The yeast two-hybrid screening was performed in Saccharomyces cerevisiae using the Matchmaker system (BD Biosciences).

The gene coding for the S-HDAg was PCR amplified from pSVL-AgS [11], using the primers 5’-ttatcaattgatgag ccggtccgagtcg-3 (forward) and 5´-ttatgtcgacctatggaaatccct ggtttccc-3’ (reverse). The resulting fragment was cleaved (MunI/BamHI) and inserted into pAS2-1 vector (BD Biosciences) in the local EcoRI/BamHI, to obtain the plasmid pAS2-1/S-HDAg. This plasmid contains the S-HDAg entire gene fused to the GAL4 DNA binding domain (BD) and was used as bait to screen a human liver cDNA library representing fusions of the GAL4 activation domain (AD) pretransformed into the Y187 yeast strain.

Screening of the library was performed according to manufacturer’s instructions. To reduce the false positives after selection, the selected yeast colonies were transferred to Trp, Leu, His and Ade dropout nutritional selection media and assayed for MEL1 activity.

Prey plasmids encoding human liver proteins that interacted with S-HDAg in the MEL1 positive yeast clones were isolated after transformation in E. coli DH5α. AD plasmids were recovered from positive yeast clones and retransformed into native AH109 yeast with the bait plasmid pAS2-1/S-HDAg to retest the interactions. Finally, the recovered plasmids were subjected to DNA sequencing.

The bacterial expression vectors pGEX-6P-2 (GE HealthCare) and pET28c (Novagen) were used to produce GST or His-tagged fusion proteins, respectively.

The fusion vector for expression of GST/S-HDAg (pGEX-6P-2/S-HDAg) was constructed by digesting plasmid pEGFP-C1-HDAg (Tavanez et al. 2002) using the restrictions enzymes BglII and SalI followed by insertion of the obtained S-HDAg cDNA fragment into the XhoI/BamHI cloning sites of pGEX-6P-2. To construct the plasmid pET28c/HuR for expression of His₆/HuR fusion proteins, the HuR cDNA cloned in pCMV_SPORT6 (Image clone # 2901220) was amplified using the following primers: 5’-ttatcaattgccatgtctaatggttatgaagaccacatgg-3’ (forward) and 5’- ttatgtcgaccgagttatttgtgggacttgttgg-3´ (reverse). The resulting fragment was cleaved with MunI and SalI and cloned into pET28c, in the EcoRI/SalI sites.

Resulting recombinant plasmids pGEX-6P-2-HDAg and pET28c/HuR were used to transform E. coli BL21 and E. coli BL21-CodonPlus-(DE3)-RP competent cells (Stratagene), respectively. The expression of both GST/S-HDAg and His₆/HuR were induced with 1 mM isopropyl-B-D-thiogalactopyranoside (IPTG) at 37º C for 2 h. Cells were then harvested by centrifugation, washed with phosphate-buffered saline (PBS) and resuspended in PBS containing 0.1mg/mL lysozyme (Sigma). After three freeze-thaw cycles the obtained samples were treated with DNase I (Roche).

For purification of recombinant S-HDAg, protein extracts were loaded onto a GSTrap FF 1mL column (GE Healthcare). After washing, the GST-tag was removed by digesting with the PreScission protease (GE Healthcare) according to the instructions of the manufacturer. Finally, the eluted protein was dialyzed against PBS and quantified using a Bradford assay based kit (BioRad).

Two different cell lines were used in this study. The human hepatocellular carcinoma cell line (Huh7) and the corresponding HDV cDNA stably transfected cell line (Huh7-D12), which constitutively expresses HDV ribonuc-leoproteins [35]. Cells were grown as monolayers in RPMI 1640 medium (Sigma) supplemented with 10% fetal bovine serum (Invitrogen). Cells were cultured at 37ºC in a humidified atmosphere containing 5% CO₂.

Blot overlay assays were performed essentilally as described [36]. Briefly, increasing amounts of bacterial extracts containing His₆/HuR recombinant protein were separated on 12% SDS/PAGE gels. The proteins were subsequently transferred to a nitrocellulose membrane that was then overlaid with 40 µg purified recombinant S-HDAg protein. After washing, bound S-HDAg was detected by immunoblot using a rabbit polyclonal anti-S-HDAg antibody (B3; [37]) followed by incubation with a secondary anti-rabbit IgG antibody conjugated with peroxidise (BioRad). Membrane development was achieved using an enhanced chemiluminescence system (ECL; GE Healthcare) according to the specifications of the manufacturer.

For coimmunoprecipitation, Huh7 or Huh7-D12 cells were collected in lysis buffer [50mM Tris-HCl (pH 7.5), 150mM NaCl, 1% Nonidet P-40 and 10% glycerol] supplemented with complete protease inhibitor cocktail (ROCHE). Cell extracts were precleared with protein G Dynabeads (Invitrogen) and incubated for 2 h at 4ºC with a polyclonal rabbit anti-HDAg antibody (B3; [37]). Protein G Dynabeads (Invitrogen) were then added and immunoprecipitates were washed three times with PBS. Antigen-antibody complexes were recovered by boiling in SDS sample buffer and analyzed by western blot as described below.

To prepare protein samples for western blot, Huh7 or Huh7-D12 cells were first harvested, washed twice with PBS, and pelleted by centrifugation. The cell pellet was re-suspended in SDS sample buffer and boiled. Western blot was performed as previously described [38].

Protein samples, either recombinant or immunoprecipitated proteins, were separated by SDS-PAGE on 12% polyacrilamide gels and electroblotted onto nitrocellulose membranes (Schleicher & Schuell). Blots were then incubated for 1 h with one of the following primary antibodies: mouse monoclonal antibody against HuR (Santa Cruz Biotechnology), rabbit polyclonal anti-HDAg antibody B3 [37], and mouse anti-His-tag monoclonal antibody (Sigma-Aldrich). When appropriate the control house-keeping protein clathrin was also detected using an anti-clathrin antibody (Santa Cruz Biotechnology), and used to normalize for the amount of each selected protein. After washing with PBS containing 0.1% Tween 20 and 2% low fat milk powder, blots were further incubated with appropriate secondary antibodies conjugated with horseradish peroxidase (BioRad). Membrane development was performed using an enhanced chemiluminescence system (ECL; GE Healthcare).

When needed the density of the bands was quantified using the ImageJ software (http://rbs.info.nih.gov/ij/).

Indirect imunofluorescence and in situ hybridization were essentially performed as earlier described [38]. In brief, for immunofluorescence, cells grown on coverslips were rinsed briefly in PBS, fixed with 3.7% formaldehyde and permeabilized with 0.5% Triton X-100 in PBS, for 10 min, at room temperature. Cells were then rinsed with PBS containing 0.05% Tween 20 (PBS-T) and incubated with primary antibodies (polyclonal rabbit anti-HDAg serum B3 and/or mouse monoclonal against HuR; Santa Cruz Biotechnology). After washing with PBS-T, cells were incubated with appropriate secondary antibodies conjugated to either FITC or Texas Red (Jackson Immunoresearch Laboratories). Finally, the coverslips were mounted in Vectashield (Vector Laboratories, UK).

In situ hybridization was performed using a digoxigenin-11-dUTP (Roche) nick translated plasmid pSVL(D3) [39] as probe. This plasmid contains a trimer of full-length genomic HDV cDNA cloned in pSVL (GE Healthcare). After fixation and permeabilization, as described above for immunofluorescence, cells were sequentially treated with 70% formamide at 73ºC followed by incubation with 50% formamide at 73ºC, for 5 min, and finally subjected to hybridization. After hybridization, digoxigenin detection was performed using a monoclonal anti-digoxigenin antibody conjugated with FITC (Roche) followed by incubation with a secondary anti-FITC antibody conjugated with Alexa-488 (Jackson Immuno Research Laboratories). Finally, the coverslips were mounted in Vectashield (Vector Laboratories, UK).

Fluorescent-labeled samples were analyzed under a Zeiss LSM 510 META confocal microscope. Green fluorescence was detected using an Argon Ion laser (488 nm) and red fluorescence was detected using a Helium-Neon laser (543 nm). Calibration of the equipment was achieved using multicolor fluorescent beads (Molecular Probes, USA) and a dual-band filter which allows simultaneous visualization of green and red fluorescence.

The quantitative colocalization analysis was performed using the JACoP plugin (Just Another Co-localization Plug-in; [40]) and the ImageJ software (http://rbs.info.nih.gov/ij/). The Mander´s overlap coefficient (OC), which expresses the correlation of pixel intensities in both analyzed channels was used for presentation of results. This coefficient ranges from 0 to 1, where 0 indicates no overlap and 1 reflects 100% colocalization between both channels of the analyzed image [41].

To knockdown endogenous expression of HuR protein in Huh7 cells, we used the pSIREN-RetroQ vector (BD Biosciences). The oligonucleotide sequences used for silencing the expression of HuR were predicted using a bioinformatic tool (http://bioinfo.clontech.com/rnaidesigner):

5´-gatccgaggcaattaccagtttcattcaagagatgaaactggtaattgcctctttttt acgcgtg 3’ (sense) and

5´-aattcacgcgtaaaaaagaggcaattaccagtttcatctcttgaatgaaactggtaa ttgcctcg-3’ (antisense), where the underlined nucleotides denote the HuR shRNA sequence targeting HuR mRNA (nucleotides 484 to 492 in HuR cDNA).

To construct plasmids used for RNAi the oligonucleotides were first allowed to anneal in vitro. This was performed by ressuspending in TE buffer followed by heating to 95ºC for 30 sec and two sequential incubations at 72º C and 37º C, both for 2 min. Finally the oligonucleotides were allowed to cool at room temperature to generate double stranded DNA (dsDNA). The obtained dsDNA was cloned into the EcoRI/BamHI digested pSIREN-RetroQ vector thus generating the pSIREN-RetroQ/HuR. To construct the control vector pSIREN-RetroQ/Luc, the oligonucleotides provided by the manufacturer for luciferase gene silencing were introduced into an identical vector according to the corresponding instructions.

For transfection, the FuGENE6 transfection reagent (Roche) was used according to the specifications of the manufacturer. First 1 µg plasmid pSVL(D3) was used to transfect Huh7 cells. After 48 hours incubation, cells were transfected again with plasmid pSIREN-RetroQ/HuR or the negative control (pSIREN-RetroQ/Luc). 24 hours post-transfection, puromycin was added to the medium (2µg/mL) to select for pSIREN-RetroQ plasmids transfected cells. Following additional 48 hours incubation, cells were harvested for subsequent western blot analysis. Western blots were performed as described before using the following primary antibodies: rabbit polyclonal anti-HDAg antibody (B3; [37]), mouse anti-His-tag monoclonal antibody (Sigma-Aldrich), rabbit polyclonal anti-clathrin antibody (Santa Cruz Biotechnology), and a mouse monoclonal antibody against HuR (Santa Cruz Biotechnology). After incubation with appropriate anti-mouse or anti-rabbit secondary antibodies conjugated with horseradish peroxidase (GE Healthcare) blots were developed using an ECL system (GE Healthcare).

To identify proteins capable of interacting with S-HDAg, this protein was fused to the C terminus of the DNA binding domain of GAL4 (GAL4BD) and used as bait in a yeast two hybrid screen of a human liver cDNA library containing approximately 1 x10⁸ clones. A total of 112 positive clones were obtained. After sequence analysis and database search, one of the positive clones was found to bear a plasmid containing a cDNA fragment encoding about 96% of the entire open reading frame (ORF) of the ELAV-like 1 (ELAVL1) protein, in frame with the GAL4 activation domain (Fig. 1A).

Fig. (1). HuR interacts with S-HDAg in a yeast two-hybrid assay. (A) Schematic representation of HuR and the AD fusion protein expressed in yeast positive clone (AD/HuR). The protein AD/HuR includes a truncated form of HuR, devoid of the first N-terminal 12 aa. GAL-AD represents the activation domain of the GAL4 transcription factor and RRMs represent the RNA recognition motifs in the HuR protein. (B) AH109 yeast cells were (co)transformed with the indicated plasmids. Interactions were tested by growth selection on dropout media lacking the amino acids Leu or Trp (SD/-Leu or SD/-Trp) to select for pGADT7/Rec or pAS2-1 plasmid constructs, respectively. Activation of reporter genes was tested on plates lacking Leu, Trp, Hist and Ade, and containing X-α-Gal (SD/-Trp/-Leu/-Hist/-Ade/X-a-Gal).

Fig. (2). In vitro and in vivo interaction of HuR and S-HDAg. (A) Overlay of bacterially expressed His6-HuR with recombinant S-HDAg. E. coli protein extracts were prepared after induction of recombinant protein expression. Increasing amounts of extracts containing His6-HuR (lanes 2 to 6), were separated by SDS/PAGE, blotted onto nitrocellulose membranes and overlayed with the same amount of purified recombinant S-HDAg. Detection of S-HDAg was performed with a specific rabbit polyclonal antibody. E. coli protein extracts lacking His6-HuR were used as a negative control (lane 1). (B) Coimmunoprecipitation of HuR with S-HDAg. Huh7-D12 cell lysates were immunoprecipitated with an anti-HDAg antibody bound to protein G beads. The immunoprecipitates were run separated on 12% SDS/PAGE gels and immunoblotted with an anti-HuR or anti-HDAg antibody. The negative controls were prepared in the absence of the anti-HDAg antibody or using Huh7 cell protein extracts.

Fig. (3). Colocalization of HuR with HDV antigens and RNA. (A-C) Double indirect immunofluorescence was used to detect HDAg and HuR proteins in cultured Huh7-D12 cells. Cells were fixed with 4% formaldehyde, permeabilized with 0.5% Triton X-100, and stained with anti-HDAg antibody (A, green staining) and anti-HuR antibody (B, red staining). In situ hybridization followed by immunofluorescence was performed to detect HDV RNA and the HuR protein. After fixation and permeabilization as above, cells were hybridized with a digoxigenin-labeled probe to detect HDV RNA (D, green staining) and with an anti-HuR antibody (E, red staining). Overlaps of images are shown in panels C and F. The JaCoP plugin and ImageJ program were used to determine the Mander´s overlap coefficient.

Fig. (4). HuR overexpression in Huh7-D12 cells. (A) Western blot analysis of HuR protein in Huh7 and Huh7-D12 cells. Total protein extracts were separated by SDS-PAGE, transferred to a nitrocellulose membrane, and incubated with anti-HuR and anticlathrin antibodies. Images are representative of three independent experiments. (B) The blots were digitalized and band intensity was calculated using the ImageJ program. Relative expression levels were obtained after normalization with the housekeeping clathrin protein.

Fig. (5). Silencing of endogenous HuR reduces expression of HDAgs in Huh7-D12 cells. (A) Knockdown of the HuR protein was performed using a specific siRNA in Huh7-D12 cells. 48 hrs after addition of puromycin, total protein extracts were prepared, separated by SDS-PAGE, and transferred to a nitrocellulose membrane. Blots were incubated with anti-HuR, anti-clathrin or anti-HDAg antibodies, and the displayed images are representative of three independent experiments. (B) Images were digitalized and band intensity was determined using the ImageJ program. Relative expression levels were obtained after normalization with the housekeeping protein clathrin. Inhibition of HDAg expression is presented as a percentage after comparison with the values obtained with the negative control vector pSIREN-RetroQ/Luc provided by the manufaturer.

Human ELAVL1 (GenBank NP_001419), also known as HuR (Hu antigen R), is a 326 amino acid protein and has been described as a RNA binding protein, comprising three RNA binding domains (RBDs), also known as RNA recognition motifs (RRMs). The sequence identified in the yeast two- hybrid screening comprises amino acids 12 to 326 of the full length HuR (Fig. 1A), including the three RRMs.

The HuR protein was previously reported to participate in post-transcriptional regulation processes and has also been proposed to be an important player in the replication cycle of several viruses, including HIV-1, HCV, and HVS [30, 31, 42]. Thus, it is possible that this protein might also involved in HDV replication. To investigate this hypothesis we first decided to confirm the results obtained in the yeast two-hybrid screening. Yeast AH109 cells were cotransformed with the bait (pAS2-1/S-HDAg) and prey plasmids isolated from the positive yeast clone (pGADT7/Rec/HuR_12-326), and the interaction between the two proteins was tested using three reporter genes (ADE2, HIS3, and MEL1). Yeast AH109 cells transformed with pAS2-1/S-HDAg or pGADT7-Rec/HuR_12-326 alone were used as controls in these experiments. In contrast with the negative controls, yeast cells expressing both BD/S-HDAg and AD/HuR_12-326 fusion proteins could effectively grow in minimal medium lacking tryptophan, leucine, histidine, and adenine (SD/-Leu/-Trp/-Hist/-Ade), and displayed strong α-galactosidase activity (Fig. 1B). These results allowed confirming the interaction of S-HDAg and HuR in the yeast two-hybrid system.

After establishing that S-HDAg and HuR interact in the yeast two-hybrid system, we decided to investigate whether the two proteins also interact in human liver cells. We first tested this interaction in vitro using a blot overlay assay. The vector pGEX-6P-2/S-HDAg was used to produce GST/S-HDAg fusion protein in E. coli cells. GST/S-HDAg protein was purified from bacterial extracts using glutathione sepharose columns and the GST tag was cleaved with PreScission Protease to obtain untagged S-HDAg. Purification of recombinant untagged S-HDAg was evaluated by western blot, and protein concentration was determined using a Bradford assay (data not shown). Expression of His₆-HuR was induced in E. coli cells transformed with plasmid pET28c/HuR. Increasing amounts of the obtained protein samples were loaded onto different wells of a 12% SDS-PAGE gel. After separation proteins were transferred to a nitrocellulose membrane which was subsequently overlaid with 40 µg of recombinant S-HDAg. Detection of bound S-HDAg was performed using a specific polyclonal antibody. The obtained results are displayed in Fig. (2A) showing that increasing amounts of the His₆-HuR protein result in a proportional increase in the amounts of bound S-HDAg, suggesting a specific interaction between the two proteins.

The in vivo interaction of HuR with S-HDAg was further investigated using a coimmunoprecipitation assay and Huh7-D12 cells. This cell line was previously reported to express constitutively both forms of the delta antigen and virus RNA [35, 38]. Since HuR is ubiquitously expressed in proliferating cells [43], the eventual interaction of the two proteins should be detectable in this system. Huh7-D12 protein extracts were prepared and incubated with a rabbit polyclonal antibody against HDAg. Immunoprecipitated protein samples were separated by SDS-PAGE, transferred to a nitrocellulose membrane, and immunobloted with an anti-HuR antibody. Similar procedures in the absence of anti-HDAg antibodies or using Huh7 cells protein extracts were performed in these experiments as negative controls. As shown in Fig. (2B), in Huh7-D12 cells protein extracts HuR was coprecipitated together with the S-HDAg being consistent with an in vivo association of the two proteins in human liver cells.

It has been previously shown that the HuR protein is involved in post-transcriptional modifications of RNAs and may modulate the replication of HIV-1, HCV, HSV, and HPV [30-33]. Since S-HDAg and HuR interact both in vitro and in vivo, it may be plausible that HuR is also involved in the control of some stages of the HDV replication cycle. To test this possibility our first approach was designed in order to investigate an eventual colocalization of HuR with the virus RNA and antigens inside the cell. This was performed by immunofluorescence and in situ hybridization followed by confocal microscopy and image analysis of Huh7-D12 cells, which constitutively express HDV RNPs [35, 38].

In accordance with previous findings [38] the delta antigens were found to be localized in the nucleus of Huh7-D12 cells, with a preferential accumulation in foci (Fig. 3A). The HuR protein also showed a nuclear localization pattern and seemed to colocalize with the HDAgs (Fig. 3B, C). The extent of colocalization of S-HDAg and HuR was analyzed quantitatively by calculating the Mander´s overlap coefficient (OC). This coefficient gives an estimate of intensity-based correlation of green (HDAg) and red (HuR) signals. The average OC value (n=10) for HuR and S-HDAg is 0.772 (std. deviation= 0,113; n=10) indicating a significant colocalization. The results obtained are consistent with the hypothesis that HuR and HDAg physically interact in Huh7-D12 cells.

Previous reports showed that HDAgs and HDV RNA colocalize in the nucleus of human liver cells [38]. Since HuR is a RNA binding protein we decided to investigate if the distribution pattern of the protein in relation to the virus RNA. To do this, a combined in situ hybridization and immunofluorescence approach was utilized. A digoxigenin-labelled probe was used to hybridize with HDV RNAs and the HuR protein was detected by immunofluorescence, as above described. Typical examples of the obtained results can be observed in Fig. (3D, F). The average OC value obtained (0.762; std. deviation = 0.074; n=10), is very similar with the one determined for HuR/S-HDAg colocalization. Taken together, our results indicate that the HuR protein interacts with S-HDAg in human liver cells, and additionally suggest that it may also interact with the virus RNA.

It is known that viruses subvert the cellular metabolism namely by recruiting host factors to locals of replication in the nucleus and by altering gene expression in infected cells. Accordingly, it has been previously reported that in human hepatoma cells, synthesis of HDV RNA and antigens is accompanied by changes in expression levels of several host proteins [21, 22]

In an attempt to investigate if the synthesis of HDV RNPs also induces alterations in the expression levels of the HuR protein we decided to compare the relative amounts of the protein in Huh7 and Huh7-D12 cells. Protein extracts from both cell lines were prepared, separated by SDS-PAGE, and analyzed by western blot using an antibody to detect the HuR protein. An antibody recognizing the housekeeping protein clathrin was used as an internal control in these experiments. The blot images were digitalized and band intensities were calculated by densitometry using the ImageJ program. The obtained results showed that the expression levels of HuR were increased 1.7-fold in Huh7-D12 cells when compared with Huh7 cells (Fig. 4). Considering that only 10-15% of Huh-D12 cells are expressing HDV RNPs at a given moment [38] this finding may be indicative of a possible involvement of the HuR protein in the virus replication cycle.

The HuR protein was previously shown to play an important role in mRNA stabilization and control of translation (reviewed in [25]). Since we found that HuR binds in vivo to S-HDAg, both proteins colocalize in Huh7 cells, and HuR is overexpressed during synthesis of HDV RNPs, we decided to investigate the effect of HuR knockdown on HDAg synthesis.

To do this we utilized a small hairpin RNA mediated knockdown technology based on a pSIREN-RetroQ vector coding for a HuR-specific short hairpin RNA (pSIREN-RetroQ/HuR). Huh7 cells were first transfected with plasmid pSVL(D3) in order to produce HDV RNA and antigens, and 48 hours later transfected again with the pSIREN-RetroQ-HuR plasmid. Puromycin was added to the medium in order to select the pSIREN-RetroQ/HuR transfected cells, and protein extracts were prepared. A pSIREN-RetroQ vector, targeting the luciferase mRNA (pSIREN-RetroQ/Luc) was used in parallel transfections and used as negative control in these experiments. The efficiency of reduction of cellular HuR protein levels by the shRNAs coded by the pSIREN-RetroQ/Hur vector was monitored by western blot. As shown in Fig. (5), the plasmid-derived expression of shRNA against HuR mRNA, reduced the expression of the HuR protein by 55% when compared with control pSIREN-RetroQ/Luc transfected cells.

We next investigated the effect of HuR silencing in the synthesis of delta antigens. This was performed by analyzing the relative expression of HDAgs by western blot of protein samples obtained from the same knockdown experiment. The results are displayed in Fig. (5) and show that knockdown of HuR resulted in a decrease of 65% and 93%, in S-HDAg and L-HDAg expression levels, respectively. These results indicate that HDAg expression is inhibited when HuR is silenced in Huh7 cells, suggesting that this protein may play an important role in HDV replication.