). At this point, we asked ourselves whether TARDBP could also bind to HBV cccDNA. Results from a ChIP assay demonstrated that TARDBP could not precipitate cccDNA, as opposed to HBV core protein, which has already been reported to bind to cccDNA 21 (Fig. S5), implying that TARDBP binds to cccDNA during active transcription and is therefore easily nicked and degraded during the extraction process.

Establishment of TARDBP-knockout (KO) NTCP-HepG2 cells by the CRISPR/Cas9 system. To
further examine the roles of TARDBP on viral replication, we established a TARDBP-deficient NTCP cell line (TARDBP KO) by the CRISPR/Cas9 system using two guide RNAs (Fig. 4a). We confirmed that KO of TARDBP did not affect the expression levels of NTCP, the protein responsible for HBV infection in this cell line, nor another known HBV transcription factor, SIRT1 (Fig. 4b). At first, we examined the core promoter transcriptional activity in the KO cells versus the parental cells. The pGL3-Cp plasmid was transfected in equal amounts C4 cells were transduced with a lentiviral vector allowing the expression of a control shRNA (shControl) or shRNA directed against TARDBP (shTardbp). After selection with puromycin, reduction of TARDBP mRNA and protein level by shRNA was confirmed by qPCR and western blotting, respectively. (b) The cells in (a) were then infected with HBV in duplicate sets for 12 days. Supernatants were collected at the indicated time points for extracellular HBV DNA analysis. (c) A separate portion of supernatants from (b) were also analyzed by CLEIA for measurement of HBs and HBe antigen levels. (d) To detect the level of the HBV core protein, one set of the cells were harvested at the end of the 12-day infection period for protein lysis and subjected to western blotting using the anti-HBc antibody. (e) For detection of the core-associated HBV DNA, the lysates in (d) were subjected to immunoprecipitation by an anti-HBc antibody followed by qPCR (left panel) and Southern blotting (right panel). (f) Total RNA was extracted from the other set of cells from after the 12-day infection period in (b) and subjected to RT qPCR to detect HBV precore, pregenomic (pg) and total mRNAs. Gene expression was normalized to that of GAPDH. The data is shown as the mean ± SD (n = 3 per bar), **P < 0.01, ***P < 0.001 and ns, non-significant.
www.nature.com/scientificreports www.nature.com/scientificreports/ to the cells, and luciferase activity was determined 48 hours after transfection. Core promoter transcriptional activity was compromised in KO cells in comparison to the parental NTCP-HepG2 cells, but interestingly, the transcriptional activity was rescued upon ectopic expression of TARDBP (Fig. 4c,d). This observation implied that failure of the cells to activate the core promoter was due to the absence of TARDBP. We next evaluated HBV infection and replicative ability in the KO cells relative to the parental cells. The cells were infected with the HBV virus as in Fig. 1, and HBV-secreted virions and proteins were measured at the end of the infection period. Consistently, HBV gene expression was drastically reduced in the TARDBP KO cells in relation to cells expressing the protein but was partially rescued by the exogenous protein ( Fig. 4e-g). Although complementation with ectopic TARDBP exhibited total rescue of the core promoter activity in KO cells, HBV gene expression after HBV infection was only partially rescued. This could be due to the challenge in re-expressing the protein in KO cells for the entire infection period. Whereas the luciferase transcriptional assays were performed 48 hours after transfection when protein expression was still intact, the longer period of time required for HBV infection may have compromised the expression of the transiently transfected plasmid. Indeed, a time series assessment confirmed that the protein was only detected until 72 hours after transfection, beyond which the expression slowly became undetectable (Fig. S2B). Our attempts to establish stably-transfected TARDBP KO cells also failed. The reason for this was not clear, but TARDBP is known to control its own homoeostasis in human cells, as ectopic expression of the protein promoted instability of both the endogenous and ectopic mRNAs 22 . Moreover, other members of the hnRNP family have been shown to possess these auto-regulatory mechanisms 23,24 . This could explain our difficulties in expressing the protein. However, it is still interesting to note that the initial presence of the protein alone was responsible for a significant restoration of the HBV gene expression in KO cells.
The RNA Recognition Motifs (RRMs) of TARDBP are crucial to activate the core promoter. Our results to this point were not sufficient to determine whether the interaction of TARDBP and core promoter is direct or indirect. As noted earlier, TARDBP has two RNA Recognition Motifs (RRM1 and RRM2) through which it is able to bind to DNA and RNA targets. It is believed that both RRMs play a role in the nucleic acid binding by the protein 25 . In addition, the protein has a glycine rich C-terminal region through which it interacts with other proteins 14 . We therefore reasoned that TARDBP could either directly bind to the DNA sequences in the core promoter region or it could indirectly activate the core promoter through interaction with transcription factors that are known to bind and activate the core promoter. If TARDBP binds to the core promoter directly, then disruption of the RRMs could inhibit its transcriptional activity. To this end, we employed three TARDBP mutants in which either or both RRMs were deleted (Fig. 5a). Expression of these proteins in NTCP-HepG2 cells was confirmed by The T23 cell line, which stably expresses HBV plasmid, was transfected with the negative control siRNA (siControl) or the TARDBP siRNA (siTARDBP) in duplicate sets for 7 days. One set of the cells was lysed, and the TARDBP protein and core protein were quantified by western blotting using specific antibodies. (b) The supernatants were harvested from one set of the cells in (a) and analyzed for extracellular HBV DNA. (c) For the intracellular HBV DNA, protein lysates harvested in A were immunoprecipated with an anti-HBc antibody, followed by Southern blotting. (d) The total RNA was extracted from the second set of cells and subjected to RT qPCR to detect the HBV mRNAs (pregenomic-pg, precore, and total). The mRNA values were normalized against the GAPDH RNA internal control. The data is shown as the mean ± SD (n = 3 per bar), **P < 0.01, ***P < 0.001. www.nature.com/scientificreports www.nature.com/scientificreports/ western blotting using the anti-FLAG antibody (Fig. 5b), as they all contain the FLAG tag at the N-terminus. As a next step, we used the luciferase reporter assay to compare core promoter transcriptional activation by the three RRM mutant proteins versus the wild type protein. The ability of TARDBP to activate transcription from the core promoter was partially disrupted by deletion of RRM1 but was totally abolished by deletion of RRM2 or both RRMs (Fig. 5c). Immunofluorescence microscopy confirmed that each of the proteins maintained nuclear localization (Fig. 5d), indicating that interference with the transcriptional effect was not due to nuclear translocation. To further confirm the significance of the two RRMs in HBV DNA binding, we used the TARDBP ΔRRM1, 2 mutant, in which both nucleic acid binding domains were deleted for the ChIP assay. We confirmed that deletion of the two RRMs abolished the ability of the protein to interact with the core promoter (Fig. 5e). Taken together, control siRNA or siRNA specific to TARDBP to confirm specificity of the antibody. (b) The T23 cells were harvested at confluency and analyzed by ChIP assay using the antibody against TARDBP or the control rabbit IgG. Immunoprecipitated DNA was analyzed in triplicate by qPCR with primers specific for each of the HBV promoter DNA sequences (Core, Sp1, Sp2 and X). The results are displayed as the ratio of the amount of DNA bound to the TARDBP antibody to that bound to the control antibody, with the amount bound to the control antibody set to one. (c) To confirm precipitation of TARDBP and core promoter by the antibody, western blotting was performed using the TARDBP antibody, whereas the presence of core promoter DNA was detected by PCR using the core specific primers. The asterisk (*) indicates the location of the IgG heavy chain bands. (d) A schematic representation of the luciferase reporter plasmid pGL3-Cp that contains the luciferase reporter whose expression is controlled by the core promoter. (e) The pGL3-CP plasmid was co-transfected with the control vector pcDNA3.1/FLAG or increasing concentrations of pcDNA3.1/TARDBP-FLAG into NTCP-HepG2 cells as indicated. The pRL-TK plasmid, which expresses Renilla luciferase, was also included to monitor the transfection efficiency. Cells were lysed at 48 hours after transfection for the luciferase assay. Firefly luciferase values were normalized with control Renilla luciferase activity. The results are expressed as relative luciferase value, which refers to differences (n-fold) from the control value, which is set at one. The protein expression of TARDBP in the lysates was confirmed by western blotting using the anti-FLAG antibody. Error bars represent the SD of three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001 and ns-non significant.
www.nature.com/scientificreports www.nature.com/scientificreports/ our results imply that the RRM2 plays a dominant role, with RRM1 having a supportive role in the protein's transcriptional effect. Moreover, deletion of both RRMs totally abrogates the interacting ability of the protein and subsequent transcriptional effect, which confirms that the DNA binding ability of the protein is necessary for it to activate the core promoter.
TARDBP binds to a sequence within the core promoter of HBV. TARDBP has been shown to bind to relatively long DNA sequences; one group demonstrated binding of the protein to a 38 bp sequence of the SP-10 gene promoter 26 , whereas another study showed that it binds to a 43 bp sequence of the HIV TAR DNA 11 , and yet another study confirmed that it binds to a 66 bp sequence of the TNF-α gene promoter 27 . The minimum sequence necessary for DNA binding by TARDBP remains unknown. Since the core promoter spans the region nt1613-nt1849 of the HBV viral genome 28 , we randomly selected an approximately 46 bp target length to create 6 overlapping DNA probes from the entire 237 bp region for the EMSA experiment. Because TARDBP has been shown to sometimes bind to the antisense sequence but not the corresponding sense sequence 26 , we also included the antisense probes for each of the 6 probes to make a total of 12 probes. The probes were labeled as sense (S) probes 1-6 and anti-sense (AS) probes 1-6 in a consecutive manner starting from the C-terminus (Fig. S3). To screen for sequences among these probes to which TARDBP could bind, EMSA was performed using a GST-purified TARDBP recombinant protein that was successfully detected by anti-GST and anti-TARDBP antibodies (Fig. 6a). In the EMSA experiment, two shifted bands, indicating a possible DNA-protein interaction, were observed for 2 of the 12 probes, namely S probe 3 and AS probe 1 (Fig. S4A, lanes 4 and 8). However, further experimentation indicated that binding to the S probe 3 was only dependent on the formation of a secondary structure whose intensity was reduced in the presence of the protein, suggesting that the protein bound to the secondary structure and not the intact probe (Fig. S4A, lane 15 and Fig. S4B lane 5). Henceforth, we decided www.nature.com/scientificreports www.nature.com/scientificreports/ to focus on the AS probe 1 that indicated direct protein binding (Fig. S4B, lane 2). The probe was the antisense strand of nt1804-nt1849 whose sequence is shown in Fig. 6b. Indeed we confirmed binding of TARDBP to the probe, as the intensity of the shifted band increased in proportion to the dose of the protein (Fig. 6c, lanes 1-5). A similar pattern was observed when the probe was incubated with decreasing concentrations of a specific competitor (Fig. 6c, lanes 7-12). To rule out the possibility of non-specific binding, we tested binding of TARDBP to the adjacent AS probe of nt1765-nt1810 whose sequence is also shown in Fig. 6b. The protein failed to bind to the probe from the adjacent sequence, showing that the interaction was specific (Fig. 6c, lane 6). To further confirm binding of TARDBP to the 46 bp element, the EMSA reaction was performed in the presence of a TARDBP antibody. Indeed the antibody did cause a supershift (Fig. 6d, lane 3). However, we noticed that two super-shifted bands were present. We reasoned that this could be due to the presence of homodimers, as TARDBP molecules are known to fold and homodimerize 14,29 . To resolve this point, we immunoprecipitated a FLAG-TARDBP cell lysate using an anti-FLAG antibody followed by immunoblotting with an anti-TARDBP antibody, as exogenous tagged-TARDBP is known to appear at a slightly higher molecular weight than the endogenous one 30 . Indeed we observed two bands, one corresponding to the exogenous (FLAG-tagged) protein and the other to the endogenous protein (Fig. 6e) confirming self-interaction of TARDBP molecules. Altogether, the observations confirm that TARDBP directly binds to the core promoter DNA. TARDBP has two critical binding sites on the core promoter of HBV. To identify the elements within the 46 bp probe of the core promoter region that represent the specific targets for TARDBP binding, we first tested binding to four overlapping truncated fragments of the probe, namely s1-s4 (Fig. S4C). The wild-type probe interacted with TARDBP as expected (Fig. S4D, lane 2); however, under the same conditions, none of the four fragments could interact with the TARDBP protein (Fig. S4D, lanes 4, 6, 8 and 10). As a next step, we used mutational analysis to scan the sequence for possible target regions based on previous reports. TARDBP has been To compare the effect of these proteins on the core promoter activity, the pGL3-CP plasmid was transfected alongside equal amounts of the control plasmid, WT TARDBP, or each of the three RRM mutant plasmids into NTCP-HepG2 cells. Luciferase activities were measured at 48 hours after transfections. Relative values were calculated as described in Fig. 3e. (d) For intracellular localization of the proteins, the plasmids were transfected as in (b) above. After 48 hours, the cells were treated with the anti-FLAG antibody and visualized by immunofluorescence microscopy. The nucleus was visualized with DAPI. (e) To confirm that the two RRMs play a role in interaction of TARDBP with the core promoter, the FLAG tagged WT and the ΔRRM1, 2 mutant proteins were expressed in T23 cells, followed by a ChIP assay as in Fig. 3b using control and anti-FLAG antibodies. The resulting DNA was analyzed with the core promoter primers by qPCR. The relative values were normalized to the amount of DNA precipitated by the control which, was set at one. Results are reported as mean ± SD, *P < 0.05, **P < 0.01, ***P < 0.001 and nsnon significant.
www.nature.com/scientificreports www.nature.com/scientificreports/ shown to bind preferentially to DNA enriched for TT, TG or GT nucleotide pairs 31 . We identified these sequences at several locations within the probe. To study their significance, we generated a series of five mutant probes (mutant 1-5) where individual nucleotides were substituted, focusing specifically on locations with TT, TG or GT pairs (Fig. 6b). The mutants were subjected to an EMSA assay alongside the wild type probe. As shown, mutation of the central part containing a five A-nucleotide repeat did not affect binding of TARDBP to the probe (Fig. 6f,  lane 5). However mutations of T or G nucleotides abolished the binding of the protein to the probe (Fig. 6f,  lanes 3, 4, 6 and 7). Consistent with previous reports, TG and GT nucleotide pairs seem to play a central role in www.nature.com/scientificreports www.nature.com/scientificreports/ TARDBP binding to the core promoter. In addition, the results also show that TARDBP has two separate binding sites on the probe, as mutations in either the N-terminal or the C-terminal sequences abolished the interaction. Taken together, our results demonstrate that TARDBP has two binding sites on the core promoter and both of them are necessary for the interaction.
TARDBP assembles protein complexes that support the HBV virus in the nucleus. As mentioned earlier, aside from the nucleic acid binding properties of TARDBP, another important aspect of the molecule is the ability to interact with other proteins via the glycine-rich domain at the C-terminus (Fig. 5a). We therefore wondered whether HBV may take advantage of this in addition to the nuclear localization of the protein to assemble protein complexes that favor viral replication. We first screened for proteins that interact with TARDBP in a HBV replicating cell line using LC-MS/MS (Fig. 7a) and identified 491 potential interaction partners (Table S3)  www.nature.com/scientificreports www.nature.com/scientificreports/ pathways indicated significant enrichment of splicing, poly(A) RNA binding, and RNA processing pathways. To identify any proteins that could play a part in the HBV lifecycle, we selected the top proteins with coverage 10 or above and searched the literature using these proteins and HBV as search terms. Following this approach, 8 proteins were identified, namely PTBP1, PUF60, SFPQ, Hsp90, PARP1, HNRNPK, NPM1, and HNRNPC (Fig. 7b), which have been reported to play various roles in transcription and post-transcriptional and nucleocapsid assembly stages of the virus [32][33][34][35][36][37][38][39] . To verify these interactions in vivo, we immunoprecipitated a nuclear lysate of the same cells expressing FLAG-tagged TARDBP by an anti-FLAG antibody, followed by western blotting to detect proteins. Out of the 8 proteins, 5 were confirmed to interact with TARDBP (Fig. 7c). Altogether our results suggest that HBV benefits from the ability of TARDBP to assemble diverse proteins that support the viral life cycle in hepatocytes.
TARDBP regulates HBV pgRNA splicing. TARDBP is well known to be involved in RNA splicing 12 .
In addition, most of the TARDBP-interacting proteins identified in Fig. 7 have been reported to be involved in mRNA splicing events [40][41][42][43][44] . We therefore investigated whether TARDBP could play a role in HBV mRNA splicing. HBV undergoes reverse transcription during its replication and only utilizes unspliced mRNA for viral gene expression 45 . In addition to the unspliced mRNAs, a series of spliced (SP) HBV RNAs have been widely described in model systems and in HBV-infected livers 46 . The most frequently detected variant is a 2.2 kb molecule termed SP1, which is generated through the removal of a 1.3 kb intron from the pgRNA at nt2447 and nt489 45 . To determine any role for TARDBP in HBV pgRNA splicing, we measured the ratio of SP1 to the WT pgRNA in HBV producing cells with or without silencing of TARDBP. We employed two sets of primers to recognize intron-internal sequences to detect the WT product and those across the exons to detect the SP form (Fig. 8a). Specificity of the primers was confirmed, as SP primers did not amplify the WT product (Fig. 8b). As shown, diminishing TARDBP in the cells resulted in 100% increase in splicing of pgRNA (Fig. 8c), which implies that TARDBP serves as an inhibitor of splicing, thereby enhancing the export of unspliced pgRNA into the cytoplasm. As a next step, we tested whether TARDBP could bind to HBV RNA. To this end, total lysates were obtained from HBV-producing cells that were over-expressing FLAG-tagged TARDBP. They were then subjected to an RNA The T23 cells, which stably express HBV, were transfected in triplicate with siRNAs against TARDBP or control. Total RNA was extracted from the cells, followed by reverse transcription. The resulting cDNAs were analyzed by PCR followed by agarose gel electrophoresis with primers specific for WT pgRNA or SP1 RNA. The results confirmed that there was no nonspecific amplification of WT products with SP1 primers and that the two products were at the expected sizes. (c) The primers in (b) were henceforth utilized in a quantitative PCR of the samples to determine the proportion of SP1 RNA, which was calculated as SP1 RNA/(SP1 pgRNA + WT pgRNA). (d) Whole cell lysate from FLAG-TARDBP or FLAG-transfected T23 cells was immunoprecipitated with an anti-FLAG antibody and subjected to isolation of RNA and qPCR analysis using primers for TARDBP, APOA2 and total HBV mRNAs. The amount of mRNA that was precipitated from the TARDBP transfected cells was calculated as a ratio relative to that bound to the control-transfected cells for each of the primers. Relative mRNA values are presented as a ratio with respect to the amount of GAPDH mRNA, which was set at one. Results are reported as mean ± SD, while **P < 0.01. (2019) 9:8462 | https://doi.org/10.1038/s41598-019-44934-5 www.nature.com/scientificreports www.nature.com/scientificreports/ immunoprecipitation assay using TARDBP antibody as the bait. The precipitated mRNAs were purified and subjected to qPCR analysis to detect HBV mRNA. TARDBP and APOA2 mRNAs served as positive controls since they are already known to bind to TARDBP 22,47 . GAPDH mRNA was detected as a negative control to exclude non-specific interactions. As expected, TARDBP and APOA2 mRNAs were enriched on the TARDBP antibody (Fig. 8d). In addition, total HBV mRNA was also shown to be enriched on the TARDBP precipitate, indicating that the mRNA was precipitated by the protein (Fig. 8d). As a next step, we attempted to identify potential TARDBP HBV RNA binding sites by examining the HBV genome for conserved TG-repeats. We downloaded aligned genome sequences from HBVdb and performed a regular expression search for (TG)+ repeats. While we found a number of clusters of repeated T or G nucleotides throughout the HBV genome, we found few conserved TG stretches with more than two repeats (Fig. S6A). However, experimentally confirmed TARDBP RNA binding sites typically indicate some degree of flexibility in this basic pattern (Fig. S6B). Therefore, we performed a fuzzy search in which gaps were penalized over mismatches for each of the 11 experimentally determined TARDBP RNA-binding motifs listed in the RBPDB database of RNA-binding protein specificities. We uploaded the highest scoring matches in BED format to the NCBI Graphics panel for the HBV reference sequence KR819180.1. We observed a number of matches within the X transcript overlapping the core promoter/enhancer II region and found other clusters within the S and core transcripts, as well as a match downstream of the polyadenylation site (Fig. S6C). These results suggest that TARDBP might interact with HBV transcripts within the 3′ non-coding regions shared by all transcripts and might interact with HBV RNA in the same region as the DNA binding site in the core promoter/enhancer II region. In the future, we will attempt to experimentally confirm these potential RNA interaction sites and determine their role in HBV replication. In all, our data shows that in addition to DNA, TARDBP is also able to interact with HBV mRNA and, consequently, plays a role in inhibition of pgRNA splicing to enhance HBV gene expression.

Discussion
In this study we demonstrate for the first time that TARDBP binds both to HBV DNA as well as RNA and functions as a transcriptional activator through binding to the core promoter and as a splicing suppressor through binding to pgRNA. TARDBP was originally identified as a cellular factor that binds strongly to double-stranded HIV-1 TAR DNA to repress transcription from the HIV-1 LTR 11 . Subsequent studies confirmed this role for the protein as a transcriptional repressor 13,19,20 . However, more recent research has demonstrated that TARDBP can indeed also work as a transcriptional activator by binding to the LPS-responsive element in the TNF-α promoter to increase TNF-α expression 27 . Our study provides several lines of evidence that TARDBP acts as a positive regulator of HBV. First, silencing of TARDBP diminished HBV gene expression. Secondly, TARDBP could activate transcription from the HBV core promoter. Thirdly, HBV benefits from the nuclear localization of TARDBP and its ability to assemble protein complexes that are already known to support various stages of viral replication. Finally, the protein served an inhibitory role on pgRNA splicing, which is likely to facilitate nuclear export of the non-canonical unspliced viral RNAs for HBV gene expression.
Known HBV-associated transcription factors either bind to the core promoter directly or regulate transcription indirectly as co-activators or co-repressors via interaction with other molecules. For instance, several host transcription factors have been reported to indirectly activate the HBV core promoter, such as SIRT1, Cyclin D2 and jumonji C domain-containing 5 (JMJD5) [48][49][50] . On the other hand, many other transcription factors bind directly to the core promoter, including CAAT enhancer-binding protein alpha (CEBPA), hepatocyte nuclear factor 4 (HNF4), chicken ovalbumin upstream promoter transcription factor (COUP TF) 51 , among others. The present study adds a new transcription factor to this list, as we observed direct binding of TARDBP to the core promoter, resulting in a positive effect on transcription. Our findings confirm prior reports that TARDBP can function both as a transcriptional repressor and as an activator. This pattern has been reported for Spindlin1, a protein that could transcriptionally repress HBV in the context of infection by binding to HBV cccDNA, whereas previous research had focused on its role in transcriptional activation 52,53 . The molecular mechanisms responsible for both transcriptional activation and repression by TARDBP remain unreported. However, it is possible that the molecular switch between repressor and activator functions can be influenced by the infection model or cell type, post-translational modifications, or interactions with different proteins. It will therefore be important to characterize the conditions that favor one transcriptional function of TARDBP over the other.
By performing mutagenesis assays, we demonstrated that deletion of the RNA/DNA binding motifs abolished the transactivation effect of TARDBP on the core promoter. This observation is consistent with previous reports that both domains are necessary for nucleic acid interactions 14 . The motifs were also crucial for the ability of TARDBP to bind to HIV TAR DNA, as deletion of the glycine rich C-terminus had no effect, whereas deletion of the ribonucleoprotein binding motifs abolished the interaction 11 . In our case, we noted that deletion of RRM1 only partially reduced the effect, whereas deletion of RRM2 completely abolished the effect. This is contrary to previous reports in which RRM1 was found to play the dominant role, while RRM2 was dispensable or played only a supportive role 12,25,31 . It is conceivable that both domains are necessary for interaction with DNA, but their respective functions might depend on an unknown factor. On the other hand, while the glycine rich C-terminal domain of TARDBP was not necessary for binding to the core promoter, it still played a separate but important role in supporting HBV replication. As the motif is important for interaction with other proteins, we found a substantial number of factors that are already known to support the HBV viral lifecycle as interacting partners of TARDBP.
Whereas most studies have emphasized the preference of TARDBP for TG-rich DNA sequences 20,26,54 , it has recently become clear that the protein can also interact with other nucleotide motifs 31 . In agreement with Furukawa et al., we showed that regions containing TG, GG, TT, and GT nucleotide pairs play important roles in binding, whereas an intervening region with AA pairs was apparently dispensable, although we do not discount the possibility that this region could serve as a spacer sequence and that the nucleotide composition could allow (2019) 9:8462 | https://doi.org/10.1038/s41598-019-44934-5 www.nature.com/scientificreports www.nature.com/scientificreports/ the DNA molecule to flex in order to facilitate the formation of protein complexes. The initial report of TARDBP binding to HIV TAR DNA actually showed that the protein bound to pyrimidine-rich sequences but not to TG repeats 11 . They also demonstrated that two separate groups of pyrimidine-rich residues were crucial for TARDBP binding to TAR DNA, as mutations in either of the residues abolished the binding 11 . Accordingly, we observed that TARDBP failed to bind to the DNA sequence when either of the two target sequences was mutated. KLF15 is another protein which has been shown to bind to the core promoter via two binding sites, for which mutation of one of the sites abolished the binding 55 . While it is not yet known how TARDBP binds to two sites simultaneously, we speculate that two molecules of TARDBP bind to the core promoter in the form of homodimers, as TARDBP molecules are known to bind to one another 29 . In support of this model, our supershift assay demonstrated that two (instead of one) shifted bands existed in the same assay condition. In this scenario, we speculate that in the sample mixture, the antibody bound to either or both of the two TARDBP molecules in the protein-DNA complex, resulting in a single and double shift, respectively. This was further confirmed by an immunoprecipitation experiment where an anti-FLAG antibody precipitated both the FLAG-tagged (exogenous) and endogenous protein.
In this study we showed that DNA binding of TARDBP to the core promoter directly up-regulates HBV transcription. However, TARDBP is a multifunctional protein that can bind to either RNA or DNA and is well known to bind to 3′ UTR regulatory regions or splicing regulatory sites in RNA 56 . TARDBP is also known to interact with a number of proteins, most of which belong to two major functional categories, RNA processing in the nucleus and protein translation in the cytoplasm 57 . Our LC-MS/MS results revealed a large number of potential interaction partners (NPM1, PARP1, Hsp90, HNRNPC, SFPQ, PTBP1, HNRNPK and PUF60), several of which have already been reported to affect HBV replication via, e.g., transcriptional regulation, RNA processing, nuclear export, or nucleocapsid assembly [32][33][34][35][36][37][38][39] . HBV faces several RNA processing challenges, including avoiding RNA degradation, preventing unwanted splicing, and exporting the non-canonical viral transcripts from the nucleus. Export of mRNA from the nucleus normally involves formation of ribonucleoprotein particles and is tightly coupled to splicing; therefore, export of non-canonical unspliced HBV RNA transcripts and pgRNA from the nucleus poses a challenge for the virus and requires recruitment of adapter proteins to facilitate transfer through the nuclear pore. Given the ability of TARDBP to alter or silence RNA splicing and the protein's capacity to interact with a large number of proteins involved in RNA metabolism, as well as the presence of both nuclear import and export signals, we speculated that TARDBP might play a role in RNA processing and export of HBV transcripts. In support of this hypothesis, we were able to detect HBV mRNA in TARDBP precipitate, suggesting an interaction between HBV RNA and TARDBP. Furthermore, in a recent study, Duriez et al. performed RNA pull-down assays followed by LC-MS analysis to detect pgRNA-interacting proteins and identified TARDBP as a potential interaction partner 58 . They observed that 15% of the RNA-interacting proteins were directly associated with splicing and noted that pgRNA splicing results in formation of HBV splicing-generated protein (HBSP), which confers a protective effect against immune-mediated hepatic injury through down-regulation of CCL2 expression. In this study, we demonstrated that TARDBP inhibits splicing of pgRNA in T23 cells and showed that TARDBP knockdown enhanced pgRNA splicing. Further research is necessary to confirm whether TARDBP plays a role in HBV splicing and/or nuclear export, but our current findings suggest that disruption of TARDBP-protein complex formation may yield seeds for anti-HBV drug development.
Briefly, we identified TARDBP as a host factor that facilitates HBV gene expression by stimulating transcription from the core promoter, assembly of protein complexes implicated in transcriptional and post-transcriptional stages of the virus life cycle, and suppressing pgRNA splicing during nuclear export. Our results provide valuable insight into the interaction between host factors and HBV. TARDBP may thus serve as an effective target for novel anti-HBV therapies.

Materials and Methods
Plasmids, Oligos, siRNA and shRNA. Full-length TARDBP was amplified from a human brain cDNA library (Clontech) and cloned into the BamHI-HindIII site of pcDNA3.1 (−) (Invitrogen) with a 5′ FLAG tag inserted through the NotI-EcoRI site. The RRM mutants of TARDBP were generated by site-directed mutagenesis and cloned into the same vector as the full length plasmid. The control pcDNA3.1/FLAG plasmid only contains the flag tag at the same location as the TARDBP plasmids. These plasmids were a kind gift from Dr. Philipp Kahle (University of Tübingen, Germany). The pGEX-TARDBP plasmid that was used for GST-tagged protein production and purification for the EMSA experiments was constructed from the above TARDBP plasmid. In brief, a fragment containing human TARDBP CDS was amplified from the pcDNA3.1/FLAG-TARDBP plasmid by PCR using the primers, 5′-TGGTTCCGCGTGGATCCATGTCTGAATATATTCGGGTAACC-3′ and 5′-AGTCGACCCGGGAATTCTACATTCCCCAGCCAGAAGACTT-3′ and ligated into the BamHI-EcoRI site of pGEX-4T-1 vector. For the luciferase experiments, the HBV core promoter (CP) plasmid was constructed by inserting the nucleotides 1613-1849 of the HBV genome (GenBank accession number KR819180) into the pGL3 basic vector (Promega) through the XhoI-HindIII site.
All oligonucleotides used as primers or probes for PCR, EMSA and RNA IP experiments were designed and ordered from Sigma-Aldrich-Genosys, Japan.
C ontrol shRNA and shRNA TARDBP (TARDBP MISSION shRNA TRCN0000016040: CCGGGCTTTGGCTCAAGCATGGATTCTCGAGAATCCATGCTTGAGCCAAAGCTTTTT) in the form of Lentiviral Transduction Particles were purchased from Sigma Aldrich, Japan.
www.nature.com/scientificreports www.nature.com/scientificreports/ Plasmids used in establishment of NTCP-HepG2 knock out (KO) cells were constructed as follows: for the knock-in donor plasmids, the left and right homology arms of the TARDBP gene (Fig. S1A) were amplified by PCR from NTCP-HepG2 C4 cells genomic DNA. Antibiotic resistance ORF (puromycin or blasticidin) together with the set sequence of P2A, DD-tag and 3xGGGGS, which was fully chemo-synthesized (GenScript) according to the published protocol 59 , was inserted between the left and right homology arms of TARDBP genomic sequence (Fig. S1B,C) preliminarily cloned into pBlueScript II SK(+) using In-Fusion (Clontech). The pX330, a human codon-optimized SpCas9 and chimeric guide RNA expression plasmid, was obtained from Addgene (catalog no. 42230). Two sgRNA sequences neighboring to the start codon of TARDBP gene (TARDBP_guide 1 and TARDBP_guide 3 (Fig. 4a) were selected by CRISPR design tool 60 . The pX330 was linearized with Bbs I digestion and two pairs of oligonucleotides for TARDBP targeting sites were annealed and ligated into the Bbs I site of pX330, respectively. Avoiding re-digestion by Cas9 nuclease after desired knock-in occurred on the genome, point mutations with no-change on the amino acid sequence of TARDBP protein were inserted within each PAM sequence of guide1 and guide 3 (Fig. S1B,C) by site-directed mutagenesis.
Cell culture and treatments. The HepG2 cell line was derived from a human hepatoma cell line. The production of the T23 cell line from HBV-stably transfected HepG2 cells has previously been described 17 . Briefly, HepG2 cells were transfected with the plasmid pTRE-HB-wt by calcium precipitation and the transfected cells were selected. Colonies were isolated, and clones that were positive for both HBs and HBe antigens were selected. Finally, one cell line named T23 was identified and used for further experiments as it was able to continuously produce high copies of HBV DNA in supernatant. The T23 cells were grown in DMEM (Gibco) supplemented with 10% (v/v) fetal bovine serum and 400 μg/ml hygromycin (InvivoGen, CA, USA). NTCP-HepG2 is another cell line derived from the HepG2 line and is stably transfected with NTCP, the membrane transporter necessary for entry of the HBV virus into hepatocytes 5 . The NTCP-HepG2 C4 clone which was used in this study was a kind gift from Dr. Watashi and is highly susceptible to HBV infection 16 . The NTCP-HepG2 cells were cultured with DMEM/F-12 + GlutaMax (Gibco) supplemented with 10% FBS, 10 mM HEPES (Gibco) and 5 μg/ml insulin (Sigma) in the presence of 400 μg/ml G418 (Gibco). All plasmid transfections were carried out using the transit-IT-LT1 transfection medium according to the manufacturer's recommended conditions (Mirus Bio LLC), while the siRNA transfections were performed using Lipofectamine RNAiMAX transfection reagent according to the manufacturer's instructions (Invitrogen). Lentiviral particles were transduced into the cells in the presence of 10 μg/mL hexamethrine bromide according to the manufacturer's recommended conditions (Sigma-Aldrich), followed by selection of positive colonies with 1 μg/ml puromycin (InvivoGen).

HBV preparation and infection. HBV infection of NTCP-HepG2 C4 cells was performed using HBV
genotype C derived from the serum of a chronic HBV patient with high viral load (≥10 9 copies/ml). The patient agreed to provide blood samples for a viral hepatitis study through written informed consent. The study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki and was approved a priori by the ethical committee of Hiroshima University. The infection of cells was performed at 200 genome equivalents (GEq)/cell in the presence of 4% PEG8000 at 37 °C and 5% CO 2 for 16 h as previously described 61 . The infectious serum was then removed from the cells followed by washing three times with PBS and incubation in fresh medium. The subsequent culture medium contained 3% DMSO (Sigma) for the rest of the culture period, as DMSO has been shown to augment susceptibility of cells to HBV infection 16 . During the infection period, the medium was changed every four days. An aliquot of the supernatant was reserved for HBV gene expression analysis prior to changing of the culture medium at the indicated time points.
Quantification of extracellular HBV DNA, HBe and HBs Antigen. DNA was extracted from 100 μL of supernatant from the HBV-infected NTCP-HepG2 C4 or T23 cells using the SMITEST EX-R&D Nucleic Acid Extraction Kit (Medical & Biological Laboratories Co, Ltd, Nagoya, Japan) and dissolved in 20 μL of H 2 O. A 1 μL volume of the DNA solution was amplified, and HBV DNA copy numbers were determined by quantitative real-time PCR (qPCR), as we reported previously 10 , using the primers used for detection of total HBV RNA described below. HBs and HBe antigen levels were assayed by Chemiluminescence enzyme immunoassay (CLEIA) using the HISCL HBsAg Assay Kit and the HISCL HBeAg Assay Kit, respectively. Quantification was achieved by the Automated Immunoassay System HISCL-5000 (All from Sysmex Corporation, Kobe, Japan) according to the manufacturer's instructions.
Immunoblotting. Cultured cells were lysed with RIPA buffer consisting of 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1% (v/v) NP40, 0.5% sodium deoxycholate, 0.1% SDS and protease inhibitor cocktail (Nacalai Tesque, Kyoto, Japan). The lysates were incubated for 30 min at 4 °C and centrifuged at 14,000 × g for 15 min at 4 °C. Supernatants were reserved and protein concentration determined by the Bradford Protein Assay (Bio-Rad, Japan). A total of 30 μg of the proteins were boiled at 95 °C for 5 min in the SDS sample buffer (Bio-Rad, Japan) and separated by SDS-PAGE. This was followed by a transfer to nitrocellulose membranes, blocking (2019) 9:8462 | https://doi.org/10.1038/s41598-019-44934-5 www.nature.com/scientificreports www.nature.com/scientificreports/ no-antibiotics medium, and after 24 hours, the cells were co-transfected with each 0.5 μg of TARDBP_guide 1/ pX330, TARDBP_guide 3/pX330, TARDBP puro-DD knock-in plasmid and TARDBP bla-DD knock-in plasmid (Fig. S1), using Lipofectamine 3000 as per the manufacturer's instructions (Life Technologies). At 24 hours post-transfection, 100 nM Shield-1 (Clontech) was added to the culture medium, and cells were subsequently always maintained in Shield-1. Beyond 24 hours, the medium was replaced with that containing 1.5 μg/mL puromycin, 1.5 μg/mL blasticidin and antibiotics. At ≥30 days post-transfection, colonies were isolated and expanded. Gene knockout was confirmed by western blotting. In the destabilizing domain (DD) knock-in strategy, the protein is fused to the DD tag such that, in the absence of the DD specific ligand (Shield-1), the protein is degraded by the proteasome. However, when Shield1 is added to the culture medium, the fusion protein should accumulate within 24 hours. Unfortunately, our attempts to add Shield-1 led to failure to detect any TARDBP protein (Fig. S2A). Henceforth, for the rescue experiments, TARDBP was expressed in TARDBP KO cells using a transiently transfected plasmid.
Chromatin immunoprecipitation (ChIP) assays. For HBV promoter analysis, the ChIP assay was performed on genomic DNA samples from the HBV-producing cell line T23 using the ChIP-IT Express Enzymatic kit according to the manufacturer's protocol (Active Motif, Tokyo, Japan). Briefly, T23 cells were grown to confluency in 10 cm dishes and cross-linked with formaldehyde. The reaction was stopped by glycine followed by cell lysis with the cold lysis buffer provided in the kit. Chromatin was sheared from the lysates with the enzymatic shearing cocktail and immunoprecipitated overnight with the magnetic beads in the presence of antibodies. The antibodies used for immunoprecipitation were the rabbit anti-TARDBP and the rabbit control IgG (MBL, Life science). Chromatin was eluted from the beads and crosslinking reversed with 5 M NaCl. The samples were treated with proteinase K and subjected to qPCR analysis by the CFX96 TM Real-Time PCR Detection system (Bio-Rad, Tokyo, Japan). The region corresponding to each of the four HBV promoters (X, Core, SP1 and SP2) was amplified using the following sets of primers, respectively: X, 5′-CCGCTCGAGTGGCTCCTCTGCCGATCCATA-3′ (forward) 5′-CCCAAGCTTGGAAAGGAGG TGTATTTCCGA-3′ (reverse) Core, 5′-CCGCTCGAGAACCACCGTGAACGCCCGCCA-3′ (forward) 5′-CCCAAGCTTACATGAGATGA TTAGGCAGAG-3′ (reverse) SP1, 5′-CCGCTCGAGAGATCTCAATCTCGGGAATCT-3′ (forward) 5′-CCCAAGCTTCCACTGCATGG CCTGAGGATG-3′ (reverse) SP2, 5′-CCGCTCGAGGATCAGGGTTCACCCCACCAC-3′ (forward) 5′-CCCAAGCTTGAGATGGGAGTA GGCTGTCTC-3′ (reverse) Each of the reagents used for chromatin extraction and purification were provided in the kit. For HBV cccDNA ChIP assay, the nuclear extract from a chimeric mouse tissue was immunoprecipitated with control, TARDBP, or HBc antibodies using the Dynabeads Co-Immunoprecipitation Kit  Dual luciferase assay. The luciferase reporter plasmid driven by the HBV core promoter, pGL3-Cp was co-transfected with the pcDNA3.1-FLAG (empty vector), the pcDNA3.1-FLAG/TARDBP or the pcD-NA3.1-FLAG/TARDBP RRM mutant plasmids into NTCP-HepG2 cells. The pRL-TK plasmid (Promega, USA) was transfected with the reporter plasmid in all wells to normalize the transfection efficiency. At 48 hours after transfections, the cells were lysed in the Passive Lysis Buffer and assayed using the Dual-Luciferase Reporter assay system (both by Promega, USA) according to the manufacturer's instructions. Firefly and renilla luciferase activity was determined using the Mithras LB 940 Multimode Microplate Reader (Berthold Technologies).
Electrophoretic mobility shift assay (EMSA). 12 overlapping (6 sense and 6 anti-sense) probes of about 46 bp in size were prepared from the HBV core promoter spanning the region nt1613-1849 of the HBV genome. The probes were labelled with biotin using the Pierce Biotin 3′ End DNA Labeling Kit according to the manufacturer's instructions. The recombinant GST-TARDBP protein used in EMSA was produced as described above. The EMSA reaction was carried out using the Pierce LightShift ™ Chemiluminescent EMSA Kit. A 20 μl EMSA reaction was performed as per the manufacturer's protocol with 2 μg (or as indicated) of TARDBP protein, 1% BSA, 100 fmol of labeled probe, and ×200 (or as indicated) amount of unlabeled oligonucleotide competitor. For the super-shift assays, the EMSA reaction was performed in the presence of 2 μg of the rabbit anti-TARDBP antibody. After 30 minutes incubation at room temperature, the reaction products were separated by electrophoresis in a 5% polyacrylamide gel containing 0.5 × TBE. The samples were then transferred to a positively charged Biodyne ™ B Nylon Membrane and detected by the Pierce Chemiluminescent Nucleic Acid Detection Module Kit as per the manufacturer's protocol. The three kits used for EMSA experiments were purchased from Thermoscientific, Japan.