Requirement of the cyclic adenosine monophosphate response element-binding protein for hepatitis B virus replication

Authors

  • Bo Kyung Kim,

    1. Department of Biochemistry, Korea University College of Medicine, Seoul, Korea
    2. Division of Brain Korea 21 Program for Biomedical Science, Korea University College of Medicine, Seoul, Korea
    Search for more papers by this author
  • Seoung Ok Lim,

    1. Department of Biochemistry, Korea University College of Medicine, Seoul, Korea
    2. Korean Institute of Molecular Medicine and Nutrition, Korea University College of Medicine, Seoul, Korea
    Search for more papers by this author
  • Yun Gyu Park

    Corresponding author
    1. Department of Biochemistry, Korea University College of Medicine, Seoul, Korea
    2. Division of Brain Korea 21 Program for Biomedical Science, Korea University College of Medicine, Seoul, Korea
    3. Korean Institute of Molecular Medicine and Nutrition, Korea University College of Medicine, Seoul, Korea
    • Department of Biochemistry, Korea University College of Medicine, 126-1, 5-ga, Anam-dong, Sungbuk-gu, Seoul 136-701, Korea
    Search for more papers by this author
    • The corresponding author certifies that all of the listed authors participated meaningfully in the study and that they have seen and approved the final manuscript.

    • fax: 82-2-923-0480.


  • Potential conflict of interest: Nothing to report.

Abstract

The cyclic adenosine monophosphate–response element (CRE)-transcription factor complex participates in the regulation of viral gene expression and pathologic processes caused by various viruses. The hepatitis B virus (HBV) enhancer I directs liver-specific transcription of viral genes and contains a CRE sequence (HBV-CRE); however, whether the HBV-CRE and CRE-binding protein (CREB) are required for the HBV life cycle remains to be determined. This study was designed to investigate the role of CREB in HBV replication and gene expression. Sequence-comparison analysis of 984 HBVs reported worldwide showed that the HBV-CRE sequence is highly conserved, indicating the possibility that it plays an important role in the HBV life cycle. The binding of CREB to the HBV-CRE site was markedly inhibited by oligonucleotides containing HBV-CRE and consensus CRE sequences in vitro and in vivo. The HBV promoter activity was demonstrated to be dependent upon the transactivation activity of CREB. Treatment with CRE decoy oligonucleotides reduced HBV promoter activity, and this was reversed by CREB overexpression. The levels of viral transcripts, DNA, and antigens were remarkably decreased in response to the overexpression of CREB mutants or treatment with the CRE decoy oligonucleotides, whereas enhancing CREB activity increased the levels of viral transcripts. In addition, introduction of a three-base mutation into the HBV-CRE led to a marked reduction in HBV messenger RNA synthesis. Conclusion: Taken together, our results demonstrate that both replication and gene expression of HBV require a functional CREB and HBV-CRE. We have also demonstrated that CRE decoy oligonucleotides and the overexpression of CREB mutants can effectively block the HBV life cycle, suggesting that interventions against CREB activity could provide a new avenue to treat HBV infection. (HEPATOLOGY 2008.)

The World Health Organization has reported that about 4 million people worldwide are acutely infected with hepatitis B virus (HBV) each year, and there are estimated to be 350 million chronic carriers worldwide.1 The HBV is a major cause of acute and chronic hepatitis, cirrhosis, and hepatocellular carcinoma.1–3 The agents that are currently available for the treatment of chronic HBV infection are interferon-α and nucleos(t)ide analogs, such as lamivudine (L-2′,3′-dideoxy-thiacytidene), adefovir dipivoxil, and entecavir.2–5 However, there are some limitations to these therapeutics: interferon-α has adverse effects, and nucleos(t)ide derivatives, although safe and efficacious, can lead to the emergence of resistant mutants.2–5 Furthermore, none of these therapeutic agents can effectively eradicate the virus.4, 5 Studies on the regulatory mechanisms of HBV gene expression and replication are, therefore, necessary for the development of promising antiviral medications. Interventions against HBV include novel nucleic acid-based approaches employing antisense RNA and DNA,6–8 ribozymes,9, 10 and RNA interference.11–13

The HBV genome is a small, circular, and incomplete double-stranded DNA molecule (3.2 kilobases [kb] in length). The viral DNA is transcribed in the nucleus of hepatocytes to produce viral transcripts including pregenomic RNA (3.6, 2.4, 2.1, and 0.7 kb). Synthesis of these transcripts is regulated by four HBV promoters: the preC/pregenomic, S1, S2, and X promoters.14 In addition, two enhancers (I and II) as well as cis-acting negative regulatory elements play important roles in the regulation of viral gene transcription, and enhancer I partially overlaps with the X promoter.14–16 Enhancer I has been reported to regulate the global and temporal expression of HBV genes in a dominant manner to enhancer II and is required for enhancer II activation.17 Indeed, enhancer I is composed of at least five different elements that bind transcription factors, called the 2C, GB, EP, E, and NF1 sites.14, 18 Among these, the E element contains various transcription factor binding sites including a cyclic adenosine monophosphate (cAMP)-response element (CRE) sequence (HBV-CRE), which is known to be the binding site for CRE-binding protein (CREB)/activating transcription factor (ATF) family members.19, 20 The transcription factor CREB is a nuclear protein that mediates cAMP-stimulated gene expression. It binds to the CRE as a homodimer or heterodimer with other transcription factors belonging to the CREB/ATF or Jun/Fos families.21 The transcriptional activity of CREB is regulated through phosphorylation at Ser133 by various protein kinases, including cAMP-dependent protein kinase and mitogen-activated protein kinase.22 HBx, an important transactivator of HBV,23 has been shown to directly interact with CREB and facilitate its binding affinity for the CRE site in HBV enhancer I,19, 20 and to augment HBV replication and gene expression.23–27 However, whether HBx enhances HBV replication and gene expression through activation of the CRE/CREB system remains largely unclear.

Although CRE sequences that are found in specific regulatory elements of various viruses often control viral replication and gene expression,28 there is no direct evidence to indicate that the HBV-CRE site and its binding proteins including CREB are required for HBV replication and the expression of its genes. We hypothesized that CREB and the HBV-CRE are required for the HBV life cycle; therefore, inhibition of CREB transactivation activity would interrupt the life cycle. Herein, we demonstrate that the transactivation activity of CREB and normal function of the HBV-CRE are indispensable for HBV replication and gene expression. In addition, to evaluate the therapeutic potential of targeting CREB, we used two types of synthetic CRE decoy oligonucleotides. Transcription factor decoy oligonucleotides contain the short consensus binding sequence for a specific transcription factor or a family of transcription factors.29 By competing with the response elements within the promoter and enhancer regions for the binding of a specific transcription factor(s), transcription factor decoys offer a means of blocking transcription factor function in a sequence-specific manner for both basic research and novel drug development. We demonstrated that the efficacy of CREB inactivation by CRE decoy oligonucleotides to inhibit HBV replication was equal to or more than that of lamivudine. These results suggest that approaches for modulating the transactivation activity of CREB may produce promising therapeutics for HBV infection.

Abbreviations

ATF, activating transcription factor; cAMP, cyclic adenosine monophosphate; CAT, chloramphenicol acetyltransferase; C/EBP, CCAAT/enhancer-binding protein; CEP-CAT, CAT reporter plasmid driven by HBV enhancer I to enhancer II; Cp-CAT, CAT reporter plasmid driven by a full-length HBV genome; CRE, cAMP-response element; C-CRE, consensus CRE; CREB, CRE-binding protein; DHBV, duck HBV; EDTA, ethylene diamine tetraacetic acid; EMSA, electrophoretic mobility shift assay; HBV, hepatitis B virus; HBeAg, HBV envelope antigen; HBsAg, HBV surface antigen; HBx, HBV X protein; IBMX, 3-isobutyl-1-methylxanthine; KCREB, DNA-binding mutant of CREB; mRNA, messenger RNA; MTT, 3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyltetrazolium bromide; PCR, polymerase chain reaction.

Materials and Methods

Cell Culture and Transfection.

The human hepatocellular carcinoma cells, HuH-7 and HepG2, were cultured in Roswell Park Memorial Institute (RPMI)-1640 medium and maintained at 37°C in a 5% CO2-air mixture incubator. HepG2.2.15 cells, HepG2 cells that constitutively produce HBV mRNAs, DNAs, and viral particles, were kindly provided by Y.H. Kim and were cultured in RPMI-1640 media containing 400 μg/mL geneticin.30 HepG2 cells that constitutively expressed a DNA-binding mutant of CREB (KCREB) were prepared by transfection with the KCREB plasmid. Stably transfected cells were selected in the presence of 400 μg/mL geneticin for 2 to 3 weeks. In transfection experiments, the total amount of DNA was adjusted to be equal by adding salmon sperm DNA, and the transfection efficiency was normalized with cotransfection of the expression plasmids encoding β-galactosidase (pSV-β-gal; Promega, Madison, WI) or green fluorescent protein (pEGFP-E1; Clontech Laboratories Inc., Mountain View, CA).

Plasmids and Reagents.

The replication-competent constructs containing the incomplete head-to-tail dimers of wild-type HBV (adwR9) and X-minus mutant HBV (HBX-21) containing a stop codon mutation in the HBx gene (codon 8), both of which were cloned in pGEM-7Zf(+) (Promega), were kindly provided by H.E. Blum.31, 32 The adwR9-mCRE construct, bearing a three-base mismatched sequence (Mut-HBV-CRE, 5′-TtgCGCAc-3′) in the HBV-CRE site of adwR9, was generated by polymerase chain reaction (PCR)-mediated site-specific mutagenesis.33 adwR9 was used as a template and the primers were as follows: forward primer containing an EcoRI site (5′-GCAGTGGAATTCCACTGCCT-3′), reverse primer containing a NcoI site (5′-AGCAGCCATGGATACGATGT-3′), and 5′-Mut-HBV-CRE and 3′-Mut-HBV-CRE primers containing a three-base mutation in the HBV-CRE site (5′-TTGCTtgCGCAcCCCCCACT-3′ and 3′-TTCACAAACGAacGCGTgGG-5′, respectively). Two separate PCRs were performed with one pair of the forward and 3′-Mut-HBV-CRE primers and the other pair of 5′-Mut-HBV-CRE and the reverse primers. Two PCR products that overlap in the sequence containing the same mutation were mixed in an equal molar ratio and amplified in the presence of the forward and reverse primers. Finally, adwR9-mCRE was produced by ligation of the EcoRI/NcoI fragment of the resulting PCR product and the same restriction enzyme-digested fragment of adwR9. The plasmid expressing wild-type CREB (CREB) and the KCREB plasmid were kindly provided by R.H. Goodman.34 KCREB has a single point mutation at Arg287 within the DNA-binding domain. An empty vector as a control plasmid was produced by removal of the HindIII/XbaI fragment (containing the region encoding DNA-binding mutant CREB) from KCREB plasmid, blunting, and ligation. M1CREB, in which the phosphorylation site Ser133 has been converted into alanine, was kindly provided by M. Montminy.35 The plasmids expressing wild-type ATF-2 and a DNA-binding mutant of ATF-2 (ATF-2-M2) were kindly provided by G. Redeuilh36 and R.G. Pestell,37 respectively. ATF-2-M2 contains an alanine to arginine substitution, which abolishes DNA binding to the CRE site. Cp-CAT and CEP-CAT are the CAT reporter plasmids driven by a full-length HBV genome (3.2 kb) and HBV enhancer I to enhancer II, respectively. Cp-CAT was a kind gift from Y. Yun,38 and CEP-CAT was generated by subcloning the NsiI fragment of Cp-CAT (containing HBV enhancer I to II linked to the CAT gene) into the cloning vector, pGEM-7Zf(+) (Promega) as described earlier.38 enI-Luc is a HBV enhancer I-luciferase reporter plasmid, which was constructed by introducing the BamHI/HindIII fragment of Cp-CAT into the luciferase reporter vector, pXP2 (American Type Culture Collection, Manassas, VA). The CAT reporter gene driven by somatostatin promoter containing a consensus CRE (C-CRE) sequence (−71 to +53 nucleotides from the somatostatin transcriptional initiation site), SS-CAT was kindly provided by R.H. Goodman.39 Lamivudine was kindly provided by J.E. Yeon. Forskolin and 3-isobutyl-1-methylxanthine (IBMX) (Calbiochem, Germany) were used to increase intracellular cAMP levels.

Treatment of Cells with CRE Decoy Oligonucleotides.

Cationic lipid N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium methyl-sulfate (DOTAP; Roche Diagnostics GmbH, Germany) was used in the oligonucleotide treatment as previously described.40 The HBV-CRE decoy, C-CRE decoy, and control oligonucleotides have a phosphorothioate backbone and contain three copies of HBV-CRE, C-CRE, and scrambled C-CRE sequences: 24-mer HBV-CRE decoy, 5′-TGACGCAATGACGCAATGACGCAA-3′ (Bioneer, Korea); 24-mer C-CRE decoy, 5′-TGACGTCATGACGTCATGACGTCA-3′; and 24-mer control, 5′-CTAGCTAGCTAGCTAGCTAGCTAG-3′, respectively. C-CRE decoy and control oligonucleotides were kindly provided by Y.S. Cho-Chung.40

Electrophoretic Mobility Shift Assay.

Nuclear extraction and electrophoretic mobility shift assay (EMSA) were performed as described previously.40 The [32P]-labeled probes were oligonucleotides containing single copies of the following sequences: C-CRE (5′-AGAGATTGCCTGACGTCAGAGAGCTAG-3′; Promega) and the HBV-CRE sequences, HBV-CRE(C) (5′-GTGTTTGCTGAcGCAACCCCCACT-3′; Bioneer) and HBV-CRE(T) (5′-GTGTTTGCTGAtGCAACCCCCACT-3′; Bioneer). The competitors were the unlabeled probes and the oligonucleotides containing single copies of the following sequences: Mut-HBV-CRE, which has the same sequence that was introduced into adwR9-mCRE (5′-GTGTTTGCTtgCGCAcCCCCCACT-3′; Bioneer); AP-1 (5′-CGCTTGATGAGTCAGCCGGAA-3′; Promega); and Oct1 (5′-TGTCGAATGCAAATCACTAGA-3′; Promega). Antibodies against CREB, ATF-2, c-Fos, and c-Jun (Santa Cruz Biotech, Santa Cruz, CA) were used for supershift formation.

CAT and Luciferase Activity Assays.

Cells were cotransfected with one of the reporter genes (Cp-CAT, CEP-CAT, or enI-Luc) and expression plasmids or oligonucleotides if required. Cells were harvested after 48 hours and cell lysates were prepared using a freeze-thaw procedure. To measure CAT activity, cell lysates were incubated with a substrate solution (0.2 μCi of [14C]-chloramphenicol, 0.53 mM acetyl-CoA, and 250 mM Tris-HCl, pH 7.8) for 90 minutes at 37°C. The reaction products were analyzed by thin-layer chromatography. The plate was autoradiographed, and the radioactivity was quantified using an Image Analyzer FLA-8000 (FujiFilm, Japan). The activities of luciferase (Promega) and β-galactosidase (BD BioSciences, Franklin Lakes, NJ) were measured using a luminometer (Berthold Detection Systems, Germany).

RNA Preparation and Northern Blotting.

Total cellular RNA was prepared using TRIzol reagent (Molecular Research Center, Cincinnati, OH). RNA was electrophoresed on 1% agarose/6.7% formaldehyde gels and transferred to nylon membranes (Amersham Biosciences, Piscataway, NJ) using a vacuum-blotting system (Amersham Biosciences). The membranes were hybridized overnight at 42°C with [32P]-labeled probes that were prepared according to the standard protocols for random-prime labeling (New England Biolabs, Beverly, MA). The [32P]-labeled probes were the 1.4-kb Aat II/EcoR I fragment of adwR9, the 1.2-kb EcoR I fragment of glutaraldehyde 3-phosphate dehydrogenase, the 0.6-kb BstX I fragment of β-galactosidase, and the 0.9-kb AatII/BglII fragment of green fluorescent protein.

DNA Preparation and Southern Blotting.

Core-associated DNA was prepared by the method described in hepatitis B and D protocols.41 Cell lysate was prepared by adding lysis buffer A (10 mM Tris, pH 7.5, 1 mM ethylene diamine tetraacetic acid [EDTA], 50 mM NaCl, 0.25% Nonidet P-40, and 8% sucrose) and incubating at 37°C for 15 minutes. Deoxyribonuclease I (1 U) was added to the supernatant (adjusted to 8 mM CaCl2 and 6 mM MgCl2), followed by an incubation at 37°C for 15 minutes. Polyethylene glycol (26% in 1.5 M NaCl and 60 mM EDTA) was added and the mixture was incubated at 4°C for 1 hour. The core particle was isolated by centrifugation at 16,000g for 4 minutes. The pellet was resuspended in lysis buffer B (10 mM Tris, pH 7.5, 8 mM CaCl2, and 6 mM MgCl2), and Deoxyribonuclease I (1 U) was added. The mixture was incubated at 37°C for 15 minutes. Lysis buffer C (25 mM Tris, pH 7.5, 10 mM EDTA, 1% sodium dodecyl sulfate, and 400 μg/mL proteinase K) was added and the mixture was incubated at 50°C for 1 hour. The DNA, prepared by phenol/chloroform extraction and ethanol precipitation, was electrophoresed on 1%-agarose gels and transferred to nylon membranes using a vacuum-blotting system. The membranes were hybridized overnight at 42°C with the [32P]-labeled 3.2-kb full-length HBV genome in adwR9.

Analysis of HBV Antigen Levels.

The levels of HBV surface antigen (HBsAg) and HBV envelope antigen (HBeAg) in the media were analyzed with a Gemini microplate reader (Molecular Devices Corporation, Sunnyvale, CA) using a MUREX kit (Abbott Laboratories, Chicago, IL).

MTT-Based Cell Viability Assay.

Cells were incubated in 3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyltetrazolium bromide (MTT) solution (5 mg/mL in phosphate-buffered saline) for 4 hours at 37°C to form formazan crystals. Culture medium was removed without disturbing the formazan crystals. The crystals were dissolved in 100% dimethyl sulfoxide on a shaking table for 10 minutes. The optical density at 560 nm was measured using a Gemini microplate reader (Molecular Devices Corp.).

Statistical Analysis.

Statistical analyses were performed by the unpaired t test or the Mann-Whitney U test, depending on the normality test using SigmaStat 2.0 (Systat Software Inc., San Jose, CA).

Results

High Conservation of the CRE Sequence in HBVs Reported Worldwide.

To determine whether HBV-CRE, the CRE sequence in HBV enhancer I, was conserved, we analyzed 984 full-length HBV sequences that were registered in GenBank using the multiple sequence alignment program, ClustalX2.0. The sequences of HBV-CRE(C) (5′-TGAcGCAA-3′) and HBV-CRE(T) (5′-TGAtGCAA-3′) constituted 88.2% and 10.7%, respectively, of the seven types of HBV-CRE sequences (Table 1). There were one-base mismatches between HBV-CRE(C) and HBV-CRE(T) and a two-base or three-base difference in HBV-CRE(C) (5′-TGACGcaA-3′) or HBV-CRE(T) (5′-TGAtGcaA-3′), compared to the C-CRE (5′-TGACGTCA-3′). The sequence of HBV-CRE(C) was the most highly conserved in HBV enhancer I, compared to the binding sites for all of the other transcription factors except STAT3; a STAT3 binding sequence (5′-AAGGCCTT-3′) occupied 94.1% of all of the STAT3 binding sequences that were present in HBV enhancer I (data not shown).

Table 1. CRE Sequences in the HBV Enhancer I Region
SymbolSequenceFrequencyRatio (%)
  1. A total of 984 HBV sequences registered in GenBank were analyzed using the multiple sequence alignment program ClustalX2.0.

HBV-CRE(C)TGACGCAA868/98488.2
HBV-CRE(T)TGATGCAA105/98410.7
OthersTGAAGCAA1/9841.1
 GGACGCAA4/984 
 TGACCCAA3/984 
 TGACGCAC2/984 
 GGATGCAA1/984 

Inhibition of CREB Binding to the HBV-CRE by CRE Decoy Oligonucleotides.

[32P]-Labeled HBV-CRE(C) and HBV-CRE(T) probes used in the EMSA contained single copies of the two corresponding CRE sequences that were mainly present in HBV enhancer I, as described above. In the EMSA, four shifted bands were observed in each control group (Fig. 1A-C,E; lane 2). In the in vitro competition assay, all the shifted bands, except the lowermost band, almost disappeared by adding 10-fold molar excess of the unlabeled probes containing a single copy of HBV-CRE(C), HBV-CRE(T), or C-CRE sequences (Fig. 1A-C), indicating that the lowermost band among the shifted bands was nonspecific. Interestingly, when the unlabeled probes were used at concentrations of less than 2.5-fold molar excess, the probe containing HBV-CRE(C) appeared to be more potent than those containing HBV-CRE(T) or C-CRE in inhibiting the DNA-protein complex formation. In contrast, complex formation was not affected by the oligonucleotides containing the three-base mismatch from HBV-CRE(C) (Mut-HBV-CRE) (Fig. 1A,B), AP-1, or Oct1 (Fig. 1C). As there are one-base to three-base mismatches between HBV-CRE(C), HBV-CRE(T), and C-CRE, we next examined the in vivo effect of CRE decoy oligonucleotides on the binding of nuclear proteins to HBV-CRE. These 24-mer HBV-CRE and C-CRE decoys contained three copies of the HBV-CRE(C) sequence (5′-TGACGCAA-3′) and the C-CRE sequence (5′-TGACGTCA-3′), respectively. Formation of the HBV-CRE(C)-protein complex was almost completely abolished in the nuclear extract that was prepared from HepG2 cells that had been pretreated with the HBV-CRE and C-CRE decoys for 2 days (Fig. 1D). Antibody supershift assay demonstrated that a supershift band was observed only with the anti-CREB antibody, but not with antibodies against other members of the CREB/ATF or Jun/Fos families, for example, ATF-2, c-Fos, and c-Jun (Fig. 1E). This result shows that the CREB protein is present in the HBV-CRE-protein complex, and that HBV-CRE is the specific binding site for CREB. To rule out the possibility that the binding capacity of the antibodies was lost, a [32P]-labeled probe containing the C-CRE sequence was incubated with nuclear extract in the absence or presence of anti-CREB and anti-ATF-2 antibodies (Fig. 1F). Three shifted bands were observed in the control lane (Fig. 1F, lane 1). Addition of anti-CREB or anti-ATF-2 antibody not only resulted in the formation of a supershifted band but also loss of the middle (Fig. 1F; lane 2) or the uppermost (Fig. 1F; lane 3) shifted band, indicating that the antibodies had good binding capacity. These results demonstrate that HBV-CRE is the specific binding site for CREB, and that oligonucleotides containing HBV-CRE or C-CRE sequences can compete with HBV-CRE sites for the binding of transcription factors in vitro and in vivo in a sequence-specific manner.

Figure 1.

CRE decoy oligonucleotides reduce HBV-CRE DNA-protein complex formation. (A-C) For the in vitro competition assay, [32P]-labeled probes were incubated with nuclear extracts of HepG2 cells in the absence or presence of (A,B) the indicated fold molar excess of unlabeled probes containing sequences of HBV-CRE(C), HBV-CRE(T), C-CRE, and three-base mismatch of HBV-CRE(C) (Mut-HBV-CRE) or (C) 10-fold molar excess of unlabeled probes containing the consensus sequence of CRE, AP-1, or Oct1. (D) For the in vivo competition assay, a [32P]-labeled probe was incubated with nuclear extracts of HepG2 cells that had been pretreated with 150 nM HBV-CRE decoy, C-CRE decoy, or control oligonucleotides for 2 days. (E,F) Existence of CREB in the HBV-CRE-protein complex. The [32P]-labeled probes were incubated with HepG2 nuclear extracts in the absence or presence of antibodies. The arrowheads point the supershifted bands caused by the indicated antibodies. Data represent one of three to five independent experiments with similar results. N.E., nuclear extract; S.S., specific shifted band; N, nonspecific shifted band; F, free-probe band; S, shifted band.

Requirement of Functional CREB for HBV Promoter and Enhancer Activities.

The HBV-CRE site is located in enhancer I of the two HBV enhancers; therefore, we constructed a luciferase reporter gene (enI-Luc), the expression of which is directed by enhancer I, to elucidate the role of CREB in the activity of HBV enhancer I. HuH-7 cells were cotransfected with the enI-Luc reporter gene and a plasmid encoding wild-type CREB (CREB) or DNA-binding mutant CREB (KCREB). The reporter activity was decreased in response to transfection with the KCREB plasmid, whereas increased in response to transfection with the CREB plasmid in a dose-dependent manner (Fig. 2A). In contrast, the activity of HBV enhancer I was not affected by the overexpression of wild-type or mutant ATF-2 (Fig. 2B), in agreement with the EMSA result in which a supershift band was not produced by the addition of the anti-ATF-2 antibody (Fig. 1E). These results demonstrate that CREB binding to the HBV-CRE site is required for HBV enhancer I activation. In addition, we used CAT reporter genes that were fused with the full-length HBV genome (Cp-CAT) or enhancers I-II (CEP-CAT), and HepG2 cells were cotransfected with the reporter genes and the C-CRE decoy oligonucleotide. The expression of both Cp-CAT and CEP-CAT was markedly inhibited by the decoy, but not by the control oligonucleotide (Fig. 2C). These findings demonstrated that the C-CRE decoy reduced HBV promoter activity in a sequence-specific manner. To verify that the decrease of HBV promoter/enhancer activity that was caused by the C-CRE decoy oligonucleotide was due to the functional inactivation of CREB, the effect of CREB overexpression on the C-CRE decoy-reduced reporter activity was analyzed. In HepG2 and HuH-7 cells, treatment with 125 nM of C-CRE decoy inhibited more than 60% of the enI-Luc and Cp-CAT activities, as compared to the control oligonucleotide (Fig. 3A-C). These reductions were dose-dependently reversed by CREB overexpression. Taken together, these data indicate that CREB activity and normal interaction between the HBV-CRE site and CREB are necessary for the enhancement of the HBV promoter and enhancer.

Figure 2.

Inhibition of CREB function decreases HBV promoter activity. (A,B) enI-Luc was used for reporter assays. CREB and ATF-2, and KCREB and ATF-2-M2 indicate cotransfection with the expression plasmids encoding wild-types, and DNA-binding mutants of CREB and ATF-2, respectively. Luciferase activity was measured 48 hours after transfection with calcium phosphate. (C) The fusion genes for the reporter assays were Cp-CAT and CEP-CAT. Cells were transfected with a CAT reporter gene and 125 nM oligonucleotides using N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium methyl-sulfate (DOTAP). CAT activity is expressed as radioactivity measured with a phosphoimager. Variations in transfection efficiency were corrected by normalizing luciferase or CAT activity to β-galactosidase activity. The autoradiographed image shows one of four independent experiments. All of the data represent means ± standard deviations obtained from four to six independent experiments. The asterisks indicate significant differences (**P < 0.001 and *P < 0.05) in the mean values compared to each control group.

Figure 3.

Restoration of decoy oligonucleotide-induced decrease in HBV promoter activity by ectopic expression of wild-type CREB. The plasmids that were used as reporter genes were (A,B) enI-Luc and (C) Cp-CAT. Cells were transfected with oligonucleotides in the absence or presence of the CREB plasmid. The assays of luciferase and CAT activity were performed 48 hours after transfection using N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium methyl-sulfate (DOTAP). CAT activity is expressed as radioactivity measured with a phosphoimager. The luciferase and CAT activities were normalized with β-galactosidase activity. All of the data are means ± standard deviations of three replicates. The autoradiographed image shows one of three independent experiments. The asterisks denote significant differences (**P < 0.001 and *P < 0.05) in the mean values between the indicated groups.

Dependence of HBV RNA Synthesis on Normal Function of CREB and the HBV-CRE.

The pregenomic RNA of HBV is a template for reverse transcription of the viral DNA genome; therefore, the levels of pregenomic RNA (3.6 kb) and other viral transcripts (2.4 and 2.1 kb) were measured in cells that constitutively (HepG2.2.15 cells) and transiently (the cells transfected with adwR9, a replication-competent HBV plasmid) produced HBV transcripts and DNAs. The synthesis of HBV mRNAs was remarkably decreased in HepG2 cells that constitutively expressed KCREB (Hep G2/KCREB) and had been transiently transfected with adwR9, as well as HepG2 cells transiently cotransfected with KCREB and adwR9 (Fig. 4A). Synthesis of HBV mRNAs was also investigated in HepG2.2.15 cells that were transiently transfected with CREB mutants (KCREB and a phospho-mutant, M1CREB). Levels of the measured viral transcripts (3.6 and 2.4/2.1 kb) were remarkably reduced by expressing mutant forms of CREB (Fig. 4B; left panel). In contrast, overexpression of wild-type CREB (Fig. 4B; central panel), as well as an increased level of intracellular cAMP that was induced by treatment with forskolin and IBMX (Fig. 4B; right panel), enhanced the synthesis of viral transcripts. To examine whether CRE decoy oligonucleotides inhibited HBV RNA synthesis, HepG2.2.15 cells were treated with various concentrations of CRE decoy and control oligonucleotides. The production of viral transcripts was manifestly inhibited in a dose-dependent manner by HBV-CRE or C-CRE decoy, but not by control oligonucleotides (Fig. 4C,D). Moreover, a mutant adwR9 (adwR9-mCRE) containing the Mut-HBV-CRE sequence instead of HBV-CRE(C), showed a remarkable reduction in the synthesis of HBV mRNAs (Fig. 4E). Taken together, these results clearly show that normal functioning of the CREB and the HBV-CRE is indispensable for the synthesis of HBV mRNAs including pregenomic RNA.

Figure 4.

Functional inhibition of CREB and mutation in the HBV-CRE site correlate with reduced synthesis of HBV mRNAs including pregenomic RNA. (A) HepG2 cells constitutively expressing KCREB (Hep G2/KCREB) were transiently transfected with adwR9, a replication-competent plasmid containing the incomplete head-to-tail dimer of wild-type HBV, and HepG2 cells were transiently cotransfected with KCREB and adwR9. (B) Cells were transfected with the mutant (KCREB and a phospho-mutant, M1CREB) or wild-type CREB plasmids, and cotreated with 10 μM forskolin and 10 μM IBMX. (C,D) Cells were transfected with HBV-CRE or C-CRE decoy oligonucleotides. (E) Cells were transfected with adwR9 or adwR9-mCRE containing Mut-HBV-CRE sequence. The empty vector was produced by removal of HindIII/XbaI fragment from KCREB, blunting, and ligation. The transfections were carried out with (A,B,E) calcium phosphate or (C,D) N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium methyl-sulfate (DOTAP). Total cellular RNA was prepared 48 hours after transfection. Levels of viral transcripts were analyzed by Northern blot analysis. Glutaraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA was used as an internal control to correct RNA loading errors, and variations in transfection efficiency were corrected by mRNA levels of β-galactosidase or green fluorescent protein (GFP). For quantification, levels of HBV and GAPDH mRNAs were measured with a phosphoimager. The [32P]-labeled probes were fragments of adwR9 (1.4-kb), GAPDH (1.2 kb), β-galactosidase (0.6 kb), and GFP (0.9 kb). Data represent one of three or four independent experiments with similar results. The asterisks designate significant differences (**P < 0.001 and *P < 0.05) in the mean values compared to each control group.

Reduction of HBV DNA Synthesis by Interfering with CREB Function.

To further examine the effect of CREB on HBV replication, the levels of relaxed circular (RC), double-stranded linear (DS), and single-stranded (SS) replicative intermediates were determined in HepG2.2.15 cells. The synthesis of HBV DNAs was significantly inhibited by the overexpression of M1CREB (Fig. 5A) or by treatment with HBV-CRE and C-CRE decoy oligonucleotides (Fig. 5B,C). At the same molar concentration (150 nM), the extent of the inhibition of HBV DNA synthesis by the C-CRE or HBV-CRE decoys was equal to or more than that caused by lamivudine, respectively (Fig. 5D). The inhibition of HBV DNA synthesis by transfection with the M1CREB plasmid and CRE decoy oligonucleotides was consistent with their inhibition of HBV RNA synthesis.

Figure 5.

Inhibition of CREB activity reduces viral DNA synthesis. HepG2.2.15 cells were transfected with (A) M1CREB, (B) HBV-CRE decoy oligonucleotide, and (C) C-CRE decoy oligonucleotide. (D) Cells were transfected with 150 nM oligonucleotides or treated with 150 nM lamivudine. The empty vector was the same one described in Fig. 4. The transfections were performed using (A) calcium phosphate or (B,C,D) N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium methyl-sulfate (DOTAP). Core-associated DNA was prepared 48 hours after transfection or lamivudine treatment. Viral DNA levels were measured by Southern blot analysis using [32P]-labeled full-length HBV DNA (3.2 kb). For quantification, level of HBV DNAs were measured with a phosphoimager. RC, DS, and SS indicate relaxed circular, double-stranded linear, and single-stranded HBV DNA, respectively. Data are from one of three or four independent experiments with similar results. The asterisks mean significant differences (*P < 0.05) in the mean values compared to each control group.

Suppression of HBsAg and HBeAg Expression by CREB Inhibition.

To evaluate the effects of CREB inactivation on HBV gene expression, we measured the levels of HBsAg and HBeAg in the culture media of HepG2.2.15 cells and HepG2 cells transfected with a replication-competent HBV construct, adwR9. The ectopic expression of KCREB and M1CREB in HepG 2.2.15 cells inhibited the secretion of HBsAg and HBeAg by 40% and 20%, respectively (Fig. 6A,B). The inhibition of HBsAg and HBeAg secretion with mutant forms of CREB appeared to be lower than that of HBV mRNA and DNA syntheses. Since it is probable that some of the viral antigens in the culture medium were synthesized before the expression of the mutant CREBs reached their inhibitory level, the levels of viral antigens were evaluated in HepG2/KCREB cells that had been transiently transfected with adwR9. As expected, inhibition of HBsAg and HBeAg secretion was increased to 80% and 50%, respectively (Fig. 6C,D). Taken together, these results demonstrate that blocking CREB activity reduces the expression and/or secretion of viral antigens.

Figure 6.

Functional inactivation of CREB decreases the secretion of HBsAg and HBeAg. (A,B) Cells were transfected with KCREB or M1CREB plasmids by calcium phosphate. After 3 days, the levels of HBsAg and HBeAg in the culture medium were measured and normalized for total cell number (counted with the Coulter counter; Beckman Coulter, Miami, FL) and total protein amount (quantified by the Bradford method). (C,D) HepG2 and HepG2/KCREB cells were transfected with adwR9 by calcium phosphate. After 5 days, levels of HBsAg and HBeAg in the culture medium were analyzed. Variations in transfection efficiency were normalized with β-galactosidase activity. The data represent means ± standard deviations obtained from five independent experiments. The asterisks indicate significant differences (**P < 0.001 and *P < 0.05) in the mean values compared to each control group.

No Dependence of CREB Inactivation-Reduced HBV Replication and Gene Expression on Interference with Cell Viability.

It was necessary to elucidate whether CRE-decoy oligonucleotides (HBV-CRE decoy and C-CRE decoy) and CREB mutants exerted their effects on the HBV replication and gene expression by affecting cell viability. Therefore, changes in cell viability of HepG2.2.15 cells, which had been transiently transfected with CREB mutants (KCREB or M1CREB) or empty vector, were measured by MTT-based assay and cell counting. CREB inactivation by transfection of mutant CREB plasmids had no effect on cell viability (Fig. 7A). Treatment of the cells with CRE-decoy oligonucleotides, as compared to sperm DNA or control oligonucleotide, showed no significant difference in the MTT-based cell viability, but resulted in 30% decrease of cell number (Fig. 7B). Even though the both CRE-decoys decreased the cell number to a similar extent, a remarkable reduction of HBV DNA synthesis was observed only in the group treated with HBV-CRE decoy, suggesting that there was no correlation between the decrease of cell number and reduction in HBV DNA synthesis. The above described results together demonstrate that the decrease of HBV replication and gene expression by CREB inactivation was not due to interference with cell viability.

Figure 7.

Decrease of HBV replication and gene expression by CREB inactivation was not due to interference with cell viability. (A) Cells were transfected with CREB mutants (KCREB or M1CREB) or empty vector using calcium phosphate. The empty vector was the same one described in Fig. 4. (B) Cells were transfected with 150 nM sperm DNA, HBV-CRE decoy, C-CRE decoy, and control oligonucleotides using N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium methyl-sulfate (DOTAP). After 48 hours, cell viability was determined using MTT assay, and cell number was counted with the Coulter counter (Beckman Coulter). Viral DNA levels were measured by Southern blot analysis using [32P]-labeled full-length HBV DNA (3.2 kb). Levels of HBV DNAs were measured with a phosphoimager. The data represent means ± standard deviations obtained from three or five independent experiments. The asterisks designate significant differences (*P < 0.05) in the mean values compared to each control group or significant differences (*P < 0.05) in the mean values between the indicated groups.

No Difference in CRE-Directed Transcription Between the Cells Containing Wild-Type and X-Minus Mutant HBV.

It is well known that HBx, one of HBV regulatory proteins, acts directly to enhance the binding of CREB to the CRE site present in HBV enhancer I, and augments HBV replication and viral gene expression. Until now, it has not yet been studied whether HBx enhances HBV replication and gene expression through activation of the CRE/CREB system. Therefore, it is interesting and meaningful to examine whether there is a difference in CRE-directed gene expression between the cells in which wild-type HBV, adwR9, and X-minus mutant HBV, HBX-21, are replicating. Thus, HepG2 and HuH-7 cells were cotransfected with the CRE-CAT reporter gene, SS-CAT, and one of the replication-competent plasmids (adwR9 and HBX-21) and empty vector. However, Fig. 8 shows that there was no difference in the CAT activities between the cells containing wild-type HBV and X-minus mutant HBV.

Figure 8.

There was no difference in CRE-directed transcription between the cells containing wild-type and X-minus mutant HBV. (A) HepG2 and (B) HuH-7 cells were cotransfected with the somatostatin-CAT reporter gene (SS-CAT) and one of empty vector (pGEM-7Zf(+); Promega), adwR9, and a replication-competent plasmid containing the incomplete head-to-tail dimer of X-minus mutant HBV (HBX-21). CAT activity is expressed as radioactivity measured with a phosphoimager. Variations in transfection efficiency were corrected by normalizing CAT activity to β-galactosidase activity. All of the data represent means ± standard deviations obtained from three or four independent experiments.

Discussion

In this study, we provided several lines of evidence indicating that normal interaction between CREB and the HBV-CRE is indispensable for HBV gene expression and viral replication, by using cells that constitutively (HepG2.2.15) and transiently (HepG2 transfected with adwR9, a replication-competent HBV plasmid) produce HBV transcripts including pregenomic RNA. First, based on sequence-comparison analysis of 984 HBVs reported worldwide, the HBV-CRE sequence is one of the most highly conserved among the transcription factor-binding sites in the HBV enhancer I region. Second, HBV-CRE is the specific binding site for CREB. A supershift band in an EMSA was produced only with an anti-CREB antibody but not with anti-ATF-2, anti-c-Fos, or anti-c-Jun antibodies. The nuclear protein binding to the HBV-CRE probe in vitro was competed only by the oligonucleotides containing wild-type HBV-CRE or C-CRE sequences, indicating that the competition was sequence-specific. Moreover, HBV-CRE as well as C-CRE decoy oligonucleotides could compete with the HBV-CRE site for the binding of transcription factors in vivo. Although ATF-2 is known to bind to the HBV-CRE sequence,19 the overexpression of wild-type or mutant ATF-2 had no influence on the activity of HBV enhancer I in agreement with the EMSA result where the anti-ATF-2 antibody did not supershift HBV-CRE DNA-protein complexes. Altogether, these results strongly suggest that CREB is the major transcription factor that binds to the HBV-CRE site. Third, the ectopic expression of wild-type or mutant CREBs resulted in increased or decreased expression, respectively, of the reporter gene containing HBV enhancer I. The C-CRE decoy remarkably reduced the expression of reporter genes containing HBV enhancer I, HBV enhancers I and II, and the full-length of HBV genome. Furthermore, the ectopic expression of wild-type CREB could completely rescue the reduction of reporter-gene expression. Fourth, and most importantly, the enhancement of CREB activity by transfection with a CREB plasmid or an increase of intracellular cAMP levels by cotreatment with forskolin and IBMX was accompanied by increased levels of HBV transcripts including pregenomic RNA. Various methods to inhibit CREB function, including the overexpression of CREB mutants and treatment with CRE decoy oligonucleotides, dramatically reduced viral replication, as evidenced by decreased levels of HBV transcripts and all viral replicative forms, as well as the secretion of HBsAg and HBeAg. However, the effect of CREB inactivation on HBV replication and gene expression was not due to interference with cell viability. In addition, introduction of a three-base mutation into the HBV-CRE sequence significantly reduced the synthesis of HBV mRNAs. Interestingly, the HBV-CRE decoy was much more potent than lamivudine, and the C-CRE decoy was comparable to the latter in terms of inhibiting HBV DNA synthesis. This finding is in parallel with the EMSA result in which the oligonucleotide containing HBV-CRE(C) seems to be more potent than the two other oligonucleotides in competing with the DNA probes.

The high conservation rate of the CRE sequence in HBV enhancer I suggests that viruses with mutations in HBV-CRE that prevent transcription factors from binding to the site are selected against, since the normal function of the CRE site and CREB are necessary for the HBV life cycle, as demonstrated in this study. This speculation is supported by a previous report showing that HBV enhancer I containing HBV-CRE is more important than enhancer II in regulating global and temporal HBV gene expression.17 Moreover, it is known that the CRE sequences found in specific regulatory elements of various viruses often control viral replication and gene expression.28

In addition to the CRE site in HBV enhancer I, there are two CRE sites in the Pre-S2/S promoter region.42 Therefore, the decrease of 2.4-kb/2.1-kb RNA species by inhibiting CREB activity might have been due to reduced activities of Pre-S2/S promoter as well as enhancer I, even though the degree of conservation of the CRE sequences in the promoter was lower than that in enhancer I (Table 2).

Table 2. CRE Sequences in the Pre-S2/S Promoter Region
SymbolSequenceFrequencyRatio (%)
  1. A total of 984 HBV sequences registered in GenBank were analyzed using the multiple sequence alignment program ClustalX2.0. *The CRE sequences of the Pre-S2/S promoter region that were described in a previous report42 were used as standard sequences.

CREB ITCTCGTCA*209/98421.2
 TATCGTCA652/98466.3
 CATCATCA44/9844.5
 Others79/9848
CREB IICCTGTGACGAAC*68/9846.9
 CCTGCACCGAAC276/98428
 CCTGCGCTGAAC143/98414.5
 CCTGTACCGAAC152/98415.4
 CCTGCACCGAAT72/9847.3
 Others273/98427.9

HBx is known to enhance the binding of CREB to the CRE site present in HBV enhancer I,19, 20 and augment HBV replication and viral gene expression.23-27 In this study, we demonstrated that normal interaction between CREB and the HBV-CRE is indispensable for HBV gene expression and viral replication. In HBV replicating cells, it is, therefore, probable that HBx exerts its effect on HBV replication and gene expression through activation of the CRE/CREB system. Although we could not observe a difference in the CRE-directed reporter activities between the cells in which wild-type HBV and X-minus mutant HBV were replicating, further studies are needed to clarify this hypothesis.

In contrast to our results, it has been reported in the duck HBV (DHBV) system that the increase of intracellular cAMP level by treatment of persistently infected hepatocytes with forskolin or dibutyryl-cAMP resulted in a slight but continuous decrease of viral replication.43 However, the transcription factor involved in the inhibition of DHBV replication by cAMP was not identified in the above report, and there has been no report until now to mention the existence of a CRE site in DHBV enhancer. The region of DHBV enhancer that has the sequence (5′-TGGCGCAAT-3′) most similar to the HBV-CRE(C) site is well known to be the CCAAT/enhancer-binding protein (C/EBP)-binding site.44 It has been demonstrated that the C/EBP site in the DHBV enhancer acts as an inhibitory element,45 and that C/EBP can act as a repressor at high concentrations or as an activator at low concentrations in the HBV.46 Moreover, it has been reported that a number of gene promoters that are activated in response to cAMP contain binding sites for C/EBP, and C/EBPs are both constitutive and cAMP-inducible.47 Therefore, it is quite likely in the DHBV system that stimulation of the cAMP signal transduction pathway results in C/EBPs activation, and in turn, reduces viral replication. However, the detail mechanism by which cAMP decreases DHBV replication still remains to be elucidated.

In our previous report, we showed that the C-CRE decoy oligonucleotide can interfere with the expression of CRE-responsive genes.40 In this study, we demonstrated that the HBV-CRE and C-CRE decoys effectively reduce HBV replication and gene expression. Since HBV-CRE is a very highly conserved transcription factor-binding site in HBV enhancer I, and only very few mutations in the sequence are allowed for viability, approaches to modulate the transactivation activity of CREB are expected to minimize the emergence of resistant mutants. However, it would be necessary to evaluate the emergence of resistant mutants as well as the anti-HBV actions of CRE-decoy oligonucleotides in future animal experiments.

In summary, our present data provide the first experimental proof that interaction between CREB and the HBV-CRE plays a critical role in the HBV life cycle and suggest that methods to suppress CREB activity may offer a new means to treat HBV infection.

Acknowledgements

We thank Y.H. Kim, H.E. Blum, R.H. Goodman, M. Montminy, G. Redeuilh, R.G. Pestell, Y. Yun, J.E. Yeon, and Y.S. Cho-Chung for HepG2.2.15 cells, plasmids, and reagents. We thank H.S. Na for statistical analysis and J. Sohn for helpful suggestions on writing the manuscript.

Ancillary