Fig 1.
siRNA-knockdown screening for ESCRT factors required for HBV replication in HepG2 cells.
(A) A cartoon illustrates the ESCRT cascade and the interactions between each component members of ESCRT-0, I, II, III, and the VPS4 ATPase complex. Of note, the exact stoichiometry of certain component members remains tentative [5]. (B) An HBV replicon plasmid containing a genomic dimer and individual siRNA against each ESCRT factor were co-transfected into HepG2 cells. Viral DNA replication, viral RNA expression and intracellular capsid particles were analyzed by Southern, Northern and Western blot analysis at day 5 post-transfection. The 28S rRNA was included as an RNA loading control. RC: relaxed circular DNA, DL: double-strand linear DNA, SS: single-strand DNA. We detected significant effect on HBV replication from 18 out of 30 tested ESCRT factors (as summarized in Table 1). Data shown here are representative of at least three independent experiments.
Table 1.
A list of tested ESCRT factors that affected HBV replication in HepG2 cells.
Fig 2.
Knockdown of HGS destabilized ESCRT-0 complex and inhibited core protein expression and HBV replication.
(A) Suppression of viral DNA replication was detected by Southern blot analysis in HepG2 and HuH-7 cells transfected with an HBV genomic dimer and siRNAs specific for HGS or STAM2. Signal intensity from RC and SS bandings was quantified by Image J software. (B) Significant reduction of virion-associated viral DNA was detected by qPCR in HepG2 cells upon depletion of ESCRT-0 components. Relative levels of HBV DNA in samples treated with si-HGS or si-STAM2 were normalized to that of the control sample treated with non-targeting siRNA (si-NT). Data here are representative of three independent experiments. (C) A co-depletion phenomenon was observed in HGS-depleted ESCRT-0 complex by SDS-PAGE analysis. Knockdown of HGS resulted in the concurrent reductions of STAM1, STAM2 and HBc protein expressions when the ESCRT-0 complex was destabilized. (D) The siRNA rescue experiment was performed by co-transfecting a constant amount of HBV replicon DNA with different doses of HGS expression vector (up to 0.2 μg per well in 6-well plate) in si-HGS (50 nM) treated HepG2 cells. Complementation of HGS efficiently restored the HGS protein level in a dose-dependent manner (lanes 2–4), as well as HBV replication. (E) Knockdown of ESCRT-0 factors inhibited HBV enhancer II and core promoter activity. A mixture of the reporter plasmid pGL3-EnhII-Cp, a control plasmid pRL-TK, and 50 nM siRNA, were co-transfected into HepG2 cells. Three days post-transfection, cell lysates were analysed by Dual-Glo Luciferase Assay System. The firefly luciferase activity driven by the HBV EnhII-Cp was always normalized first to an internal control of the renilla luciferase activity. The value of the Y-axis represents the relative luciferase activity in each siRNA-treated sample over the control siRNA-treated sample. Data are representative of three independent experiments.
Fig 3.
Overexpressed HGS suppressed HBV transcription and replication.
(A) Overexpression of HGS reduced HBV DNA, RNA, and capsid particles. Plasmid DNAs of an HBV replicon pCHT-9/3091 and an HGS expression vector were co-transfected into HepG2 and HuH-7 cells (2:1 ratio). Intracellular core particle-associated viral DNA, RNA and capsid particles were examined at day 5 post-transfection. (B) Overexpression of HGS inhibited HBV EnhII-Cp activity. The reporter plasmid of pGL3-EnhII-Cp (as depicted in Fig 2E), control plasmid pRL-TK, and two doses of HGS expression vector (1X: 25 ng, 4X: 100 ng), were co-transfected into HepG2 and HuH-7 cells. Three days post-transfection, cell lysates were subjected to Western blot and reporter analysis. Increasing doses of the HGS protein further suppressed HBV EnhII-Cp activity.
Fig 4.
Overexpressed HGS significantly suppressed HBV replication in vivo.
(A) The expression of HGS protein can be detected in the liver sample of hydrodynamically injected mice at 1 dpi by Western blot analysis. Six to eight week old BALB/c mice were tail-vein injected with 14 μg plasmid DNA of an HBV replicon (adr, dimer) and 6 μg DNA of an HGS expression vector or a control plasmid. (B) Immunohistochemistry analysis detected less HBc protein in sectioned liver in the presence of Myc-HGS at 1 dpi. (C) Secreted HBsAg (1000X dilution) and (D) HBeAg (10X dilution) in mouse sera were reduced upon HGS co-injection at 1 and 3 dpi by ELISA. Data are representative of three independent experiments. (E) (left panel) HGS reduced HBV replication in the hydrodynamically injected liver by Southern blot analysis. Each lane represents a liver sample from each mouse. (right panel) HGS reduced HBV DNA in the mouse sera by Southern blot analysis. Data are representative of two independent experiments. (F) HGS reduced the formation of HBV virions and HBsAg particles from pooled mouse sera by 1% native agarose gel electrophoresis at 3 dpi. (G) Liver injurieswere similar between the control and HGS-expressing mice at 1 dpi (p = 0.19 by the Student’s t-test), as measured by the serum alanine aminotransferase (ALT) levels.
Fig 5.
Overexpressed HGS stimulated the release of HBV naked capsid particles.
(A) HGS stimulated the secretion of naked capsids in a dose-dependent manner, but not virions or HBsAg particles by native agarose gel electrophoresis. HBV replicon pCHT-9/3091 and HGS expression vector DNA were mixed at different ratios before co-transfection into HuH-7 cells. Secreted HBV particles in supernatants were collected by ultracentrifugation at day 3 post-transfection. The ratios of extracellular capsids over intracellular capsids or virions were quantified by Image J analysis and normalized to the control group. (B) Co-transfection of an HBV replicon plasmid pCHT-9/3091 and an HGS expression vector suppressed HBsAg expression, but strongly stimulated the secretion of naked capsids by ELISA assay (see text for details). Data are representative of two independent experiments. (C) Overexpression of HGS (0.3 μg) in a stable HBV-producing cell line Qs21 significantly reduced virion secretion without an apparent effect on the secretion of naked capsids. The ratio of naked capsids/virions (1.61) suggests that HGS could help maintain the level of naked capsids despite the concurrent reduction in virions. (D) Knockdown of endogenous HGS in HepG2 cells suppressed the secretion of naked capsids, as well as reduced the synthesis of intracellular viral DNA and capsid particles. The ratio of naked/intra capsid particles was quantified by Image J analysis and normalized to the si-NT control group. Data are representative of two independent experiments. (E) Unlike HGS, overexpression of the other two ESCRT-0 factors STAM1 and STAM2, stimulated HBsAg expression and virion formation without an apparent enhancement of the secretion of naked capsids. (F) ALIX, an ESCRT accessory factor, could contribute to the HGS-facilitated secretion of naked capsids. Co-transfection of HGS expression vector (0.3 μg) and si-ALIX (50 nM) in HuH-7 cells reduced the level of secreted naked capsids at day 5 post-transfection in two independent transfection experiments (compare lanes 3 and 4).
Fig 6.
HGS stimulated the secretion of naked capsids regardless of the genome maturation status.
(A) (upper panel) HGS can promote the secretion of empty naked capsids in a viral polymerase-independent manner. The polymerase-null missense mutant 2310 is defective in AUG initiation of HBV polymerase. While this mutant is normal in the expression of HBsAg and HBc, it is unable to encapsidate pregenomic RNA due to lack of the polymerase. As shown in the lower panel, this polymerase-null mutant could not replicate itself. (One of the duplicated control samples in native gel was partially lost) (B) HGS co-transfection with wild type HBV into HuH-7 cells increased the secretion of DNA-containing naked capsids of at day 3 and 5 post-transfection. In contrast, HBsAg particles, total virions, as well as virion-associated viral DNA, were all reduced by HGS. Data are representative of two independent experiments. (C) HGS can stimulate the secretion of naked capsids containing viral RNA of HBV polymerase mutant Y63D by Northern blot analysis, despite its suppressive effects on HBsAg particles, total virions, and intracellular capsids. Data are representative of two independent experiments. (D) The arginine-rich domain (ARD) of HBc 147–183 is dispensable for the HGS-enhanced secretion of naked capsids (compare lanes 1–4). The result here also suggests that ARD could play a role essential to the secretion of empty virions (compare lanes 1 and 3).
Fig 7.
ESCRT-0 protein HGS associated with HBV core protein through an ubiquitin-independent recognition.
(A) The association between HGS and HBc proteins was identified by the co-IP assay at day 3 post-transfection. The HBV replicon plasmid pCHT-9/3091 and an HGS expression vector (2:1 ratio) were co-transfected into HuH-7 cells. Deletion of the known ubiquitin interaction motif (UIM, aa 257–277) of HGS exhibited no effect on its association with HBc. (B) There are only two lysine residues K7 and K96 in HBc which could serve as candidate ubiquitination sites. Lysine to alanine substitutions at these two sites (K7A and K96A) did not abrogate the association between HBc and HGS, with or without UIM. (C) A lysine-free HBc double mutant K7A/K96A maintained the associations with HGS and HGS dUIM in the co-IP assay, suggesting that the association between HBc and HGS is ubiquitination-independent. (D) ARD-truncated mutant HBc 1–147 containing the N-terminal capsid assembly domain was necessary and sufficient for its association with wild type HGS and HGS dUIM mutant. (E) HGS can efficiently promote the secretion of naked capsids in an ubiquitin-independent manner in HuH-7 cells at day 5 post-transfection. No difference in the enhanced secretion of naked capsids was detected between wild type and mutant HBV, or between wild type and mutant HGS, despite the fact that intracellular capsids were all significantly diminished upon wild type and mutant HGS overexpression.
Fig 8.
HBc co-localized with HGS mostly near the cell periphery, but not within the punctate structure of HGS.
(A) Confocal microscopic analysis of the Flag-HGS protein alone revealed two distinct patterns of subcellular localization in HepG2 cells at day 2 post-transfection. Approximately 70% of the HGS-transfected cells (anti-Flag) exhibited a punctate structure in the cytoplasm (right panel). The remaining 30% of cells exhibited a diffuse HGS distribution in the cytoplasm with a more enriched concentration toward the cell periphery. (B) Confocal microscopic analysis of the co-localization pattern between HBc (anti-HBc, Dako) and Flag-HGS (anti-Flag). An HBV replicon plasmid pCHT-9/3091 and an HGS expression vector were co-transfected into HepG2 cells. Co-localization was observed mainly near the cell periphery, but less often with the punctate structure. Quantification of the relative abundance of the co-localization patterns (periphery vs. puncta) was shown in the right panel. Data above are representative of three independent experiments. (C) A similar HBc and HGS protein co-localization pattern was observed between wild type HGS and mutant HGS dUIM in HuH-7 cells.