Identification of an effective siRNA target site and functional regulatory elements, within the hepatitis B virus posttranscriptional regulatory element

The functional elements within the PRE (nucleotides 1151-1684, of Accession number AM282986) were analyzed using results of CDS-Plotcon and Alidot programmes provided by the database HBVRegDB [23]. In brief, a set of 32 completed HBV genomes were analysed by the CDS-Plotcon programme [38] to specifically detect regulatory elements that are present within the coding sequence. They were also analysed using the Alidot programme [39] to determine conserved RNA secondary structures.

The siRNA target sites within the PRE sequence were predicted using siExplorer [40], siDirect [41], and siRNA target designer [42] programmes. Characteristics of desired siRNAs identified by Reynolds et al (2004) [22] were also taken into consideration. Nucleotide blast (blastn) was performed to check specificity of predicted siRNA target sites against human genomic and human transcript databases (Table 1 and additional file 1). The selected sequences were chosen to target PRE positions 1317-1337 and 1329-1349.

An intronless reporter vector [pBasic (-IN)] was constructed by removing the chimeric intron of the pGL3MS2site/Basic reporter construct [43]. To construct a splicing luciferase reporter vector (pSpliceLuc), the parental vector, pGL3MS2site/Basic was modified to place the luc+ gene between the splicing donor (SD) and splicing acceptor (SA) sites. This modification involved two-step cloning. The first cloning involved amplification of the SD (forward: SV40+ 5'-ctgactaattttttttatttatgc-3', reverse: SDR_AflII 5'-gaattccttaagccttaaacctgtcttgtaacc-3') and then the amplification of the SA sequence (forward: SAF_XhoI 5'-gaattcctcgagagaccaatagaaactgggc-3', reverse: SAR_EcoRV 5'-gaattcgatatccctgtggagagaaaggcaaagtg-3').

Three pairs of primers were specifically designed to amplify three sub-sections of the HBV PRE: (i) full length HBV PRE (forward: HBVPRE_1151F 5'-tctagagctagcttgct cggcaacggcctggtctgtg-3', reverse: HBVPRE_1684R 5'-gccggcctcgaggacattgctgaga gtccaagagtcc-3'); (ii) HBV PRE 1399-1684 (forward: HBVPRE_1399F 5'-tctagagc tagctggatccttcgcgggacgtcctttg-3', reverse: HBVPRE_1684R 5'-gccggcctcgagga cattgctgagagtccaagagtcc-3') and (iii) HBV PRE 1485-1584 (forward: HBV PRE_1485F 5'-tctagagctagctcgtccccttctccgtct-3', reverse: HBV PRE_1584R 5'-gccgg cctcgaggtgcacacggaccggcagat-3'). The Amplification was performed from a clone containing the complete HBV genome (a gift from M-H Lin, National Taiwan University) using Expand™High Fidelity DNA polymerase (Roche). Short fragments of the PRE 1292-1321 and HBV PRE 1411-1433 were created by annealing synthetic oligonucleotides: (i) forward: HBVSL_alpha oligoF 5'-ctagcgttttgctcgcagccggtctggggcaaagcc-3', reverse: HBVSL_alpha oligoR 5'-tcgaggctttgccccagaccggctgcgagcaaaacg-3'; and (ii) forward: HBVSL_beta oligoF 5'-ctagcgggacgtcctttgtttacgtcccc-3', reverse: HBVSL_beta oligoR 5'-tcgaggggacgtaaacaaaggacgtcccg-3'.

Selected siRNA target sites at positions 1317-1337 and 1329-1349 were converted into shRNA template oligonucleotides by adding the loop sequence, the target site's antisense strand and a termination signal for the RNA pol III (Figure 2A). For cloning purposes, overhangs of half restriction enzyme sites for BamHI and HindIII were flanked at the 5'-end and the 3'-end of the template respectively. Sequences of the shRNA templates for targeting HBV PRE 1317-1337 and 1329-1349 were as follows: PRE1317 5'-gatccagctcatcggaactgacaattcaagagattgtcagttccgatgagctttttttggaaa-3'; 5AtPRE1317 5'-agcttttccaaaaaaagctcatcggaactgacaatctcttgaattgtcagttccgatgagctg-3'; PRE 1329 5'-gatccgctgacaattctgtcgtcctttcaagagaaggacgacagaattgtcagttttttggaaa-3'; AtPRE1329 5'-agcttttccaaaaaactgacaattctgtcgtccttctcttgaaaggacgacagaattgtcagcg-3'. Then, the two complementary hairpin siRNA oligonucleotides (each contains 1 μg/μL) were annealed together and ligated with the cut pSilencer 3.0-H1 promoter vector (Ambion).

Cell lysates were separated on 4-12% Bis-Tris gel (NuPAGE^® Novex gel, InvitrogenTM Life Technologies) and electrophoretically transferred onto polyvinylidene difluoride (PVDF) membrane (Hybond-P, Amersham Pharmacia Biotech). Blots were blocked with 5% of skim milk in TBS-T buffer for 1 h and subsequently incubated at 4°C for overnight with appropriate diluted primary antibodies, anti-Luc (1:500, Roche), anti-GAPDH (1:2,500, Ambion) and anti-PABP (1:10,000, Abcam). Then blot was incubated with diluted horseradish peroxidase-conjugated secondary antibody, goat anti-mouse (1:10,000, BIO-RAD) at room temperature for 1 h. For chemiluminescent detection, the immuno-blot was applied with the ECL Plus reagents (Amersham Pharmacia Biotech) and exposed to X-ray films (HyperfilmTM, Amersham Bioscience) for 15 s - 10 min at room temperature. All exposed films were then processed and qualified by imaging densitometry (Molecular Analyst software).

HuH-7, HepG2 and COS-7 cells were cultured in 75 cm³ sterile tissue culture flasks (Greiner Bio-One) at 37°C with 5% CO₂ in DMEM supplemented (Invitrogen) with 10% heat inactivated FBS (10% v/v) (Invitrogen) and 1% L-glutamine (Invitrogen). Prior to transfection, cells were seeded on 24-well plates (Greiner Bio-One) with a cell density approximately 1X 10⁴ cells/mL and incubated for 24 h. All transfection was performed using FuGENE6 (Roche). The ratio between FuGENE6 (μL) and DNA (μg) was 3:1. For deletion analysis, 195 ng of the deletion series of HBV PRE reporter plasmids were transiently co-transfected in quadruplicate with 5 ng of the internal control plasmid (phRL-SV40: Promega, a plasmid expressing humanized Renilla luciferase protein). pUC18 was used to top up the total amount of DNA if required. For evaluating effect of siRNA expression plasmids, cells were co-transfected in triplicate with 95 ng pSpliceLuc/fPRE or pBasic (-IN)/fPRE, 5 ng of phRL-SV40 and various amounts (0 ng, 300 ng, 600 ng and 900 ng) of the siRNA expression plasmids. For studying the siRNA effect of the pShRNA PRE 1317-1337 on the luciferase protein, cells were transiently co-transfected in triplicate with 45 ng of pSpliceLuc/fPRE or pBasic (-IN)/fPRE, 5 ng of the phRL-SV40 and either 60 ng or 300 ng of the pShRNA PRE 1317-1337 construct. The pSilencer-Negative was used to make up the total of plasmid DNA to 350 ng. The pSilencer-GAPDH was also included in the experiment as the positive control for the siRNA effect on the GAPDH protein.

Forty-eight h post-transfection, cells were lysed with 100 μL of 1 × passive lysis buffer (Promega). Cell debris and nuclei were removed by centrifugation and the supernatant was collected. The luciferase activity assay was performed as described by Tanguay and Gallie (1996)

A plasmid expressing covalently closed HBV genome (EMBL:AM282986) was constructed in pGEM-T Easy Vector (Promega, Madison, WI) through a T-A cloning strategy. Serial 10-fold dilutions of the cccDNA standard plasmid from 10 to 10¹⁰ copies/μL were detected by real-time PCR assay and used to prepare the standard curve for quantitation of HBV cccDNA. The standard curve of HBV cccDNA was then constructed by plotting the logarithm of the initial plasmid concentration against the threshold cycle (Ct) obtained from each dilution. The standard plasmid DNA for quantitation was included in each run as an external standard. HBV cccDNA was amplified and quantified in real-time PCR assay using the primers and probe as described previously [44]. The forward primer was HBV_CCC_F1 (5'-actcttggactc cagcaatg-3'); the reverse primer was HBV_CCC_R1 (5'-ctttatacgggtcaatgtcca-3') and the cccDNA specific probe was FAM-ttcaagcctccaagctgtgccttg-BHQ1. The optimized real-time PCR reaction mixture comprised 1 μL of DNA template, 0.75 μM final concentration of each primer, 0.25 μL of the probe, 5 μL of 2 × Platinum qPCR Super Mix-UDG (Invitrogen, California, USA), additional 2.5 mM MgCl₂, and nuclease-free water to a final volume of 10 μL. The real-time PCR assay was carried out in a Rotor Gene RG-3000 (Corbett Research, Australia) under the following conditions: initial denaturing step at 95°C for 10 min, followed by 45 cycles of 95°C for 15 s and 61.5°C for 1 min. Then the Rotor-Gene Software Version 6.0 (Corbett Research) was used for data acquisition and analysis of the HBV cccDNA level. The result was indicated in term of relative quantitation by comparative threshold (delta-delta Ct) method (2^{- ΔΔCt}). The amount of target gene in the sample, normalized to an endogenous housekeeping gene (reference gene) and relative to the normalized calibrator, is then given by 2 ^{- ΔΔCt}, where

ΔCt (sample) = Ct (target gene of sample) - Ct (reference gene of sample)

ΔCt (calibrator) = Ct (target gene of calibrator) - Ct (reference gene of calibrator)

Ratio (folds of difference) of sample: calibrator = 2 ^{- ΔΔCt}

In this study, the reference gene was beta-globin, the target gene was the cccDNA of HBV co-transfected with siRNA, and the calibrator was cells transfected with only the HBV plasmid.

The HBV post-transcriptional regulatory element is highly conserved among the mammalian hepadnaviridae [23]. As the sequence of the PRE also encodes the P protein, the conservation may partly be due to constraints on the encoded protein. This study therefore analyzed functional core elements of the PRE using results generated by the CDS-plotcon programme, which scores conservation beyond what is required for coding [38] as well as using programmes for predicting conserved RNA secondary structure [39]. CDS-plotcon indicated four conserved elements within the functional PRE at nucleotide positions 1151-1310, 1390-1450, 1515-1610 and 1540-1684 (EMBL: AM282986, Figure 1B). Three of these potential conserved regulatory elements were predicted to form local RNA secondary structures by Alidot (Figure 1C). The results of both programmes therefore suggested three putative functional conserved elements at nucleotide positions 1151-1410, 1411-1433 and 1510-1620 (Dashed boxes, Figure 1).

Notably, the previously identified regulatory elements: the human La binding site [30], SRE-1 [24] and HBV SL alpha [32] are shown to be part of a large secondary RNA structure within the predicted functional element at the position 1151-1410 whereas the reported HBV SL beta [32] and the PRE III [34, 35] are part of the identified functional elements at the position 1411-1433 and 1520-1620 respectively (Figure 1). In addition, known regulatory elements at the DNA level are also found to be part of the region identified as the functional elements (high mountain peak) generated by the CDS-plotcon such as the DNA enhancer 1 (nucleotide position 900-1310), the X promoter (nucleotide position 950-1310), the DNA primer binding DR1 (nucleotide position 1590-1600) and the DNA enhancer 2 (nucleotide position 1636-1744) [45, 46].

Taken together, the results support that the PRE is an important regulatory region that contains multiple functional conserved elements.

Although several sites in the HBV genome have previously been able to be targeted by siRNA, there is no report of targets within this region of PRE. To predict siRNA target sites, the sequence was analysed using three programmes, siExplorer [40], siDirect [41] and siRNA target designer [42]. In addition, the criteria of ideal siRNAs reported by Reynolds et al (2004) [22] were also taken into consideration. The predicted siRNA target sites were checked for specificity using nucleotide blast (blastn) against human genomic and transcript databases. Predicted siRNA target sequences that had more than 85% identity to human genomic DNA or transcripts were designated as non-specific siRNA targets and not used. As a result, an overlapping region of predicted siRNA target sites was detected by three different approaches (Table 1). Selected siRNA target sites were chosen to cover this region, at nucleotide position 1317-1337 and 1329-1349 (Figure 1A). These two predicted siRNA sites were found within the identified putative functional PRE element (nucleotide position 1151-1410) and they are part of a large predicted conserved RNA secondary structure (Figure 1). Selected siRNA target sites were converted into shRNA template oligonucleotides (Figure 2A) and ligated with the cut siRNA expression vector (pSilencer 3.0-H1, Ambion). The generated siRNA expression plasmids were designated as pShRNA PRE 1317-1337 and pShRNA PRE 1329-1349. Subsequently, various amounts (0 ng, 60ng, 300 ng, 600 ng and 900 ng) of the generated siRNA expression plasmids were transiently co-transfected in triplicate with 95 ng of luciferase reporter vector (pSpliceLuc/fPRE or pBasic (-IN)/fPRE) and 5 ng of Renilla expression plasmid (phRL-SV40) using FuGENE6. The experiment also included the positive control shRNA plasmid (pSilencer-GAPDH, Ambion), which targets the human GAPDH mRNA and the negative control plasmid (pSilencer-Negative, Ambion) a scambled sequence that is not found in the human genome. Cells were harvested at different time points (1 day, 2 days, 3 days post-transfection).

The result shows that the pShRNA PRE 1317-1337 could specifically and significantly reduce the level of luciferase activity at the day 2-time point (Figure 2B, p < 0.001) even with a low amount (60 ng) of the siRNA expression plasmid (Figure 2C). In contrast, the presence of pShRNA PRE 1329-1349 in different amounts showed no effect on luciferase expression at any time point (Figure 2B) although it was selected by similar criteria and position to the effective siRNA target site 1317-1337 (Table 1 and Figure 1). Therefore, the results suggest that specific properties of siRNA target sites are more significant than others for effective targeting. The level of pBasic (-IN)/fPRE was also significantly reduced by this siRNA (60 ng, by 43% and 300 ng, by 79%).

This experiment was carried out to evaluate whether the anti-HBV PRE 1317-1377 could also inhibit HBV replication in infected cells. This was done by measuring the level of HBV covalently closed circular DNA (cccDNA) in HepG2 cells that were transiently co-transfected with 30 ng of a HBV clone that expresses HBV and with 0 ng, 100 ng or 600 ng of the pShRNA PRE 1317-1377. Forty eight h post-transfection, cells were harvested to analyze the level of cccDNA using quantitative real time PCR. The results indicated that the plasmid expressing siRNA anti-HBV PRE 1317-1377 significantly reduced the level of cccDNA in transiently infected cells (Figure 3, p < 0.001).

Previous reports showed that new formation of cccDNA in transfected cells was directly controlled by the expression of HBV transcripts [47, 48]. As this siRNA target site (HBV PRE 1317-1337) is present in all HBV transcripts, it is possible that any or all HBV transcripts were reduced by the siRNA, resulting in the reduction of level of cccDNA.

To investigate the functional core elements of the PRE, a deletion series of the PRE was designed based on these predictions (CDS-plotcon and Alidot) and previous reports on PRE regulatory elements, HBV SL alpha (nucleotide position 1291-1321), PRE III (nucleotide position 1485-1584) (Figure 1). Each PRE subsection was then specifically generated by PCR and then inserted into two different luciferase (luc+) reporter constructs digested at NheI and XhoI sites: (i) the splicing luc+ reporter construct (pSpliceLuc), and (ii) the intronless luciferase reporter construct [pBasic (-IN)] (Figure 4A). Notably, the pSpliceLuc construct has the luc+ gene within an intron, thus it could be used to study whether the PRE could enhance the unspliced luc+ gene expression. On the other hand, the pBasic (-IN) reporter construct was designed to study functional nuclear export of PRE by imitating the natural context of the intronless HBV S transcript where PRE is located the 3' UTR of the gene. Subsequently, a deletion series of the PRE reporter plasmids were transiently co-transfected in quadruplicate with the phRL-SV40 vector in HuH-7 and COS-7 cells.

The full length (fPRE) significantly enhanced unspliced luc+ gene expression in both HuH-7 (Figure 4B, p < 0.001) and COS-7 cells (data not shown). This result suggests that the fPRE either inhibit splicing or enhance nuclear export, or both. Surprisingly, the PRE sub-section 1399-1684 significantly inhibited unspliced luc+ gene expression (p < 0.001) whereas PRE sub-section 1292-1321 (HBV SL alpha), HBV PRE 1411-1433 (HBV SL beta) and HBV PRE 1485-1584 (PRE III) did not individually enhance the expression of unspliced luc+ (Figure 4B). This result was consistent with previous published results, using the cat reporter system (pDM138), this construct has a design similar to the unspliced luc+ reporter construct used in this study. The previous report indicated that duplication of HBV SL beta and HBV SL alpha was required to enhance the level of CAT activity in the cat reporter system (pDM138) [26]. Indeed, six copies of HBV PRE III were reported to increase the CAT activity (pDM138 reporter system) in the same level as the full-length the PRE [49].

On the other hand, the results from the intronless luc+ construct showed that none of the PRE sub-sections including the fPRE were able to enhance the expression of the intronless luc+ gene. Interestingly, the PRE sub-section 1399-1684 also significantly inhibited the intronless luc+ gene whereas PRE alpha, PRE beta had no effect on the expression of intronless luc+ gene (Figure 4C). Previously using northern blot analysis and primer extension, the PRE has been shown to significantly increase cytoplasmic export of the HBV S RNA [27, 50]. It has also been reported to function in the context of heterologous genes by enhancing expression of intronless transcripts of β-globin and c- myc [27, 50‐52]. In this study, the PRE failed to increase activity of the intronless Luc+ protein (Figure 4C). Therefore, experimental results from this study provide evidence suggesting that the ability of the PRE to enhance expression of intronless transcripts is not applicable to all intronless genes. It is possible that the PRE may not have an effect on the highly expressed gene luc+ whereas it did on poorly expressed reporters (e.g. cat). Therefore, the results of PRE deletion analysis from the intronless luc+ system might not be able to conclusively evaluate the identified functional elements. Subsequent studies should be conducted to test the function of these PRE elements in a natural context using pgRNA (C and P proteins) or surface (S) protein.

Interestingly, the PRE sub-section 1399-1684 significantly reduced luciferase activity (Figure 4B, p < 0.001). This result was also observed in the intronless luc+ reporter system (Figure 4C, p < 0.001). The result may suggest either that the PRE sub-section 1399-1684 contains a novel inhibitory element or that the PRE sub-section 1151-1398 is an important element for the function of the PRE. Taken together, the PRE appears to contain multiple weak regulatory elements, but some aspects regarding the function of the PRE are still unclear.