Inhibition of hepatitis C virus replication by single-stranded RNA structural mimics

Abstract

AIM: To examine the effect of hepatitis C virus (HCV) structural mimics of regulatory regions of the genome on HCV replication.

METHODS: HCV RNA structural mimics were constructed and tested in a HCV genotype 1b aBB7 replicon, and a Japanese fulminant hepatitis-1 (JFH-1) HCV genotype 2a infection model. All sequences were computer-predicted to adopt stem-loop structures identical to the corresponding elements in full-length viral RNA. Huh7.5 cells bearing the BB7 replicon or infected with JFH-1 virus were transfected with expression vectors generating HCV mimics and controls. Cellular HCV RNA and protein levels were quantified by real-time polymerase chain reaction and Western blotting, respectively. To evaluate possible antisense effects, complementary RNAs spanning a mimic were prepared.

RESULTS: In the BB7 genotype 1b replicon system, mimics of the polymerase (NS-5B), X and BA regions inhibited replication by more than 90%, 50%, and 60%, respectively. In the JFH-1 genotype 2 infection system, mimics that were only 74% and 46% identical in sequence relative to the corresponding region in JFH-1 inhibited HCV replication by 91.5% and 91.2%, respectively, as effectively as a mimic with complete identity to HCV genotype 2a. The inhibitory effects were confirmed by NS3 protein levels. Antisense RNA molecules spanning the 74% identical mimic had no significant effects.

CONCLUSION: HCV RNA structural mimics can inhibit HCV RNA replication in replicon and infectious HCV systems and do so independent of close sequence identity with the target.

Key Words: Hepatitis C virus, Japanese fulminant hepatitis virus, Complementarity, RNA sequence, Hybridization

INTRODUCTION

Chronic hepatitis C virus (HCV) infection is a leading cause of chronic hepatitis, liver cirrhosis and hepatocellular carcinoma[1-3]. Current therapeutic options are limited and associated with significant adverse effects[4,5]. Accordingly, there is a strong impetus to develop novel therapeutic options. Because HCV replicates through the interaction of its RNA-dependent RNA polymerase and its RNA template to generate progeny RNA, physical contact between the enzyme and genomic RNA is required[6-8]. We hypothesized that mimics could be created with sufficient resemblance to the natural genomic structure to compete with the HCV genome, resulting in inhibition of replication. HCV RNA has a number of cis-acting replication elements (CREs) whose function could potentially be inhibited by structural RNA mimics[9]. For example, structural mimics based on the HCV internal ribosome entry site (IRES) were recently shown to inhibit HCV translation in vitro and in replicon cell culture[10]. Synthesis of (+)-strand RNA molecules was inhibited by RNA aptamers specific for the (-)-strand IRES domain corresponding to the 3’-terminal end of the negative strand of HCV RNA[11,12]. In addition to the IRES[13-15], HCV RNA bears CREs in the positive-strand NS5B[6,8,16,17] and X region[18,19] as well as in the negative strand 3’-terminal region[13,20]. The function of each CRE is assumed to depend on the ability of these structures to bind host factors[21-24] and viral non-structural proteins[8,14,15,25,26]. The objective of the current study was to determine to what extent computer-predicted single-stranded HCV RNA structures could inhibit HCV replication.

MATERIALS AND METHODS

Cloning of HCV RNA structural mimics

For replicon studies, structural mimics were constructed based on RNA sequences of the HCV 1b subgenomic replicon, BB7[27,28] (Table 1). Table 1 shows the regions and the nucleotides from which the mimic sequences were derived. Mimics 5B and X were designed to be identical to the (+) strand of the NS5B and X regions (Figure 1A and D, respectively). Mimics BA (Figure 1B) and EC (Figure 1C) were designed to be identical to the (-) strand of the 3’-terminus. The HCV RNA structural mimic sequences were predicted to adopt stem-loop structures identical to the corresponding CRE in full-length viral RNA, as determined by the Mfold ver 3.2 web server[29,30]. For Japanese fulminant hepatitis virus (JFH-1) infection studies, mimic sequence identities corresponding to HCV JFH-1 genotype 2a were designed to vary from 100% (5B-100), 74% (5B-74) and 46% (5B-46), the latter achieved by changes in the stem regions. The 5B-74 mimic was identical to the mimic (NS5B) used in genotype 1b replicon studies. To construct 5B-46, base-pair exchanges were made at positions 2-15, 18, 30-43, 49-53, 57-60, 75-78, 89-94 in 5B-74 (Figure 2). Regions in the stems or loops in which changes were predicted by Mfold ver 3.2 to alter the secondary structure were left intact. DNA fragments containing each mimic sequence flanked by BamHI and HindIII sites were generated by polymerase chain reaction (PCR) amplification of sequences from pHCV rep1bBB7 (Apath, LLC, St. Louis, MO, USA). As a negative control, an unrelated [hepatitis B virus (HBV) encapsidation signal] was amplified by PCR from plasmid adwR9 (from Dr. T Jake Liang, NIH, Bethesda, MD, USA). The region cloned included HBV sequences flanking a 60 nucleotide (nt) element, so that the total insert length approximated that of the HCV RNA structural mimics. Each insert was cloned into a T7 transcription vector, pENT7 by BamHI/HindIII digestion and ligation. pENT7 is a derivative of pENTR4 (Invitrogen Co., Carlsbad, CA), in which a T7 expression cassette (BamHI and HindIII sites downstream of the consensus T7 RNA polymerase promoter) had been inserted in the multiple cloning site.

Table 1 HCV RNA structural mimics.

Structural analogs	Source BB7 sequence (nt)	Source JFH sequence (nt)	Replicon strand	JFH strand	Stem-loop domains
NS-5B (5B-74)	9216-9310		(+) strand		NS5B coding region
X	9508-9605		(+) strand		X region
EC	107-222		(-) strand		3'-terminus
BA	1-104		(-) strand		3'-terminus
5B-100		9284-9378		(+) strand	NS5B coding region

HCV: Hepatitis C virus; JFH: Japanese fulminant hepatitis.

Figure 1 Predicted folding of RNA sequences designed as mimics of hepatitis C virus (HCV) genotype 1b regions. HCV RNA sequences were predicted by Mfold ver 3.2 to adopt stem-loop structures identical to the corresponding sequences in full-length viral RNA. A: 5B-74 corresponding to NS-5B (+) strand; B: BA region 3’-terminus of (-) strand; C: EC region 3’-terminus of (-) strand; D: X region of (+) strand.

Figure 2 Diagram showing the nucleotide replacements (in bold) in 5B-74 made to create the 5B-46 structural mimic and predicted stem-loop structures identical to the corresponding sequences in 5B-74.

For expression of HCV RNA structural mimics as polymerase II transcripts in mammalian cells, each insert was subcloned into pSilencer 4.1-CMV puro (Ambion) by BamHI/HindIII digestion and ligation. To confirm fidelity, clones containing pSilencer 4.1-CMV puro with HCV RNA structural mimics were sequenced with primers as recommended by the manufacturer: 5'-AGGCGATTAAGTTGGGTA-3', 5'-CGGTAGGCGTGTACGGTG-3'.

Cloning of a 5B-100 structural mimic

As a positive control for HCV JFH-1 genotype 2a inhibition, a 5B-100 mimic was constructed to be 100% identical to the HCV JFH-1 genotype 2a NS5B coding region (Table 1). This mimic was also predicted using web server Mfold ver. 3.2 to adopt stem-loop structures identical to the corresponding CRE in full-length viral RNA.

Cells

Huh7.5 cells, a human hepatoma cell line, bearing the BB7 HCV genotype 1b replicon were obtained from Apath (St. Louis, MO). Huh7.5 cells, which support JFH-1 replication, were obtained from Dr. Charles M. Rice, Rockefeller University, NY, USA[16,31,32]. Cells were maintained in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1X antibiotic/antimycotic solution (Invitrogen). Cells were passaged every 3-4 d to maintain 75% confluence.

Transfection of plasmids for in situ generation of HCV structural mimics

Huh-7.5 cells containing the BB7 replicon, or cells carrying persistent infection with JFH-1 HCV RNA were seeded at a density of 10⁵ cells in 2 mL of growth medium per well in 6-well plates, 6 d before transfection. For transfection, 25 μg of each plasmid encoding a structural mimic were transfected into Huh7.5 cells, 95% confluent, with 15 μL Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. Plasmid DNA was diluted in 250 μL of Opti-MEM I Reduced Serum Medium (Invitrogen) without serum. Lipofectamine 2000 was diluted to the appropriate amount in 250 μL of Opti-MEM I medium. After 5 min incubation, diluted DNA was combined with diluted Lipofectamine 2000, mixed and incubated for 20 min at 25°C. Complexes in 500 μL medium were added. After 6 h of incubation at 37°C under 50 mL/L CO₂, 1.5 mL of DMEM with 15% FBS were added to each well, and cells were incubated at 37°C under CO₂ for 48 and 72 h. Cells were harvested and lysates assayed for NS3 protein levels by Western blotting. Total RNA was isolated and HCV RNA levels were determined by real-time reverse transcriptase PCR (RT-PCR).

HCV infection system

A cDNA from hepatitis C genotype 2a virus, JFH-1 strain (from Dr. Takaji Wakita, National Institute of Infectious Diseases, Tokyo, Japan)[33] has been shown to replicate autonomously without drug selection or adaptive mutations[34-37]. This cDNA was used to transfect the full-length JFH-1 genome into Huh7.5 cells to produce infectious HCV[38].

Real-time RT-PCR

RNA was isolated from BB7 replicon cells with Trizol (Invitrogen), and treated with RNase-free DNase (Promega). One microgram of DNase-treated total RNA was reverse transcribed using an iScript cDNA Synthesis Kit (BioRad Laboratories, Hercules, CA, USA). After incubation at 25°C for 5 min, at 42°C for 30 min and at 85°C for 5 min, the resulting cDNA was quantified by real-time RT-PCR with SYBR GREEN according to the manufacturer’s protocol (Roche Applied Science, Indianapolis, IN, USA) using HCV genotype 1b specific primers: forward primer: 5'-CTGTCTTCACGCAGAAAGCG-3' and reverse primer: 5'-CACTCGCAAGCACCCTATCA-3'. Homo sapiens lactate dehydrogenase A (LDHA) mRNA levels in each sample were simultaneously quantified to normalize values of HCV RNA. The primer sequence for LDHA forward primer was: 5'-TAATGAAGGACTTGGCAGATGAACT-3' and reverse primer: 5'-ACGGCTTTCTCCCTCTTGCT-3'. Assays were performed in triplicate and results expressed as mean ± SD of HCV replication as a percent of unrelated control.

For the JFH infection system, RNA replication in transfected cells was quantified by RT-PCR using HCV genotype 2a specific forward primer: 5'-TAGGAGGGCCCATGTTCAAC-3', reverse primer 5'-CCCCTGGCTTTCTGAGATGAC-3'. The PCR conditions were: 2 min at 50°C, 10 min at 95°C and 15 s at 95°C. After 40 cycles, final extension was performed at 60°C for 1 min. Assays were performed in triplicate and results expressed as mean ± SD of HCV replication as a percent of unrelated control.

Western blotting analyses

Total protein lysates of Huh7.5 cells stably infected with JFH-1 were evaluated by Western blotting. Cells were harvested in RIPA buffer (50 mmol/L Tris-HCl, pH 7.4, 1% NP-40, 0.25% Na-deoxycholate, 150 mmol/L NaCl, 1 mmol/L EDTA) supplemented with protease inhibitors (Roche). Protein concentrations were determined with Bio-Rad assay as described by the manufacturer. Forty micrograms of protein was resolved by SDS-PAGE, and transferred to Hybond nitrocellulose membranes (Amersham Pharmacia, Piscataway, NJ, USA). Membranes were blocked with 5% nonfat milk in PBS, incubated with a 1:1000 dilution of the monoclonal mouse HCV NS3 antibody (Biodesign International, Saco, ME, USA), washed with PBS/0.05% Tween 20 and incubated with horseradish peroxidase (HRP)-conjugated goat anti-mouse antibody (Pierce, Rockford, IL, USA) at 1:20 000 dilution. Bound antibodies were detected with SuperSignal chemiluminescent substrate (Pierce) and quantified by Gel Imaging software GeneTools (SynGene, Frederick, MD, USA). To ensure comparable loading of samples, blots were incubated with a 1:1000 dilution of a polyclonal rabbit α-tubulin antibody (Abcam Inc., Cambridge, MA, USA), and detected by HRP-conjugated goat anti-rabbit secondary antibody using procedures as described above.

Evaluation of possible antisense effects

Because the 5B-74 structural mimic was single-stranded, and retained significant (74%) sequence identity to JFH-1 genotype 2a sequence, it was possible that the observed inhibition of replication was a result of antisense effects. To assess this possibility, 4 complementary RNAs 21-28 nt long, spanning the 5B-74 mimic were synthesized (IDT-Integrated DNA Technologies, Inc., Coralville, IA, USA). The antisense sequences were: 5B-C1 (nt 5-21), 5'-CCGGCUGCGUCCCAGUUGGAU-3', 5B-C2 (nt 22-43), 5'-UUAUCCAGCUGGUUCGUUGCUG-3', 5B-C3 (nt 44-71), 5'-GUUACAGCGGGGGAGACAUAUAUCACAG-3' and 5B-C4 (nt 72-95), 5'-CCUGUCUCGUGCCCGACCCCGCUG-3'.

One day before transfection, JFH-1-infected Huh7.5 cells were plated in growth medium without antibiotics, to be 95% confluent at transfection. Oligomer-Lipofectamine 2000 complexes were prepared by dilution of 50 pmol or 500 pmol RNA oligomer in 250 μL of Opti-MEM I without serum. Five microliters of Lipofectamine 2000 was diluted in 250 μL Opti-MEM mixed gently and incubated for 5 min at 25°C. The diluted oligomer was combined with diluted Lipofectamine 2000, mixed and incubated for 20 min at 25°C. Complexes were added to each well containing cells and medium. After 6 h, 1.5 mL of 15% FBS diluted in DMEM was added and the cells were incubated at 37°C in a CO₂ incubator for 48 h. HCV RNA replication was determined by real-time RT-PCR using HCV specific primers. Assays were performed in triplicate and results expressed as mean ± SD of HCV replication as a percent of unrelated control.

RESULTS

For replicon studies, structural RNA mimics were constructed based on sequences of the HCV 1b subgenomic replicon, BB7. Because HCV RNA replication involves generation of both positive and negative strands, structural mimics were designed for both strands. These targets were selected based on previous reports that HCV RNA bears CREs in the positive-strand NS5B coding region[6,8,16,17] and X region[18,19] as well as in the negative strand 3’-terminal region[13,39].

The results of transfection of RNA mimics into BB7 replicon cells quantified by real time RT-PCR are shown in Figure 3. The 5B mimic inhibited HCV replication by more than 90%. Mimics to X and BA regions decreased replication by 50%, and 60%, respectively. The EC mimic actually had no inhibitory effect, and in fact appeared to increase replication, but the effect was not statistically significant. The fact that this sequence, also identical to the natural HCV sequence, failed to inhibit replication indicates that the observed inhibition by the NS5B mimic did not result from a nonspecific effect of HCV RNA. The inhibition by the NS5B plasmid was found to be dose-dependent up to 25 μg (data not shown).

Figure 3 HCV RNA levels in the BB7 replicon cell lysates at 48 h post-transfection with structural mimics. BB7 replicon cells were transfected with plasmids generating structural mimics, and HCV RNA levels were assayed by real-time reverse transcriptase polymerase chain reaction (RT-PCR). The columns and bars represent the mean ± SD, respectively of 3 independent triplicate transfections. Statistically significant differences relative to controls are indicated by (^aP < 0.05).

Because the NS5B structural mimic appeared to be the most effective, further experiments were performed to determine whether this mimic might be effective against a different genotype in JFH-1-infected Huh-7.5 cells. It was found that the NS5B, now named 5B-74, mimic which was only 74% identical to the genotype 2a sequences of NS5B was as effective in suppression of HCV replication of JFH-1 genotype 2a as 5B-100 which was 100% identical to the genotype 2a (Figure 4).

Figure 4 Effects of transfection of NS5B mimics 5B-74, 5B-46 and 5B-100 on Japanese fulminant hepatitis-1 (JFH-1) genotype 2a viral infection in Huh7. 5 cells. Cells stably infected with JFH-1 genotype 2a virus were transfected with plasmids generating structural mimics, and HCV RNA levels were assayed by real-time RT-PCR. Assays were performed in triplicate and results expressed as mean ± SD of HCV replication percents of unrelated hepatitis B virus (HBV) control.

The intent of the design of the structural mimics was to create single-stranded versions of the natural structure to compete for protein interactions. However, it is possible that the observed inhibitory effects could have been simply the result of non-specific effects. To explore this possibility, the mimic 5B-46 was prepared in which bases were changed at positions 2-15, 18, 30-43, 49-54, 57-60, 75-78, and 89-94 (Figure 2), with a sequence identity of the stems of only 27.5%, and overall identity of 46%, relative to JFH-1 (genotype 2a) NS5B.The 5B-74 mimic decreased HCV replication by 91%. Furthermore, in spite of a sequence identity of only 46%, the novel mimic 5B-46, decreased HCV replication in JFH-1-infected cells by 90%, not significantly different from the mimic 5B-74 with 74% identity, or the mimic 5B-100 with complete identity to the HCV genotype 2a (Figure 4).

To assess viral protein levels, Western blottings of the NS3 protease were performed. Levels of neither the housekeeping protein tubulin (lane 2, Figure 5B) nor the NS3 protease (lane 2, Figure 5A) were affected by the unrelated HBV control mimic. However, the 5B-74 mimic decreased NS3 protease levels by 90% (lane 3, Figure 5A), while tubulin levels remained unchanged (lane 3, Figure 5B), the latter arguing against a non-specific effect or toxicity by the mimics. Quantification of the Western blottings confirmed the strong inhibitory effects of the 5B-46 mimic as shown in Figure 5C.

Figure 5 Western blotting analysis of NS3 protease gene expression in JFH-1 infected Huh7. 5 cells after transfection of 5B-74 mimic and controls. A: HCV NS3; B: Tubulin, 72 h after transfection. Lane 1: Levels prior to addition of mimic; Lane 2: HBV (unrelated) control; Lane 3: 5B-46 mimic; C: Quantification of Western blottings of NS3 protein in JFH-1-infected cells at baseline, after 72 h exposure to 5B-46, and to HBV unrelated control. Bound NS3 antibodies were detected with SuperSignal chemiluminescent substrate and quantitated by Gel Imaging software GeneTools.

In the 5B-46 mimic, although most of the stem regions were altered, the sequences in the loops were left unchanged in order to retain secondary structure. Therefore, an antisense effect of the loop regions was still possible. To evaluate this possibility, short single-stranded fragments were prepared spanning the 5B-74 mimic (Figure 6). Analysis by Mfold ver 3.2 did not generate any stem-loop structures corresponding to regions in the native 5B-74 (data not shown). None of these sequences had any significant effects on HCV RNA levels compared to control (Figure 7), supporting the conclusion that the 5B-74 and 5B-46 mimics inhibited HCV replication not by sequence complementarity, but by conformational attributes.

Figure 6 Diagram of single-stranded fragments spanning the entire length of the 5B-74 mimic used as antisense controls.

Figure 7 HCV replication after transfection of single-stranded RNA fragments corresponding to specific regions of the 5B-74 mimic in JFH-1-infected Huh7. 5 cells. Cells stably infected with JFH-1 HCV were transfected separately with single-stranded fragments spanning the entire length of the 5B-74 mimic. HCV RNA levels were assayed by real-time RT-PCR. Assays were performed in triplicate and results expressed as mean ± SD of HCV replication as percent of unrelated (HBV) control.

DISCUSSION

Previous studies have shown that overexpression of RNA analogs in infected cells can result in sequestration of essential factors. For example, suppression of human immunodeficiency virus-1 replication in cell models has been demonstrated by expression of structural mimics of the TAR element[40], RRE structure[41], and primer binding sequence[42]. HCV RNA has a number of CREs that could potentially be inhibited by structural RNA mimics. Structural mimics based on the HCV IRES were recently shown to inhibit HCV translation in vitro and in replicon models[10]. However, no effect on RNA replication was observed in the latter, possibly because of the bicistronic nature of the replicon.

It has been shown previously that double-stranded RNA can induce synthesis of cytokines including interferons which have antiviral effects. However, stimulation of interferon production has been reported to be dependent on the length of the RNA, with duplexes greater than 30 bp having been found to be most efficient[43]. Many studies have shown that smaller siRNAs of 20-21 base pairs (bp) have purely RNAi activity. In the current study, for the 5B-74, 5B-46, and 5B-100 mimics, the longest double-stranded regions were only 9 bp, followed by 8, 6 and 3 bp. Furthermore, the EC mimic was also predicted to have 9, 8, 6 and 3 bp regions. Yet, EC had essentially no significant inhibitory activity. Therefore, while double-stranded RNA stimulation of interferon production cannot be entirely excluded, if it occurs, it is not likely to account for the observed inhibitory effects of the 5B mimics.

Recently, cell culture systems for in vitro replication and infectious virus production have been established based on full-length HCV genotype 2a RNA[36,38,44,45]. This infectious HCV possesses 21% differences compared to the HCV 1b subgenomic replicon, BB7. Nucleotide sequences involved in the kissing-loop interaction are conserved between JFH-1 and BB7. However, mutations in other regions may affect this interaction by disrupting the RNA secondary structure[46]. It is interesting to note that small RNAs corresponding to 2 domains of HCV 2a, which differed from those in HCV 1b in 13 of 44 nucleotides and 6 of 48 nucleotides, respectively, inhibited the replication of replicon RNA from pLMH14 (HCV 1b) by 48.3% and 37.1%, respectively indicating that HCV 2a-derived RNAs could cross-interact with HCV 1b components despite the sequence differences between the elements of the 2 viruses[47]. Based on this finding, and the fact that most of the nucleotide substitutions in one domain from HCV 2a were compensatory bp changes, Zhang et al[47] hypothesized that the structure rather than the primary sequence was important for its function.

Viral replication has been shown to require the interaction of host factors with viral RNA and/or proteins. For HCV, several RNA-binding proteins, including L autoantigen[48], NS1-associated protein 1[49], polypyrimidine tract binding protein[50] and cyclophilin B[22] have been shown to bind to HCV RNA. Therefore, while we have focused on effects of mimics on HCV replication, because other vital HCV activities also depend on protein-RNA interactions, it is possible that the strategy for inhibition of replication by structural mimicry could be extended to other targets.

COMMENTS

Background

Chronic hepatitis C virus (HCV) infection is a leading cause of chronic hepatitis, liver cirrhosis and hepatocellular carcinoma. Current therapeutic options are limited and associated with significant adverse effects. Accordingly, there is a strong impetus to develop novel therapeutic options. Because HCV replicates through the interaction of its RNA-dependent RNA polymerase and its RNA template to generate progeny RNA, physical contact between the enzyme and genomic RNA is required. The authors hypothesized that mimics could be created with sufficient resemblance to the natural genomic structure to compete with the HCV genome, resulting in inhibition of replication.

Research frontiers

Previous studies have shown that overexpression of RNA analogs in infected cells can result in sequestration of essential factors. For example, suppression of human immunodeficiency virus-1 replication in cell models has been demonstrated by expression of structural mimics of the TAR element, RRE structure, and primer binding sequence. HCV RNA has a number of cis-acting replication elements that could potentially be inhibited by structural RNA mimics. Structural mimics based on the HCV internal ribosome entry site were recently shown to inhibit HCV translation in vitro and in replicon models.

Innovations and breakthroughs

The novel feature of this research is the discovery that stretches of RNA can be made that have little sequence similarity to HCV. However, because of the retention of the predicted structure, they have been found to be highly potent and specific agents for inhibition of HCV replication.

Applications

Specific sequences inhibiting HCV replication without substantial HCV sequence identity may be a novel therapeutic option to circumvent the appearance of resistance arising from mutation of the viral genome.

Terminology

Single-stranded RNA structural mimics are a class of polynucleotides that can inhibit replication of the HCV and are designed to mimic the shape of natural viral nucleic acids, without requiring the matching of nucleic acid sequences.

Peer review

The study of Smolic et al described the antiviral effect of HCV structural mimics in replicon and infectious models. It is a very straightforward and concise paper. The results provided clear evidence of the inhibition of viral replication by these mimics, likely via conformational attributes. However, several issues regarding the mechanism of action should be addressed to further strengthen this study.