Critical role of the 36-nucleotide insertion in hepatitis B virus genotype G in core protein expression, genome replication, and virion secretion

Abstract

Frequent coinfection of hepatitis B virus genotype G with genotype A suggests that genotype G may require genotype A for replication or transmission. In this regard, genotype G is unique in having a 12-amino-acid extension in the core protein due to a 36-nucleotide insertion near the core gene translation initiation codon. The insertion alters base pairing in the lower stem of the pregenome encapsidation signal, which harbors the core gene initiator, and thus has the potential to affect both core protein translation and pregenomic RNA encapsidation. Genotype G is also unusual for possessing two nonsense mutations in the precore region, which together with the core gene encode a secreted nonstructural protein called hepatitis B e antigen (HBeAg). We found that genotype G clones were indeed incapable of HBeAg expression but were competent in RNA transcription, genome replication, and virion secretion. Interestingly, the 36-nucleotide insertion markedly increased the level of core protein, which was achieved at the level of protein translation but did not involve alteration in the mRNA level. Consequently, the variant core protein was readily detectable in patient blood. The 12-amino-acid insertion also enhanced the genome maturity of secreted virus particles, possibly through less efficient envelopment of core particles. Cotransfection of genotypes G and A did not lead to mutual interference of genome replication or virion secretion. Considering that HBeAg is an immunotolerogen required for the establishment of persistent infection, its lack of expression rather than a replication defect could be the primary determinant for the rare occurrence of genotype G monoinfection.

FIG. 1.

(A) Predicted secondary structure of the ɛ signal for genotypes A and G. The core gene translation initiation codon and the 5′ end of the HBV sequence in the CMV-core constructs (see panel B) are indicated. Also shown for genotype G are the first 3 nt of the 36-nt insertion, the double nucleotide change at the loop that prevents pg RNA encapsidation (ɛ⁻ mutant), and a G to A change of the initiation codon that abolishes core protein translation (core⁻ mutant). (B) HBV genes, mRNAs, and constructs used for the present study. pc, precore; pg, pregenomic; sg, subgenomic; pol, polymerase; L, M, S, large, middle, and small envelope proteins, respectively. There are three in-frame translation initiation sites in the pre-S/S gene, leading to the expression of L protein from the 2.4-kb sg RNA and M and S proteins from the 2.1-kb sg RNA. The two initiation sites in the precore/core gene lead to the expression of HBeAg from pc RNA and core protein/polymerase from the slightly shorter pg RNA. For the same reason, CMV-core constructs produce core protein, whereas CMV-precore constructs generate HBeAg. The 1.1-mer constructs cannot express HBeAg, because the 3.5-kb mRNA produced mimics pg RNA, not pc RNA.

FIG. 2.

Identification of hybridization conditions for unbiased detection of both genotypes A and G. pTriExD1 and pTriExG1 were linearized by XhoI digestion, while an EcoRI dimer of genotype A (clone 4B) in the pUC18 vector was linearized with HindIII. Increasing amounts of digested DNA (30 pg, 100 pg, 300 pg, and 1,000 pg) were separated in an agarose gel, and the blot was hybridized successively with probes of genotype A, genotype G, and A/G at a 1:1 ratio. The blots were washed at 60°C for a total of 2 h with high-salt (2× SSC-0.1% SDS) or low-salt (0.5× SSC-0.1% SDS) buffer.

FIG. 3.

HBV RNA transcription in transfected Huh7 cells. The 1.1-mers were cloned in the pTriEx vector with transcription of the 3.5-kb pg RNA under control of the chicken actin promoter. The dimers were cloned into pUC18 vector, with transcription of all viral mRNAs under endogenous promoters. Huh7 cells were cotransfected with HBV constructs and SEAP cDNA, and 8 μg of RNA extracted at day 3 posttransfection was separated in a denaturing agarose gel. (A) Hybridization of the blot with a mixed probe of genotypes A and G. (B) Hybridization of the same blot with a SEAP probe following stripping of the HBV probe.

FIG. 4.

Core protein expression from 1.1-mer genomes (A), tandem dimers (B), and CMV-core expression constructs (C). Cells were harvested at day 5 posttransfection, and core protein was detected from lysates by a rabbit polyclonal antibody (Dako), except for clone 4B and its derivatives (panels B and C, right). For these samples, a mouse monoclonal antibody was used instead. The predicted core protein size is 183 aa for genotypes C and D, 185 aa for genotype A, and 195 aa for genotype G. For panel C, two independent clones of G1d36 and 4Bins36 were analyzed. The lower part of panel A shows intracellular core particles that have been separated in a native agarose gel and probed with the rabbit polyclonal antibody.

FIG. 5.

(A to C) Relative contribution of the 36-nt insertion in the genome versus the 12-aa insertion in the core protein to HBV DNA replication and virion secretion. The 1.1-mer genomes were used. The core-negative (core⁻) mutants of G1 and G1d36 (1 μg) were cotransfected with same amount of core-expressing constructs: ɛ-negative (ɛ⁻) mutants or CMV-core (lanes 7 to 14). For controls, 1 μg of the ɛ-negative mutants, CMV-core constructs, or parental clones was transfected with 1 μg of pcDNA3.1 vector DNA (lanes 1 to 6). About half of the cell lysate was used for detection of DNA replication (A), while another aliquot was used for successive detection of core protein and GAPDH (C). The supernatant was employed for detection of secreted virus particles (B). The EcoRI/RsrII digest of an EcoRI dimer served as size markers of 3.2, 1.7, and 1.5 kb. RC, relaxed circular; PDS, partially double stranded; SS, single stranded. (D) Impact of the 36-nt/12-aa insertion on release of HBV as virions or naked core particles. Pooled supernatants from transfected Huh7 cells were divided into two equal parts, one for immunoprecipitation with anti-S antibody and another for precipitation with anti-core antibody. The env⁻ mutant of genotype A did not express envelope proteins and therefore failed to secrete virions.

FIG. 6.

The 36-nt insertion in the core gene increased core protein translation without up-regulating its steady-state mRNA level. (A) Pulse-chase experiment. Huh7 cells were transfected with the indicated constructs and labeled with [³⁵S]methionine for 3 h. One set of the samples was harvested immediately, while another set was cultured for an additional 68 h in the absence of labeled methionine. Core protein was immunoprecipitated from cell lysates and revealed by SDS-PAGE and fluorography. Positions of prestained molecular size markers are also indicated (they underestimate the sizes of the endogenous proteins by at least 10%). The high-molecular-weight bands unique to pTriex constructs have the size of core protein dimers. (B) Primer extension assay. RNA (10 μg) extracted from day 3 posttransfection was annealed with an end-labeled antisense primer of genotype A (lanes 2 to 8) or genotype G (lanes 9 to 13), and the negative-stranded cDNA was synthesized with AMV Reverse Transcriptase. The product was heated and separated in a 5% denaturing acrylamide gel. End-labeled, HaeIII-digested φX174 DNA served as molecular size markers. The predicted sizes of the primer extension products are indicated.

FIG. 7.

HBV DNA replication and virion secretion from the 1.1-mer constructs (A) and dimer constructs (B and C). Cells and culture supernatants were harvested at day 5 posttransfection. Core particles extracted from cell lysates were used for the detection of DNA replication, while virus particles were immunoprecipitated from culture supernatants with anti-S antibodies prior to Southern blot analysis. For the lower part of panel B, virus particles secreted from clone 2A contained primarily single-stranded genome. For panel C, SphI dimers were cotransfected with CMV-core constructs at a 1:1 ratio. In addition to DNA replication and virion secretion (upper and middle panels), core protein expression was also monitored (lower panel). The rabbit polyclonal antibody used (Dako) failed to detect core protein expressed from clone 4B.

FIG. 8.

Lack of significant interference between genotypes A and G with regard to genome replication and virion secretion. (A) Cartoon view of the EcoRI site of genotype A and the BglI/XhoI sites of genotype G. Shown are linear and RC forms with the single-stranded region repaired. EcoRI digestion converts the linear form of genotype A into two bands of 1.8 kb and 1.4 kb, whereas double digestion with BglI and XhoI produces 2.2-kb and 1-kb bands for the linear form of genotype G. The heating step prior to electrophoresis melts the base pairing between the 5′ ends of the positive and negative strands in the RC DNA, thus generating the same migration pattern as the linear DNA. The numbers 1 and 2 inside boxes refer to DR1 and DR2 sequences. (B) HBV DNA associated with intracellular core particles. (C) HBV DNA of extracellular virions. Huh7 cells grown in 10-cm dishes were transfected with a total of 10 μg of 1.1-mer genomes at the ratios indicated. DNA harvested at day 5 posttransfection was treated with Klenow fragment in the presence of dNTP. A one-third aliquot of the DNA was digested with EcoRI, while another aliquot with treated with BglI and XhoI. Digested and undigested DNA was heated at 85°C for 10 min before Southern blot analysis with mixed A/G probe. For intracellular DNA (B), the smaller fragments of EcoRI and BglI/XhoI digestion (1.3 kb and 1.0 kb) were less distinct.

FIG. 9.

Core protein and HBeAg in the culture supernatant of transfected Huh7 cells (A to C) and patient blood (D). (A) “HBeAg” from cells transfected with 1.1-mer viral genomes. “HBeAg” was measured from 15 μl of culture supernatant harvested at day 5 posttransfection. Values (with A1 arbitrarily set at 1) were averaged from three transfection experiments. (B) Immunoprecipitation-Western blot analysis of core protein and HBeAg secreted to the culture supernatant. Clones K81 and 5.4 were defective in HBeAg expression due to frameshift mutations in the precore region and the core gene, respectively. (C) Impact of the 36-nt insertion on HBeAg expression from two genotype A clones as an EcoRI dimer (4B) or CMV-precore construct (6.2). The left and middle panels show HBeAg values of the insertion mutants relative to the parental constructs, which were arbitrarily set at 1. Results are based on three transfection experiments. The right panel shows immunoprecipitation-Western blot analysis. (D) Immunoprecipitation-Western blot analysis of core protein and HBeAg in sera of patients infected with genotype A, G, or both. The culture supernatant of transfected Huh7 cells (sup) served as a positive control. The HBeAg values shown at the bottom were measured from 1.5 μl of serum sample or culture supernatant.