Point Mutations Upstream of Hepatitis B Virus Core
http://www.100md.com
病菌学杂志 2006年第2期
The Liver Research Center and Brown Medical School, Providence, Rhode Island 02903
ABSTRACT
The pregenomic RNA directs replication of the hepatitis B virus (HBV) genome by serving both as the messenger for core protein and polymerase and as the genome precursor following its packaging into the core particle. RNA packaging is mediated by a stem-loop structure present at its 5' end designated the signal, which includes the core gene initiator AUG. The precore RNA has a slightly extended 5' end to cover the entire precore region and, consequently, directs the translation of a precore/core protein, which is secreted as e antigen (HBeAg) following removal of precore-derived signal peptide and the carboxyl terminus. A naturally occurring G1862T mutation upstream of the core AUG affects the bulge of the signal and generates a "forbidden" residue at the –3 position of the signal peptide cleavage site. Transfection of this and other mutants into human hepatoma cells failed to prove their inhibition of HBeAg secretion but rather revealed great impairment of genome replication. This replication defect was associated with reduced expression of core protein and could be overcome by a G1899A covariation, or by nonsense or frameshift mutation in the precore region. All these mutations antagonized the G1862T mutation on core protein expression. Cotransfection of the G1862T mutant with a replication-deficient HBV genome that provides core protein in trans also restored genome replication. Consistent with our findings in cell culture, HBV genotype A found in African/Asian patients has T1862 and is associated with much lower viremia titers than the European subgroup of genotype A.
INTRODUCTION
The hepatitis B virus (HBV) primarily infects the liver and causes chronic hepatitis, liver cirrhosis, and hepatocellular carcinoma worldwide. It is an enveloped DNA virus with a small double-stranded genome of 3.2 kb. Inside hepatocytes, viral genomic and subgenomic RNAs are transcribed from a covalently closed circular DNA template in the nucleus and exported to the cytoplasm for translation into viral proteins. The core protein assembles into the core particle, packaging both the pregenomic RNA and DNA polymerase. Subsequently, the DNA polymerase synthesizes the negative-strand DNA via reverse transcription from the RNA template, followed by RNA degradation and synthesis of positive-strand DNA. The core particle with the double-stranded DNA genome is enveloped by host-derived lipids and viral envelope proteins and secreted as an infectious virus particle (for a review, see reference 7). Since the 3.5-kb pregenomic RNA is also the mRNA for the expression of both core protein and DNA polymerase, it is the sole component required for genome replication. Consequently, increased transcription of pregenomic RNA will lead to enhanced HBV replication, as exemplified by the naturally occurring core promoter mutants (3, 20).
The AUG initiator of the core gene (position 1901 to 1903) is located approximately 80 nucleotides (nt) downstream of the 5' end of its mRNA, while the initiation codon of polymerase is 400 nt further downstream. Therefore, core protein translation can proceed directly by ribosomal scanning, whereas translation of polymerase requires a specific mechanism of translational termination and reinitiation or ribosomal shunting (9, 23). Nevertheless, the 5' end of the pregenomic RNA also functions as its encapsidation signal (the signal), which forms a stem-loop structure consisting of two base-paired regions, a 6-nt bulge, and a 6-nt loop (Fig. 1C) (10, 21, 31). The core AUG is located near the 3' end of this stem-loop as part of the lower stem. Considering that RNA secondary structure impedes the passage of the scanning 40S ribosome (12), whether translation of the core protein is regulated negatively by the signal or adjusted temporarily according to the changing functions of the pregenomic RNA remains to be determined.
The signal is not only required for the packaging of pregenomic RNA but also involved in the initiation of reverse transcription. The polymerase employs an N-terminal tyrosine residue as a primer to generate the first three nucleotides (5'-GAA-3') of the negative-strand DNA, using the UUC sequence at the 3' bulge of the signal as the template (17, 32) (Fig. 1C). Next, the negative-strand DNA is dislodged from the bulge and transferred to the UUC motif about 3.2 kb downstream in the 3' direct repeat 1 region, where reverse transcription will resume. This long-range template switch is probably facilitated by an RNA secondary structure (28).
Transcription initiation mediated by the core promoter is imprecise, and a fraction of the 3.5-kb RNA is about 30 nt longer than pregenomic RNA, thus enabling it to cover the intact precore region (positions 1814 to 1900) for an additional 29 amino acid codons (Fig. 1A). This subset of the 3.5-kb RNA, termed the precore RNA, does not express DNA polymerase and cannot serve as the pregenome because the secondary structure of the signal is melted by translating ribosome (16). The resultant precore/core protein is targeted to the lumen of the endoplasmic reticulum by its N-terminal signal peptide of 19 residues, which is clipped off by the signal peptidase (4, 22, 24) (Fig. 1B). The protein is secreted as e antigen (HBeAg) following further removal of a C-terminal basic sequence (Fig. 1A). The HBeAg is not essential for HBV replication in vitro but potentially important for the establishment of persistent infection in vivo (5, 30). Following the rise of antibody against HBeAg in hepatitis B patients, however, HBV variants with reduced or abolished expression of HBeAg often become the dominant viral species, possibly due to the negative selection for viral strains expressing HBeAg (34).
In this regard, a G1862T substitution is frequently detected in patients following seroconversion to anti-HBe (14, 33). This point mutation converts residue 17 of the precore peptide, the –3 position of the signal peptidase cleavage site, to phenylalanine. The G1862A and G1862C mutations have also been reported, albeit at much reduced frequency (14). These mutations convert the –3 position to isoleucine and leucine, respectively. Since a bulky residue such as phenylalanine, but not leucine or isoleucine, at the –3 position is considered "forbidden" for signal peptidase cleavage (18), the G1862T mutation has been speculated to reduce HBeAg expression, thus accounting for its prevalence at the anti-HBe stage of infection. The G1862T mutation also affects the bulge of the signal, at a position that is 39 nt upstream of the core gene initiator (Fig. 1C). By site-directed mutagenesis and transfection experiments, we failed to validate inhibition of HBeAg secretion by any of the substitutions introduced into the –3 position of the signal peptidase cleavage site but rather observed a severe replication defect of the corresponding mutants. This defect was associated with reduced core protein expression and could be overcome by compensatory mutations in the precore region or by core protein provided in trans.
MATERIALS AND METHODS
Mutant HBV constructs. Table 1 lists the single mutants and one double mutant used in this study. Other double and triple mutants are not listed, since they combine the properties of the single or double mutants. Mutant constructs generated for this study are based on clone N4, which is a chimera between two naturally occurring core promoter mutants of the European subgroup of genotype A. Specifically, it contains the backbone of clone 3.4 with a 0.8-kb AvrII-EcoRV fragment derived from clone 4B (11, 20). This construct maintains high replication capacity and efficient HBeAg expression as well as robust expression of HBsAg (hepatitis B surface antigen, viral envelope proteins) and secretion of virus particles. Each mutation was created by overlap extension PCR using Roche High-Fidelity PCR system (1, 2, 11, 20). The PCR products were double digested with either RsrII-BspE1 (positions 1750 to 2328) or RsrII-ApaI (positions 1750 to 2509) for cloning back to the N4 construct. The PCR-derived region of each mutant was verified by sequence analysis. Each mutant construct was converted into a tandem dimer version in the EcoRI site of the pUC18 vector (20). The plasmid DNA was purified by a commercial kit (Marligen), further extracted with phenol and chloroform/isoamyl alcohol, and dissolved in Tris-EDTA buffer. Before transfection, DNA concentrations of the constructs were verified by digestion of 500 ng of DNA with HindIII followed by gel electrophoresis.
The 4B genome lacking core protein expression (core– 4B genome) and with defective polymerase expression and nonfunctional encapsidation signal (Pol–/– 4B genome) were employed in the cotransfection experiments. They have been described previously (1, 2). The core– 4B dimer genome contains a C2044G nonsense mutation in the core gene to ablate core protein expression (1). It is still competent in the expression of polymerase and viral envelope proteins, and its pregenomic RNA can be encapsidated if core protein is provided by another plasmid. The Pol–/– 4B genome contains a C2589T nonsense mutation in the polymerase gene to prevent polymerase expression as well as a G1879T/T1880A double mutation in the loop of the signal, thus abolishing packaging of pregenomic RNA (2).
Transfection. The Huh7 human hepatoma cells were cultured in Dulbecco's modified Eagle medium supplemented with 10% calf serum. The HBV dimers were transfected into Huh7 cells via the TransIT Transfection system (Mirus) as described previously (2). Serum-free minimal essential medium (150 μl) was mixed with 4 μl of transit-LT1 reagent, vortexed, and incubated at room temperature for 15 min. Following mixing with 1.5 μg of DNA and a further incubation for 15 min, the complex was added dropwise to cells grown in six-well plates. After overnight incubation, cells were washed, replenished with fresh medium, and harvested 4 days later. For the transcomplementation assay, 1.1 μg of HBV mutants generated in this study was cotransfected with 0.4 μg of either pcDNA3.1/zeo(–) vector, core– 4B dimer, or Pol–/– 4B dimer.
Analysis of HBV DNA replication and virion secretion. The details of these assays have been described previously (1, 2, 11, 20). Briefly, transfected plasmid DNA was eliminated from cell lysate by nuclease treatment, and core particles were precipitated by polyethylene glycol solution. Core particles were disintegrated by digestion with proteinase K in the presence of sodium dodecyl sulfate (SDS), and HBV DNA was extracted with phenol, precipitated with ethanol, and dissolved in Tris-EDTA buffer. DNA was separated on a 1.5% agarose gel, with ethidium bromide (1 μg/ml) present in both the gel and the running buffer. After transfer to a membrane, HBV DNA was detected with a randomly primed probe. Extracellular viral particles (both naked core particles and enveloped Dane particles) were concentrated by ultracentrifugation through a 20% sucrose cushion, and DNA was extracted for Southern blot analysis.
Measurement of HBeAg and HBsAg expression. HBeAg was measured by the EBK 125I radioimmunoassay (DiaSorin) in earlier experiments using 4 to 10 μl of culture supernatant, and by the ETI-EBK plus enzyme immunoassay (DiaSorin) in later experiments when the radioimmunoassay kit was discontinued. HBsAg was measured by the Auszyme monoclonal HBsAg kit (Abbot Laboratories) with 4 to 10 μl of culture supernatant. The volumes of culture supernatants used will not cause signal saturation.
Western blot analysis of core protein, HBsAg, and HBeAg. Huh7 cells grown in each well of the six-well plates were scraped at day 5 posttransfection, and the cell pellet was lysed with 100 μl of buffer (10 mM HEPES, pH 7.5, 100 mM NaCl, 1 mM EDTA, 1% NP-40) supplemented with protease inhibitor cocktail (Roche). Proteins from 40 μl of lysate were separated on a 0.1% SDS-12% polyacrylamide gel and transferred to polyvinylidene difluoride membranes. The membranes were blocked at room temperature with 3% bovine serum albumin (BSA) in Tris-buffered saline-Tween (TBST) buffer, and incubated at 4°C for 3 overnights with a mouse monoclonal anti-core antibody (14E11; Abcam) diluted 1:2,000 in 3% BSA-TBST. After 40 min of washing at room temperature with TBST buffer, the blots were incubated at room temperature for 1 h with horseradish peroxidase-conjugated anti-mouse antibodies (Amersham) diluted 1:30,000 in 3% BSA-TBST. The blots were washed with TBST for 40 min, and signals were revealed by enhanced chemiluminescence (ECL). To control for transfection efficiency, the blots were stripped with Restore Western Blot stripping buffer (Pierce) at 37°C for 20 min, rinsed with water and TBST buffer, and blocked again with 3% BSA in TBST buffer. The blots were incubated at 4°C overnight with a goat polyclonal anti-HBs antibody (Dako) diluted 1:8,000 in 3% BSA-TBST, washed, and incubated at room temperature for 1 h with horseradish peroxidase-conjugated anti-goat antibodies (Amersham) diluted 1:50,000 in TBST. The blots were washed again, and signals were revealed by ECL. Alternatively, HBsAg was detected by a 1:5,000 dilution of a polyclonal horse anti-HBs (Ad/Ay) antibody (Abcam), and a 1:5,000 dilution of rabbit anti-horse immunoglobulin G conjugated to horseradish peroxidase (Abcam).
For immunoprecipitation of secreted HBeAg, culture supernatants collected at day 5 posttransfection were precleared of cell debris by centrifugation at 3,000 rpm for 10 min. A 1-ml aliquot was incubated overnight at 4°C with 0.5 μl of anti-core antibody (Dako), followed by addition of 5 μl of protein G beads and further incubation for 3 to 4 h. The immune complex was collected by low-speed centrifugation, washed once with TBST buffer, and separated in 15% polyacrylamide gels. The protein blot was blocked with 3% BSA-TBTS and incubated overnight at 4°C with a 1:2,000 dilution of rabbit polyclonal antibody against core antigen (Dako). After washing with TBST buffer, the blot was incubated with a 1:2,000 dilution (in 3% BSA-TBST) of anti-rabbit antibodies conjugated with horseradish peroxidase (Amersham). The blot was washed with TBST, and signals were revealed by ECL.
RESULTS
Mutations at the –3 position of the signal peptidase cleavage site do not reduce HBeAg secretion. The amino terminus of HBeAg has been mapped to S20 of the precore peptide, thus establishing A19 and V17 as the –1 and –3 positions, respectively (Fig. 1B) (24, 26). This finding is consistent with the preference of the endoplasmic reticulum (ER) signal peptidase for small and neutral amino acids at the –1 and –3 positions (18). Bulky or charged residues at these two positions often impair signal peptide processing (6). The V17F substitution in the precore peptide (pcV17F) as a result of the G1862T mutation generates a bulky residue at the –3 position, whereas the pcV17L substitution does not. We generated these two naturally occurring variants as well as the artificial pcV17D and pcA19D mutants, which contain charged residues at the –3 and –1 positions, respectively. The DNA constructs were transfected to Huh7 cells, and HBeAg at day 5 posttransfection was measured with a commercial radioimmunoassay. To correct for variations in transfection efficiency, the HBeAg values were normalized against HBsAg values (Fig. 2A). No reduction in the secretion of HBeAg was observed in association with any mutation at the –3 position (results of three independent experiments are shown in Fig. 2B). The pcA19D mutation at the –1 site incurred, at most, modest (20 to 30%) reduction of HBeAg expression. Similar results were obtained when culture supernatants were harvested at earlier time points (data not shown). Results obtained by radioimmunoassay were further validated by immunoprecipitation-Western blot analysis. Since the E77Q mutation present in the core gene of N4-based constructs greatly reduces the sensitivity of HBeAg detection by a commercial polyclonal antibody (Dako) (K. Kim and S. Tong, unpublished), we introduced a Q77E mutation back into these constructs. As a result, no significant change in either the amount or sizes of HBeAg was observed, even when both pcV17F and pcA19D mutations were present (Fig. 2C).
The G1862T, T1863A, and C1869A mutants are impaired in genome replication. The pcV17F, pcV17L, pcV17D, and pcA19D mutations in the precore peptide correspond to G1862T, G1862C, T1863A, and C1869A mutations at the nucleotide level (Table 1). Of these, the T1863A mutation changes the 3'-most position of the priming site for reverse transcription, whereas the C1869A mutation disrupts a base pair in the upper stem of the signal (Fig. 3A). Since the signal is essential for the packaging of pregenomic RNA and, hence, genome replication (10, 21, 31), we analyzed the replication capacity of the mutants. Figure 3C shows Southern blot analysis of HBV DNA associated with intracellular core particles and extracellular viral particles. The transfection efficiencies were similar among the various constructs, as indicated by values of HBsAg in the culture supernatant (Fig. 3B). DNA replication was severely impaired not only for the T1863A and C1869A mutants but also for the G1862T mutant (Fig. 3C). Combining the G1862T and C1869A mutations led to nearly complete loss of replication capacity.
Introduction of the G1899A mutation overcomes the replication defect of these mutants. The G1862T mutation is frequently associated with a G1888A or G1899A covariation (8, 14, 25, 27, 33). While the G1888A mutation is silent at the amino acid level, the G1899A mutation induces a pcG29D substitution in the terminal residue of the precore peptide (Table 1). Interestingly, both mutations convert a wobble U:G base pair in the signal into a U:A pair (upper stem for G1888A and lower stem for G1899A) (Fig. 4A). Mutagenesis and transfection experiments revealed that, while the G1888A mutant maintained the wild-type level of genome replication and virion secretion, introduction of the G1862T mutation into this mutant severely impaired DNA replication (Fig. 4B). The G1899A mutation modestly increased genome replication, and the G1899A/G1862T double mutant replicated much better than the G1862T single mutant. The G1899A mutation could also markedly enhance the replication of the T1863A, C1869A, and G1862C mutants (Fig. 5), suggesting a general mechanism of rescue.
Replication impact of the G1899A mutation is independent of its improvement of base pairing in the signal. As already mentioned, the G1899A mutation converts a U:G pair in the lower stem of the signal, between U1855 and G1899, into a U:A pair. In this regard, a T1855C mutation will convert the same U:G pair into an even stronger C:G pair. Transfection experiments revealed wild-type replication capacity of the T1855C single mutant but much reduced replication of the T1855C/G1862T double mutant reminiscent of the G1862T single mutant (Fig. 4B). On the other hand, the T1855C/G1899A double mutant retained a high replication capacity, despite the disruption of the base pair involved. Furthermore, introduction of the G1862T mutation into this double mutant did not markedly reduce genome replication (Fig. 4B, last lane). Thus, the ability of the G1899A mutation to rescue replication of the G1862T mutant is not mediated by, and is independent of, stronger base pairing in the lower stem of the signal.
Since G1899 occupies the –2 position of core gene AUG codon (Fig. 1C), the G1899A mutation may function through modulation of core protein expression. In this regard, a cytosine at the –2 position is considered optimal for translation initiation (13). We therefore generated the G1899C mutant alone or together with G1862T. Indeed, the G1899C/G1862T double mutant retained relatively high replication capacity, much better than the G1862T single mutant (Fig. 6).
Nonsense and frameshift mutations in the precore region could also rescue replication of the G1862T mutant. Incidentally, we found that replication of the G1862T mutant could be rescued by mutations at other positions. A C1865T mutation did not markedly reduce genome replication, despite its location at the priming site for reverse transcription, and the C1865T/G1862T double mutant was found to replicate as efficiently (Fig. 6). Interestingly, the C1865T mutation converts precore codon 18 into a TAA termination codon, which has the potential to increase core protein expression by translational termination and reinitiation from precore RNA (29). To verify the significance of the nonsense mutation, we created a missense mutant of the same codon, C1865G (pcQ18E at the amino acid level). Although the C1865G single mutant had only a modest decline in replication, the C1865G/G1862T double mutant was severely impaired in DNA replication. On the other hand, a frameshift mutation (1840insC) introduced into the precore region upstream of the signal, which would terminate precore-core protein translation in the vicinity of core AUG, could rescue the replication of the G1862T mutant as well (Fig. 6).
Replication capacities of the mutants correlate well with their levels of core protein expression. The ability of the G1899A mutation to rescue replication of several mutants, its juxtaposition with core gene AUG codon, and the ability of precore nonsense and frameshift mutations to rescue the G1862T mutant strongly suggested coupling of core protein expression with the replication capacity of the constructs. Indeed, Western blot analysis revealed reduced core protein expression by the G1862T mutation (Fig. 7A, upper panel). In contrast, the G1899A mutation enhanced core protein expression. This was not due to higher transfection efficiency of the G1899A mutant, as revealed by reprobing of the same blot with anti-HBs antibodies (lower panel). The presence of both mutations resulted in core protein expression similar to that of the parental clone, N4. Introduction of the G1899A mutation could also override the low core protein expression associated with the T1863A, G1862C, or C1869A mutant (Fig. 7B). Similarly, other mutations capable of rescuing the G1862T mutant, such as C1865T and G1899C, could maintain core protein expression when combined with the G1862T mutation. In contrast, the T1855C, C1865G, or G1888A mutation could not rescue core protein expression when coupled with the G1862T mutation (Fig. 7A, upper panel).
Replication of the G1862T, T1863A, and C1869A mutants could be rescued by core protein provided in trans. If the G1862T, T1863A, and C1869A mutants harbor no defect other than core protein expression, then providing core protein from another plasmid should suffice to relieve their replication block. To test this hypothesis, constructs with normal, increased, or reduced replication capacity were cotransfected with either pcDNA vector, core– 4B dimer, or Pol–/– 4B dimer, using a ratio of 1.1 μg/0.4 μg. Clone 4B is a core promoter mutant with extremely high levels of genome replication (20). The core– 4B dimer was derived from 4B by a nonsense mutation in the core gene, whereas the Pol–/– 4B dimer contains a nonsense mutation in the polymerase gene as well as a double mutation in the loop of the signal that abolishes packaging of pregenomic RNA. The Pol–/– 4B dimer produced high levels of core protein even at this ratio of cotransfection (data not shown). As illustrated in Fig. 8A, the low replicating G1862T single or double mutants were not rescued by cotransfection with a core– 4B dimer but by the Pol–/– 4B dimer, such that their differences with high replicating clones were largely abolished. The Pol–/– 4B dimer could also rescue the replication of the T1863A and C1869A mutants and even the G1862T/C1869A double mutant, which had negligible genome replication (Fig. 8B). On the other hand, the T1880G mutant with a nonfunctional signal (2) could rescue the core– 4B dimer but not the Pol–/– 4B dimer. Therefore, the defect of the G1862T, T1863A, and C1869A mutants indeed lies at core protein expression rather than the pregenome or DNA polymerase.
DISCUSSION
Effect of mutations at the –3 and –1 positions of the signal peptide cleavage site on HBeAg secretion. The N terminus of the precore peptide serves as a signal peptide for the cotranslational translocation of the precore-core protein to the ER lumen, leading to HBeAg secretion (4, 22, 24). A signal peptide is composed of an N-terminal basic sequence, a central hydrophobic region for membrane insertion, and a polar but noncharged C terminus for cleavage (18, 19). Residues –1 and –3 interact with specific P1 and P3 pockets in the ER signal peptidase and are critical for cleavage. Aromatic and charged residues introduced into these positions often abolish cleavage or alter the cleavage site (6). Molecular epidemiological studies have revealed, with the exception of the African/Asian subgroup of genotype A (genotype Aa), the rise of the G1862T mutation only at the anti-HBe stage of infection (14, 25, 27, 33). The corresponding V17F substitution in the precore peptide generates a "forbidden" residue at the –3 position. In the current study, we found that HBeAg secretion was not impaired in constructs with forbidden residues at the –3 position (pcV17F and pcV17D) relative to the wild-type virus or the pcV17L variant (Fig. 2). The pcA19D mutation at the –1 position only mildly impaired HBeAg production. Most strikingly, the pcV17F/pcA19D double mutant displayed no further reduction in HBeAg secretion. However, computer predictions of the double mutant suggested a possible shift of the cleavage site from residue 19 to residue 18, which would put the pcV17F and pcA19D mutations at the flexible –2 and +1 positions, with pcQ18 and pcT16 now occupying the –1 and –3 positions. Similarly, introduction of forbidden amino acids into the –1 positions caused a shift of the cleavage site for the signal peptide of Escherichia coli maltose-binding protein (6). At any rate, our observations suggest that single and even double mutations near the signal peptide cleavage site may not constitute an effective means of terminating HBeAg secretion. These findings are consistent with the fact that genotype Aa, which often has T1862 as its wild-type sequence, continues to express HBeAg (25, 27).
Do mutations at nucleotides 1862, 1863, and 1869 just down regulate core protein expression The G1862T, T1863A, and C1869A mutants are hampered in genome replication and core protein expression. In this regard, HBV genome replication is driven by the pregenomic RNA, which serves as the pregenome and the messenger for core protein and polymerase. Our preliminary primer extension analysis of total cellular RNA revealed no reduction in the level of pregenomic RNA in cells transfected with the G1862T mutant, although we do not know whether expression of polymerase is impaired. The ability of the Pol–/– 4B dimer to rescue replication of all three of the mutants (Fig. 8) argues against a defect in polymerase expression and suggests intact function of the signal (with the exception of T1863A mutation, see below). Alternatively, one may argue that mutations at positions 1862, 1863, and 1869 do reduce polymerase expression, yet DNA polymerase is not the limiting factor for RNA packaging. However, the Pol–/– 4B dimer did not markedly enhance replication of clone N4, the parental construct.
Nucleotides 1865 to 1863 serve as the priming site for reverse transcription (Fig. 1) (17). In this regard, the T1863A mutant produced progeny viral DNA of reduced size (Fig. 3 and 5), suggesting aberrant translocation leading to shorter products. No aberrant products were observed for the C1865G or C1865T mutant. These findings are consistent with the greater importance of nucleotide 1863 than 1865 in determining the specificity of the template switch. Nucleotide 1863 defines the 3'-most nucleotide of the nascent minus DNA strand (5'-GAA-3') (Fig. 1C). The T1863A mutation will produce a minus DNA strand of 5'-GAT-3'. The mismatch of its 3'-most nucleotide with direct repeat 1 is very likely to prevent DNA elongation, thus promoting annealing elsewhere to generate shorter DNA products.
How do mutations at positions 1862, 1863, 1869, and 1899 modulate core protein expression The G1899A mutation enhances core protein expression. Although this mutation affects a base pair in the lower stem of the signal, we demonstrated that its biological properties were not mediated by, and is independent of, enhanced stability of the signal. Thus, its effect may be related to its occupation of the –2 position of core gene AUG codon. Indeed, the G1899C mutation of the same position could also overcome the replication defect of the G1862T mutant. The G1862T, T1863A, and C1869A mutations affect the –39, –38, and –32 positions of core gene AUG. These mutations are located at the bulge (G1862T, T1863A) and upper stem (C1869A) of the signal, respectively. Whether they work through structural changes to the signal remains unclear. It is unknown whether the stem-loop structure of the signal is formed during core protein translation, and if so, whether it facilitates or hinders core protein expression. The signal binds viral polymerase through the loop, upper stem, and part of the bulge (nucleotides 1860 and 1861).
Clinical relevance of the replication effect of G1862T and G1899A mutations. Genotype Aa patients manifest much lower viremia titers than patients of genotype Ae (European genotype A) (27). For HBeAg-positive patients, the mean viremia titer was 5 x 105 genome copies/ml in genotype Aa patients compared to 4 x 108 copies/ml in genotype Ae patients, with a difference of 800-fold. In the HBeAg-negative group of patients, the value was 7 x 102 versus 6 x 103 copies/ml. Based on the markedly down regulation of genome replication by the G1862T mutation, we propose that T1862 contributes to the reduced replication capacity of genotype Aa. Since not all genotype Aa isolates contain T1862, it will be of great interest to correlate the viremia titers in genotype Aa patients with polymorphism at this position (T1862 versus G1862).
The G1899A mutation is highly prevalent at the anti-HBe stage of chronic infection, often subsequent to the G1886A nonsense mutation in the precore region that abolishes HBeAg expression (15). It is also common in patients containing the G1862T mutation (14, 33). We found that the G1899A mutation largely relieves the inhibitory effect of the G1862T mutation on genome replication, as a consequence of restored core protein expression. The G1862T mutant of genotype B has been implicated in several cases of fulminant hepatitis in China (8). Interestingly, all such isolates harbored the G1899A mutation but no G1896A nonsense mutation or core promoter mutations (8). Since enhanced replication of the HBV genome is considered critical in the induction of fulminant hepatitis (3), the presence of the G1899A mutation in such strains may not be a mere coincidence.
ACKNOWLEDGMENTS
This work was supported by grants AI54535, DK62857, p20RR15578, CA35711, DK66950, and CA109733 from the National Institutes of Health and from Life Span Research Funds.
REFERENCES
Ahn, S., A. Kramvis, S. Kawai, H. Spangenberg, J. Li, G. Kimbi, M. Kew, J. Wands, and S. Tong. 2003. Sequence variation upstream of precore translation initiation codon reduces hepatitis B virus e antigen production. Gastroenterology 125:1370-1378.
Bang, G., K. Kim, M. Guarnieri, F. Zoulim, S. Kawai, J. Li, J. Wands, and S. Tong. 2005. Effect of mutating the two cysteines required for HBe antigenicity on hepatitis B virus DNA replication and virion secretion. Virology 332:216-224.
Baumert, T., S. Rogers, K. Hasegawa, and T. Liang. 1996. Two core promoter mutations identified in a hepatitis B virus strain associated with fulminant hepatitis result in enhanced viral replication. J. Clin. Investig. 98:2268-2276.
Bruss, V., and W. Gerlich. 1988. Formation of transmembraneous hepatitis B e-antigen by cotranslational in vitro processing of the viral precore protein. Virology 163:268-275.
Chen, H., M. Kew, W. Hornbuckle, B. Tennant, P. Cote, J. Gerin, R. Purcell, and R. Miller. 1992. The precore gene of the woodchuck hepatitis virus genome is not essential for viral replication in the natural host. J. Virol. 66:5682-5684.
Fikes, J., G. Barkocy-Gallagher, D. Klapper, and P. Bassford. 1990. Maturation of Escherichia coli maltose-binding protein by signal peptidase I in vivo. J. Biol. Chem. 265:3417-3423.
Ganem, D., and R. Schneider. 2001. Hepadnaviridae: the viruses and their replication, p. 2923-2970. In D. Knipe and P. Howley (ed.), Fields virology, 3rd ed. Lippincott-Raven Publishers, Philadelphia, Pa.
Hou, J., Y. Lin, J. Waters, Z. Wang, J. Min, H. Liao, J. Jiang, J. Chen, K. Luo, and P. Karayiannis. 2002. Detection and significance of a G1862T variant of hepatitis B virus in Chinese patients with fulminant hepatitis. J. Gen. Virol. 83:2291-2298.
Jean-Jean, O., T. Weimer, A. de Recondo, H. Will, and J. Rossignol. 1989. Internal entry of ribosomes and ribosomal scanning involved in hepatitis B virus P gene expression. J. Virol. 63:5451-5454.
Junker-Niepmann, M., R. Bartenschlager, and H. Schaller. 1990. A short cis-acting sequence is required for hepatitis B virus pregenome encapsidation and sufficient for packaging of foreign RNA. EMBO J. 9:3389-3396.
Khan, N., M. Guarnieri, S. H. Ahn, J. Li, Y. Zhou, G. Bang, K. H. Kim, J. R. Wands, and S. Tong. 2004. Modulation of hepatitis B virus secretion by naturally occurring mutations in the S gene. J. Virol. 78:3262-3270.
Kozak, M. 1986. Influence of mRNA secondary structure on initiation by eukaryotic ribosomes. Proc. Natl. Acad. Sci. USA 83:2850-2854.
Kozak, M. 1987. At least six nucleotides preceding the AUG initiator codon enhance translation in mammalian cells. J. Mol. Biol. 196:947-950.
Kramvis, A., S. Bukofzer, M. Kew, and E. Song. 1997. Nucleic acid sequence analysis of the precore region of hepatitis B virus from sera of southern African black adult carriers of the virus. Hepatology 25:235-240.
Li, J., S. Tong, L. Vitvitski, F. Zoulim, and C. Trepo. 1990. Rapid detection and further characterization of infection with hepatitis B virus variants containing a stop codon in the distal pre-C region. J. Gen. Virol. 71:1993-1998.
Nassal, M., M. Junker-Niepmann, and H. Schaller. 1990. Translational inactivation of RNA function: discrimination against a subset of genomic transcripts during HBV nucleocapsid assembly. Cell 63:1357-1362.
Nassal, M., and A. Rieger. 1996. A bulged region of the hepatitis B virus RNA encapsidation signal contains the replication origin for discontinuous first-strand DNA synthesis. J. Virol. 70:2764-2773.
Nielsen, H., J. Engelbrecht, S. Brunak, and G. von Heijne. 1997. Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites. Protein Eng. 10:1-6.
Paetzel, M., A. Karla, N. Strynadka, and R. Dalbey. 2002. Signal peptidases. Chem. Rev. 102:4549-4579.
Parekh, S., F. Zoulim, S. Ahn, A. Tsai, J. Li, S. Kawai, N. Khan, C. Trepo, J. Wands, and S. Tong. 2003. Genome replication, virion secretion, and e antigen expression of naturally occurring hepatitis B virus core promoter mutants. J. Virol. 77:6601-6612.
Pollack, J., and D. Ganem. 1993. An RNA stem-loop structure directs hepatitis B virus genomic RNA encapsidation. J. Virol. 67:3254-3263.
Roossinck, M., S. James, S. Loukin, and A. Siddiqui. 1986. Expression of hepatitis B viral core region in mammalian cells. Mol. Cell. Biol. 6:1393-1400.
Sen, N., F. Cao, and J. Tavis. 2004. Translation of duck hepatitis B virus reverse transcriptase by ribosomal shunting. J. Virol. 78:11751-11757.
Standring, D., J. Ou, F. Masiarz, and W. Rutter. 1988. A signal peptide encoded within the precore region of hepatitis B virus directs the secretion of a heterogeneous population of e antigens in xenopus oocytes. Proc. Natl. Acad. Sci. USA 85:8405-8409.
Sugauchi, F., H. Kamada, S. Acharya, S. Shrestha, M. Gamutan, M. Khan, R. Gish, Y. Tanaka, T. Kato, E. Orito, R. Ueda, Y. Miyakawa, and M. Mizokami. 2004. Epidemiological and sequence difference between two subtypes (Ae and Aa) of hepatitis B virus genotype A. J. Gen. Virol. 85:811-820.
Takahashi, K., S. Kishimoto, K. Ohori, H. Yoshizawa, A. Machida, H. Ohnuma, F. Tsuda, E. Munekata, Y. Miyakawa, and M. Mayumi. 1991. Molecular heterogeneity of e antigen polypeptides in sera from carriers of hepatitis B virus. J. Immunol. 147:3156-3160.
Tanaka, Y., I. Hasegawa, T. Kato, E. Orito, N. Hirashima, S. Acharya, R. Gish, A. Kramvis, M. Kew, N. Yoshihara, S. Shrestha, M. Khan, Y. Miyakawa, and M. Mizokami. 2004. A case-control study for differences among hepatitis B virus infections of genotypes A (subtypes Aa and Ae) and D. Hepatology 40:747-755.
Tang, H., and A. McLachlan. 2002. A pregenomic RNA sequence adjacent to DR1 and complementary to epsilon influences hepatitis B virus replication efficiency. Virology 303:199-210.
Tong, S., B. Brotman, J. Li, L. Vitvitski, D. Pascal, A. Prince, and C. Trepo. 1991. In vitro and in vivo replication capacity of the precore region defective hepatitis B virus variants. J. Hepatol. 13(Suppl. 4):S68-S73.
Tong, S., C. Diot, P. Gripon, J. Li, L. Vitvitski, C. Trepo, and C. Guguen-Guillouzo. 1991. In vitro replication competence of a cloned hepatitis B virus variant with a nonsense mutation in the distal pre-C region. Virology 181:733-737.
Tong, S., J. Li, L. Vitvitski, and C. Trepo. 1992. Replication capacities of natural and artificial stop codon mutants of hepatitis B virus: relevance of pregenome encapsidation signal. Virology 191:237-245.
Wang, G., and C. Seeger. 1993. Novel mechanism for reverse transcription in hepatitis B viruses. J. Virol. 67:6507-6512.
Zhang, X., F. Zoulim, F. Habersetzer, S. Xiong, and C. Trepo. 1996. Analysis of hepatitis B virus genotypes and pre-core region variability during interferon treatment of HBe antigen negative chronic hepatitis B. J. Med. Virol. 48:8-16.
Zhang, Y., and J. Summers. 1999. Enrichment of a precore-minus mutant of duck hepatitis B virus in experimental mixed infection. J. Virol. 73:3616-3622.(Michael Guarnieri, Kyun-H)
ABSTRACT
The pregenomic RNA directs replication of the hepatitis B virus (HBV) genome by serving both as the messenger for core protein and polymerase and as the genome precursor following its packaging into the core particle. RNA packaging is mediated by a stem-loop structure present at its 5' end designated the signal, which includes the core gene initiator AUG. The precore RNA has a slightly extended 5' end to cover the entire precore region and, consequently, directs the translation of a precore/core protein, which is secreted as e antigen (HBeAg) following removal of precore-derived signal peptide and the carboxyl terminus. A naturally occurring G1862T mutation upstream of the core AUG affects the bulge of the signal and generates a "forbidden" residue at the –3 position of the signal peptide cleavage site. Transfection of this and other mutants into human hepatoma cells failed to prove their inhibition of HBeAg secretion but rather revealed great impairment of genome replication. This replication defect was associated with reduced expression of core protein and could be overcome by a G1899A covariation, or by nonsense or frameshift mutation in the precore region. All these mutations antagonized the G1862T mutation on core protein expression. Cotransfection of the G1862T mutant with a replication-deficient HBV genome that provides core protein in trans also restored genome replication. Consistent with our findings in cell culture, HBV genotype A found in African/Asian patients has T1862 and is associated with much lower viremia titers than the European subgroup of genotype A.
INTRODUCTION
The hepatitis B virus (HBV) primarily infects the liver and causes chronic hepatitis, liver cirrhosis, and hepatocellular carcinoma worldwide. It is an enveloped DNA virus with a small double-stranded genome of 3.2 kb. Inside hepatocytes, viral genomic and subgenomic RNAs are transcribed from a covalently closed circular DNA template in the nucleus and exported to the cytoplasm for translation into viral proteins. The core protein assembles into the core particle, packaging both the pregenomic RNA and DNA polymerase. Subsequently, the DNA polymerase synthesizes the negative-strand DNA via reverse transcription from the RNA template, followed by RNA degradation and synthesis of positive-strand DNA. The core particle with the double-stranded DNA genome is enveloped by host-derived lipids and viral envelope proteins and secreted as an infectious virus particle (for a review, see reference 7). Since the 3.5-kb pregenomic RNA is also the mRNA for the expression of both core protein and DNA polymerase, it is the sole component required for genome replication. Consequently, increased transcription of pregenomic RNA will lead to enhanced HBV replication, as exemplified by the naturally occurring core promoter mutants (3, 20).
The AUG initiator of the core gene (position 1901 to 1903) is located approximately 80 nucleotides (nt) downstream of the 5' end of its mRNA, while the initiation codon of polymerase is 400 nt further downstream. Therefore, core protein translation can proceed directly by ribosomal scanning, whereas translation of polymerase requires a specific mechanism of translational termination and reinitiation or ribosomal shunting (9, 23). Nevertheless, the 5' end of the pregenomic RNA also functions as its encapsidation signal (the signal), which forms a stem-loop structure consisting of two base-paired regions, a 6-nt bulge, and a 6-nt loop (Fig. 1C) (10, 21, 31). The core AUG is located near the 3' end of this stem-loop as part of the lower stem. Considering that RNA secondary structure impedes the passage of the scanning 40S ribosome (12), whether translation of the core protein is regulated negatively by the signal or adjusted temporarily according to the changing functions of the pregenomic RNA remains to be determined.
The signal is not only required for the packaging of pregenomic RNA but also involved in the initiation of reverse transcription. The polymerase employs an N-terminal tyrosine residue as a primer to generate the first three nucleotides (5'-GAA-3') of the negative-strand DNA, using the UUC sequence at the 3' bulge of the signal as the template (17, 32) (Fig. 1C). Next, the negative-strand DNA is dislodged from the bulge and transferred to the UUC motif about 3.2 kb downstream in the 3' direct repeat 1 region, where reverse transcription will resume. This long-range template switch is probably facilitated by an RNA secondary structure (28).
Transcription initiation mediated by the core promoter is imprecise, and a fraction of the 3.5-kb RNA is about 30 nt longer than pregenomic RNA, thus enabling it to cover the intact precore region (positions 1814 to 1900) for an additional 29 amino acid codons (Fig. 1A). This subset of the 3.5-kb RNA, termed the precore RNA, does not express DNA polymerase and cannot serve as the pregenome because the secondary structure of the signal is melted by translating ribosome (16). The resultant precore/core protein is targeted to the lumen of the endoplasmic reticulum by its N-terminal signal peptide of 19 residues, which is clipped off by the signal peptidase (4, 22, 24) (Fig. 1B). The protein is secreted as e antigen (HBeAg) following further removal of a C-terminal basic sequence (Fig. 1A). The HBeAg is not essential for HBV replication in vitro but potentially important for the establishment of persistent infection in vivo (5, 30). Following the rise of antibody against HBeAg in hepatitis B patients, however, HBV variants with reduced or abolished expression of HBeAg often become the dominant viral species, possibly due to the negative selection for viral strains expressing HBeAg (34).
In this regard, a G1862T substitution is frequently detected in patients following seroconversion to anti-HBe (14, 33). This point mutation converts residue 17 of the precore peptide, the –3 position of the signal peptidase cleavage site, to phenylalanine. The G1862A and G1862C mutations have also been reported, albeit at much reduced frequency (14). These mutations convert the –3 position to isoleucine and leucine, respectively. Since a bulky residue such as phenylalanine, but not leucine or isoleucine, at the –3 position is considered "forbidden" for signal peptidase cleavage (18), the G1862T mutation has been speculated to reduce HBeAg expression, thus accounting for its prevalence at the anti-HBe stage of infection. The G1862T mutation also affects the bulge of the signal, at a position that is 39 nt upstream of the core gene initiator (Fig. 1C). By site-directed mutagenesis and transfection experiments, we failed to validate inhibition of HBeAg secretion by any of the substitutions introduced into the –3 position of the signal peptidase cleavage site but rather observed a severe replication defect of the corresponding mutants. This defect was associated with reduced core protein expression and could be overcome by compensatory mutations in the precore region or by core protein provided in trans.
MATERIALS AND METHODS
Mutant HBV constructs. Table 1 lists the single mutants and one double mutant used in this study. Other double and triple mutants are not listed, since they combine the properties of the single or double mutants. Mutant constructs generated for this study are based on clone N4, which is a chimera between two naturally occurring core promoter mutants of the European subgroup of genotype A. Specifically, it contains the backbone of clone 3.4 with a 0.8-kb AvrII-EcoRV fragment derived from clone 4B (11, 20). This construct maintains high replication capacity and efficient HBeAg expression as well as robust expression of HBsAg (hepatitis B surface antigen, viral envelope proteins) and secretion of virus particles. Each mutation was created by overlap extension PCR using Roche High-Fidelity PCR system (1, 2, 11, 20). The PCR products were double digested with either RsrII-BspE1 (positions 1750 to 2328) or RsrII-ApaI (positions 1750 to 2509) for cloning back to the N4 construct. The PCR-derived region of each mutant was verified by sequence analysis. Each mutant construct was converted into a tandem dimer version in the EcoRI site of the pUC18 vector (20). The plasmid DNA was purified by a commercial kit (Marligen), further extracted with phenol and chloroform/isoamyl alcohol, and dissolved in Tris-EDTA buffer. Before transfection, DNA concentrations of the constructs were verified by digestion of 500 ng of DNA with HindIII followed by gel electrophoresis.
The 4B genome lacking core protein expression (core– 4B genome) and with defective polymerase expression and nonfunctional encapsidation signal (Pol–/– 4B genome) were employed in the cotransfection experiments. They have been described previously (1, 2). The core– 4B dimer genome contains a C2044G nonsense mutation in the core gene to ablate core protein expression (1). It is still competent in the expression of polymerase and viral envelope proteins, and its pregenomic RNA can be encapsidated if core protein is provided by another plasmid. The Pol–/– 4B genome contains a C2589T nonsense mutation in the polymerase gene to prevent polymerase expression as well as a G1879T/T1880A double mutation in the loop of the signal, thus abolishing packaging of pregenomic RNA (2).
Transfection. The Huh7 human hepatoma cells were cultured in Dulbecco's modified Eagle medium supplemented with 10% calf serum. The HBV dimers were transfected into Huh7 cells via the TransIT Transfection system (Mirus) as described previously (2). Serum-free minimal essential medium (150 μl) was mixed with 4 μl of transit-LT1 reagent, vortexed, and incubated at room temperature for 15 min. Following mixing with 1.5 μg of DNA and a further incubation for 15 min, the complex was added dropwise to cells grown in six-well plates. After overnight incubation, cells were washed, replenished with fresh medium, and harvested 4 days later. For the transcomplementation assay, 1.1 μg of HBV mutants generated in this study was cotransfected with 0.4 μg of either pcDNA3.1/zeo(–) vector, core– 4B dimer, or Pol–/– 4B dimer.
Analysis of HBV DNA replication and virion secretion. The details of these assays have been described previously (1, 2, 11, 20). Briefly, transfected plasmid DNA was eliminated from cell lysate by nuclease treatment, and core particles were precipitated by polyethylene glycol solution. Core particles were disintegrated by digestion with proteinase K in the presence of sodium dodecyl sulfate (SDS), and HBV DNA was extracted with phenol, precipitated with ethanol, and dissolved in Tris-EDTA buffer. DNA was separated on a 1.5% agarose gel, with ethidium bromide (1 μg/ml) present in both the gel and the running buffer. After transfer to a membrane, HBV DNA was detected with a randomly primed probe. Extracellular viral particles (both naked core particles and enveloped Dane particles) were concentrated by ultracentrifugation through a 20% sucrose cushion, and DNA was extracted for Southern blot analysis.
Measurement of HBeAg and HBsAg expression. HBeAg was measured by the EBK 125I radioimmunoassay (DiaSorin) in earlier experiments using 4 to 10 μl of culture supernatant, and by the ETI-EBK plus enzyme immunoassay (DiaSorin) in later experiments when the radioimmunoassay kit was discontinued. HBsAg was measured by the Auszyme monoclonal HBsAg kit (Abbot Laboratories) with 4 to 10 μl of culture supernatant. The volumes of culture supernatants used will not cause signal saturation.
Western blot analysis of core protein, HBsAg, and HBeAg. Huh7 cells grown in each well of the six-well plates were scraped at day 5 posttransfection, and the cell pellet was lysed with 100 μl of buffer (10 mM HEPES, pH 7.5, 100 mM NaCl, 1 mM EDTA, 1% NP-40) supplemented with protease inhibitor cocktail (Roche). Proteins from 40 μl of lysate were separated on a 0.1% SDS-12% polyacrylamide gel and transferred to polyvinylidene difluoride membranes. The membranes were blocked at room temperature with 3% bovine serum albumin (BSA) in Tris-buffered saline-Tween (TBST) buffer, and incubated at 4°C for 3 overnights with a mouse monoclonal anti-core antibody (14E11; Abcam) diluted 1:2,000 in 3% BSA-TBST. After 40 min of washing at room temperature with TBST buffer, the blots were incubated at room temperature for 1 h with horseradish peroxidase-conjugated anti-mouse antibodies (Amersham) diluted 1:30,000 in 3% BSA-TBST. The blots were washed with TBST for 40 min, and signals were revealed by enhanced chemiluminescence (ECL). To control for transfection efficiency, the blots were stripped with Restore Western Blot stripping buffer (Pierce) at 37°C for 20 min, rinsed with water and TBST buffer, and blocked again with 3% BSA in TBST buffer. The blots were incubated at 4°C overnight with a goat polyclonal anti-HBs antibody (Dako) diluted 1:8,000 in 3% BSA-TBST, washed, and incubated at room temperature for 1 h with horseradish peroxidase-conjugated anti-goat antibodies (Amersham) diluted 1:50,000 in TBST. The blots were washed again, and signals were revealed by ECL. Alternatively, HBsAg was detected by a 1:5,000 dilution of a polyclonal horse anti-HBs (Ad/Ay) antibody (Abcam), and a 1:5,000 dilution of rabbit anti-horse immunoglobulin G conjugated to horseradish peroxidase (Abcam).
For immunoprecipitation of secreted HBeAg, culture supernatants collected at day 5 posttransfection were precleared of cell debris by centrifugation at 3,000 rpm for 10 min. A 1-ml aliquot was incubated overnight at 4°C with 0.5 μl of anti-core antibody (Dako), followed by addition of 5 μl of protein G beads and further incubation for 3 to 4 h. The immune complex was collected by low-speed centrifugation, washed once with TBST buffer, and separated in 15% polyacrylamide gels. The protein blot was blocked with 3% BSA-TBTS and incubated overnight at 4°C with a 1:2,000 dilution of rabbit polyclonal antibody against core antigen (Dako). After washing with TBST buffer, the blot was incubated with a 1:2,000 dilution (in 3% BSA-TBST) of anti-rabbit antibodies conjugated with horseradish peroxidase (Amersham). The blot was washed with TBST, and signals were revealed by ECL.
RESULTS
Mutations at the –3 position of the signal peptidase cleavage site do not reduce HBeAg secretion. The amino terminus of HBeAg has been mapped to S20 of the precore peptide, thus establishing A19 and V17 as the –1 and –3 positions, respectively (Fig. 1B) (24, 26). This finding is consistent with the preference of the endoplasmic reticulum (ER) signal peptidase for small and neutral amino acids at the –1 and –3 positions (18). Bulky or charged residues at these two positions often impair signal peptide processing (6). The V17F substitution in the precore peptide (pcV17F) as a result of the G1862T mutation generates a bulky residue at the –3 position, whereas the pcV17L substitution does not. We generated these two naturally occurring variants as well as the artificial pcV17D and pcA19D mutants, which contain charged residues at the –3 and –1 positions, respectively. The DNA constructs were transfected to Huh7 cells, and HBeAg at day 5 posttransfection was measured with a commercial radioimmunoassay. To correct for variations in transfection efficiency, the HBeAg values were normalized against HBsAg values (Fig. 2A). No reduction in the secretion of HBeAg was observed in association with any mutation at the –3 position (results of three independent experiments are shown in Fig. 2B). The pcA19D mutation at the –1 site incurred, at most, modest (20 to 30%) reduction of HBeAg expression. Similar results were obtained when culture supernatants were harvested at earlier time points (data not shown). Results obtained by radioimmunoassay were further validated by immunoprecipitation-Western blot analysis. Since the E77Q mutation present in the core gene of N4-based constructs greatly reduces the sensitivity of HBeAg detection by a commercial polyclonal antibody (Dako) (K. Kim and S. Tong, unpublished), we introduced a Q77E mutation back into these constructs. As a result, no significant change in either the amount or sizes of HBeAg was observed, even when both pcV17F and pcA19D mutations were present (Fig. 2C).
The G1862T, T1863A, and C1869A mutants are impaired in genome replication. The pcV17F, pcV17L, pcV17D, and pcA19D mutations in the precore peptide correspond to G1862T, G1862C, T1863A, and C1869A mutations at the nucleotide level (Table 1). Of these, the T1863A mutation changes the 3'-most position of the priming site for reverse transcription, whereas the C1869A mutation disrupts a base pair in the upper stem of the signal (Fig. 3A). Since the signal is essential for the packaging of pregenomic RNA and, hence, genome replication (10, 21, 31), we analyzed the replication capacity of the mutants. Figure 3C shows Southern blot analysis of HBV DNA associated with intracellular core particles and extracellular viral particles. The transfection efficiencies were similar among the various constructs, as indicated by values of HBsAg in the culture supernatant (Fig. 3B). DNA replication was severely impaired not only for the T1863A and C1869A mutants but also for the G1862T mutant (Fig. 3C). Combining the G1862T and C1869A mutations led to nearly complete loss of replication capacity.
Introduction of the G1899A mutation overcomes the replication defect of these mutants. The G1862T mutation is frequently associated with a G1888A or G1899A covariation (8, 14, 25, 27, 33). While the G1888A mutation is silent at the amino acid level, the G1899A mutation induces a pcG29D substitution in the terminal residue of the precore peptide (Table 1). Interestingly, both mutations convert a wobble U:G base pair in the signal into a U:A pair (upper stem for G1888A and lower stem for G1899A) (Fig. 4A). Mutagenesis and transfection experiments revealed that, while the G1888A mutant maintained the wild-type level of genome replication and virion secretion, introduction of the G1862T mutation into this mutant severely impaired DNA replication (Fig. 4B). The G1899A mutation modestly increased genome replication, and the G1899A/G1862T double mutant replicated much better than the G1862T single mutant. The G1899A mutation could also markedly enhance the replication of the T1863A, C1869A, and G1862C mutants (Fig. 5), suggesting a general mechanism of rescue.
Replication impact of the G1899A mutation is independent of its improvement of base pairing in the signal. As already mentioned, the G1899A mutation converts a U:G pair in the lower stem of the signal, between U1855 and G1899, into a U:A pair. In this regard, a T1855C mutation will convert the same U:G pair into an even stronger C:G pair. Transfection experiments revealed wild-type replication capacity of the T1855C single mutant but much reduced replication of the T1855C/G1862T double mutant reminiscent of the G1862T single mutant (Fig. 4B). On the other hand, the T1855C/G1899A double mutant retained a high replication capacity, despite the disruption of the base pair involved. Furthermore, introduction of the G1862T mutation into this double mutant did not markedly reduce genome replication (Fig. 4B, last lane). Thus, the ability of the G1899A mutation to rescue replication of the G1862T mutant is not mediated by, and is independent of, stronger base pairing in the lower stem of the signal.
Since G1899 occupies the –2 position of core gene AUG codon (Fig. 1C), the G1899A mutation may function through modulation of core protein expression. In this regard, a cytosine at the –2 position is considered optimal for translation initiation (13). We therefore generated the G1899C mutant alone or together with G1862T. Indeed, the G1899C/G1862T double mutant retained relatively high replication capacity, much better than the G1862T single mutant (Fig. 6).
Nonsense and frameshift mutations in the precore region could also rescue replication of the G1862T mutant. Incidentally, we found that replication of the G1862T mutant could be rescued by mutations at other positions. A C1865T mutation did not markedly reduce genome replication, despite its location at the priming site for reverse transcription, and the C1865T/G1862T double mutant was found to replicate as efficiently (Fig. 6). Interestingly, the C1865T mutation converts precore codon 18 into a TAA termination codon, which has the potential to increase core protein expression by translational termination and reinitiation from precore RNA (29). To verify the significance of the nonsense mutation, we created a missense mutant of the same codon, C1865G (pcQ18E at the amino acid level). Although the C1865G single mutant had only a modest decline in replication, the C1865G/G1862T double mutant was severely impaired in DNA replication. On the other hand, a frameshift mutation (1840insC) introduced into the precore region upstream of the signal, which would terminate precore-core protein translation in the vicinity of core AUG, could rescue the replication of the G1862T mutant as well (Fig. 6).
Replication capacities of the mutants correlate well with their levels of core protein expression. The ability of the G1899A mutation to rescue replication of several mutants, its juxtaposition with core gene AUG codon, and the ability of precore nonsense and frameshift mutations to rescue the G1862T mutant strongly suggested coupling of core protein expression with the replication capacity of the constructs. Indeed, Western blot analysis revealed reduced core protein expression by the G1862T mutation (Fig. 7A, upper panel). In contrast, the G1899A mutation enhanced core protein expression. This was not due to higher transfection efficiency of the G1899A mutant, as revealed by reprobing of the same blot with anti-HBs antibodies (lower panel). The presence of both mutations resulted in core protein expression similar to that of the parental clone, N4. Introduction of the G1899A mutation could also override the low core protein expression associated with the T1863A, G1862C, or C1869A mutant (Fig. 7B). Similarly, other mutations capable of rescuing the G1862T mutant, such as C1865T and G1899C, could maintain core protein expression when combined with the G1862T mutation. In contrast, the T1855C, C1865G, or G1888A mutation could not rescue core protein expression when coupled with the G1862T mutation (Fig. 7A, upper panel).
Replication of the G1862T, T1863A, and C1869A mutants could be rescued by core protein provided in trans. If the G1862T, T1863A, and C1869A mutants harbor no defect other than core protein expression, then providing core protein from another plasmid should suffice to relieve their replication block. To test this hypothesis, constructs with normal, increased, or reduced replication capacity were cotransfected with either pcDNA vector, core– 4B dimer, or Pol–/– 4B dimer, using a ratio of 1.1 μg/0.4 μg. Clone 4B is a core promoter mutant with extremely high levels of genome replication (20). The core– 4B dimer was derived from 4B by a nonsense mutation in the core gene, whereas the Pol–/– 4B dimer contains a nonsense mutation in the polymerase gene as well as a double mutation in the loop of the signal that abolishes packaging of pregenomic RNA. The Pol–/– 4B dimer produced high levels of core protein even at this ratio of cotransfection (data not shown). As illustrated in Fig. 8A, the low replicating G1862T single or double mutants were not rescued by cotransfection with a core– 4B dimer but by the Pol–/– 4B dimer, such that their differences with high replicating clones were largely abolished. The Pol–/– 4B dimer could also rescue the replication of the T1863A and C1869A mutants and even the G1862T/C1869A double mutant, which had negligible genome replication (Fig. 8B). On the other hand, the T1880G mutant with a nonfunctional signal (2) could rescue the core– 4B dimer but not the Pol–/– 4B dimer. Therefore, the defect of the G1862T, T1863A, and C1869A mutants indeed lies at core protein expression rather than the pregenome or DNA polymerase.
DISCUSSION
Effect of mutations at the –3 and –1 positions of the signal peptide cleavage site on HBeAg secretion. The N terminus of the precore peptide serves as a signal peptide for the cotranslational translocation of the precore-core protein to the ER lumen, leading to HBeAg secretion (4, 22, 24). A signal peptide is composed of an N-terminal basic sequence, a central hydrophobic region for membrane insertion, and a polar but noncharged C terminus for cleavage (18, 19). Residues –1 and –3 interact with specific P1 and P3 pockets in the ER signal peptidase and are critical for cleavage. Aromatic and charged residues introduced into these positions often abolish cleavage or alter the cleavage site (6). Molecular epidemiological studies have revealed, with the exception of the African/Asian subgroup of genotype A (genotype Aa), the rise of the G1862T mutation only at the anti-HBe stage of infection (14, 25, 27, 33). The corresponding V17F substitution in the precore peptide generates a "forbidden" residue at the –3 position. In the current study, we found that HBeAg secretion was not impaired in constructs with forbidden residues at the –3 position (pcV17F and pcV17D) relative to the wild-type virus or the pcV17L variant (Fig. 2). The pcA19D mutation at the –1 position only mildly impaired HBeAg production. Most strikingly, the pcV17F/pcA19D double mutant displayed no further reduction in HBeAg secretion. However, computer predictions of the double mutant suggested a possible shift of the cleavage site from residue 19 to residue 18, which would put the pcV17F and pcA19D mutations at the flexible –2 and +1 positions, with pcQ18 and pcT16 now occupying the –1 and –3 positions. Similarly, introduction of forbidden amino acids into the –1 positions caused a shift of the cleavage site for the signal peptide of Escherichia coli maltose-binding protein (6). At any rate, our observations suggest that single and even double mutations near the signal peptide cleavage site may not constitute an effective means of terminating HBeAg secretion. These findings are consistent with the fact that genotype Aa, which often has T1862 as its wild-type sequence, continues to express HBeAg (25, 27).
Do mutations at nucleotides 1862, 1863, and 1869 just down regulate core protein expression The G1862T, T1863A, and C1869A mutants are hampered in genome replication and core protein expression. In this regard, HBV genome replication is driven by the pregenomic RNA, which serves as the pregenome and the messenger for core protein and polymerase. Our preliminary primer extension analysis of total cellular RNA revealed no reduction in the level of pregenomic RNA in cells transfected with the G1862T mutant, although we do not know whether expression of polymerase is impaired. The ability of the Pol–/– 4B dimer to rescue replication of all three of the mutants (Fig. 8) argues against a defect in polymerase expression and suggests intact function of the signal (with the exception of T1863A mutation, see below). Alternatively, one may argue that mutations at positions 1862, 1863, and 1869 do reduce polymerase expression, yet DNA polymerase is not the limiting factor for RNA packaging. However, the Pol–/– 4B dimer did not markedly enhance replication of clone N4, the parental construct.
Nucleotides 1865 to 1863 serve as the priming site for reverse transcription (Fig. 1) (17). In this regard, the T1863A mutant produced progeny viral DNA of reduced size (Fig. 3 and 5), suggesting aberrant translocation leading to shorter products. No aberrant products were observed for the C1865G or C1865T mutant. These findings are consistent with the greater importance of nucleotide 1863 than 1865 in determining the specificity of the template switch. Nucleotide 1863 defines the 3'-most nucleotide of the nascent minus DNA strand (5'-GAA-3') (Fig. 1C). The T1863A mutation will produce a minus DNA strand of 5'-GAT-3'. The mismatch of its 3'-most nucleotide with direct repeat 1 is very likely to prevent DNA elongation, thus promoting annealing elsewhere to generate shorter DNA products.
How do mutations at positions 1862, 1863, 1869, and 1899 modulate core protein expression The G1899A mutation enhances core protein expression. Although this mutation affects a base pair in the lower stem of the signal, we demonstrated that its biological properties were not mediated by, and is independent of, enhanced stability of the signal. Thus, its effect may be related to its occupation of the –2 position of core gene AUG codon. Indeed, the G1899C mutation of the same position could also overcome the replication defect of the G1862T mutant. The G1862T, T1863A, and C1869A mutations affect the –39, –38, and –32 positions of core gene AUG. These mutations are located at the bulge (G1862T, T1863A) and upper stem (C1869A) of the signal, respectively. Whether they work through structural changes to the signal remains unclear. It is unknown whether the stem-loop structure of the signal is formed during core protein translation, and if so, whether it facilitates or hinders core protein expression. The signal binds viral polymerase through the loop, upper stem, and part of the bulge (nucleotides 1860 and 1861).
Clinical relevance of the replication effect of G1862T and G1899A mutations. Genotype Aa patients manifest much lower viremia titers than patients of genotype Ae (European genotype A) (27). For HBeAg-positive patients, the mean viremia titer was 5 x 105 genome copies/ml in genotype Aa patients compared to 4 x 108 copies/ml in genotype Ae patients, with a difference of 800-fold. In the HBeAg-negative group of patients, the value was 7 x 102 versus 6 x 103 copies/ml. Based on the markedly down regulation of genome replication by the G1862T mutation, we propose that T1862 contributes to the reduced replication capacity of genotype Aa. Since not all genotype Aa isolates contain T1862, it will be of great interest to correlate the viremia titers in genotype Aa patients with polymorphism at this position (T1862 versus G1862).
The G1899A mutation is highly prevalent at the anti-HBe stage of chronic infection, often subsequent to the G1886A nonsense mutation in the precore region that abolishes HBeAg expression (15). It is also common in patients containing the G1862T mutation (14, 33). We found that the G1899A mutation largely relieves the inhibitory effect of the G1862T mutation on genome replication, as a consequence of restored core protein expression. The G1862T mutant of genotype B has been implicated in several cases of fulminant hepatitis in China (8). Interestingly, all such isolates harbored the G1899A mutation but no G1896A nonsense mutation or core promoter mutations (8). Since enhanced replication of the HBV genome is considered critical in the induction of fulminant hepatitis (3), the presence of the G1899A mutation in such strains may not be a mere coincidence.
ACKNOWLEDGMENTS
This work was supported by grants AI54535, DK62857, p20RR15578, CA35711, DK66950, and CA109733 from the National Institutes of Health and from Life Span Research Funds.
REFERENCES
Ahn, S., A. Kramvis, S. Kawai, H. Spangenberg, J. Li, G. Kimbi, M. Kew, J. Wands, and S. Tong. 2003. Sequence variation upstream of precore translation initiation codon reduces hepatitis B virus e antigen production. Gastroenterology 125:1370-1378.
Bang, G., K. Kim, M. Guarnieri, F. Zoulim, S. Kawai, J. Li, J. Wands, and S. Tong. 2005. Effect of mutating the two cysteines required for HBe antigenicity on hepatitis B virus DNA replication and virion secretion. Virology 332:216-224.
Baumert, T., S. Rogers, K. Hasegawa, and T. Liang. 1996. Two core promoter mutations identified in a hepatitis B virus strain associated with fulminant hepatitis result in enhanced viral replication. J. Clin. Investig. 98:2268-2276.
Bruss, V., and W. Gerlich. 1988. Formation of transmembraneous hepatitis B e-antigen by cotranslational in vitro processing of the viral precore protein. Virology 163:268-275.
Chen, H., M. Kew, W. Hornbuckle, B. Tennant, P. Cote, J. Gerin, R. Purcell, and R. Miller. 1992. The precore gene of the woodchuck hepatitis virus genome is not essential for viral replication in the natural host. J. Virol. 66:5682-5684.
Fikes, J., G. Barkocy-Gallagher, D. Klapper, and P. Bassford. 1990. Maturation of Escherichia coli maltose-binding protein by signal peptidase I in vivo. J. Biol. Chem. 265:3417-3423.
Ganem, D., and R. Schneider. 2001. Hepadnaviridae: the viruses and their replication, p. 2923-2970. In D. Knipe and P. Howley (ed.), Fields virology, 3rd ed. Lippincott-Raven Publishers, Philadelphia, Pa.
Hou, J., Y. Lin, J. Waters, Z. Wang, J. Min, H. Liao, J. Jiang, J. Chen, K. Luo, and P. Karayiannis. 2002. Detection and significance of a G1862T variant of hepatitis B virus in Chinese patients with fulminant hepatitis. J. Gen. Virol. 83:2291-2298.
Jean-Jean, O., T. Weimer, A. de Recondo, H. Will, and J. Rossignol. 1989. Internal entry of ribosomes and ribosomal scanning involved in hepatitis B virus P gene expression. J. Virol. 63:5451-5454.
Junker-Niepmann, M., R. Bartenschlager, and H. Schaller. 1990. A short cis-acting sequence is required for hepatitis B virus pregenome encapsidation and sufficient for packaging of foreign RNA. EMBO J. 9:3389-3396.
Khan, N., M. Guarnieri, S. H. Ahn, J. Li, Y. Zhou, G. Bang, K. H. Kim, J. R. Wands, and S. Tong. 2004. Modulation of hepatitis B virus secretion by naturally occurring mutations in the S gene. J. Virol. 78:3262-3270.
Kozak, M. 1986. Influence of mRNA secondary structure on initiation by eukaryotic ribosomes. Proc. Natl. Acad. Sci. USA 83:2850-2854.
Kozak, M. 1987. At least six nucleotides preceding the AUG initiator codon enhance translation in mammalian cells. J. Mol. Biol. 196:947-950.
Kramvis, A., S. Bukofzer, M. Kew, and E. Song. 1997. Nucleic acid sequence analysis of the precore region of hepatitis B virus from sera of southern African black adult carriers of the virus. Hepatology 25:235-240.
Li, J., S. Tong, L. Vitvitski, F. Zoulim, and C. Trepo. 1990. Rapid detection and further characterization of infection with hepatitis B virus variants containing a stop codon in the distal pre-C region. J. Gen. Virol. 71:1993-1998.
Nassal, M., M. Junker-Niepmann, and H. Schaller. 1990. Translational inactivation of RNA function: discrimination against a subset of genomic transcripts during HBV nucleocapsid assembly. Cell 63:1357-1362.
Nassal, M., and A. Rieger. 1996. A bulged region of the hepatitis B virus RNA encapsidation signal contains the replication origin for discontinuous first-strand DNA synthesis. J. Virol. 70:2764-2773.
Nielsen, H., J. Engelbrecht, S. Brunak, and G. von Heijne. 1997. Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites. Protein Eng. 10:1-6.
Paetzel, M., A. Karla, N. Strynadka, and R. Dalbey. 2002. Signal peptidases. Chem. Rev. 102:4549-4579.
Parekh, S., F. Zoulim, S. Ahn, A. Tsai, J. Li, S. Kawai, N. Khan, C. Trepo, J. Wands, and S. Tong. 2003. Genome replication, virion secretion, and e antigen expression of naturally occurring hepatitis B virus core promoter mutants. J. Virol. 77:6601-6612.
Pollack, J., and D. Ganem. 1993. An RNA stem-loop structure directs hepatitis B virus genomic RNA encapsidation. J. Virol. 67:3254-3263.
Roossinck, M., S. James, S. Loukin, and A. Siddiqui. 1986. Expression of hepatitis B viral core region in mammalian cells. Mol. Cell. Biol. 6:1393-1400.
Sen, N., F. Cao, and J. Tavis. 2004. Translation of duck hepatitis B virus reverse transcriptase by ribosomal shunting. J. Virol. 78:11751-11757.
Standring, D., J. Ou, F. Masiarz, and W. Rutter. 1988. A signal peptide encoded within the precore region of hepatitis B virus directs the secretion of a heterogeneous population of e antigens in xenopus oocytes. Proc. Natl. Acad. Sci. USA 85:8405-8409.
Sugauchi, F., H. Kamada, S. Acharya, S. Shrestha, M. Gamutan, M. Khan, R. Gish, Y. Tanaka, T. Kato, E. Orito, R. Ueda, Y. Miyakawa, and M. Mizokami. 2004. Epidemiological and sequence difference between two subtypes (Ae and Aa) of hepatitis B virus genotype A. J. Gen. Virol. 85:811-820.
Takahashi, K., S. Kishimoto, K. Ohori, H. Yoshizawa, A. Machida, H. Ohnuma, F. Tsuda, E. Munekata, Y. Miyakawa, and M. Mayumi. 1991. Molecular heterogeneity of e antigen polypeptides in sera from carriers of hepatitis B virus. J. Immunol. 147:3156-3160.
Tanaka, Y., I. Hasegawa, T. Kato, E. Orito, N. Hirashima, S. Acharya, R. Gish, A. Kramvis, M. Kew, N. Yoshihara, S. Shrestha, M. Khan, Y. Miyakawa, and M. Mizokami. 2004. A case-control study for differences among hepatitis B virus infections of genotypes A (subtypes Aa and Ae) and D. Hepatology 40:747-755.
Tang, H., and A. McLachlan. 2002. A pregenomic RNA sequence adjacent to DR1 and complementary to epsilon influences hepatitis B virus replication efficiency. Virology 303:199-210.
Tong, S., B. Brotman, J. Li, L. Vitvitski, D. Pascal, A. Prince, and C. Trepo. 1991. In vitro and in vivo replication capacity of the precore region defective hepatitis B virus variants. J. Hepatol. 13(Suppl. 4):S68-S73.
Tong, S., C. Diot, P. Gripon, J. Li, L. Vitvitski, C. Trepo, and C. Guguen-Guillouzo. 1991. In vitro replication competence of a cloned hepatitis B virus variant with a nonsense mutation in the distal pre-C region. Virology 181:733-737.
Tong, S., J. Li, L. Vitvitski, and C. Trepo. 1992. Replication capacities of natural and artificial stop codon mutants of hepatitis B virus: relevance of pregenome encapsidation signal. Virology 191:237-245.
Wang, G., and C. Seeger. 1993. Novel mechanism for reverse transcription in hepatitis B viruses. J. Virol. 67:6507-6512.
Zhang, X., F. Zoulim, F. Habersetzer, S. Xiong, and C. Trepo. 1996. Analysis of hepatitis B virus genotypes and pre-core region variability during interferon treatment of HBe antigen negative chronic hepatitis B. J. Med. Virol. 48:8-16.
Zhang, Y., and J. Summers. 1999. Enrichment of a precore-minus mutant of duck hepatitis B virus in experimental mixed infection. J. Virol. 73:3616-3622.(Michael Guarnieri, Kyun-H)