Potential conflict of interest: Nothing to report.
This work was supported by National Institutes of Health grants R01CA55578 (to T.S.B.Y.), R21RR024229 (to T.S.B.Y.), and P01CA123328 (to T.S.B.Y. and J.H.J.O.), a VA Merit Review Award (to T.S.B.Y.), and an American Cancer Society postdoctoral fellowship (to T.T.).
Hepatitis B virus (HBV) is a small DNA virus that requires cellular transcription factors for the expression of its genes. To understand the molecular mechanisms that regulate HBV gene expression, we conducted a yeast one-hybrid screen to identify novel cellular transcription factors that may control HBV gene expression. Here, we demonstrate that Krüppel-like factor 15 (KLF15), a liver-enriched transcription factor, can robustly activate HBV surface and core promoters. Mutations in the putative KLF15 binding site in the HBV core promoter abolished the ability of KLF15 to activate the core promoter in luciferase assays. Furthermore, the overexpression of KLF15 stimulated the expression of HBV surface antigen (HBsAg) and the core protein and enhanced viral replication. Conversely, small interfering RNA knockdown of the endogenous KLF15 in Huh7 cells resulted in a reduction in HBV surface- and core-promoter activities. In electrophoretic mobility shift and chromatin immunoprecipitation assays, KLF15 binds to DNA probes derived from the core promoter and the surface promoter. Introduction of an expression vector for KLF15 short hairpin RNA, together with the HBV genome into the mouse liver using hydrodynamic injection, resulted in a significant reduction in viral gene expression and DNA replication. Additionally, mutations in the KLF15 response element in the HBV core promoter significantly reduced viral DNA levels in the mouse serum. Conclusion: KLF15 is a novel transcriptional activator for HBV core and surface promoters. It is possible that KLF15 may serve as a potential therapeutic target to reduce HBV gene expression and viral replication. (HEPATOLOGY 2011;)
Hepatitis B virus (HBV) is an enveloped hepatotropic virus that can cause liver cirrhosis and hepatocellular carcinoma. This virus chronically infects approximately 350 million people worldwide and causes approximately 500,000 to 1 million deaths annually. HBV is a small DNA virus with a circular and partially double-stranded genome of approximately 3.2 kilobases. The HBV genome contains four genes: S, C, X, and P. The S gene codes for the large, middle, and major surface antigens (HBsAgs), which are three related viral envelope proteins. The C gene codes for the precore protein, which is the precursor of the e antigen found in the sera of patients with HBV, and the core protein, which is the viral capsid protein. The P gene codes for the viral DNA polymerase, and the X gene codes for a regulatory protein. The expression of these HBV genes is controlled by four promoters and two enhancers that depend on host factors for transcriptional regulation.1, 2 The preS1 promoter controls the expression of the large surface antigen, and the major surface promoter (also known as the surface promoter) controls the expression of the middle and small surface antigens. The core promoter dictates the expression of the HBV e antigen, core protein, and DNA polymerase. The X promoter controls the transcription of the X RNA. After its synthesis, the core protein packages the core RNA, which is larger than the genome size and is also known as the pregenomic RNA (pgRNA), to form the core particle. The pgRNA then serves as the template to direct the synthesis of the partially double-stranded viral DNA genome, using the viral DNA polymerase that is also packaged. The core RNA plays a pivotal role in the HBV life cycle and its increased expression has been shown to enhance viral replication.3, 4
The identification of host factors that interact with the HBV DNA genome has made significant contributions to our understanding of mechanisms that regulate HBV gene expression. Indeed, both liver-enriched and ubiquitous transcription factors, such as hepatocyte nuclear factor 1 (HNF1), HNF3, HNF4, CCAAT enhancer-binding protein (C/EBP), chicken ovalbumin upstream promoter transcription factor (COUP-TF), nuclear transcription factor Y (NF-Y), and specificity protein 1 (Sp1), have been shown to regulate the expression of the S and C genes.5-13 The liver specificity of the preS1 promoter, the major surface promoter, and the core promoter is attributed to the need of liver-enriched transcription factors for their activities.10, 14-18
In this study, we used a yeast one-hybrid screen to identify additional transcription factors that could activate the major surface promoter. Using cDNA libraries prepared from the human hepatoma cell line, Huh7, and mouse liver, we identified several members of the Krüppel-like factor (KLF) family as potential activators of the surface promoter (T. Tan and T.S.B. Yen, unpublished data). KLF family members are characterized by their three carboxy-terminal C2H2 zinc fingers and share a high degree of homology with Sp1-like proteins. At least 21 Sp1/KLF proteins have been identified in the human genome. They have highly conserved DNA-binding domains, but show significant variations in the transactivation domain in their amino terminus.19, 20 Krüppel-like factor 15 (KLF15) has been shown to regulate the expression of a number of genes involved in many aspects of physiological homeostasis, including glucose uptake and adipogenesis.21-25 Moreover, KLF15 is highly expressed in the human liver.25 These observations led us to hypothesize that KLF15 might be a potential activator for HBV gene expression. Indeed, our results indicate that KLF15 can activate the expression of the HBV S and C genes both in vitro and in vivo. Our results thus uncovered previously unrecognized functions of KLF15 in HBV gene expression.
The reporter plasmids, pCP, pS1-Luc,26 pCCD1 (cyclin D1 luciferase construct),27 and pRL-TK, as well as the expression plasmids, pHBV1.3D28 and pXGH,29 have been previously described. The reporter plasmids, pS1Z1/Z2mut-Luc and pS1M2mut-Luc, were made by polymerase chain reaction (PCR) amplifying from pS1Z1+2mutCAT and pS1M2mutCAT and cloning into pS1-Luc digested with BglII and HindIII. The plasmids, pAAV-HBV1.2, pCP1.3x/Luc, and pKLF15, were obtained from P.J. Chen (National Taiwan University, Taipei, Taiwan), Y. Shaul (The Weizmann Institute of Science, Rehovot, Israel), and S. Gray (Harvard Medical School, Boston, MA), respectively. pLive-SEAP (secreted alkaline phosphatase) (Mirus Bio, Madison, WI), which expresses secreted human placental alkaline phosphatase, was used to monitor the efficiency of plasmid delivery after hydrodynamic injections. pCPm1, pCPm2, and pCP2m were generated from pCP with primer pairs CPm1-s/as, CPm2-s/as, and CP2m-s/as, respectively, and pAAV-HBV1.2-CPm2 was generated from pAAV-HBV1.2 with the primer, CPm2-60 (Table 1), using the QuikChange Lightning site-directed mutagenesis kit (Stratagene, La Jolla, CA).
Table 1. Sequences of Primers and Probes in This Study
Nucleotide Sequence (5′ to 3′)
Cell Transfection, Messenger RNA Analysis, and Protein Assays.
HepG2 and Huh7 cells were cultured at 37°C in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and penicillin-streptomycin in 7% CO2. Cells in a 12-well plate were transfected with 800 ng of DNA plasmids, using 2.4 μL of FugeneHD (Roche Diagnostics, Indianapolis, IN), and harvested at specific time points after transfection. pRL-TK, which expresses the Renilla luciferase reporter under the control of the herpes simplex virus thymidine kinase promoter, or pXGH, which expresses the human growth hormone (hGH) reporter under the control of the mouse metallothionein promoter, was used for cotransfection to monitor transfection efficiency. The hGH enzyme-linked immunosorbent assay (ELISA) kit (Roche Diagnostics, Indianapolis, IN) was used to detect hGH in the culture medium.
Stealth Select RNA interference (RNAi) short interfering RNA (siRNA) (Invitrogen, Carlsbad, CA) was used for RNAi studies in Huh7 cells. For experiments involving cotransfection of siRNA (50 nM) and pHBV1.3D, Lipofectamine 2000 (Invitrogen) was used according to the manufacturer's protocol. For RNAi experiments in luciferase assays, siRNA (50 nM) was transfected using Lipofectamine RNAiMAX (Invitrogen). Luciferase reporter plasmids were transfected 24 hours after siRNA transfection. Luciferase activities were measured using the Dual-Glo Luciferase assay system (Promega, Madison, WI).
RNA extraction and reverse transcription were performed using the RNeasy Mini kit and the SuperScript III first-strand synthesis system (Invitrogen), respectively. The SYBR green master mix (Roche Diagnostics) was used for quantitative real-time PCR (qRT-PCR) for analysis of the KLF15 RNA. Mouse 36B4 RNA (Table 1) was also analyzed to serve as an internal control. The primer pairs for KLF15 and 36B4 RNA analysis are shown in Table 1.
To measure HBV core protein levels, transfected HepG2 cells in a 12-well plate were lysed with 200 μL of RIPA buffer (50 mM Tris HCl, pH 7.5, 150 mM NaCl, 1 mM ethylene diamine tetraacetic acid [EDTA], 1% Nonidet P-40, 0.1% sodium dodecyl sulfate [SDS], 10% glycerol, and 0.5% deoxycholic acid). Samples were subjected to electrophoresis in a 15% SDS-PAGE (polyacrylamide gel electrophoresis) gel and then transferred to a polyvinylidene fluoride membrane (Millipore, Billerica, MA). One-half of the membrane was probed with the rabbit anti-HBcAg antibody (1:300 dilution; US Biological, Marblehead, MA), followed by incubation with horseradish peroxidase (HRP)-conjugated goat anti-rabbit antibody (1:3000 dilution; Santa Cruz Biotechology, Santa Cruz, CA). The other half of the membrane was probed with the mouse antiactin primary antibody (1:40,000 dilution; Calbiochem, San Diego, CA) and the HRP-conjugated goat anti-mouse immunoglobulin M (IgM) secondary antibody (1:5000 dilution). The enhanced chemiluminescence (ECL) Plus Western blotting detection system (Amersham Biosciences, Pittsburgh, PA) was used to develop the signals. HBsAg levels in culture media and mouse sera were measured by the HBs enzyme immunoassay (EIA) kit (International Immuno-Diagnostics, Foster City, CA).
Expression of Recombinant KLF15 and Electrophoretic Mobility Shift Assays.
pKLF15, which expresses mouse KLF15 with a C-terminal FLAG tag, was transfected into 293T cells using FugeneHD. The culture medium was changed to Opti-MEM 10 hours posttransfection. After further incubation for 48 hours, cells were harvested for protein purification using EZview Red ANTI-FLAG M2 Affinity Gel (Sigma-Aldrich, St. Louis, MO), according to the manufacturer's instruction. Purified recombinant KLF15 (rKLF15) protein was analyzed by Western blot using anti-FLAG M2 (Sigma-Aldrich) and anti-KLF15 (ab2647; Abcam, Cambridge, UK) antibodies. Double-stranded synthetic oligonucleotides were prepared by annealing the two DNA strands in 10 × buffer (200 mM Tris, 100 mM MgCl2, and 250 mM NaCl), followed by cooling from 65°C to 37°C. The double-stranded oligonucleotides were labeled with [γ-32P]ATP (PerkinElmer, Waltham, MA), using T4 polynucleotide kinase (Roche Diagnostics). Unincorporated [γ-32P]ATP was removed using the Auto-Seq G-50 dye terminator removal kit (Amersham). Nucleotide sequences of the DNA probes used in this study, CP35, CLCK1, MLTF, and SP70, are listed in Table 1.
For the electrophoretic mobility shift assay (EMSA) reaction, 2 μL of rKLF15 (100 ng/μL) were mixed with 25 fmol of labeled probe and 4 μL of 5× gelshift buffer (Promega), in a total volume of 20 μL, and incubated at 37°C for 30 minutes. The reaction mixtures were loaded on a 6% polyacrylamide gel and subjected to electrophoresis in 0.5× Tris/Borate/EDTA (TBE) buffer at 200 V for 2∼3 hours. The gel was dried and analyzed by a Typhoon phosphorimager (GE Healthcare, Waukesha, WI). For the supershift assay using the anti-KLF15 antibody, rKLF15 was incubated with a labeled probe at 37°C for 30 minutes, followed by incubation with either anti-KLF15 or control antibody at room temperature for 40 minutes.
Chromatin Immunoprecipitation Assay.
The chromatin immunoprecipitation (ChIP) assay was conducted using the EZ-Magna ChIP G Chromatin Immunoprecipitation Kit (Millipore). Briefly, HepG2 cells in 10-cm dishes were cotransfected with 4.8 μg of pKLF15 and 1 μg of the reporter construct, pCP, or pS1-Luc or the mutated constructs, pCP-2m, pS1Z1/Z2mut-Luc, and pS1M2mut-Luc. Forty-eight hours after transfection, cells were crosslinked with formaldehyde and harvested for immunoprecipitation. An aliquot of the cell lysates was saved to serve as the input DNA control. After the reversal of crosslinking with 5 M of NaCl, ChIP samples were subjected to PCR using the primer pair, HBV1644F/HBV1805R (for the core promoter), and primer pair RV3/HBV22R (for the surface promoter). Antibodies used in ChIP assays included KLF15 (Abcam), NF-Y (Thermo Fisher Scientific, Wilmington, MA), Sp1 (Abcam), rabbit control IgG, and goat control IgG (Abcam) antibodies.
HBV DNA Assay.
HepG2 cells cotransfected with pHBV1.3D and pKLF15 or its control vector pcDNA3.1 were harvested in 600 μL of lysis buffer (50 mM Tris-HCl, pH 7.0, and 0.5% Nonidet P-40) 96 hours after transfection. Ninety microliters of cell lysates or culture medium were mixed with 1 μL of TURBO DNase (Ambion, Austin, TX) and 10× DNase buffer and incubated at 37°C for 30 minutes. After, DNase was inactivated by heating at 75°C for 10 minutes. The mixtures were subsequently processed with the virus extraction column (QIAamp MinElute Virus Spin Kit; Qiagen, Germantown, MD), following the manufacturer's instruction. Viral genome thus purified was quantified by RT-PCR, using the SYBR green master mix and the HBV DNA-F/R primer pair (Table 1). To extract the encapsidated viral DNA from the mouse serum, 25 or 100 μL of mice serum was used.
Experiments involving mice were approved by the Institutional Animal Care and Use Committee (IACUC) of the San Francisco VA Medical Center. Male C57BL/6J mice were purchased from the Jackson Laboratory (Bar Harbor, ME) and used at 6-7 weeks of age. To study the effects of KLF15 on HBV viral protein expression and DNA replication, 5 μg of pAAV-HBV1.2, 5 μg of pLive-SEAP, and 30 or 50 μg of miR RNAi constructs were injected into the mouse tail vein by hydrodynamic injection in a volume of phosphate-buffered saline (PBS) equivalent to 8% of the mouse body weight. Mouse sera were assayed for HBsAg and capsid-associated HBV DNA at the indicated time points after injection. The SensoLyte FDP SEAP Reporter Gene Assay kit (AnaSpec, Fremont, CA) was used to detect SEAP activity in mouse sera.
For immunofluorescence staining of the HBV core antigen, mouse livers were fixed with 4% formalin overnight, cryoprotected in 30% sucrose, and sectioned at a thickness of 10 μm, using Leica cryostat (Leica Microsystems, Buffalo Grove, IL), and mounted on Superfrost glass slides (Thermo Fisher Scientific). Sections were incubated with the primary antibody (anti-HBc; US Biological) overnight, followed by incubation with the goat anti-rabbit secondary antibody conjugated with Alexa Fluor 568 (Invitrogen). Slides were subsequently counterstained with 4′,6-diamidino-2-phenylindole (DAPI). Images were captured using a Zeiss LSM 510 Meta Confocal Microscope (Carl Zeiss GmbH, Jena, Germany). For Western blot analysis of the HBV core protein, approximately 120 mg of the mouse liver was rinsed with cold buffer A (50 mM Tris-HCl, pH 7.0, 2 mM EDTA, and 150 mM NaCl) and homogenized in buffer B (50 mM Tris-HCl, pH 7.0, 10% glycerol, 5 mM MgCl2, 0.2 mM EDTA, 1 mM dithiothreitol, and 1 × protease inhibitor cocktail). The homogenates were centrifuged at 15,000g for 30 minutes twice to pellet the cell debris. Next, 150 μg of total proteins were analyzed in 15% SDS-PAGE, using the same protocol described above for HepG2 cell lysates.
BLOCK-iT Pol II miR RNAi expression vectors (Invitrogen) were used to knock down the expression of KLF15 in mice. To analyze the expression level of KLF15 in miR RNAi-transfected hepatocytes, mice were anesthetized and their livers were perfused with collagenase 3 days after hydrodynamic injection to obtain hepatocytes, which were subsequently sorted by flow cytometry to separate transfected (i.e., green fluorescent protein [GFP]-positive) hepatocytes from untransfected (i.e., GFP-negative) hepatocytes. To analyze the effect of KLF15 on viral gene expression, 10 μg of pAAV-HBV1.2 or pAAV-HBV1.2-CPm2 and 5 μg of pLive-SEAP were coinjected into mice through the tail veins. All of the plasmids used for hydrodynamic injection were prepared using the EndoFree plasmid preparation kit (Qiagen).
The Student t test and Mann-Whitney U test were used to analyze data. A value of P < 0.05 was regarded as statistically significant.
KLF15 Transactivates HBV Surface and Core Promoters.
To identify host factors that can promote HBV gene expression, we initiated a yeast one-hybrid assay to screen for transcription factors that could bind the HBV major surface promoter. Multiple screens pulled out the previously identified NF-Y transcription factor, as well as a few members of the KLF family of transcription factors12 (T. Tan and T.S.B. Yen, unpublished data). We chose to focus on KLF15 for our studies, as this transcription factor is enriched in the liver.25 Using a luciferase reporter, which was driven by the HBV surface promoter, we found that KLF15 increased luciferase activity in a dose-dependent manner by up to 80-fold (Fig. 1A and Supporting Fig. 1). This transactivation effect of KLF15 was specific to the HBV surface promoter, because it had little effect on the cyclin D1 promoter (Fig. 1B). Previously, we and others found that the NF-Y binding site (CCAAT box, designated as the M2 site) and two flanking Sp1 factor binding sites (Z1/Z2 sites) are critical for HBV surface promoter activity.1, 10, 12, 30 As shown in Fig. 1C, the transactivation effect of KLF15 on the surface promoter was dramatically reduced to approximately four-fold by the mutations in the Z1/Z2 site and completely abolished by the mutation in the M2 site (Fig. 1C and Supporting Fig. 1). These results indicate that KLF15 is a potent activator for the HBV surface promoter, and that its optimal activity on the surface promoter requires intact Z1/Z2 and M2 sites.
Because the HBV core promoter is activated by Sp1 and/or Sp1-like factor,34 we thought KLF15 might also be involved in its regulation. To determine whether KLF15 could also activate the HBV core promoter, two different core promoter reporters, pCP1.3x and pCP, were used for the studies. pCP1.3x was generated from an HBV genomic DNA fragment, in which a luciferase open-reading frame substituted the core open-reading frame in the parental construct, whereas the pCP contains only a 162–base pair HBV core promoter fragment. In both reporter constructs, the expression of the luciferase was under the control of the core promoter. As shown in Fig. 2A,B, KLF15 could also activate the core promoter in a dose-dependent manner similar to the effect on the surface promoter (Fig. 1). Notably, we identified a sequence within the 162–base pair core promoter in pCP that matched exactly the KLF15 consensus binding sequence (GGGGNGGNG) reported by Uchida et al.25 Moreover, this sequence matched an Sp1 or Sp1-like factor binding site (C region, site 3) identified by McLachlan's group in the HBV core promoter.34 To determine whether this consensus sequence could be recognized by KLF15, we generated two mutant luciferase reporters, pCPm1 and pCPm2, in which two guanosine residues in the KLF15 consensus sequence were changed to thymidine (Fig. 2C). These two constructs were designed to disrupt possible KLF15 binding to the core promoter. In addition, the CPm2 sequence was designed to maintain the overlapping HBV X (HBx) protein-coding sequence. Hence, the exact same mutations can be introduced into the HBV genome to study their effects on HBV gene expression without the confounding effect of HBx mutations.28, 31, 32 Consistent with our predictions, mutations in the KLF15 consensus sequence abolished the ability of KLF15 to transactivate CPm1 and CPm2 (Fig. 2D). Therefore, these results support the notion that KLF15 activates the HBV core promoter via the consensus KLF15-binding sequence embedded in this promoter.
KLF15 Increases HBV Gene Expression and Viral DNA Replication.
To further investigate the possible effect of KLF15 on HBV gene expression from the HBV genome, we cotransfected pKLF15 or its control vector with pHBV1.3D, which contains the 1.3-mer HBV genome, into HepG2 and Huh7 cells. Our results showed that the coexpression of KLF15 led to a seven-fold increase of the HBsAg level in the culture medium of HepG2 cells (Fig. 3A). Such an increase in HBsAg production was even more prominent in Huh7 cells, which was up to nearly 20-fold (Fig. 3B). Similarly, KLF15 also increased the core protein expression level in HepG2 cells (Fig. 3C). When the culture medium and cell lysates from transfectants were analyzed for encapsidated HBV DNA by RT-PCR, we found that the coexpression of KLF15 increased the extracellular encapsidated HBV DNA level by approximately two-fold and also slightly increased the intracellular encapsidated HBV DNA level (Fig. 3D). Taken together, our results indicated that, in the context of the HBV genome, KLF15 could enhance the expression of HBsAg and the core protein, as well as HBV DNA replication.
Suppression of Endogenous KLF15 Expression Reduces HBsAg Production and HBV Surface and Core Promoter Activities.
To determine whether endogenous KLF15 would also regulate HBV gene expression, we used siRNA to reduce the expression of endogenous KLF15. As shown in Fig. 4A, the transfection of KLF15 siRNA into Huh7 cells resulted in an approximately 70% reduction of the KLF15 messenger RNA (mRNA) level. This reduction of KLF15 expression led to an approximately 50% reduction in HBsAg expression from the HBV genome (Fig. 4B). Consistent with this result, KLF15 siRNA also reduced the luciferase activities of the HBV core promoter and the surface promoter by approximately 50% and 30%, respectively (Fig. 4C and 4D). Thus, the results shown in Fig. 4 indicated that endogenous KLF15 also positively regulates HBV surface and core promoters.
KLF15 Binds to HBV Core and Surface Promoters in EMSAs and ChIP Assays.
To characterize the mechanism by which KLF15 binds to core and surface promoters, the FLAG-tagged KLF15 protein was expressed in 293T cells and purified with an anti-FLAG affinity gel (Fig. 5A). Although the crude cell lysates also contained Sp1 and NF-Y that are known to interact with the S promoter (Fig. 5A, lane 1), these two protein factors were found only in the unbound fraction (Fig. 5A, lane 2) and not in the affinity-purified KLF15 fraction (Fig. 5A, lane 3), indicating the specificity of this purification. Using EMSAs, we showed that rKLF15 was able to bind to labeled core promoter probe CP35 (Fig. 5B). KLF15-DNA binding was specific, as the addition of 100-fold nonlabeled CP35 disrupted the protein-DNA complex (Fig. 5B, lane 2). It has been very well demonstrated that a functional KLF15 binding site is present in the CLCK1 gene promoter.25 As shown in the same panel, this KLF15-DNA complex could also be removed by the unlabeled CLCK1 oligonucleotide (lane 3), but not by a nonspecific oligonucleotide MLTF (lane 4), indicating the specificity of this binding (Fig. 5B). Moreover, the addition of the anti-KLF15 antibody resulted in a supershift, whose intensity positively correlated with the amount of the anti-KLF15 antibody used (lanes 5 and 6). It is noteworthy that the addition of the control antibody could increase the binding of KLF15 to the DNA probe. In addition, despite the appearance of the supershifted signal, the intensity of the original KLF15-DNA complex did not diminish accordingly, which was also observed in another study using the same antibody.24 The reason why the addition of the antibodies increased the binding of KLF15 to the DNA probe is unclear. It may have been related to stabilization by protein (antibody or bovine serum albumin [BSA] in the antibody storage buffer) or other components in the antibody storage buffer. To determine whether KLF15 binding to CP35 would be specific, we synthesized CP35-2m, which had mutations in the two potential KLF15-binding sites (Supporting Fig. 1). As shown in Fig. 5C, KLF15-DNA complex was decreased by the nonlabeled CP35 competitor in a dose-dependent manner, whereas CP35-2m failed to compete for KLF15 binding (Fig. 5C). To further confirm that the binding of KLF15 to DNA would depend on the KLF15 consensus sequence embedded in the core promoter, we performed ChIP assays using cells cotransfected with pKLF15 and the core promoter reporter, pCP, or its mutant, pCP-2m. Our results showed that the anti-KLF15 antibody could efficiently precipitate pCP, but not pCP-2m (Fig. 5F, upper panel), indicating that KLF15 could, indeed, bind to pCP, and this binding was dependent on the intact KLF15 consensus sequence.
To determine whether KLF15 could also bind to the surface promoter, we performed similar EMSA assays. As shown in Fig. 5D, the incubation of rKLF15 with the labeled surface promoter probe, SP70, resulted in a bandshift, which could be competed off by nonlabeled SP70, but not by a nonspecific competitor. Similarly, a supershift band could be observed when the anti-KLF15 antibody was added in the binding reaction (Fig. 5E). Consistent with these results, ChIP assays showed that KLF15 was able to bind to the S promoter DNA. Further analysis indicated that mutations in the Sp1 sites (i.e., Z1/Z2 mutant; Fig. 5F, middle panel), which prevented the binding of Sp1, reduced the binding of KLF15 to the S promoter by 42% (Supporting Fig. 2). In contrast, mutations in the NF-Y site (i.e., M2 mutant; Fig. 5F, bottom panel) had essentially no effect on the binding of KLF15 to the S promoter, despite suppressing NF-Y binding. Together, the results in Fig. 5 indicated that KLF15 could bind to the HBV core and surface promoters and further suggested the partial overlap of the KLF15 sites with the Sp1 sites or the presence of a cryptic KLF15-binding site elsewhere in the S promoter.
Suppression of KLF15 expression in the liver decreases viral gene expression and DNA replication in a HBV mouse model. To determine whether KLF15 would also regulate HBV gene expression in vivo, we performed a hydrodynamic injection to introduce the HBV genome into the mouse liver and used RNAi to suppress KLF15 expression. First, we validated the silencing effect of KLF15-specific short-hairpin RNA (shRNA), which was expressed using the Invitrogen BLOCK-it miR RNAi vector that also contains an embedded EmGFP cassette, in Huh7 cells. As shown in Fig. 6A, four independent KLF15 miR RNAi constructs (Cons1-4) could each reduce the KLF15 mRNA level by approximately 60% at 48 hours after transfection. Next, we introduced the KLF15 RNAi construct 4 with pAAV-HBV1.2 into the mouse liver using the hydrodynamic injection. Because the transfection efficiency of this injection procedure ranges from 10% to 40%, transfected (i.e., GFP-positive) and nontransfected (i.e., GFP-negative) hepatocytes were separated by cell sorting after liver perfusion. Similar to Huh7 cells, the KLF15 mRNA level in KLF15 RNAi construct-transfected hepatocytes was reduced by approximately 60% when it was compared with the nontransfected hepatocytes. Such a reduction was not observed if the KLF15 RNAi construct was replaced with the control RNAi construct (Fig. 6B). These results indicated that the KLF15 RNAi construct could also reduce the expression of KLF15 in mouse hepatocytes. To determine the effect of KLF15 knockdown on HBV gene expression, we performed immunofluorescence staining on mouse liver tissue sections. In mice coinjected with the control RNAi construct and pAAV-HBV1.2 (Fig. 6C), almost all the cells positive for GFP were also positive for HBcAg, indicating the successful cotransfection of the same hepatocytes by the RNAi construct and the HBV genome. However, when the control RNAi construct was replaced by the KLF15 RNAi construct, the HBcAg signal in GFP-positive cells was greatly diminished. This immunofluorescence staining result was further confirmed by Western blot analysis of the core protein. As shown in Fig. 6D, the liver of mice injected with the KLF15 RNAi construct had a lower level of the core protein than the liver of mice injected with the control RNAi construct. These results indicated that the knockdown of KLF15 expression could result in the suppression of core protein expression.
After the codelivery of pAAV-HBV1.2 and RNAi constructs (KLF15 construct 4 or the control vector) into the mouse liver through hydrodynamic injection, we monitored the level of HBsAg in the serum of injected mice. pAAV-HBV1.2 contains the 1.2-mer HBV genome in an AAV vector. This construct was previously shown to lead to a high replication level of HBV in the mouse liver.33 Our results indicated that mice with a KLF15 knockdown had consistently lower levels of HBsAg than control mice (Fig. 7A and B). This reduction of the HBsAg level was more dramatic with 50 than with 30 μg of KLF15 RNAi construct (data not shown), indicating a dose-dependent effect of the KLF15 RNAi construct on HBsAg expression in mice. We also examined the effect of KLF15 knockdown on the replicated HBV DNA level in the mouse serum. KLF15 knockdown also reduced the HBV DNA level in the serum (Fig. 7C). Similar to HBsAg profiles, this reduction effect was more prominent with 50 than with 30 μg of KLF15 RNAi construct.
Mutations in the KLF15 Binding Site in HBV Core Promoter Reduced Viral DNA Production in Mice.
To further confirm the effect of KLF15 on HBV replication, we generated an HBV genome with the CPm2 mutations that abolished the stimulatory effect of KLF15 on the core promoter (Fig. 2D). The replication efficiency of this HBV mutant plasmid in mice was then compared with that of the wild-type plasmid by hydrodynamic injection. As shown in Fig. 8, mice injected with the mutant genome had significantly lower levels of viral DNA in the sera than those injected with the wild-type genome (Mann-Whitney U = 27.0, P = 0.030, two-tailed). These results demonstrated the importance of the KLF15 response element in the core promoter in HBV replication.
In this study, we demonstrated that the transcription factor, KLF15, could activate HBV major surface and core promoters (Figs. 1 and 2). The overexpression of KLF15 in hepatoma cell lines increased, whereas the suppression of KLF15 expression with RNAi reduced, the activities of HBV surface and core promoters (Fig. 4). Consistent with these results, EMSAs and ChIP assays showed that KLF15 could bind to core and surface promoters (Fig. 5). The role of KLF15 in HBV gene expression was also confirmed in vivo using a mouse model, as we demonstrated that RNAi knockdown of KLF15 expression in the mouse liver could lead to a significant reduction in the expression of HBV core protein and HBsAg (Figs. 6 and 7), as well as HBV DNA copy number in mouse sera (Fig. 8). Therefore, KLF15 is important for modulating HBV gene expression and viral load.
By performing mutagenesis studies, we demonstrated that mutations in the two Sp1-binding sites in the surface promoter (i.e., the Z1/Z2 mutant) could reduce the transactivation effect of KLF15 on this promoter (Fig. 1C). This observation is consistent with our ChIP assay results, which showed that these mutations reduced the binding of KLF15 to the surface promoter (Fig. 5F). Because the mutations in the Sp1 sites reduced, but did not abolish, the binding of KLF15 to the surface promoter, it is likely that the KLF15 binding sites partially overlap with the Sp1 sites. The possibility that there are cryptic KLF15 sites elsewhere in the surface promoter cannot be ruled out, at present. Interestingly, however, results from the ChIP assays showed that mutating the CCAAT site did not affect KLF15 binding to the surface promoter (Fig. 5), but yet it abolished the effect of KLF15 on this promoter (Fig. 1C). It is conceivable that KLF15 needs to cooperatively interact with NF-Y, which binds to CCAAT,12 to exert its effect on the S promoter. Using similar approaches, we also found that mutations in the consensus KLF15 sequence in the core promoter could abolish the effects of KLF15 on the core promoter (Fig. 2C and D). This KLF15 binding site in the core promoters was initially thought to be the Sp1-binding site.34 However, as we demonstrated by EMSA and ChIP assays (Fig. 5), it is also recognized by KLF15. It is possible that KLF15 and Sp1 work synergistically to modulate gene transcription as has been documented.24 Finally, mutations in the putative KLF15-binding site in the core promoter reduced HBV DNA copy numbers in mouse sera, indicating the importance of this KLF15 site in HBV gene expression and replication (Fig. 8).
KLF factors regulate various important cellular functions, including differentiation, apoptosis, cell proliferation, and metabolism.19 KLF15 activates the expression of genes involved in glucose metabolism and adipogenesis, including the insulin-sensitive glucose transporter, GLUT4, and peroxisome proliferator-activated receptor gamma.22, 35 It is expressed in multiple tissues, including the liver.25 Hepatic expression of KLF15 is increased upon fasting and decreased upon feeding.36 Interestingly, Shaul et al. have shown, in a mouse model, that food deprivation induces the expression of HBV genes, which is reversible upon refeeding.37 Perhaps, part of the HBV activation observed by Shaul et al. is attributable to the fasting-induced activation of KLF15. KLF15−/− mice are viable and show hypoglycemia only upon fasting.23 Therefore, inhibition of KLF15 should be amenable as a potential HBV therapeutic modality.
We thank Drs. P.J. Chen, Y. Shaul, and S. Gray for plasmids. This article is dedicated to Dr. T.S. Benedict Yen, who was an inspiring mentor.