Avian and Mammalian Hepadnaviruses Have Distinct Transcription Factor Requirements for Viral Replication

Hepadnavirus replication occurs in hepatocytes in vivo and inhepatoma cell lines in cell culture. Hepatitis B virus (HBV)replication can occur in nonhepatoma cells when pregenomic RNAsynthesis from viral DNA is activated by the expression of thenuclear hormone receptors hepatocyte nuclear factor 4 (HNF4)and the retinoid X receptor ${alpha}$ (RXR ${alpha}$ ) plus peroxisome proliferator-activatedreceptor ${alpha}$ (PPAR ${alpha}$ ) heterodimer. Nuclear hormone receptor-dependentHBV replication is inhibited by hepatocyte nuclear factor 3(HNF3). In contrast, HNF3 and HNF4 support duck hepatitis Bvirus (DHBV) replication in nonhepatoma cells, whereas the RXR ${alpha}$ -PPAR ${alpha}$ heterodimer inhibits HNF4-dependent DHBV replication. HNF3 andHNF4 synergistically activate DHBV pregenomic RNA synthesisand viral replication. The conditions that support HBV or DHBVreplication in nonhepatoma cells are not able to support woodchuckhepatitis virus replication. These observations indicate thatavian and mammalian hepadnaviruses have distinct transcriptionfactor requirements for viral replication.

INTRODUCTION

Top

Introduction

Hepadnaviruses are enveloped viruses that replicate in the liverof the host (11, 24, 35). The mammalian hepadnaviruses includehepatitis B virus (HBV), chimpanzee hepatitis B virus (13),orangutan hepatitis B virus (45), wooly monkey hepatitis B virus(16), woodchuck hepatitis virus (WHV) (40), and ground squirrelhepatitis virus (22). The avian hepadnaviruses include duckhepatitis B virus (DHBV) (23), heron hepatitis B virus (37),Ross goose hepatitis B virus (3), snow goose hepatitis B virus(3), and stork hepatitis B virus (28).

The hepadnaviruses all contain a small partially double-strandedDNA genome of 3.0 to 3.3 kb in length and replicate by reversetranscription of a pregenomic RNA that contains all the geneticinformation of the virus (11, 24, 35, 38, 47). Hepadnavirusreplication is believed to be largely restricted to the liverbecause virus entry into hepatocytes is dependent on the presenceof a receptor that is predominantly expressed on this cell type.However, it is likely that additional steps in the viral lifecycle also contribute to the predominantly hepatocyte-specifictropism of hepadnaviruses.

HBV replication is restricted to hepatocytes in part becausethe liver-enriched nuclear hormone receptors hepatocyte nuclearfactor 4 (HNF4) and retinoid X receptor ${alpha}$ (RXR ${alpha}$ ) plus peroxisomeproliferator-activated receptor ${alpha}$ (PPAR ${alpha}$ ) are essential for pregenomicRNA synthesis (42). Therefore, viral transcription is a criticaldeterminant of HBV tropism. The contribution of transcriptionalregulation to the tropism of other hepadnaviruses is largelyunknown due to the lack of extensive characterization of thefactors regulating viral RNA synthesis. In WHV, HNF1 and HNF4bind to recognition sites located in the nucleocapsid promoterregion of the viral genome and probably contribute to the expressionof the pregenomic RNA and viral replication in the liver (10,44). The DHBV genome contains an enhancer element upstream fromthe pregenomic RNA initiation site (8, 18, 33). This regulatoryelement may contribute to both the level and tissue-specificexpression of pregenomic RNA synthesis and consequently to viralreplication (8, 18, 33). The presence of binding sites for theliver-enriched HNF1, HNF3, and CCAAT/enhancer binding protein(C/EBP) transcription factors may influence the tissue-specificproperties of this regulatory element (7, 17, 19).

In this study, the possible role of liver-enriched transcriptionfactors in determining the tropism of additional hepadnaviruseswas examined by determining their ability to support DHBV andWHV replication in nonhepatoma cells. As with HBV, HNF4 supportedDHBV replication in nonhepatoma cells. However, contrary tothe findings with HBV, HNF3 can also support DHBV replicationin nonhepatoma cells. HNF3 can also synergistically activateDHBV pregenomic RNA synthesis and viral replication when expressedwith HNF4. RXR ${alpha}$ -PPAR ${alpha}$ failed to support DHBV replication and inhibitedHNF4-dependent DHBV DNA synthesis. This is the opposite of theeffect of RXR ${alpha}$ -PPAR ${alpha}$ on HBV synthesis, where it was the most efficientactivator of viral replication. WHV replication was not observedunder any of the conditions that supported HBV and DHBV replication.These observations indicate that these three hepadnaviruseshave distinct transcription factor requirements for viral replication.

MATERIALS AND METHODS

Top

Materials and Methods

Plasmid construction. The steps in the cloning of the plasmid constructs used in thetransfection experiments were performed by standard techniques(32). HBV DNA sequences in these constructions were derivedfrom plasmid pCP10, which contains two copies of the HBV genome(subtype ayw) cloned into the EcoRI site of pBR322 (9). Twotandem copies of the HBV genome were cloned into the EcoRI siteof pSP65 to generate the pHBV2E construct. Two tandem copiesof the replication-competent DHBV 16 genome were cloned intothe EcoRI site of pSP65 to generate the pSPDHBV2X5.1 construct(20, 27). The pCMVDHBV construct contains the cytomegalovirus(CMV) immediate-early promoter (region -522 to -4) (1) locateddirectly upstream of the DHBV 16 sequence from nucleotide coordinates2527 to 3021 plus 1 to 3021 plus 1 to 19. In this construct,the expression of the DHBV 3.3-kb pregenomic RNA is controlledby the CMV immediate-early promoter. Two tandem copies of thereplication-competent WHV 2 genome were cloned into the HindIIIsite of pSP65 to generate the pSPWHV2X5.2 construct (15, 34).

The pMTHNF1 ${alpha}$ , pMTHNF1ß, pCMVHNF3 ${alpha}$ , pCMVHNF3ß,pCMVHNF4, pRS-hRXR ${alpha}$ and pCMVPPAR ${alpha}$ -G vectors express HNF1 ${alpha}$ , HNF1ß,HNF3 ${alpha}$ , HNF3ß, HNF4, RXR ${alpha}$ , and PPAR ${alpha}$ -G polypeptides, respectively,from the rat HNF1 ${alpha}$ , mouse HNF1ß, rat HNF3 ${alpha}$ , rat HNF3ß,rat HNF4, human RXR ${alpha}$ , and mouse PPAR ${alpha}$ -G cDNAs, respectively, usingthe mouse metallothionein I promoter (pMT), the CMV immediate-earlypromoter (pCMV), or the Rous sarcoma virus long terminal repeat(pRS) (4, 21, 26, 29-31). The PPAR ${alpha}$ -G polypeptide contains amutation in the PPAR ${alpha}$ cDNA that changes Glu²⁸² to Gly and maydecrease the affinity of the receptor for the endogenous ligand.Consequently, this mutation increases the peroxisome proliferator-dependent(i.e., clofibric acid-dependent) activation of transcriptionfrom a peroxisome proliferator response element-containing promoter(26) and was used in this study to demonstrate the peroxisomeproliferator-dependent transcriptional transactivation of thenucleocapsid promoter.

Cells and transfections. The human hepatoma HepG2 cell line and the mouse NIH 3T3 fibroblastcell line were grown in RPMI 1640 medium and 10% fetal bovineserum at 37°C in 5% CO₂-air. Transfections for viral RNAand DNA analysis were performed as previously described (25)with 10-cm plates containing approximately 10⁶ cells. DNA andRNA isolation was performed 3 days posttransfection. The transfectedDNA mixture was composed of 15 µg of pHBV2E DNA, pSPDHBV2X5.1DNA, pCMVDHBV DNA, or pSPWHV2X5.2 DNA plus 1.5 µg of theliver-enriched transcription factor expression vectors pMTHNF1 ${alpha}$ ,pMTHNF1ß, pCMVHNF3 ${alpha}$ , pCMVHNF3ß, pCMVHNF4,pRS-hRXR ${alpha}$ , and pCMVPPAR ${alpha}$ -G as required (4, 21, 26, 29-31, 42).Controls were derived from cells transfected with pHBV2E DNA,pSPDHBV2X5.1 DNA, or pSPWHV2X5.2 DNA and the pCMV expressionvector lacking a liver-enriched transcription factor cDNA insert(30). All-trans retinoic acid and clofibric acid at 1 µMand 1 mM, respectively, were used to activate the nuclear hormonereceptors RXR ${alpha}$ and PPAR ${alpha}$ (42).

Characterization of HBV, DHBV, and WHV transcripts and viral replication intermediates. Transfected cells from a single plate were divided equally andused for the preparation of total cellular RNA and viral DNAreplication intermediates as described previously (39) withminor modifications. For RNA isolation (5), the cells were lysedin 1.8 ml of 25 mM sodium citrate (pH 7.0)-4 M guanidinium isothiocyanate-0.5%(vol/vol) Sarcosyl-0.1 M 2-mercaptoethanol. After addition of0.18 ml of 2 M sodium acetate (pH 4.0), the lysate was extractedwith 1.8 ml of water-saturated phenol plus 0.36 ml of chloroform-isoamylalcohol (49:1). After centrifugation for 30 min at 3,000 rpmin a Sorval RT6000, the aqueous layer was precipitated with1.8 ml of isopropanol. The precipitate was resuspended in 0.3ml of 25 mM sodium citrate (pH 7.0)-4 M guanidinium isothiocyanate-0.5%(vol/vol) Sarcosyl-0.1 M 2-mercaptoethanol and precipitatedwith 0.6 ml of ethanol. After centrifugation for 20 min at 14,000rpm in an Eppendorf 5417C microcentrifuge, the precipitate wasresuspended in 0.3 ml of 10 mM Tris hydrochloride (pH 8.0)-5mM EDTA-0.1% (wt/vol) sodium lauryl sulfate and precipitatedwith 45 µl of 2 M sodium acetate plus 0.7 ml of ethanol.

For the isolation of viral DNA replication intermediates, thecells were lysed in 0.4 ml of 100 mM Tris hydrochloride (pH8.0)-0.2% (vol/vol) NP-40. The lysate was centrifuged for 1min at 14,000 rpm in an Eppendorf 5417C microcentrifuge to pelletthe nuclei. The supernatant was adjusted to 6.75 mM magnesiumacetate plus 200 µg of DNase I per ml and incubated for1 h at 37°C to remove the transfected plasmid DNA. The supernatantwas readjusted to 100 mM NaCl, 10 mM EDTA, 0.8% (wt/vol) sodiumlauryl sulfate, and 1.6 mg of pronase per ml and incubated foran additional 1 h at 37°C. The supernatant was extractedtwice with phenol, precipitated with 2 volumes of ethanol, andresuspended in 30 or 100 µl of 10 mM Tris hydrochloride(pH 8.0)-1 mM EDTA for DHBV DNA and HBV DNA, respectively. RNA(Northern) and DNA (Southern) filter hybridization analyseswere performed with 10 µg of total cellular RNA and 30µl of viral DNA replication intermediates, respectively,as described before (32).

RESULTS AND DISCUSSION

Top

Results and Discussion

DHBV transcripts and replication intermediates in human hepatoma cells. The transcripts and replication intermediates synthesized froma replication-competent DHBV genome were initially characterizedin the human hepatoma cell line HepG2 and compared with theviral products present in infected duck liver (Fig. 1). ThepCMVDHBV construct directed the expression of the DHBV 3.3-kbpregenomic RNA from the cytomegalovirus immediate-early promoterin HepG2 cells (Fig. 1A, lane 2). The pSPDHBV2X5.1 construct,which contains a dimer of the DHBV genome, also directed theexpression of the DHBV 3.3-kb RNA in HepG2 cells (Fig. 1A, lane3). These transcripts comigrated with the DHBV 3.3-kb RNA presentin infected duck liver (Fig. 1A, lane 1). Expression of theDHBV 3.3-kb pregenomic RNA was associated with the synthesisof DHBV viral replication intermediates (Fig. 1B). The pCMVDHBVand pSPDHBV2X5.1 constructs supported similar levels of DHBVreplication intermediates in HepG2 cells (Fig. 1B, lanes 2 and3). The relative levels of the DHBV relaxed circular and single-strandedreplication intermediates present in the transfected HepG2 cellswere similar to those observed in the infected duck liver (Fig.1B, lane 1).

View larger version (40K):
[in this window]
[in a new window]
FIG. 1. Transcription and replication of DHBV in duck liver and HepG2 cells. RNA and DNA were isolated from infected duck liver (lane 1), HepG2 cells transiently transfected with the pCMVDHBV construct (lane 2), and HepG2 cells transiently transfected with the pSPDHBV2X5.1 construct (lane 3). (A) RNA (Northern) filter hybridization analysis of DHBV transcripts. The locations of the DHBV 3.3-, 2.1-, and 1.8-kb transcripts are indicated. The glyceraldehyde 3-phosphate dehydrogenase (GAPDH) transcript was used as an internal control for RNA loading in each lane. (B) DNA (Southern) filter hybridization analysis of DHBV replication intermediates. The locations of the DHBV relaxed circular (RC) and single-stranded (SS) DNA replication intermediates are indicated.

HepG2 cells transfected with the pCMVDHBV construct synthesizedvery low levels of the subgenomic RNAs encoding the envelopepolypeptides (Fig. 1A, lane 2). The pSPDHBV2X5.1 construct encodeda significant level of the DHBV 2.1-kb transcript but a verylow level of the DHBV 1.8-kb RNA in HepG2 cells (Fig. 1A, lane3). This contrasted with the infected duck liver, where theDHBV 1.8-kb RNA was expressed at a higher level than the DHBV2.1-kb RNA (Fig. 1A, lane 1). The reason for these findingsprobably reflects, in part, the difference in the relative levelsof transcription factors that regulate DHBV transcription inHepG2 cells and duck liver.

Identification of liver-enriched transcription factors required for DHBV replication in mouse fibroblasts. HBV and DHBV can replicate in hepatoma cells transfected withgreater-than-genome-length constructs (2, 6, 12, 36, 41, 43,48). However, HBV and DHBV replication cannot be detected innonhepatoma cells (Fig. 2, lane 1) (42). In the case of HBV,it has been demonstrated that expression of the nuclear hormonereceptors HNF4 and RXR ${alpha}$ -PPAR ${alpha}$ activates pregenomic RNA synthesis(Fig. 3B, lanes 6 and 7) and viral replication in nonhepatomacells (Fig. 2B, lanes 6 and 7) (42). As DHBV is often used asa model system to investigate the mechanisms involved in regulatingvarious steps in the hepadnavirus life cycle, it was of interestto determine the similarities and differences in the transcriptionalregulation of HBV and DHBV replication.

View larger version (60K):
[in this window]
[in a new window]
FIG. 2. Effect of liver-enriched transcription factors on DHBV and HBV replication in mouse NIH 3T3 fibroblasts. Cells were transiently transfected with (A) the pSPDHBV2X5.1 DNA construct or (B) the pHBV2E DNA construct plus liver-enriched transcription factors as indicated. All-trans retinoic acid and clofibric acid at 1 µM and 1 mM, respectively, were used to activate the nuclear hormone receptors RXR ${alpha}$ and PPAR ${alpha}$ . The locations of the DHBV and HBV relaxed circular (RC) and single-stranded (SS) DNA replication intermediates are indicated.

View larger version (66K):
[in this window]
[in a new window]
FIG. 3. Effect of liver-enriched transcription factors on DHBV, HBV, and WHV transcription in mouse NIH 3T3 fibroblasts. Cells were transiently transfected with (A) the pSPDHBV2X5.1 DNA construct, (B) the pHBV2E DNA construct, or (C) the pSPWHV2X5.2 DNA construct plus liver-enriched transcription factors as indicated. All-trans retinoic acid and clofibric acid at 1 µM and 1 mM, respectively, were used to activate the nuclear hormone receptors RXR ${alpha}$ and PPAR ${alpha}$ . The locations of the DHBV 3.3-, 2.1-, and 1.8-kb transcripts are indicated. The locations of the HBV 3.5-, 2.1-, and 0.7-kb transcripts are indicated. The locations of the WHV 2.1- and 0.7-kb transcripts are indicated. The glyceraldehyde 3-phosphate dehydrogenase (GAPDH) transcript was used as an internal control for RNA loading in each lane.

The role of liver-enriched transcription factors in regulatingDHBV replication in mouse NIH 3T3 fibroblasts was investigatedand compared with their effect on HBV replication (Fig. 2).C/EBP ${alpha}$ , C/EBPß, C/EBP ${delta}$ , and HNF6 do not support detectableDHBV or HBV replication (H. Tang and A. McLachlan, unpublisheddata) (42). HNF1 ${alpha}$ , HNF1ß, and RXR ${alpha}$ -PPAR ${alpha}$ also failedto support detectable DHBV replication in mouse fibroblasts(Fig. 2A, lanes 2, 3, and 7). This result is significantly differentfrom those observed with HBV, in which viral replication wasactivated to the greatest extent by RXR ${alpha}$ -PPAR ${alpha}$ (Fig. 2B, lane7). Although both HBV and DHBV replication could be activatedby HNF4, RXR ${alpha}$ -PPAR ${alpha}$ not only failed to activate DHBV replicationbut also inhibited HNF4-dependent DHBV replication (Fig. 2A,lane 12). Therefore, DHBV replication is negatively regulatedby RXR ${alpha}$ -PPAR ${alpha}$ , which is the opposite of the effect of this transcriptionfactor on HBV replication (Fig. 2, lanes 7 and 12). The nuclearhormone receptor binding sites in the DHBV genome responsiblefor modulating viral replication have not been defined.

HNF3 also modulated HBV and DHBV replication in mouse fibroblastsin very different manners. HNF3 ${alpha}$ and HNF3ß supportedDHBV replication (Fig. 2A, lanes 4 and 5), whereas HNF3 couldnot support detectable HBV replication (Fig. 2B, lanes 4 and5). In addition, HNF3 synergistically activated DHBV replicationin combination with HNF4 (Fig. 2A, lanes 10 and 11). In contrast,HNF3 inhibited nuclear hormone receptor-dependent HBV replication(Fig. 2B, lanes 10 and 11) (42). Therefore, it is apparent thatthe transcriptional regulation of HBV and DHBV replication innonhepatoma cells is significantly different for these two hepadnaviruses.Although HNF4 activated both HBV and DHBV replication, DHBVreplication was also activated by HNF3 and repressed by RXR ${alpha}$ -PPAR ${alpha}$ ,whereas HBV replication was activated by RXR ${alpha}$ -PPAR ${alpha}$ and repressedby HNF3 under identical conditions in mouse fibroblasts. Theseobservations indicate that avian and mammalian hepadnaviruseshave distinct transcription factor requirements for viral replication.

Effect of liver-enriched transcription factors on DHBV transcription in mouse fibroblasts. The absence of detectable DHBV replication in mouse fibroblastsreflects the failure of the DHBV 3.3-kb pregenomic RNA to betranscribed from the nucleocapsid promoter (Fig. 3A, lane 1).Transcription of the DHBV 3.3-kb pregenomic RNA and viral replicationin mouse fibroblasts are dependent on the expression of HNF3and HNF4 (Fig. 2A and 3A, lanes 4 to 6 and 8 to 11). The levelof DHBV 3.3-kb pregenomic RNA synthesis in the presence of HNF3 ${alpha}$ ,HNF3ß, or HNF4 was very low (Fig. 3A, lanes 4 to 6).In the presence of HNF3 ${alpha}$ plus HNF4 or HNF3ß plus HNF4,the DHBV 3.3-kb pregenomic RNA was readily detectable (Fig.3A, lanes 10 and 11), and the level correlated with the observedlevel of viral replication (Fig. 2A, lanes 10 and 11).

The liver-enriched transcription factors also appeared to influencethe expression of the DHBV subgenomic transcripts (Fig. 3A).HNF1 and RXR ${alpha}$ -PPAR ${alpha}$ expression increased the level of expressionof the DHBV 2.1-kb RNA (Fig. 3A, lanes 2, 3, and 7). HNF3 increasedthe level of expression of the DHBV 1.8-kb RNA (Fig. 3A, lanes4 and 5). This observation is consistent with the presence ofan HNF3 binding site in the DHBV major surface antigen promoter,which is a critical regulatory element within this promoter(46).

Avian and mammalian hepadnaviruses have distinct transcription factor requirements for viral replication. Hepadnavirus replication is dependent on transcription of theviral pregenomic RNA. The synthesis of pregenomic RNA is controlledby the activity of the nucleocapsid promoter. The regulatorysequence elements that control the level of transcription fromthe nucleocapsid promoter for the different hepadnaviruses havebeen characterized to various extents. For HBV, nuclear hormonereceptors are essential for transcription of pregenomic RNAand viral replication (42). HNF3 antagonizes nuclear hormonereceptor-dependent viral replication (42). The observation thatHNF3 mediates DHBV replication in mouse fibroblasts is consistentwith the observation that the nucleocapsid promoter regulatorysequences contain recognition elements that bind this transcriptionfactor (7, 19). However HBV also contains HNF3 recognition elementswithin the nucleocapsid promoter (14), but replication of thishepadnavirus is negatively regulated by HNF3 in mouse fibroblasts(42). Therefore, the presence of HNF3 recognition sites withinthe nucleocapsid promoter is not sufficient to predict the rolethat this transcription factor might have in viral replication.The presence of additional HNF3 sites in the viral genome ortheir relationship to other transcription factor regulatoryelements must influence the effect of HNF3 on viral transcriptionand replication.

DHBV replication is positively regulated by HNF4 but negativelyregulated by RXR ${alpha}$ -PPAR ${alpha}$ (Fig. 2A). The nuclear hormone receptorrecognition sequences that might mediate these alterations inviral replication in the DHBV genome have not been identified.However, the opposing effects of these transcription factorscontrast with the similar effects they have on HBV replication(Fig. 2B) (42). Therefore, it is apparent that although thesame transcription factors are the major regulators of HBV andDHBV pregenomic RNA synthesis and replication in mouse fibroblasts,the effects of HNF3 and RXR ${alpha}$ -PPAR ${alpha}$ are distinct.

Transcription and viral DNA synthesis derived from a replication-competentWHV genome were also examined in mouse fibroblasts (Fig. 3C).WHV subgenomic 2.1-kb RNA was transcribed in mouse fibroblastsin the absence of liver-enriched transcription factors (Fig.3C, lane 1). RXR ${alpha}$ -PPAR ${alpha}$ expression increased the level of expressionof the WHV 2.1-kb and 0.7-kb RNAs (Fig. 3C, lane 7). However,WHV pregenomic RNA synthesis and viral replication were notobserved with any of the combinations of transcription factorsdemonstrated to activate HBV or DHBV replication (Fig. 3C) (H.Tang and A. McLachlan, unpublished data). Consequently it appearsthat avian and mammalian hepadnaviruses have evolved distinctmodes of transcriptional regulation as they have adapted totheir highly specific host organisms.

ACKNOWLEDGMENTS

We are grateful to Jesse Summers (University of New Mexico Schoolof Medicine, Albuquerque) for plasmids pSPDHBV2X5.1 and pSPWHV2X5.2and DHBV-infected duck liver, Stefan Wieland (The Scripps ResearchInstitute, La Jolla, Calif.) for plasmid pCMVDHBV, Eric F. Johnson(The Scripps Research Institute) for plasmids pCMVHNF4 and pCMVPPAR ${alpha}$ -G,Ronald M. Evans (The Salk Institute, La Jolla, Calif.) for plasmidpRS-hRXR ${alpha}$ , Robert Costa (University of Illinois, Chicago) forplasmids pCMVHNF3 ${alpha}$ and pCMVHNF3ß, Riccardo Cortese(Instituto di Ricerche di Biologia Molecolare, Rome, Italy)for plasmid pB1.1 (rat HNF1 ${alpha}$ cDNA), and Gerald R. Crabtree (StanfordUniversity, Stanford, Calif.) for plasmid 28-1 (mouse HNF1ßcDNA).

This work was supported by a postdoctoral fellowship from theWest China University of Medical Sciences of the People's Republicof China to H.T. and Public Health Service grant AI30070 fromthe National Institutes of Health.

FOOTNOTES

* Corresponding author. Mailing address: Department of Cell Biology, The Scripps Research Institute, 10550 N. Torrey Pines Rd., La Jolla, CA 92037. Phone: (858) 784-8097. Fax: (858) 784-2513. E-mail: mclach{at}scripps.edu.

Publication number 14796-CB from The Scripps Research Institute.

REFERENCES

Top