Transcription Repression of Human Hepatitis B Virus Genes by Negative Regulatory Element-binding Protein/SON*

A negative regulatory element (NRE) is located immediately upstream of the upstream regulatory sequence of core promoter and second enhancer of human hepatitis B virus (HBV). NRE represses the transcription activation function of the upstream regulatory sequence of core promoter and the second enhancer. In this study, we described the cloning and characterization of an NRE-binding protein (NREBP) through expression cloning. NREBP cDNA is 8266 nucleotides in size and encodes a protein of 2386 amino acids with a predicted molecular mass of 262 kDa. Three previously described cDNAs, DBP-5, SONB, and SONA, are partial sequence and/or alternatively spliced forms of NREBP. The genomic locus of theNREBP/SON gene is composed of 13 exons and 12 introns. The endogenous NREBP protein is localized in the nucleus of human hepatoma HuH-7 cells. Antibody against NREBP protein can specifically block the NRE binding activity present in fractionated nuclear extracts in gel shifting assays, indicating that NREBP is the endogenous nuclear protein that binds to NRE sequence. By polymerase chain reaction-assisted binding site selection assay, we determined that the consensus sequence for NREBP binding is GA(G/T)AN(C/G)(A/G)CC. Overexpression of NREBP enhances the repression of the HBV core promoter activity via NRE. Overexpression of NREBP can also repress the transcription of HBV genes and the production of HBV virions in a transient transfection system that mimics the viral infection in vivo.

From the Institute of Microbiology and Immunology, School of Life Science, National Yang-Ming University, Shih-Pai, Taipei 112, Taiwan, Republic of China A negative regulatory element (NRE) is located immediately upstream of the upstream regulatory sequence of core promoter and second enhancer of human hepatitis B virus (HBV). NRE represses the transcription activation function of the upstream regulatory sequence of core promoter and the second enhancer. In this study, we described the cloning and characterization of an NRE-binding protein (NREBP) through expression cloning. NREBP cDNA is 8266 nucleotides in size and encodes a protein of 2386 amino acids with a predicted molecular mass of 262 kDa. Three previously described cDNAs, DBP-5, SONB, and SONA, are partial sequence and/or alternatively spliced forms of NREBP. The genomic locus of the NREBP/SON gene is composed of 13 exons and 12 introns. The endogenous NREBP protein is localized in the nucleus of human hepatoma HuH-7 cells. Antibody against NREBP protein can specifically block the NRE binding activity present in fractionated nuclear extracts in gel shifting assays, indicating that NREBP is the endogenous nuclear protein that binds to NRE sequence. By polymerase chain reaction-assisted binding site selection assay, we determined that the consensus sequence for NREBP binding is GA(G/T)AN(C/ G)(A/G)CC. Overexpression of NREBP enhances the repression of the HBV core promoter activity via NRE.

Overexpression of NREBP can also repress the transcription of HBV genes and the production of HBV virions in a transient transfection system that mimics the viral infection in vivo.
Infection of hepatitis B virus (HBV) 1 causes acute and chronic hepatitis and is closely associated with the development of cirrhosis and hepatocellular carcinoma. HBV is a small enveloped DNA virus with a partially double-stranded 3.2-kb genome. The genome contains four partially overlapping open reading frames (ORFs) coding for the surface, core, polymerase, and X proteins. The transcription of these open reading frames is under the control of four promoters (two surface promoters, one core promoter, and one X promoter) and two enhancers (enhancer I and enhancer II). Core promoter produces two 3.5-kb RNAs: the precore and pregenomic RNAs. Precore RNA encodes precore protein and e antigen. Pregenomic RNA not only serves as the mRNA that encodes core and polymerase proteins but also can be packaged into nucleocapsids along with viral polymerase, serving as the template for reverse transcription. Regulated expression of pregenomic RNA plays a pivotal role in the control of the viral replication cycle. The core promoter can be divided into two elements: the basal core promoter (BCP) and the core upstream regulatory sequence (CURS). CURS can activate the adjacent downstream BCP activity in cis. Interestingly, the CURS is also colocalized with the second enhancer (ENII) in the HBV genome (1). The ENII can activate the surface and X promoters in a position-and orientation-independent manner (2). The CURS/ENII displays a differentiated liver cell specificity (3), which is the combined effect of several liver-enriched transcription factors, such as CCAAT/enhancer-binding protein (4 -6), FTF (7)(8)(9), HNF4 (10,11), HNF3 (12,13), and HNF1 (14,15).
We have previously identified a negative regulatory element (NRE) located upstream of CURS/ENII. NRE can effectively abolish the transcription stimulatory function of both CURS and ENII from a nearby upstream position (16). The minimal essential sequence required for the NRE function has been mapped (16,17). A trans-acting factor present in the nuclear extracts derived from a human hepatoma cell line can specifically bind to this region (16).
In this paper, we describe the cloning and characterization of an NRE binding protein, NREBP. NREBP bears strong homology to previously described cDNAs: DBP-5, SONA, and SONB. Antiserum raised against NREBP recombinant protein can specifically abolish the nuclear extracts to form an NRE-protein complex in gel shifting assays. Recombinant NREBP can specifically interact with wild type but not mutated NRE sequence. PCR-assisted binding site selection assay reveals that the optimal sequence for binding to NREBP is a perfect match for the NRE sequence. Overexpression of NREBP enhances the transcription repression of core promoter mediated by NRE. Overexpression of NREBP also represses the viral replication and gene expression in a HBV replication system.

EXPERIMENTAL PROCEDURES
Isolation of cDNA Clones-A ZAPII cDNA library was made from messenger RNA isolated from human hepatoma HepG2 cells according to the manufacturer's recommendation. This library was screened with concatemerized double-stranded synthetic oligonucleotides of wild type NRE sequence. The oligonucleotides we used contained the wild type NRE sequence, gatctGAGACCACCGTGAACGCCCATCAGg, and its complement, which corresponded to the sequence from nucleotide 1613 to 1636 of the HBV genome (uppercase letters). A BglII and a BamHI overhang (lowercase letters) were introduced at the 5Ј-and 3Ј-end of this oligonucleotide.
Plasmids-The HBV sequence used in the study is of the adw subtype. Numbering of the HBV sequence begins at the unique EcoRI site, which is nt 1. Plasmids pSV2CAT, pNRE-CP-CAT, pCP-CAT, pBCP-CAT, and pHBV3.6 were described previously (1,16). The recombinant GST-BP15 and S.Tag-BP15 expression plasmids, pGST-BP15 and pET-BP15, were generated by cloning of a 1.6-kb EcoRI fragment containing the BP15 cDNA into the EcoRI site of pGEX-3X vector (Amersham Pharmacia Biotech) and pET29b vector (Novagen), respectively. The plasmid pCMV-f:BP15 was generated by cloning of the same 1.6-kb EcoRI fragment into the EcoRI site downstream of a CMV immediate early promoter of the pFLAG-CMV2 expression vector (Eastman Kodak Co.). The plasmid pCMV-f:A1 was generated by cloning of a 5-kb BamHI-XhoI fragment containing the A1 cDNA into BglII and SalI sites downstream of a CMV immediate early promoter of the pFLAG-CMV2 vector. All of these constructs were verified by DNA sequencing using appropriate sequencing primers.
Affinity Purification of Anti-BP15 Polyclonal Antibody-Expression of recombinant S.Tag-BP15 protein tagged with S.Tag was induced by isopropyl-1-thio-␤-D-galactopyranoside and then purified by S-proteinagarose chromatography according to the manufacturer's recommendation. Purified recombinant S.Tag-BP15 protein was separated on 10% SDS-PAGE and transferred to nitrocellulose membrane (Amersham Pharmacia Biotech). To identify the recombinant S.Tag-BP15 protein, a side strip of membrane was cut out and then blocked in TBST buffer (100 mM Tris-HCl, pH 8, 1.5 M NaCl, and 0.1% Tween 20) containing 1% gelatin at room temperature for 30 min. This membrane strip was then incubated with S-protein alkaline phosphatase conjugate at a 1:2500 dilution at room temperature for 30 min, washed four times with TBST buffer for 4 min at room temperature, and developed with developer solution. The part of the nitrocellulose membrane containing the recombinant S.Tag-BP15 protein was carved out and incubated in renaturation buffer containing 5% nonfat milk at 4°C overnight. This membrane strip was then washed with PBS three times for 10 min each. To affinity-purify the rabbit antiserum, crude antiserum was incubated with renatured membrane strip at room temperature for 3.5 h. The membrane strip was washed at room temperature once with phosphatebuffered saline containing 1% Tween 20 for 10 min and then washed twice with phosphate-buffered saline for 10 min. Bound antibody was eluted in 0.2 M glycine, pH 2.8, at room temperature for 2 min. An equal volume of 10% bovine serum albumin was added to the eluent, which was then dialyzed with phosphate-buffered saline overnight.
Western Blotting and Immunofluorescence-For Western blot, differentiated human hepatoma HuH-7 and human embryonic kidney 293T cells, respectively, were transiently transfected with pFLAG-CMV2, pCMV-f:BP15, and pCMV-f:A1. Total cell lysates were collected for Western blotting with purified anti-BP15 antibody as previously described (19). Similarly, 25 g of proteins of HuH-7 crude nuclear extracts and 15 g of proteins from 0.4 M NaCl step-eluted, fractionated nuclear extracts were used for Western blotting with purified anti-BP15 antibody. Proteins were typically separated on 8% SDS-PAGE. HuH-7 cells cultured on slides were transiently transfected with pFLAG-CMV2, pCMV-f:BP15, and pCMV-f:A1 plasmids, respectively. Immunofluorescence was performed as previously described (19). The endogenous NREBP protein was detected by purified anti-BP15 antibody, while overexpressed f:BP15 and f:A1 proteins were detected by anti-FLAG antibody.
Preparation of Crude and Fractionated Nuclear Extracts and Gel Shifting Assay-Preparation of crude and fractionated nuclear extracts from cultured cells was performed as previously described (1). Gel shifting analysis was performed as previously described (6,16). For blocking experiments, an appropriate amount of anti-GST or anti-GST-BP15 antisera, respectively, was preincubated with the 0.4 M NaCl step-eluted nuclear extracts on ice for 25 min before the addition of labeled probe.
Southwestern Analysis-The probe we used was concatemerized double-stranded oligonucleotides containing the wild type or mutant NRE (5Ј-gatctTCTCAACAAGTGAACGCCCATCAGg-3Ј and its complement with mutation from nt 1613 to 1621) sequence of HBV. The probe was end-labeled with [␥-32 P]ATP and T4 polynucleotide kinase. Recombinant GST-BP15 protein or affinity-purified f:BP15 protein with anti-FLAG M2 affinity gel (Sigma) used for the Southwestern assay was resolved by 8% SDS-PAGE and then transferred to nitrocellulose membrane. The membrane was renatured and blocked in renaturation buffer containing 5% nonfat milk for 8 h at 4°C. The membrane was then incubated in DNA binding buffer (20 mM HEPES, pH 7.9, 100 mM KCl, 0.5 mM EDTA, 0.5 mM dithiothreitol, 0.25 mM MgCl 2 , 0.25% nonfat milk, and 0.1 mg/ml salmon sperm DNA) containing radiolabeled probe at 10 6 cpm/ml at room temperature for 6 h. After washing three times with the binding buffer for 30 min, the membrane was subject to autoradiography.
PCR-assisted Binding Site Selection Assay-Binding site selection assay was a modification of the assay described by Pollock and Treisman (20). The sequence of degenerate template is 5Ј-TCTGCAGTCAC-TAGCANNNNNNNNNNNNNNNNNNACTGAGCATGCATGCT-3Ј.Primers for PCR amplification are 5Ј-TCTGCAGTCACTAGCA-3Ј and 5Ј-AGCATGCATGCTCAGT-3Ј. Recombinant GST-BP15 protein was immobilized on nitrocellulose membrane, renatured in renaturation buffer containing 5% nonfat milk overnight at 4°C, and then incubated with 200 ng of concatemerized double-stranded degenerate template oligonucleotides in a DNA binding buffer at room temperature for 1 h. After wash with the binding buffer, bound DNA was eluted with elution buffer (0.5 M ammonium acetate, 5 mM EDTA, and 0.5% SDS) at room temperature for 30 min. After phenol/chloroform extraction, bound DNA was purified by ethanol precipitation and then amplified by PCR. After five rounds of binding, elution, and amplification, PCR products were cloned. Individual clones were then isolated and sequenced.
Transfection and CAT Assay-For reporter gene experiment, transfection of human hepatoma cell lines HepG2 and HuH-7 was performed with 2.7 g of reporter plasmid, 0.5 g of CH101, 0.91 g of pFLAG-CMV2, or 2 g of pCMV-f:A1 expression plasmid and carrier plasmid as previously described (19). Plasmid pCH101 contained a ␤-galactosidase gene driven by the SV40 promoter and enhancer and served as a control for normalizing transfection efficiency. Plasmid pFLAG-CMV2 was used to calculate the relative activity. Each set of experiments was performed with two different preparations of plasmids and repeated two to three times for each preparation. The CAT activity was measured by a PhosphorImager and normalized against ␤-galactosidase activity.
For HBV replication experiment, 25 g of the plasmid HBV 3.6, which contained more than a unit length of HBV viral genome (1), was transfected into HuH-7 cells in a 15-cm plate with either no plasmid, 15.9 g of pFLAG-CMV2, or 35 g of pCMV-f:A1 and carrier plasmid, respectively. At day 3 posttransfection, total RNAs were collected for Northern blotting analysis, while culture media were collected for endogenous DNA polymerase activity.
Northern Blotting Analysis-Purified inserts containing BP9, BP15, 6-1, GAPDH, and ␤-actin cDNAs were labeled with the random priming method and used as probes, respectively. Preparation of total cellular RNA from human hepatoma HepG2 and HuH-7 and Northern hybridization using BP9, BP15, and 6-1 probes were performed as previously described (19). Membranes containing ϳ2 g of poly(A) ϩ RNA from different adult human tissues (CLONTECH) were used for Northern blotting analysis with BP9, BP15, and 6-1 as probes. The same blots were reprobed with ␤-actin probe.
In the HBV replication experiment, total cellular RNA was collected at day 3 posttransfection. Forty g of total RNA of each sample was analyzed by Northern hybridization. The probe was a PstI fragment of pHBV3.6 containing the HBV sequence from nt 25 to 1989. The same blot was then reprobed with GAPDH to ensure equal loading of RNA samples.
Assay for Endogenous DNA Polymerase Activity-To assay for endogenous DNA polymerase activity, the culture supernatant containing virions and core particles was collected 3 days after transient transfection, treated with 1% Nonidet P-40 for 4 h at room temperature, and centrifuged at 17,000 ϫ g for 30 min at 4°C. The supernatant was then centrifuged at 227,000 ϫ g for 1 h at 4°C. The pellet from the second centrifugation, which contains HBV viral core particles, was resuspended in TNE buffer (10 mM Tris-HCl (pH 7.5), 50 mM NaCl, and 0.1 mM EDTA) and assayed for endogenous polymerase activity as previously described (19).

RESULTS
Cloning of NREBP cDNA-NRE is a sequence element that represses the transcription activation function of the adjacent ENII and core promoter in the HBV genome. Mutational analysis reveals the minimal essential sequence, nt 1613-1621, within NRE that is required for its function (16,17). This sequence is bound specifically by nuclear proteins derived from differentiated human hepatoma HepG2 cells (16). To search for NRE-binding protein(s), we performed expression cloning using labeled concatemers of oligonucleotides of NRE sequence to screen a cDNA library made from HepG2. Three overlapping cDNAs, BP4, BP9, and BP15, were obtained ( Fig. 1). Sequence analysis reveals a long open reading frame (ORF) shared by these cDNAs. The gene encoded by these cDNAs is referred to as NREBP. The longest BP9 cDNA is 5145 nt in size.
Northern analysis using BP9 cDNA as probe showed that NREBP transcripts had apparent sizes of 8.3 kb and greater than 10 kb (see below). To obtain full-length NREBP cDNA, we used sequence from the 3Ј-end of BP9 to screen an adult human liver cDNA library. A cDNA clone, A1, was obtained that was 5539 nt in size. A1 represented the 3Ј portion of the NREBP transcript, since it contained both translation stop codon and a poly(A) tract at its 3Ј-end. 5Ј rapid amplification of cDNA ends was performed, and an additional cDNA clone, 6-1, was obtained that represented further 5Ј extension. The combined sequence information from BP9, A1, and 6-1 reveals a NREBP cDNA that is 8266 nt in size. The translation start codon is located at nt 50 within the sequence context of GCCAUGGCG, which conforms to the Kozak consensus sequence, (A/G)C-CAUGG. The stop codon is located at nt 7208. An ORF from nt 50 to 7207 encoding a protein of 2386 amino acids is predicted. The predicted molecular mass of NREBP protein is 262 kDa. An upstream in-frame stop codon is noted at nt 29 of the 5Ј-end of the cDNA, indicating that we have cloned the entire ORF for NREBP.
NREBP protein does not appear to contain any known DNA binding motif. Interestingly, seven different repeats are pres-ent in the amino acid (aa) sequence (Fig. 1 A search of the data base reveals that the genomic region containing the NREBP gene has been completely sequenced (DDBJ accession number AP000046). The NREBP gene spans ϳ35 kb and contains 13 exons and 12 introns. The sizes of exons and introns and splicing junction sequences are shown in Table I. The combined NREBP cDNA contains exons 1-12, exon 13a, and intron 3. The NREBP gene has been mapped to chromosome 21q22.1 (21), which is a critical region for an autosomal dominant familial "aspirin-like" platelet disorder associated with development of acute myelogenous leukemia and in close proximity to the Down syndrome critical region (22)(23)(24)(25)(26)(27).
Expression Pattern of NREBP Transcript-To examine whether NREBP was expressed in human hepatoma cell lines, total RNAs isolated from human hepatoma HuH-7 and HepG2 cells were used for Northern hybridization with BP9 ( Fig. 2A), BP15, and 6 -1 (data not shown) as probes, respectively. A 8.3-kb NREBP transcript was detected. Furthermore, poly(A) ϩ RNAs from a variety of human tissues were subject to Northern analysis with BP9 (Fig. 2B) and BP15 (data not shown) as probes, respectively. Two transcripts were found, one at 8.3 kb and the other at greater than 10 kb. The 8.3-kb transcript was the major one. NREBP transcripts could be detected in all tissues. The highest expression was seen in leukocyte and heart, followed by lymph node, spinal cord, ovary, testis, thymus, spleen, pancreas, placenta, and brain. Weaker expression was seen in colon, small intestine, prostate, kidney, skeletal muscle, liver, and lung.
Localization and Expression of NREBP Protein-To study the localization and expression of the NREBP protein, polyclonal anti-BP15 antiserum was generated after immunizing rabbits with recombinant GST-BP15 protein. This antiserum Three cDNAs (BP9, BP4, and BP15) were obtained by expression cloning with concatemers of NRE sequence. A1 cDNA was obtained by library screening with a probe derived from the 3Ј portion of the BP9 sequence. 6-1 cDNA was obtained by 5Ј rapid amplification of cDNA ends. These cDNAs assemble into NREBP cDNA totaling 8266 nt in length. NREBP cDNA encodes a protein of 2386 amino acids with the translation start codon and stop codon at nt 50 and 7208, respectively. Seven different repeats of this protein are shown on the NREBP ORF. The relationship of each cDNA to the combined NREBP cDNA sequence is shown with the position of nucleotides on the top and amino acids on the bottom. was then affinity-purified using a recombinant S.Tag-BP15 protein. The localization of the endogenous NREBP protein in HuH-7 cells was examined by immunofluorescence using purified anti-BP15 antibody. As shown in the top panel of Fig. 3, the endogenous NREBP protein was localized in the nucleus. Two truncated forms of NREBP protein (f:BP15 and f:A1), tagged with a FLAG epitope at their N terminus, were overexpressed in HuH-7 cells by transient transfection. Transfection with an insertless, parental vector, pFLAG-CMV2, was performed, which serves as negative control. Both f:BP15 and f:A1 proteins in transfected HuH-7 cells were localized in the nucleus as revealed by immunofluorescence using anti-FLAG antibody (Fig. 3, middle and bottom panels). HuH-7 cells transfected with a parental vector did not produce any signal (data not shown).
To identify the endogenous NREBP protein by Western blotting, crude nuclear extracts from HuH-7 cells were loaded onto a heparin-Sepharose column and step-eluted with NaCl at increasing concentrations. Each fraction was then collected and tested for binding to the NRE sequence. As previously observed, the NRE-binding protein was present in the 0.4 M fraction (16). The DNA-protein complex appeared to be specific for NRE, since it could be competitively abolished by wild type NRE sequence but not by mutant NRE or random sequence (16). When crude and 0.4 M NaCl fraction of fractionated nuclear extracts were examined by Western blotting with purified anti-BP15 antibody, several proteins were detected. The apparent molecular mass of the largest one was about 257 kDa, which was consistent with the expected molecular mass of protein encoded by the full-length NREBP, 262 kDa (Fig. 4). Proteins of smaller sizes may represent processed or degraded products, translation products from alternatively spliced transcripts, or products derived from internal initiation. No signal was seen with the preimmune serum (data not shown).
The expression of f:BP15 and f:A1 proteins was examined by Western blotting using purified anti-BP15 antibody. The expression of f:BP15 in HuH-7 transient transfectants could be detected, although the apparent molecular mass was larger than expected (138 and 130 kDa instead of 63.5 kDa; Fig. 5A). The same result was obtained in 293T cells after transient transfection (Fig. 5B). The expression of f:A1 in HuH-7 transient transfectants could not be detected by Western blotting, because the level of f:A1 expression was significantly lower than that of f:BP15 (Fig. 5A). However, the expression of f:A1 in 293T transient transfectants could be detected, although the apparent molecular mass of f:A1 was 217 and 197 kDa, which is larger than the expected molecular mass of 165.5 kDa (Fig.  5B). That the apparent molecular mass of f:BP15 and f:A1 was larger than expected is probably due to the high content of proline residues in NREBP. HuH-7 and 293T cells transfected with an insertless parental control vector did not produce any signal by Western blotting analysis.
NREBP Is Indeed the Endogenous NRE-binding Protein-To  test whether NREBP is the endogenous protein that binds to the NRE sequence, we performed gel shifting assays. When the 0.4 M fraction of fractionated nuclear extracts was preincubated with anti-BP15 antiserum, the formation of DNA-protein complex was abolished (Fig. 6, lane 4). Control anti-GST antiserum did not have any effect (lane 3). This result indicates that NREBP is indeed the endogenous protein that binds to NRE sequence in 0.4 M fractionated nuclear extracts.
Binding Specificity of NREBP Protein-To test whether NREBP protein can bind to the NRE sequence specifically, the f:BP15 protein produced by the transiently transfected HuH-7 cells was purified by affinity column chromatography coated with anti-FLAG antibody and then transferred to nitrocellulose membranes after SDS-PAGE. After denaturation and renaturation, Southwestern binding assays were performed using radioactively labeled wild type or mutant NRE sequence as probe. The amount of f:BP15 protein produced was verified with Western blotting (data not shown). As shown in Fig. 7, proteins derived from f:BP15 transfectants could interact with wild type NRE sequence but not mutant NRE sequence. Proteins derived from cells transfected with a parental control vector could not bind either wild type or mutant NRE sequence.
This study shows that the middle portion of NREBP, as represented by BP15, can interact with wild type NRE sequence specifically. Similar binding results were obtained using bacterially derived recombinant protein, GST-BP15, by Southwestern binding assays (data not shown). However, the binding of this GST-BP15 protein to the NRE sequence could not be detected in gel shifting assays (data not shown).
Recombinant GST-BP15 protein was subsequently used in a PCR-assisted binding site selection assay to examine the binding specificity of recombinant NREBP protein. Recombinant GST-BP15 protein was first immobilized on nitrocellulose filters and incubated in a DNA binding reaction with a pool of oligonucleotides. These oligonucleotides had random sequences in the middle flanked by primer binding sites suitable for PCR amplification. The binding experiment was performed in a condition identical to that of the Southwestern assay. After extensive wash, bound oligonucleotides were eluted and amplified by PCR. Amplified products were subject to binding in reiteration. After five rounds of binding, elution and amplification, the PCR products were cloned, and multiple individual clones containing the PCR products were sequenced. Fig. 8 shows the DNA sequences of the 38 independent clones that were obtained from the above procedure. They are listed in decreasing abundance, which may parallel their binding affinity toward NREBP. From this study, the consensus sequence for optimal NREBP binding is determined to be GA(G/T)AN(C/G)(A/G)CC. This sequence is a perfect match for the NRE sequence in the HBV genome, GAGACCACC.
NREBP Mediates Transcription Repression by NRE-We have previously shown that the core promoter of HBV can be functionally dissected into two regions: a basal core promoter, BCP, and a core promoter upstream regulatory sequence, CURS. CURS can activate the BCP activity. This stimulatory activity can, however, be partially abolished by the nearby NRE sequence from an upstream position (16). To test the ability of NREBP to mediate this transcription repression effect, we performed co-transfection experiments with an NREBP expression construct and constructs containing a reporter gene driven by different core promoter elements. We chose to use f:A1 construct, since numerous attempts to engineer a full-length NREBP expression construct failed, probably because of the sequence repeats present in the ORF. An empty vector, pFLAG-CMV2, was used as a control. The reporter gene constructs we used were a CAT reporter gene driven by the BCP (pBCP-CAT), BCP plus CURS (pCP-CAT), and BCP plus CURS plus NRE (pNRE-CP-CAT). As shown in Fig. 9, the overexpression of f:A1 reduced the promoter activity of pNRE-CP-CAT to about 13.5 and 12.2% in HuH-7 and HepG2 cells, respectively. The promoter activity of neither pCP-CAT nor pBCP-CAT was significantly affected by f:A1. This result strongly suggests that NREBP can further augment the transcription repression effect of NRE on the core promoter of HBV.
NREBP Represses HBV Gene Expression and Replication-To test the effect of NREBP on the HBV gene expression and replication, we resorted to transient transfection with more than a unit length of HBV genome, pHBV3.6, into HuH-7 cells. Viral gene expression and production of mature virions that closely mimic viral infection in vivo have been seen after transfection (1). The effect of NREBP on the transcription and replication of HBV was tested by co-transfecting an NREBP expression plasmid pCMV-f:A1 or an empty vector pFLAG-CMV2 with pHBV3.6. To avoid the competition among promoters, no plasmid was co-transfected to normalize the transfec-tion efficiency. To circumvent this, the experiment was repeated four times. Three days after transfection, the amount of the 2.4-kb large surface, 2.1-kb middle and major surface, and 3.5-kb precore and pregenomic transcripts was measured by Northern hybridization (Fig. 10A, top panel). The expression of GAPDH was used as an RNA loading control (Fig. 10A,  bottom panel). Overexpression of NREBP reduced the expression of 3.5-kb RNAs to 20%, as well as 2.4-and 2.1-kb RNAs to 40%. The production of mature virions and core particles was quantified by an endogenous DNA polymerase activity assay. As shown in Fig. 10B, NREBP reduced the production of virions and core particles to 43.5%. These results strongly suggest that NREBP represses the gene expression and replication of HBV.

DISCUSSION
The NRE of HBV represses the activity of the nearby core promoter and ENII by 10 -20-fold (16,17). In this paper, we describe the cloning and characterization of a transcription factor, NREBP. NREBP is a 2386-aa protein. The middle portion of NREBP can interact with wild type NRE sequence specifically, but not with mutant NRE sequence. Furthermore, the consensus sequence for NREBP binding, GA(G/T)AN(C/ G)(A/G)CC, is a perfect match for the NRE sequence in the HBV genome, GAGACCACC. Overexpression of an NREBP protein lacking the N-terminal 892 amino acids can further enhance this repression mediated by NRE. Overexpression of this truncated NREBP protein can also repress the transcription of HBV genes and the production of HBV virions in a transient transfection system that mimics the viral infection in vivo. The abolishment of the formation of DNA-protein complex between NRE and nuclear proteins by antibody against NREBP demonstrates that NREBP is the endogenous NRE binding protein present in the cells. Taken together, NREBP appears to function as a transcription repressor at NRE.
It has been reported that the sequence of HBV from nt 1605 to 1625, named as NRE ␥, can repress the core promoter activity by 2-3-fold in HuH-7 cells (28). This sequence can be bound by a transcription factor, RFX1. The consensus sequence for RFX1 binding is from nt 1605 to 1617. Overexpression of RFX1, FIG. 8. Optimal NREBP binding sites. Optimal NREBP binding sites were identified by PCR-assisted binding site selection. Alignment of the sequences of 38 individually cloned oligonucleotides after selection is shown. HBV sequence containing the NRE site is shown at the top. The essential sequence of NRE is underlined. Nucleotides shared with the HBV sequence at a given position are shown in capital letters. F and R represent the binding sites for forward and reverse primers used in PCR amplification, respectively. The numbers of clones bearing each individual sequence are shown on the right. The consensus sequence for NREBP binding is shown at the bottom.
FIG . 9. Activation of NRE function by overexpression of NREBP: a reporter gene assay. HuH-7 and HepG2 cells were transfected with pNRE-CP-CAT, pCP-CAT, or pBCP-CAT in the presence of a pCH101 and a pCMV-f:A1 or a pFLAG-CMV2 vector. This diagram shows the relative activity of core promoter achieved by overexpression of f:A1 for different core promoter constructs. The relative activity (percentage) was calculated by taking the normalized CAT activity of the transient transfectants co-transfected with pFLAG-CMV2 vector as 100%.
typically a transcriptional activator, can activate transcription in an NRE ␥-dependent manner. The authors speculated that the formation of a heterodimeric RFX1-MIBP1 complex, which had been shown to possess transcription repression activity in other promoters, might mediate the transcription repression function of NRE ␥ (29). The essential sequence of NRE identified by our group, which is from nt 1613 to 1621, is distinct from NRE ␥ (16,17). NRE represses the core promoter activity by 10 -20-fold, and the sequence from nt 1606 to 1612 is not required for its negative regulatory function (16). NREBP, moreover, can bind to a smaller region within the NRE ␥ sequence. Gel shifting assays show that the NRE binding activity from 0.4 M NaCl step-eluted nuclear extracts can be abolished specifically by anti-NREBP antiserum. This result indicates that the NRE binding activity from the 0.4 M NaCl step-eluted nuclear extracts is NREBP.
Examination of the predicted protein structure of NREBP reveals no known DNA binding motif. Several repeats are noted to be present in NREBP. Repeat 2 contains a leucine zipper-like structure. This raises an intriguing possibility that NREBP may interact with other leucine zipper protein(s). Alternatively, NREBP may form homodimers. Leucine zipper interaction may not be essential for the transcription repression function of NREBP, since overexpression of A1 protein, which is devoid of the leucine zipper domain, can still repress transcription mediated through NRE.
Interestingly, the NREBP bears strong homology to previously described cDNAs: DBP-5, SONA, and SONB (EMBL accession numbers X63071, X63753, and X63751, respectively) (Fig. 11). DBP-5 cDNA, 4972 nt in size, was obtained through expression cloning from a human B cell cDNA library using a segment of the HLA-DR gene promoter as the probe (30). The sequence of DBP-5, from nt 1 to 4967, is almost identical (99.6%) to the NREBP cDNA sequence from nt 3290 to 8257 (Fig. 11A). The sequences of NREBP and DBP-5 were colinear except for the deletion of one single nucleotide after nt 3585 in DBP-5 (corresponding to nt 6875 in NREBP), which leads to a frameshift in the predicted ORF. The sequences of the corresponding genomic region and SONA cDNA (see below) do not carry this deletion. This discrepancy, therefore, is probably due to a sequencing error in DBP-5. Furthermore, the initially assigned start codon of DBP-5 (30) is probably only an internal ATG, since DBP-5 appears to be a 5Ј-truncated cDNA. DBP-5 cDNA therefore appears to encode a truncated 1306-aa protein instead of the originally assigned 1179 amino acids. The 1306-aa DBP-5 protein corresponds to aa 1081-2386 of NREBP (Fig. 11B).
SONA and SONB cDNAs were isolated by hybridization screening of a human embryonic cDNA library probed with the rat gene K51, which was obtained initially based upon its cross-hybridization to the v-mos gene under a nonstringent condition (31)(32)(33)(34). SONB appears to be a partial cDNA corresponding to the middle portion of NREBP cDNA. The sequence of SONB from nt 1 to 3373 is colinear and nearly identical (99.7%, not including the 240-nt insertion) to that of NREBP from nt 2174 to 5303 (Fig. 11A); however, there is a 240-nt insertion (from nt 112 to 351 of SONB) in SONB. The relative position of this insertion site is located between nt 2284 and 2285 (within intron 3) of the NREBP cDNA. The reported genomic sequence has a 120-nt insertion instead. This addition results in an increase in the number of repeats in the repeat 1 region; there are 13 repeats in NREBP and 21 repeats in the SONB sequence, while the reported genomic sequence has 17 repeats. SONB cDNA encodes a 1124-aa protein corresponding to the middle portion of NREBP from aa 709 to 1751 (Fig. 11B).
The sequence of SONA from nt 7 to 5676 is partially identical (99.7%, not including intron 3 and the 53-nt deletion in exon 13) to the sequence of NREBP from nt 1708 to 8266 (Fig. 11A). SONA, however, lacks the sequences corresponding to the segments from nt 2108 to 2945 (intron 3) and from 7151 to 7203 (5Ј-terminal 53 nt of exon 13a) in NREBP (Fig. 11A). Compared with the NREBP and genomic sequences, SONA appears to be a product of alternative splicing (Table I). The SONA cDNA starts from within exon 3, continues through exons 4 -12 and ends in exon 13b. Exon 13b lacks the 5Ј-terminal 53 nt of exon 13a. Because of the absence of the third intron, there is a shift in the ORF. It is not clear if SONA represents a full-length or nearly full-length product of this alternatively spliced transcript. If SONA uses the same translation start codon as NREBP, SONA will yield a C-terminal truncated product as a stop codon (at nt 416 of SONA) is introduced from frameshift because of the splicing out of the third intron. If SONA uses a downstream translation start codon at nt 415 or a more distal location for internal initiation instead, it will yield an N-terminal truncated product of NREBP (Fig. 11B). If the latter is the case, the C-terminal portion of SONA protein will also be different from that of NREBP protein, since the omission of the 53 nt in exon 13 (corresponding to nt 7151-7203 of NREBP) FIG. 10. Activation of NRE function by overexpression of NREBP: HBV gene expression and replication. The plasmid pHBV3.6, which contains more than a unit length of HBV viral genome, was transfected into HuH-7 cells with no plasmid, pFLAG-CMV2 (Flag), or pCMV-f:A1 (f:A1), respectively. A, Northern blot analysis of HBV transcripts. The intracellular total RNAs were collected at day 3 posttransfection and analyzed by Northern hybridization with the HBV DNA from nt 25 to 1989 as the probe (top panel). The same blot was reprobed with GAPDH (G3PDH; bottom panel) to ensure equal loading of RNA samples. B, production of virions. Media from the transfectants were collected 3 days after transfection to assay for the production of HBV virions and core particles. The amount of virions and core particles produced was quantified by the endogenous DNA polymerase activity assay. L and NC represent linear and nicked circular forms, respectively, of HBV DNA.
generates an additional frameshift of the ORF (Fig. 11B). In either case, SONA does not contain repeat 1 or 2. SONA may therefore encode a protein functionally distinct from NREBP. The sequence at the extreme 5Ј-end of SONA cDNA (nt 1-6) is not found in either the cDNA or genomic sequence of NREBP. Examination of the sequence from the corresponding region in the genomic locus of NREBP shows that the sequence difference is not likely to be the result of alternative splicing. The possibility that this sequence is introduced as a cloning artifact remains to be ruled out.
The sequence of the region at chromosome 21q22.1 containing the NREBP/SON gene is completely known. Based on our results, the NREBP/SON gene spans ϳ35 kb and contains 13 exons and 12 introns. Interestingly, comparison of the cDNA and genomic sequence of NREBP reveals the insertion of a segment of 120 nt in the genomic locus that is located between nt 2284 and 2285 of the NREBP cDNA sequence. SONB contains a 240-nt insertion at this location compared with the cDNA sequence of NREBP. These extra sequences represent an increase in the number of repeats in the repeat 1 region of NREBP. The protein encoded by our NREBP cDNA, the sequenced genomic region of NREBP, and the SONB cDNA therefore will have 13, 17, and 21 repeats in the repeat 1 region, respectively. The increase in the number of repeats is not the result of alternative splicing and most likely represents genetic polymorphism. The effect of the number of repeats on the transcriptional regulation function of NREBP requires further study. Some other minor sequence difference and/or poly-morphism among NREBP, DBP-5, SONB, and SONA cDNAs and the NREBP/SON gene are not discussed here.
Although the mechanism of transcription activation has been well described, relatively little is known about how the transcription repressor works. Transcription repressors can be passive or active repressors (35)(36)(37)(38). Passive repressors can counteract the activator function by the following mechanisms: (a) direct competition of the same DNA-binding sites (39 -45); (b) interference of overlapping or neighboring activator-binding sites (46 -51); (c) prevention of the translocation of the transcription activators (52); (d) titration away of limiting protein factors required for transcription activator function (53,54); (e) modification of the DNA-binding property of the transcription activators (55); (f) blocking of the DNA binding activity of the transcription activators through protein-protein interactions (50,56,57); or (g) masking or alteration of the function of the activation domain of transcription activators (58,59). Active repressors can directly inhibit the assembly or the activity of the general transcription machinery (60 -66). Another mode of transcription repression is through the recruitment of histone deacetylase complexes by repressors or corepressors. Histone deacetylase complexes will alter the chromatin structure through increased histone deacetylation (reviewed in Refs. [67][68][69]. We have previously shown that NRE can repress the transcription activation function of the ENII (16,17). NRE is located 25 nt upstream of the ENII, and the repression function of NRE depends on its close proximity to the ENII. The spatial requirement of the function of NRE on the ENII favors the FIG. 11. Comparison of the nucleotide and predicted amino acid sequences of NREBP, DBP-5, SONA, and SONB cDNAs. A, nucleotide sequence comparison. NREBP, DBP-5, SONB, and SONA cDNAs contain 8266, 4972, 3373, and 5676 nt, respectively. The sequence of DBP-5 cDNA from nt 1 to 4967 is colinear and almost completely identical to the NREBP sequence from nt 3290 to 8257. A deletion of one single nucleotide after nucleotide 3585 in DBP-5 (corresponding to nt 6875 of NREBP) is marked as an asterisk. The SONB sequence from nt 1 to 3373 is colinear and almost identical to the NREBP sequence from nt 2174 to 5303. SONB has additional 240-nt sequence, which is marked as a stippled box (from nt 112 to 351 of SONB). The position of this insertion is between nt 2284 and 2285 of NREBP. The SONA sequence from nt 7 to 406, from nt 407 to 4614, and from nt 4615 to 5676 is almost identical to the NREBP sequence from nt 1708 to 2107, from nt 2946 to 7150, and from nt 7204 to 8266, respectively. SONA lacks the NREBP sequence from nt 2108 to 2945 (intron 3) and from nt 7151 to 7203 (5Ј-terminal 53 nt of exon 13a) due to the alternative splicing. B, predicted amino acid sequence comparison. NREBP encodes a 2386-aa protein containing repeats 1-7. DBP-5 cDNA encodes a 1306-aa protein containing repeats 4 -7. SONB encodes a 1124-aa protein containing repeats 1-5. SONB protein has 21 copies of repeat 1 in contrast to the 13 copies in NREBP protein. SONA encodes a 1523-aa protein, which contains repeats 3-7 if the nt 415 AUG is taken as the translation start codon. The absence of 53 nt in exon 13 in SONA generates a second frameshift so that the C-terminal portion of the SONA protein (from aa 1401 through 1523; gray) is distinct from that of NREBP. See "Discussion" for details. notion that NREBP functions as a passive repressor. The exact repression mechanism of NREBP remains to be elucidated.