Reactive oxygen species suppress hepatitis C virus RNA replication in human hepatoma cells
Article first published online: 5 JAN 2004
Copyright © 2004 American Association for the Study of Liver Diseases
Volume 39, Issue 1, pages 81–89, January 2004
How to Cite
Choi, J., Lee, K. J., Zheng, Y., Yamaga, A. K., Lai, M. M.C. and Ou, J.-h. (2004), Reactive oxygen species suppress hepatitis C virus RNA replication in human hepatoma cells. Hepatology, 39: 81–89. doi: 10.1002/hep.20001
- Issue published online: 5 JAN 2004
- Article first published online: 5 JAN 2004
- Manuscript Accepted: 6 OCT 2003
- Manuscript Received: 2 JUL 2003
- American Cancer Society. Grant Number: PF-01-037-01-MBC
- National Institutes of Health
Hepatitis C virus (HCV) is a positive-stranded RNA virus that causes severe liver diseases, such as cirrhosis and hepatocellular carcinoma. HCV uses an RNA-dependent RNA polymerase to replicate its genome and an internal ribosomal entry site to translate its proteins. HCV infection is characterized by an increase in the concentrations of reactive oxygen species (ROS), the effect of which on HCV replication has yet to be determined. In this report, we investigated the effect of ROS on HCV replication, using a bicistronic subgenomic RNA replicon and a genomic RNA that can replicate in human hepatoma cells. The treatment with peroxide at concentrations that did not deplete intracellular glutathione or induce cell death resulted in significant decreases in the HCV RNA level in the cells. This response could be partially reversed by the antioxidant N-acetylcysteine. Further studies indicated that such a suppressive response to ROS was not due to the suppression of HCV protein synthesis or the destabilization of HCV RNA. Rather, it occurred rapidly at the level of RNA replication. ROS appeared to disrupt active HCV replication complexes, as they reduced the amount of NS3 and NS5A in the subcellular fraction where active HCV RNA replication complexes were found. In conclusion, our results show that ROS can rapidly inhibit HCV RNA replication in human hepatoma cells. The increased ROS levels in hepatitis C patients may therefore play an important role in the suppression of HCV replication. (HEPATOLOGY 2004;39:81–89.)
Hepatitis C virus (HCV) is a positive-sensed, single-stranded RNA virus of the Flaviviridae family.1 Currently, it is estimated that there are more than 170 million people who are infected by HCV worldwide.2 About 80% of HCV infection results in chronic infection that can lead to severe liver diseases such as cirrhosis and hepatocellular carcinoma.. HCV genome is about 9.6 kilobase (kb) in length and consists of the 5′ untranslated region (UTR), the structural (C, E1, E2), and the nonstructural (NS) (p7, NS2, NS3, NS4A/B, NS5A/B) protein-coding regions, and the 3′ UTR. Translation is mediated by the internal ribosomal entry site (IRES), located at the 5′ end of the genome, and it produces a polyprotein that is subsequently cleaved by viral and host proteases to generate individual viral proteins. Some of the HCV proteins can also be synthesized from alternate reading frames through ribosomal frameshift.3–6
HCV infection is associated with elevated levels of circulating reactive oxygen species (ROS) in patients.7–14 ROS are normal products of cell metabolism,15 and their syntheses can be heightened during inflammation.16 Recent studies have revealed a complex relationship between redox chemistry and various viral infections. For example, ROS can negatively regulate hepatitis B virus replication in liver cells without affecting the cell metabolism17 but enhance the replication of human immunodeficiency virus by activating nuclear factor kappa B.18 Viral proteins such as hepatitis B virus X protein and human immunodeficiency virus tat protein have also been shown to regulate the cellular redox status.19–22 HCV NS3 protein can also activate the ROS generation by activating NADPH oxidase of monocytes.23 In addition, HCV core and NS5A proteins have been found to induce oxidative stress in cells.24, 25 Interestingly, transgenic mice carrying the HCV core protein gene show signs of increased oxidative stress in the absence of inflammation.26 Such elevated ROS concentrations are thought to play an important role in the pathogenesis of HCV.26 In spite of these observations, the effect of ROS on HCV replication has yet to be determined.
In this report, we have used a subgenomic HCV RNA replicon as well as a self-replicating HCV genomic RNA to investigate the effect of ROS on HCV RNA replication. Our results indicate that ROS, at concentrations that do not affect cell viability, significantly decreased both subgenomic and genomic HCV RNA levels. This decrease occurred rapidly at the level of RNA replication and was associated with a disappearance of NS5A from the subcellular fraction that contained the HCV replication complex.
Materials and Methods
Subgenomic and Genomic HCV Replicon Constructs.
The HCV subgenomic replicon of genotype 1b27 (Genbank accession No. AJ242652) was constructed by serial ligation of DNA oligonucleotides and by polymerase chain reaction (PCR). HindIII/NotI and ScaI/XbaI sites were engineered at the 5′ and 3′ends of the subgenomic replicon sequence for cloning into a pUC19-derived plasmid vector that contained a modified T7 promoter. This replicon carried an S1179I adaptive mutation in the NS5A region that enhanced HCV RNA replication in cell culture.28 The final sequences were confirmed by DNA sequencing.
The hybrid genomic HCV replicon was produced by fusing nucleotides (NT) 1–3092 of the H77c sequence29 to NT 1839–7990 of the subgenomic replicon sequence via StuI (NT 1839) site within the NS3 region. This fusion creates a 1a/1b hybrid HCV genome. The NS3 sequence of this hybrid differs from the subgenomic replicon sequence by a single serine-to-alanine substitution at the eighth amino acid. This hybrid sequence was cloned into the modified pUC19 plasmid, as mentioned above. The replicon RNAs were synthesized as described previously by others.27 The hybrid genomic plasmid and pH77c were linearized with XbaI prior to RNA synthesis.
Electroporation and Cell Culture.
About 5 × 106 Huh7 human hepatoma cells were rinsed with Dulbecco's Modified Eagle's Medium (DMEM) without serum, mixed briefly with 5 μg of the subgenomic replicon RNA in 0.4 mL of this medium, and then electroporated at 220 V and 975 microfarads (μF). Cell colonies were selected with DMEM containing 0.7 mg/mL G418. Two pooled cell clones (Sg-PC1 and 2) were obtained and subsequently maintained in DMEM containing 0.5 mg/mL of G418. For most of the experiments, G418 was removed from the medium one day prior to cell treatments. For transient replication experiments, 1 × 107 Huh7 cells were electroporated with 10–40 μg of subgenomic, hybrid, or H77c RNA. Then, 1 million cells were seeded onto 60-mm cell culture dishes in DMEM containing 10% fetal bovine serum. Cells were collected at different time points, as indicated in Results.
Northern Blot Analysis.
Total RNA was extracted from cells, using Trizol reagent (Invitrogen, Carlsbad, CA), following the manufacturer's protocol. Four to 15 μg RNA were then subjected to Northern blot analysis using a 32P-labeled double-stranded DNA probe, prepared from NT 3669 to 6016 of the subgenomic replicon or NT 279 to 3092 of H77c. The latter was used to detect the hybrid and H77c RNA. For the detection of the negative strand RNA, a sense riboprobe, containing the 5′ UTR and the neomycin region of the subgenomic replicon construct, was used. For these and all other experiments, RNA and protein bands were quantified by densitometry, using SigmaScan (Jandel Scientific, San Rafael, CA). The intensities of the bands of interest were normalized against the control glyceraldehyde 3-phosphate dehydrogenase (GAPDH) RNA.
HCV RNA was quantified by real-time reverse transcription (RT) PCR, as described by Lanford et al.,30 except the reverse primer sequence was 5′ CGGGTTGATCCAAGAAAGGA 3′ (NT 210 to 191) and the fluorogenic probe, labeled with 6-FAM and BHQ-1 (Biosearch Technologies, Novato, CA). ABI 7900 Sequence Detector and Taqman Gold RT-PCR Kit (Applied Biosystems, Foster City, CA) were used. For the analysis of negative-stranded HCV RNA, reverse transcription was carried out with the forward primer prior to the PCR amplification.
3H Labeling of Newly Synthesized HCV RNA.
One million Huh7 or Sg-PC2 cells were labeled with 100 μCi of 3H-uridine (35–50 Ci/mmol; ICN, Costa Mesa, CA) either in the presence or absence of 200 μg/ml actinomycin D (SigmaAldrich, St. Louis, MO) for 5–6 hours and then lysed for the RNA isolation. RNA was analyzed on a 1% formaldehyde agarose gel, which was then treated with 1M sodium salicylate for 20–30 minutes for fluorography. To analyze the effect of hydrogen peroxide (H2O2) on nascent HCV RNA, cells were treated with 0, 20, 50, or 100 μM H2O2 for the duration of 3H-labeling. To investigate the effect of H2O2 on HCV RNA stability, cells were pulse-labeled with 3H-uridine for 5–6 hours as described above, rinsed twice with phosphate-buffered saline (PBS), and chased with a hundredfold excess of nonlabeled uridine (0.1 mmol/L) with either 0 or 100 μM H2O2 for 3 or 6 hours.
Subcellular Membrane Fractionation.
Four 100-mm dishes of Sg-PC2 cells with or without the H2O2 treatment were rinsed with PBS and then with the hypotonic buffer (25 mmol/L HEPES, pH7, 0.2 M sucrose) twice. Cells were scraped off the plate in 0.5 mL hypotonic buffer, sat on ice for 10 minutes and lysed by passing through a 26-gauge needle four times. The cell lysates were centrifuged at 16,000 xg for two minutes, and the cytoplasmic supernatants were then fractionated on a discontinuous sucrose gradient, which contained 1.5 mL 2 M sucrose, 3.4 mL 1.3 M sucrose, 3.4 mL 1 M sucrose, and 2.2 mL 0.6 M sucrose in 25 mmol/L HEPES, pH 7, at 40K rpm for two hours using a Beckman SW40Ti rotor following our previous procedures.31, 32 The interface between 2 and 1.3 M sucrose solutions, which contained Golgi membranes, and that between 1 and 0.6 M sucrose solutions, which contained the membranes derived from the rough endoplasmic reticulum (RER), were isolated (approximately 1 mL each), treated with Nonidet P-40 to a final concentration of 0.2%, diluted with 1 mL of 25 mmol/L HEPES, pH 7, and then concentrated with Centricon (Amicon Corp., Billerica, MA). The samples were then analyzed by Western blot using mouse anti-NS5A (Biodesign, Saco, ME), mouse anti-NS3 (Austral Biologicals, San Ramon, CA), goat anti-GRP78 (Santa Cruz Biotechnology, Santa Cruz, CA), and rabbit anti-human albumin primary antibodies (Boehringer-Mannheim Biochemicals, Indianapolis, IN). The horseradish peroxidase-conjugated wheat germ agglutinin (WGA) (ICN ImmunoBiologicals, Costa Mesa, CA) was also used to analyze the proteins associated with the Golgi membranes.
For the analysis of the HCV RNA replication complex, the membranes isolated from the interfaces were dialyzed at 4 °C twice against 25 mmol/L HEPES, pH7 and pelleted at 16,000 xg for 30 minutes. The membrane pellet was then resuspended in 50 μL of the following reaction mixture: 50 mmol/L HEPES, pH7, 5 mmol/L MgCl2, 0.5 mmol/L MnCl2, 10 mmol/L KCl, 1 mmol/L each of adenosine triphosphate, guanosine triphosphate and uridine triphosphate, 10 μM cytidine triphosphate (CTP), 5 μg actinomycin D, and 30 μCi α-32P-CTP; it was then incubated at 30 °C for 90 minutes.33 The reaction was stopped by extraction with phenol and chloroform, and the 32P-labeled HCV RNA was precipitated with ethanol in the presence of transfer RNA carrier and analyzed by gel electrophoresis and autoradiography.
Results and Discussion
Sg-PC1 and Sg-PC2 were two different pooled cell clones that contained HCV subgenomic RNA replicon. As shown in Figures 1A and 1B, both positive- and negative-stranded HCV RNAs of approximately 8 kb were readily detected by Northern blot analysis in these two cell lines. The HCV RNA could be labeled by 3H-uridine in the presence of actinomycin D, indicating that its synthesis was independent of the host RNA transcription (Fig. 1C).
To test the effect of ROS on HCV replication, one of the pooled cell clones, Sg-PC2, was treated with various concentrations of H2O2, t-butylhydroperoxide (tBOOH), or 2-tert-butyl(1,4)hydroquinone (TBHQ) for 24 hours, and the HCV RNA was analyzed by Northern blot. H2O2 exposure induced significant decreases in the HCV RNA level (Fig. 2A). The results were confirmed by real-time RT-PCR for both positive and negative stranded HCV RNAs (Fig. 2B), and these changes in the viral RNA level were accompanied by similar decreases in the HCV protein level (Fig. 2C). The tBOOH induced similar reduction of the viral RNA level (Fig. 2A). The TBHQ is a redox-cycling hydroquinone, which generates a low level of ROS in cells.34 It also decreased the HCV RNA level in these cells (Fig. 2A). These results indicated that ROS could reduce HCV RNA levels in Sg-PC2 cells. The negative effects of ROS could be reversed by N-acetylcysteine (NAC), a thiol antioxidant compound, as the concurrent treatment of Sg-PC2 cells with NAC partially alleviated the negative effect of H2O2 on the HCV RNA level (Fig. 2D).
In all of the experiments shown in Figure 2, the expression levels of the GAPDH RNA were not affected, indicating that the effect of ROS on HCV RNA was not caused by the nonspecific cytotoxicity of ROS. The lack of cytotoxicity of ROS at the concentrations used was further confirmed by the trypan blue exclusion assay, by the lack of any increase in the amount of lactate dehydrogenase released into the medium, and by the lack of changes of the intracellular redox status based on the glutathione levels (data not shown).
Studies using the subgenomic replicons shown in Figure 2 are limited because the translation of the HCV nonstructural proteins is driven by the encephalomyocarditis virus IRES instead of the HCV IRES and because none of the structural protein genes are present. Therefore, we sought to determine the effect of H2O2 on the genomic HCV RNA replication. The H77c sequence, an infectious HCV clone,29 from the beginning of the NS3 coding sequence to the 3′ end, was replaced with the corresponding sequence of the subgenomic replicon sequence to generate a full-length, hybrid HCV sequence (Fig. 3A). This hybrid RNA was synthesized in vitro and electroporated into Huh7 cells. The total cellular RNA was then isolated at different time points and analyzed by Northern blot for HCV RNA.
As shown in the left panel of Figure 3B, the subgenomic HCV RNA replicated transiently in Huh7 hepatoma cells, as previously demonstrated by others. The RNA level dropped transiently at about 16 hours and rose at 24 hours, indicating initial degradation of the input RNA, followed by the active HCV RNA replication.35 The RNA concentration peaked at around 48 hours and then gradually decreased. The full-length genomic H77c RNA did not replicate, as expected,28 and served as a negative control (Fig. 3B, middle panel). The H77c/replicon genomic hybrid replicated transiently in cell culture (Fig. 3B, right panel). Its RNA level once again dropped transiently at 16 hours, started increasing at 24 hours, and peaked at around 72 hours. The NS5A and core protein were also detected in cells that were transfected with the hybrid RNA four days after the transfection (data not shown).
To examine whether H2O2 could also decrease the full-length HCV RNA level, Huh7 cells were transiently transfected with the hybrid HCV genomic RNA. After 48 hours, these cells were treated with various concentrations of H2O2, and the amount of the HCV RNA was analyzed after 24 hours. The H2O2 at 50 or 100 μM significantly reduced the genomic HCV RNA level to the undetectable level (Fig. 4).
To determine the mechanism by which peroxide decreased the HCV RNA level in hepatoma cells, Sg-PC2 cells were incubated with 3H-uridine and actinomycin D (see Fig. 1C) with and without H2O2 for 5–6 hours. As shown in Figure 5A, H2O2 dose-dependently decreased the amount of the HCV RNA that was synthesized during this time. The viral RNA fell to an almost undetectable level with 100 and 200 μM H2O2 treatments. The 3H-labeling of ribosomal RNA was not affected (Fig. 5A, left panel), again indicating the specificity of this suppression. Next, we examined whether this suppression, observed within 5–6 hours of H2O2 exposure, occurred at the level of RNA stability. The HCV RNA was labeled for 5–6 hours with 3H-uridine in the presence of actinomycin D and then chased with a hundredfold excess of nonlabeled uridine in the presence of 0 or 100 μM H2O2 for 3 or 6 hours. There was no appreciable decline in the control (i.e., 0 μM H2O2) HCV RNA signal after 3 hours of chase, although some reduction was observed after 6 hours (Fig. 5B). Most importantly, however, there was no apparent change in the 3H-RNA signal with 100 μM H2O2 treatment (Fig. 5B). Therefore, the rapid decline in the rate of 3H-RNA synthesis observed within 5–6 hours of the 100 μM peroxide treatment shown in Figure 5A was not due to an accelerated degradation of the viral RNA. The HCV IRES activity was not affected by H2O2 (data not shown).
To further investigate whether H2O2 directly suppressed HCV RNA synthesis, Sg-PC2 cells were first treated with 100 μM H2O2 for 30 minutes, 3 hours, or 6 hours. Then, the cytoplasmic lysates were prepared, and the RNA replication assay was carried out in vitro as described by Ali et al.36 All peroxide-treated lysates showed a reduction of the HCV RNA replication rate (Fig. 6, top panel). This reduction was apparent as early as 30 minutes after treatment. Neither the RNA loading, based on the amount of 18S ribosomal RNA in each lane (Fig. 6, second panel), nor the amount of the HCV RNA template, present in the lysates immediately prior to the in vitro RNA replication assay (Fig. 6, third panel), could account for the rapid suppression of HCV RNA replication by H2O2. There was also no change in the amount of NS5A in each of the cell lysates prior to the assay up to six hours of treatment (Fig. 6, bottom panel). Note that the NS5A protein level did decrease at 24 hours after the peroxide treatment (Fig. 2C). However, this decrease was most likely a result of the decreased HCV RNA replication rather than being the cause for the decreased replication. Therefore, the suppressive response to H2O2 occurred rapidly, at the level of RNA replication.
To further investigate the molecular mechanism of suppression of HCV RNA replication by ROS, we examined whether the subcellular localization of HCV NS5A protein was affected by the peroxide treatment. Sg-PC2 cells were treated with 100 μM H2O2 for 3 hours, and then subjected to membrane fractionation experiments, as previously described.31, 32 In the control (untreated Sg-PC2 cells), NS5A protein was detected both in Golgi and RER fractions (Fig. 7A). The peroxide treatment significantly reduced the amount of NS5A in the Golgi fraction and altered the ratio of NS5A in Golgi and RER from roughly 30:70 to about 4:96 (Fig. 7A). Similar changes were observed with NS3 protein (Fig. 7A). In contrast, peroxide did not affect the subcellular localization of albumin. To ensure that this gradient faithfully separated Golgi and RER membranes, each fraction was also analyzed with the anti-GRP78 antibody and WGA. GRP78 is an endoplasmic reticulum (ER)-associated protein and WGA binds to proteins with complex-type glycans in the trans-Golgi. As shown in the same figure, GRP78 was found in the RER fraction while WGA was primarily bound to proteins in the Golgi fraction.
To understand how the loss of NS5A and NS3 from the Golgi fraction might affect HCV RNA replication, each of the fractions in Figure 7A was subjected to in vitro HCV replication assays. Although the majority of these proteins were found in the ER fraction in these experiments (Fig. 7A), HCV RNA replicating activity was found mostly in the Golgi fraction (Fig. 7B). H2O2 treatment suppressed the HCV RNA replication in this fraction (Fig. 7B). As H2O2 also reduced the HCV protein level in the Golgi fraction, the suppression of HCV RNA replication by H2O2 in this fraction was likely due to the loss of active HCV RNA replication complexes.
The finding that active HCV RNA replication complexes were mostly identified in the Golgi fraction may be due to the association of the HCV replication complexes with the Golgi membranes as suggested by Serafino et al.37 Alternatively, lipid rafts and membrane webs, which have also been reported to contain HCV replication complexes,38–40 might cosediment with the Golgi membranes in our experiments. It should be noted that lipid rafts are also known to form in the Golgi complex.41 The close association of NS5A and NS3 with the HCV RNA replication activity is in support of the previous observations that these proteins are important components of the HCV replication complexes.42, 43 The reduction in the amount of NS5A/NS3 in the Golgi fraction indicates a possible disruption of the replication complexes or the perturbation of intracellular transport of these proteins by ROS. If it is the latter, the effect was highly specific, as the transport of albumin as well as proteins containing the complex glycans was not affected (Fig. 7).
In this report, we showed that ROS, within the biologically relevant concentration range, could suppress HCV RNA replication in Huh7 cells. The rapid response of HCV RNA replication to ROS is suggestive of a mechanism that involves signaling. ROS are known to affect diverse signaling pathways.44 These signaling events could either rapidly disrupt existing active HCV replication complexes or suppress the formation of new complexes in Huh7 cells, to result in a reduced amount of NS5A and NS3 and, hence, HCV RNA replicating activity in the Golgi fraction. It will be interesting to identify the molecular targets of ROS that eventually affect HCV replication complexes. Finally, an important corollary to our findings might be that the antioxidants, which are currently being investigated as potential adjunct therapy for various liver diseases,45 might in fact facilitate HCV replication by counteracting ROS in these patients, as suggested in Figure 2D. Due to the large number of HCV patients, further investigation in this research area is warranted.
The authors acknowledge Anne Strohecker for her assistance with the construction of the HCV subgenomic replicon, and Dr. Stanley Lemon for pRL-HL construct. The authors thank Dr. Henry Jay Forman and Dr. Rui-Ming Liu for critical reading of this manuscript.
- 2World Health Organization. Hepatitis C: global prevalence (update). Wkly Epidemiol Rec 2000; 75: 18–19.
- 11Successful interferon therapy reverses enhanced hepatic iron accumulation and lipid peroxidation in chronic hepatitis C. Am J Gastroenterol 2000; 95: 1041–1050., , , , , .Direct Link:
- 12Hepatocellular mitochondrial alterations in patients with chronic hepatitis C: ultrastructural and biochemical findings. Am J Gastroenterol 1999; 94: 2198–2205., , , , , , , et al.Direct Link:
- 15Free Radicals in Biology and Medicine. Vol 31. 3rd ed. New York: Clarendon Press, 1999., .
- 44Oxidative Stress and Signal Transduction. Vol 15. New York: Chapman & Hall, 1997., .