hnRNPA2B1 induces HBV cccDNA degradation by recruiting APOBEC3B

Abstract

Hepatitis B virus (HBV) covalently closed circular DNA (cccDNA) constitutes a viral persistence reservoir that sustains chronic infection. Although the DNA damage response (DDR) facilitates cccDNA biogenesis, its role in regulating cccDNA stability remains unclear. By intersecting published cccDNA-associated proteomic datasets with known DDR-related host factors, we identified heterogeneous nuclear ribonucleoprotein A2/B1 (hnRNPA2B1) as a novel restriction factor that binds cccDNA and suppresses HBV replication by promoting cccDNA degradation. Mechanistically, hnRNPA2B1 interacted with the G-quadruplex (G4) structure of cccDNA, with preference for G4-1, G4-7, and G4-10, and leads to the recruitment of the cytidine deaminase APOBEC3B by its prion-like domain (PrLD), thereby inducing C>T and G>A hypermutations and initiating cccDNA decay. Notably, HBV counteracts this defense mechanism through HBx-mediated hnRNPA2B1 polyubiquitination and proteasomal degradation, revealing a viral evasion strategy that perpetuates cccDNA persistence. These findings reveal a G4-dependent surveillance axis wherein hnRNPA2B1 directs APOBEC3B-mediated cytidine deamination to destabilize cccDNA while identifying HBx-induced hnRNPA2B1 ubiquitination as a viral countermeasure. This mechanistic duality not only elucidates a critical virus-host interaction governing cccDNA persistence but also provides a promising therapeutic target for the treatment of HBV infection.

Graphical Abstract

Graphical Abstract.

Introduction

Hepatitis B virus (HBV) chronically infects over 250 million individuals worldwide, establishing a persistent infection that drives liver cirrhosis and hepatocellular carcinoma (HCC) [1]. Central to this persistence is covalently closed circular DNA (cccDNA), which functions as a viral reservoir by forming a minichromosome within the hepatocyte nuclei [2]. cccDNA serves as the exclusive transcriptional template for all HBV transcripts, including pregenomic RNA (pgRNA), and is not directly targeted by current nucleos(t)ide analogs (NAs) and interferon-alpha (IFN-α) therapy [2, 3]. Although therapeutic strategies targeting cccDNA through degradation, lethal mutations, or functional silencing are under active preclinical investigation [4], their clinical translation remains a significant challenge owing to our incomplete understanding of cccDNA biology.

The establishment and maintenance of cccDNA reservoirs are governed by a dynamic equilibrium between de novo synthesis and degradation, with the host-encoded machinery critically orchestrating both pathways. Upon infection, HBV uses host DNA damage response (DDR) systems, including TDP2-mediated polymerase excision, FEN-1-dependent flap resolution, and DNA ligase-mediated nick sealing [5], to repair incoming virion-derived relaxed circular DNA (rcDNA) into cccDNA, a process tightly monitored by the ATR-CHK1/MRN complex [6, 7]. In addition, cccDNA is also derived from the intracellular recycling of nucleocapsids that are transported back to the nucleoplasm and release rcDNA to form cccDNA [2]. The resulting cccDNA reservoir is dynamically maintained through competing host mechanisms: while DNA repair machinery continuously replenishes the pool via de novo synthesis, restriction factors such as APOBEC3 cytidine deaminases, particularly APOBEC3A/B, counterbalance cccDNA stability [8].

The APOBEC3 family, comprising seven members, A3A-A3H, catalyzes the deamination of cytidine to uridine (C>U) or thymidine (C>T) nucleotides, thereby inducing hypermutation and suppressing viral replication. Early investigations established APOBEC3-mediated restriction of HBV replication by targeting replicative intermediates (rcDNA or pgRNA) [9–11], and subsequent findings identified cccDNA as a substrate for A3A- and A3B-induced cytidine deamination. This activity introduces apurinic/apyrimidinic (AP) sites, ultimately leading to cccDNA degradation [8]. Interestingly, a recent study demonstrated the active involvement of heterogeneous nuclear ribonucleoproteins (hnRNPs) in regulating APOBEC3s mutational activities [12]. hnRNPA2B1, a member of the hnRNP family, is a nuclear DNA-interacting protein that recognizes G-quadruplex (G4) structures [13–15]. hnRNPA2B1 not only acts as a regulator of genome stability by orchestrating homologous recombination (HR) repair and double-strand DNA break repair [16–18], but also activates the TBK1-IRF3-IFN-α/β signaling pathway upon detection of viral DNA fragments, including those of HBV [19]. Notably, HBV cccDNA itself harbors ten predicted G4 structures [20], raising the possibility that hnRNPA2B1 may directly engage cccDNA through its G4-sensing capacity. However, whether hnRNPA2B1 is involved in APOBEC3-dependent cccDNA restriction remains unclear.

In this study, based on whole-cell screening of hepatic cccDNA-associated proteins, we identified hnRNPA2B1 as a G4-dependent cccDNA-binding protein that recruits APOBEC3B to trigger deamination-mediated cccDNA degradation. Furthermore, we demonstrated that HBV employs HBx to hijack the CRL4-DDB1 ubiquitin ligase complex, promoting hnRNPA2B1 ubiquitination and proteasomal turnover. In summary, these findings provide novel insights into host-pathogen conflict centered on cccDNA maintenance and turnover, thereby establishing a mechanistic foundation for the therapeutic targeting of cccDNA stability.

Materials and methods

Cell culture

Human hepatocellular carcinoma (HCC) cell lines Huh7, HepG2, HepAD38, and HepG2^NTCP were cultured in Dulbecco’s modified Eagle medium (DMEM; Gibco, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin (PS). Human embryonic kidney cells HEK293T were maintained under identical DMEM conditions (10% FBS, 1% PS). HLCZ01 cells [21] were propagated in DMEM/F-12 medium (Gibco) enriched with 10% FBS, 1% PS, 40 ng/ml dexamethasone, and 10 ng/ml epidermal growth factor (EGF), with cultures maintained on collagen-coated plates. HepaRG^NTCP cells were grown in Williams’ Medium E (Gibco) containing 10% FBS, 1% PS, 0.023 IU/ml insulin, 5 μg/ml hydrocortisone, and 80 μg/ml gentamicin. All cell lines were incubated at 37°C in a humidified 5% CO₂ atmosphere.

Human studies

Twenty-eight HCC patient para-tumor tissues were obtained from patients admitted to Shandong Provincial Hospital. The tissue specimens were snap-frozen in liquid nitrogen and stored at −80°C for RNA, DNA, and protein extraction. The use of resected human samples was approved by the Medical Ethical Committee of School of Basic Medical Sciences of Shandong University (Approval Number: ECSBMSSDU2024-1-127), and all participants were informed and provided written informed consent. A summary of the clinical information for the 28 HBV-related HCC patients is available in Supplementary Table S4.

Plasmids, proteins, and siRNAs

Expression plasmids encoding FLAG-tagged hnRNPA2B1, APOBEC3A, APOBEC3B, APOBEC3C, APOBEC3D, APOBEC3F, APOBEC3G, and Uracil Glycosylase Inhibitor (UGI) were generated by cloning their respective coding sequences into the pCMV3×FLAG 7.1 vector. The prokaryotic expression plasmids of hnRNPA2B1 were constructed by cloning them into pET-22b-mEGFP vectors. The recombinant His-GFP-hnRNPA2B1 proteins were purified by Ni-NTA beads (SA005100, Tiandirenhe, Jiangsu, China). HA-HBx, HA-HBc, HA-HBsAg, HA-APOBEC3B, and HA-hnRNPA2B1 expression plasmids were constructed by cloning their coding sequence into pCAGGS-HA vector. The truncates of hnRNPA2B1, HA-APOBEC3B-E255A, and HBx-R96E were generated by KOD Kit (SMK-101, Toyobo, Japan). The minicircle-HBV (MC-HBV) and MC-HBVΔHBx were generated as described previously [22]. MC-HBVΔG4 was generated by introducing mutation into G4s to disrupt the formation of G4 structure. The pHBV1.3 and pHBV1.3-mer X-null expressing plasmids were purchased from Addgene (USA). The small interfering RNAs (siRNAs) specific for APOBEC3B were synthesized from GenePharma (Shanghai, China) (Supplementary Table S3).

Antibodies and reagents

Rabbit anti-hnRNPA2B1 polyclonal antibody (A1162, Abclonal, Wuhan, China), rabbit anti-HBc polyclonal antibody (B0586, Dako, Copenhagen, Denmark), rabbit anti-APOBEC3B polyclonal antibody (14559-1-AP, Proteintech, Chicago, IL, USA), mouse anti-β-actin monoclonal antibody (66009-1Ig, Proteintech), mouse anti-FLAG-tag monoclonal antibody (M185-3, MBL, Nagoya, Japan), mouse anti-HA monoclonal antibody (M180-3, MBL), mouse anti-ubiquitin monoclonal antibody (sc-8017, Santa), and rabbit anti-HBx polyclonal antibody (ab39716, Abcam). The streptavidin dynabeads were purchased from Invitrogen (11205D, Thermo Fisher Scientific, Waltham, MA, USA). The anti-His magnetic beads (P2135), T5 exonuclease (D7082S), puromycin, proteasome inhibitor MG132 (S1748), and Coomassie blue staining solution (P0017F) were purchased from Beyotime Biotechnology (Shanghai, China). The nucleic acid labeling kits (Biotin, MIR 3400, and Cy5, MIR 3725) were purchased from Mirus Bio LLC (Madison, WI, USA). The polyethylene glycol (PEG)-8000 (89510) was purchased from Sigma (Sigma–Aldrich). The protein synthesis inhibitor cycloheximide (CHX) (S7418) and Braco-19 trihydrochloride were purchased from Selleck (Shanghai, China). Chromatin immunoprecipitation kit was purchased from Millipore (Merck KGaA, Darmstadt, Germany).

Establishment of tet-on, knockout, and knockdown cell lines

The coding sequence of hnRNPA2B1 was cloned into the pLVX-TetOne-Puro vector to establish doxycycline (Dox)-inducible hnRNPA2B1 expression construct. The sgRNAs targeting HNRNPA2B1 (Supplementary Table S3) were cloned into the LentiCRISPR v2 vector for HNRNPA2B1 knockout. shRNAs targeting HNRNPA2B1 or SMC5 were cloned into the pLVX-puro-GFP lentiviral vector for the knockdown of respective genes. To produce lentivirus, HEK293T cells were plated in 100-mm culture dishes and co-transfected with 9 μg lentivirus plasmid, 2.25 μg pRSV-REV, 4.5 μg pMDLg, and 2.7 μg pVSV-G plasmids using PEI transfection reagent. Viral supernatants were collected 72 h post-transfection and stored at −80°C until use. For lentivirus transduction, HepG2^NTCP, HepaRG^NTCP, or HLCZ01 cells were seeded in collagen-coated six-well plates and inoculated with lentivirus in the presence of 4 μg/ml polybrene (C0531, Beyotime Biotechnology) for 12 h, washed three times with PBS, and then the fresh medium was added and grown in the presence of 1 μg/ml puromycin for 2 weeks.

HBV infection

HBV viral stocks were prepared using a centrifugal filtration system (Millipore) from the supernatants of HepAD38 cells and of Huh7 cells that had been transfected with either pHBV1.3 or pHBV1.3-mer X-null. Viral titers were quantified using a commercial HBV DNA quantitative fluorescence diagnostic kit (Sansure Biotech, Changsha, China) and aliquots of concentrated HBV stocks were stored at −80°C until use. The cells were infected with HBV at 500 genomes per cell in medium supplemented with 2% dimethyl sulfoxide (DMSO) and 4% polyethylene glycol 8000 (PEG8000; Sigma–Aldrich, USA) for 24 h. Following incubation, the cells were washed three times with phosphate buffer saline (PBS) and cultured in a maintenance medium containing 2.5% DMSO for 5–7 days prior to downstream experiments.

Chromatin immunoprecipitation (ChIP)

Cells were fixed with 1% formaldehyde for 10 min at room temperature, followed by neutralization with 0.125 M glycine. After nuclei isolation, chromatin was fragmented to 200–500 bp through sonication. Target protein–DNA complexes were captured through overnight immunoprecipitation at 4°C using specific antibodies, followed by 2-h incubation with protein A/G magnetic beads. Sequential washing was conducted under rotation using 1 ml of each solution: low-salt buffer, high-salt buffer, LiCl buffer, and TE buffer (5 min per wash). DNA was recovered through 2-h proteinase K digestion at 62°C, and enriched DNA fragments were subsequently quantified by quantitative polymerase chain reaction (qPCR).

Protein expression and purification

His-tagged recombinant hnRNPA2B1 protein was expressed in Escherichia coli strain Rosetta (DE3) (CB108; TianGen, Beijing, China). Briefly, competent Rosetta (DE3) cells were transformed with the hnRNPA2B1 expression plasmid. A single transformed colony was inoculated into LB medium and incubated at 37°C with shaking (200 rpm) until mid-log phase (OD₆₀₀ = 0.6–0.8). Then recombinant protein production was induced by adding 0.1 mM IPTG (ST098, Beyotime) for 16–20 h at 16°C. Lysis and sonication of the bacteria (lysis buffer: 25 mM Tris–HCl pH 7.5, 500 mM NaCl). The suspension was sonicated to shear DNA until it reached the turbidity typical of a protein solution. The supernatant was purified by a gravity column and centrifuged by an ultrafilter.

MC-HBV biotinylation and pull-down assay

MC-HBV was biotinylated using the Label IT Biotin Reagent (MIR 3400, Mirus Bio LLC) at 37°C for 1 h, followed by purification via a G50 microspin column in accordance with the manufacturer’s protocol. For streptavidin-pull-down, 1 μg biotinylated MC-HBV and unlabeled MC-HBV were combined with 5 μg recombinant His-hnRNPA2B1 protein and incubated at 4°C overnight. Subsequently, the mixture was incubated for an additional 1 h with 20 μl Dynabeads M-280 Streptavidin magnetic beads (11205D, Thermo Fisher Scientific). Following three washes with PBST, the complexes were denatured by boiling in 25 μl of 1× sodium dodecyl sulfate (SDS) loading buffer (P0015L, Beyotime Biotechnology). The resulting protein samples were subjected to western blot analysis. For αHis-pull-down, MC-HBV (200 ng) was incubated with His-hnRNPA2B1 protein (5 µg) and anti-His magnetic beads (20 µl slurry) in a protein–DNA binding buffer [20 mM Tris–HCl pH 8.0, 150 mM KCl, 5 mM MgCl₂, 1 mM DTT, 10% glycerol, and 1 mg/ml BSA] for 12 h at 4°C under gentle rotation. The beads were washed three times with PBST for 5 min each time. Elution of protein–DNA complexes was performed using a buffer containing 1% SDS, 50 mM NaHCO₃, 100 mM NaCl, 1 mM EDTA, and 10 mM Tris–HCl pH 6.5, supplemented with 1 mg/ml proteinase K, followed by incubation at 62°C for 2 h. The purified DNA was detected by qPCR using cccDNA-specific primers.

Electrophoretic mobility shift assay (EMSA)

The FAM-labeled oligonucleotides (detailed in Supplementary Table S2) were heated at 95°C for 5 min in 20 mM Tris, pH 7.5 buffer with 150 mM KCl and annealed to 25°C at a rate of −1°C/min. The binding reactions were performed in 20 μl mix containing 10 mM Tris–HCl, 50 mM KCl, 1 mM DTT, 1 mM MgCl₂, 5% glycerol, 50 nM FAM-labeled oligonucleotides, and recombinant proteins were incubated at room temperature for 1 h. The sample was loaded on 10% non-denaturing polyacrylamide gel at 4°C.

Enzyme-linked immunosorbent assay (ELISA) and HBV-DNA detection

HBsAg and HBeAg levels and HBV-DNA levels in the cell culture supernatant were quantified using commercially available ELISA Kit (InTec, Inc., Xiamen, China) and HBV DNA Quantitative Fluorescence Diagnostic Kit (Sansure Biotech, Changshan, China) following the manufacturer’s instructions, respectively.

CHX chase assay

Huh7 cells were transfected with HBx plasmids for 36 h and further incubated in the presence of 500 μg/ml CHX at 37°C for 0–12 h. Subsequently, at each indicated time point, cells were lysed to track the temporal degradation of hnRNPA2B1 via western blot. To quantify the rate of degradation, band intensities from western blot were normalized and plotted over time, and the hnRNPA2B1 protein half-life was determined using nonlinear regression analysis in GraphPad Prism 7.0.

Transfection and RT-qPCR

The cells were grown overnight until reaching 70%–80% confluency, then transfected with plasmids using Lipofectamine 2000 (Thermo Fisher Scientific) per the manufacturer’s protocol. After 6 h of transfection, the cell medium was changed. Total RNA was extracted from cells using TRIzol reagent (DP424, TianGen), and 1 μg of RNA was reverse-transcribed into complementary DNA (cDNA) with a PrimeScript RT Reagent Kit (2641A, Takara, Kyoto, Japan). Reverse transcription qPCR (RT-qPCR) was performed using SYBR mix (FP205-02, Vazyme, Nanjing, China) according to the manufacturer’s protocol by the indicated primers (Supplementary Table S3). Beta Cytoskeletal Actin (β-actin) was used as an internal control. Relative gene expression was normalized to β-actin using the 2^–ΔΔCt method.

Southern blot

DNA from cells was extracted by the Hirt procedure as described [23]. DNA samples were electrophoresed on a 1% agarose gel and subsequently transferred via capillary action onto Amersham Hybond-N+ membrane (GE Healthcare). Following UV-induced crosslinking, the membrane underwent prehybridization. A digoxigenin-conjugated HBV genomic DNA probe was prepared in accordance with the standardized protocol of DIG-High Prime DNA Labeling and Detection Starter Kit (Roche Diagnostics GmbH, Mannheim, Germany) and applied for hybridization at 68°C under controlled conditions for 16 h. Post-hybridization, sequential incubations were performed: first in blocking buffer (30 min, ambient temperature), followed by anti-digoxigenin antibody solution (30 min, 37°C). Target DNA-probe complexes were ultimately visualized through chemiluminescent detection following substrate exposure.

AP site quantification and APE1 digestion

AP site quantification was performed with DNA Damage Quantification Kit-AP Site Counting (DK02, Dojindo Laboratories, China) according to the manufacturer’s instructions using 10 μl of purified DNA (100 μg/ml). One microgram of Hirt-extracted DNA was digested with 10 units of apurinic/apyrimidinic endonuclease redox effector factor-1 (APE1; M0282S, New England Biolabs, Inc., Ipswich, MA, USA).

cccDNA quantification by qPCR, differential DNA denaturation PCR (3D-PCR), and next-generation sequencing

cccDNA in cells or liver paracancerous tissues collected from HBV-associated HCC patients was extracted via the Hirt DNA method and digested with T5 Exonuclease as described [23]. cccDNA was quantified by qPCR using cccDNA-qF/R primers (Supplementary Table S3) and normalized to the copy number of mitochondrial DNA ND1 gene. To investigate deamination of HBV cccDNA, cccDNA was amplified by HBVccc2760fw and HBVccc156rev primers, and 1/50-diluted PCR products were used as template to further amplify with cccDNA-3D-PCR-F/R primers under gradient denaturing temperatures. Amplicons from 3D-PCR were visualized on a 2% agarose gel and extracted for sequencing on an Illumina MiSeq machine by PE150 via commercial sequencing service (Tsingke Biotechnology, Qingdao, China). The C-to-T and G-to-A mutation events in 10⁵ nucleotide base pairs and frequency ratio in all mutation events were calculated from the sequencing data.

To test the deamination of host genome, the total genomic DNA was extracted and amplified using primers targeting ARID2, PAX5, and TP53 (Supplementary Table S3). The amplicons were gel-purified using the MN NucleoSpin Gel Extraction Kit. Equal amounts of purified PCR products were subjected to nested PCR with under denaturation temperature gradient (95°C–87°C). The final PCR products were analyzed by gel electrophoresis.

Immunofluorescence staining

Cells grown on coverslips were fixed in 4% paraformaldehyde for 15 min and then permeabilized with 0.5% Triton X-100 for 15 min at room temperature. The cells were subsequently incubated with blocking buffer (5% BSA in PBS) for 1 h at room temperature and then stained with the indicated primary antibodies overnight at 4°C. After washing with PBS, bound antibodies were detected using appropriate secondary antibodies. Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI). Images were captured by LSM980 confocal microscope system (Zeiss) and the co-localization was evaluated using the ImageJ software.

Co-immunoprecipitation (Co-IP)

Cells were lysed in IP buffer containing 20 mM Tris-Cl (pH 8.0), 150 mM NaCl, 2 mM EGTA, 1% NP-40, and protease inhibitor cocktails. The lysates were centrifuged at 12 000 rpm for 5 min. Equal numbers of supernatants were incubated with target-specific antibodies overnight at 4°C with gentle agitation, then incubated with protein A/G magnetic beads for 2 h at 4°C. Beads were washed 4 times with beads washing buffer (20 mM Tris-Cl pH 7.5, 300 mM NaCl). After washing, bound proteins were eluted from the beads by boiling in 1× sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) loading buffer at 95°C for 5 min. The eluted proteins were loaded onto an SDS–PAGE gel for western blot analysis.

Statistical analysis

All statistical analyses were performed using GraphPad Prism 7.0 software. Unpaired Student’s t-test, one-way analysis of variance (ANOVA), and two-way ANOVA were performed to determine the statistical significance of differences between groups. Data represent the means ± SEM. Significance levels are indicated by asterisks: *P < .05; **P < .01; NS: non-significant.

Results

hnRNPA2B1 directly binds HBV cccDNA through G-quadruplex structures

To investigate how the host DNA repair apparatus regulates cccDNA stability, we first performed a Venn analysis to screen candidate cccDNA-binding proteins by intersecting two recently published cccDNA-associated proteomic datasets from HBV-infected HepG2^NTCP cells [24, 25] with known DDR-related host factors (Supplementary Table S1). This analysis revealed five overlapping candidates: hnRNPA2B1, RAD50, XRCC6, PARP1, and USP7 (Fig. 1A). Among these, hnRNPA2B1, a newly characterized nuclear DNA sensor with established DNA-binding capacity [19], has emerged as the primary candidate for further investigation because of its reported role in recognizing non-canonical nucleic acid structures. To validate the physical interaction between hnRNPA2B1 and cccDNA, we performed cccDNA-specific ChIP assay in two HBV-permissive cell models, HLCZ01 and HepG2^NTCP cells. Both systems demonstrated a significant accumulation of hnRNPA2B1 in cccDNA (Fig. 1B). Spatial colocalization was confirmed by confocal microscopy in HepG2 cells transfected with Cy5-labeled MC-HBV (a cccDNA surrogate), which showed a distinct overlap between endogenous hnRNPA2B1 and Cy5-MC-HBV signals (Fig. 1C). The direct interaction between hnRNPA2B1 and cccDNA was further verified through in vitro pull-down assay using purified His-tagged recombinant hnRNPA2B1 protein (His-hnRNPA2B1) and biotinylated MC-HBV (Bio-MC-HBV). As shown in Fig. 1D, streptavidin bead capture showed a significant enrichment of MC-HBV upon incubation with His-hnRNPA2B1.

Figure 1.

hnRNPA2B1 is identified as a cccDNA-binding factor. (A) Venn analysis of potential cccDNA-binding proteins and the reported DNA repair-related proteins. hnRNPA2B1 was identified as the overlapping candidate among three independent datasets. (B) ChIP-qPCR analysis of hnRNPA2B1-cccDNA interaction in HBV-infected HLCZ01 and HepG2^NTCP cells with cccDNA-specific primers. Immunoprecipitation efficiency was validated by western blot using anti-hnRNPA2B1. (C) Co-localization of hnRNPA2B1 (green) with Cy5-labeled MC-HBV (red) in HepG2 cells detected by confocal microscopy. The intensity profiles correspond to the white lines drawn in the merged images. Scale bars, 5 μm. (D) Coomassie staining of purified His-tagged hnRNPA2B1 proteins (left panel). His-hnRNPA2B1 proteins were incubated with MC-HBV and their interaction was analyzed by αHis-pull-down (right panel). (E) Schematic diagram of G4s’ position and mutation in cccDNA. (F) ChIP analysis of hnRNPA2B1 interaction with MC-HBV or MC-HBVΔG4 in MC-HBV-transfected Huh7 cells. The immunoprecipitation efficiency was measured by western blot. Recombinant His-hnRNPA2B1 proteins were incubated with MC-HBV or MC-HBVΔG4, and their binding was detected by αHis-pull-down assay (G) and streptavidin bead pull-down assay (H). (I) Competitive pull-down analysis of hnRNPA2B1-cccDNA interaction. His-hnRNPA2B1 proteins were incubated with Bio-MC-HBV together with increasing amounts of MC-HBV or MC-HBVΔG4, and the binding of hnRNPA2B1 with Biotin-MC-HBV was analyzed by streptavidin-pull-down. (J) HBV-infected HepG2^NTCP cells were treated with 20 μM BRACO-19 or DMSO. The interaction between hnRNPA2B1 and cccDNA was analyzed by ChIP-qPCR. Immunoprecitation efficiency was validated by western blot. (K) EMSA was performed using 50 nM purified hnRNPA2B1 incubated with 50 nM of each indicated G4 sequence. (L) ChIP analysis of hnRNPA2B1 binding to wild-type MC-HBV or triple G4 mutant MC-HBVΔG4-1,7,10 in Huh7 cells transfected for 48 h. Immunoprecipitation efficiency was measured by western blot. Results are representative of two (B–D, I, J, L) or three (F–H, K) independent experiments; each dot represents a biological repeat. P-values were determined by unpaired t-test (B, D) or two-way ANOVA (F, G, J, L); **P < .01.

Given that hnRNPA2B1 is a known G4 DNA-binding protein [13, 15] and 10 G4 motifs are present in HBV cccDNA [20], we hypothesized that G4 structures mediate the interaction between hnRNPA2B1 and cccDNA. To test this, we generated an MC-HBV mutant (MC-HBVΔG4) by disrupting all ten G4 motifs in the MC-HBV genome (Fig. 1E). ChIP assay showed significantly reduced recruitment of hnRNPA2B1 to cccDNA in Huh7 cells transfected with MC-HBVΔG4 compared to that in wild-type MC-HBV-transfected cells (Fig. 1F). In vitro binding assay corroborated these findings: His-hnRNPA2B1 displayed a significantly weakened interaction with MC-HBVΔG4 compared to wild-type MC-HBV, as demonstrated by both α-His affinity purification (Fig. 1G) and streptavidin pull-down assays (Fig. 1H). Further competitive inhibition assay showed that untagged wild-type MC-HBV effectively blocked His-hnRNPA2B1/Bio-MC-HBV interaction, whereas MC-HBVΔG4 showed no inhibitory effect (Fig. 1I). To further confirm it, we constructed a G4-binding-deficient hnRNPA2B1 mutant (A2B1-mG4bd) that carries L28A, F30A, R89A, R147A, and F157A mutations [15]. ChIP assay showed this mutant lost the ability to bind cccDNA (Supplementary Fig. S1B). Importantly, ChIP analysis in HBV-infected HepG2^NTCP cells also revealed that treatment with BRACO-19, a G4 stabilizer ligand [20], enhanced the enrichment of hnRNPA2B1 on cccDNA (Fig. 1J). This data further supports the role of G4 motifs as functional mediators of hnRNPA2B1 binding in the HBV lifecycle.

To systemically determine the binding preferences of hnRNPA2B1, we performed EMSA using purified His-hnRNPA2B1 proteins and pre-folded, 5′-FAM-labeled oligonucleotides corresponding to each G4 motif (Supplementary Fig. S1A and Supplementary Table S2). The results revealed that hnRNPA2B1 binds with notably stronger affinity to G4-1, G4-7, and G4-10 compared to other G4 motifs (Fig. 1K). Consistent with this, when we introduced mutations specifically disrupting G4-1, G4-7, and G4-10 in the HBV minicircle (MC-HBVΔG4-1,7,10), ChIP assays showed a significant reduction in hnRNPA2B1 recruitment relative to the wild-type construct (Fig. 1L). Together, these findings suggest that hnRNPA2B1 directly binds to HBV cccDNA by recognizing its G4 structure.

hnRNPA2B1 inhibits HBV replication and reduces cccDNA levels

Given that cccDNA serves as the sole template for viral replication, we tested whether hnRNPA2B1 regulates HBV replication by controlling cccDNA abundance. In HBV-infected HLCZ01 cells engineered for tetracycline-inducible hnRNPA2B1 expression (HLCZ01^A2B1-Teton), cccDNA formation was validated by Southern blot analysis of Hirt-extracted DNA samples incubated at 85°C, with or without EcoRI digestion (Supplementary Fig. S2A). Doxycycline (Dox)-induced hnRNPA2B1 overexpression caused a dose-dependent reduction in cccDNA levels, as measured by Southern blot and qPCR (Fig. 2A and Supplementary Fig. S2B). This decrease coincided with the reduced production of viral antigens (HBsAg, HBeAg, and HBc), diminished preC/pgRNA transcripts, and lower HBV DNA levels (Fig. 2A). Consistent observations were made in HBV-infected HepG2^NTCP cells with Dox-inducible hnRNPA2B1 expression, where cccDNA levels were significantly reduced, along with concomitant decreases in viral antigens, transcripts, and HBV DNA (Fig. 2B and Supplementary Fig. S2C). Conversely, shRNA-mediated hnRNPA2B1 knockdown in HBV-infected HLCZ01 cells significantly increased cccDNA abundance and enhanced all the measured viral replication markers (Fig. 2C and Supplementary Fig. S2D). To validate these findings using an orthogonal model, we generated hnRNPA2B1-knockout (A2B1-KO) HepaRG^NTCP cells using CRISPR/Cas9. Following HBV infection, A2B1-KO HepaRG^NTCP cells exhibited significantly elevated cccDNA accumulation, accompanied by enhanced production of viral markers (Fig. 2D and Supplementary Fig. S2E). Collectively, these results demonstrated that hnRNPA2B1 limits HBV replication by reducing cccDNA levels.

Figure 2.

hnRNPA2B1 inhibits HBV replication. HLCZ01^A2B1-Teton and HepG2^NTCP-A2B1 cells were infected with HBV for 7 days, followed by Dox-induced expression of hnRNPA2B1 for 3 days (A and B). HLCZ01 cells with hnRNPA2B1 knockdown (shA2B1) (C) or HepaRG^NTCP cells with hnRNPA2B1 knockout (A2B1-KO) (D) were infected with HBV at 500 Geq for 7 days. cccDNA levels were detected by Southern blot, mitochondrial DNA (mtDNA) was used as the loading control (left panels). HBsAg and HBeAg levels in culture supernatants were quantified by ELISA; supernatant HBV DNA and intracellular preC/pgRNA were analyzed by qPCR; intracellular HBc protein levels and hnRNPA2B1 expression were measured by western blot (right panels). The results are representative of three (A–D) independent experiments, where each dot represents a biological repeat. P-values were determined by one-way ANOVA (A) or unpaired t-test (B–D); *P < .05, **P < .01.

hnRNPA2B1 promotes cccDNA decay through cytidine deamination-associated editing

As hnRNPs function as cofactors of APOBEC3-mediated cytidine deamination of cccDNA, we hypothesized that hnRNPA2B1 exploits its DNA-binding activity to induce cccDNA destabilization via cytidine deamination, thereby contributing to antiviral defense. The results of 3D-PCR analysis of cccDNA from HBV-infected HLCZ01^A2B1-Teton cells showed that Dox-induced hnRNPA2B1 overexpression progressively lowered the denaturation temperature required for amplification (Fig. 3A), indicative of accumulating sequence mismatches. Notably, deep sequencing of 3D-PCR products showed that hnRNPA2B1 induction dose-dependently increased cccDNA mutation events and the frequency of C>T and G>A transitions (Fig. 3B–D). Following cytidine deamination, uracil residues are excised by DNA glycosylases through cleavage of their N-glycosidic residues, generating AP sites, which can be cleaved by AP endonuclease 1 (APE1) [26]. Further digestion with APE1 led to a more profound decrease in cccDNA levels in the Dox-treated HLCZ01^A2B1-Teton cells (Supplementary Fig. S3A). To exclude the potential interference of rcDNA cytidine deamination in cccDNA, we inhibited HBV polymerase with entecavir (ETV) in HBV-infected HLCZ01^A2B1-Teton cells, thereby blocking rcDNA formation and subsequent conversion to cccDNA (Supplementary Fig. S3B). Under these rcDNA-depleted conditions, hnRNPA2B1 expression still reduced cccDNA abundance and decreased the denaturation temperature of cccDNA amplification (Supplementary Fig. S3B). Deep sequencing of 3D-PCR products showed that hnRNPA2B1 increased C>T and G>A mutation events in established cccDNA (Supplementary Fig. S3C). This suggests that hnRNPA2B1 primarily promotes cccDNA deamination rather than targeting rcDNA-to-cccDNA formation. Furthermore, aldehyde reactive probe-based AP site quantitation in Dox-treated HLCZ01^A2B1-Teton cells revealed no significant increase in the total genomic AP sites relative to untreated controls (Supplementary Fig. S3D). Consistently, 3D-PCR detected no mutational instability in the host TP53, PAX5, and ARID2 genes upon hnRNPA2B1 induction (Supplementary Fig. S3E), suggesting that the destabilization phenotype induced by ectopic hnRNPA2B1 expression specifically targets cccDNA without inducing detectable damage to genomic DNA. This pattern was conserved in HBV-infected HepaRG^A2B1-Teton cells, where ectopic hnRNPA2B1 expression reduced cccDNA abundance (Fig. 3E) and thermostability (Fig. 3F). Sequencing of these cccDNA fragments confirmed heightened mutation loads dominated by C>T and G>A substitutions (Fig. 3G–I), mirroring the mutational landscape observed in HLCZ01 cells.

Figure 3.

hnRNPA2B1 reduces the stability of cccDNA by cytidine deamination. (A–D) HBV-infected HLCZ01^A2B1-Teton cells were treated with Dox at indicated doses. cccDNA stability was analyzed by 3D-PCR using denaturation temperature gradient of 85°C–83°C followed by agarose gel electrophoresis (A). cccDNA fragments amplified at 85°C were sequenced, and the C>T and G>A events (B and C) and mutation frequency (D) were analyzed. (E–I) The differentiated HepaRG^A2B1-Teton cells were infected with HBV for 5 days, followed by 1 μg/ml Dox treatment. hnRNPA2B1 expression and cccDNA levels were detected by western blot and qPCR (E). cccDNA stability was analyzed by 3D-PCR using denaturation temperature gradient of 85°C–83°C (F); the C>T and G>A events (G and H) and mutation frequency (I) of cccDNA fragments amplified at 85°C were analyzed by DNA-seq. (J) HBV-infected HLCZ01^A2B1-Teton cells were treated with 1 μg/ml Dox. Deep-sequencing analysis of the genome-wide distribution of C-to-T and G-to-A conversion events on cccDNA. Results are representative of three (A–I) independent experiments; each dot represents a biological repeat. P-values were determined by unpaired t-test (E, G, I) or one-way ANOVA (B, D); *P < .05, and **P < .01.

To understand the comprehensive impact of hnRNPA2B1 on cccDNA deamination, we designed primers (Supplementary Fig. S3F and Supplementary Table S3) to amplify four overlapping regions covering the entire cccDNA for deep sequencing. This provided a full landscape of C>T and G>A transitions. The results confirm that hnRNPA2B1 overexpression increases these mutation events across the cccDNA genome. Importantly, the mutation frequency was notably higher in genomic regions surrounding the G4-1, G4-7, and G4-10 motifs (Fig. 3J), the primary A2B1-binding sites, indicating localized enrichment of deamination near these structures. Collectively, our data demonstrated that hnRNPA2B1 induces cytidine deamination and AP site formation in HBV cccDNA, leading to its degradation.

To establish the clinical relevance of our findings, we analyzed the correlation of hnRNPA2B1 with HBV cccDNA and its stability in paracancerous liver tissues collected from 28 HBV-associated HCC patients (Supplementary Table S4). hnRNPA2B1 expression was detected by western blot and normalized to β-actin (Fig. 4A), and a significant inverse correlation between hnRNPA2B1 abundance and both cccDNA and preC/pgRNA levels was observed (Fig. 4B). To probe cccDNA integrity directly, we performed 3D-PCR on tissue-derived cccDNA using denaturation temperature gradients. In tissues with higher hnRNPA2B1 expression, cccDNA exhibited decreased thermolability, as evidenced by the preferential amplification at lower denaturation temperatures (Fig. 4C). Furthermore, the ratio of cccDNA amplification products at 84.2°C versus 85°C denaturation temperatures showed a positive correlation with hnRNPA2B1 expression levels (Fig. 4D). These temperature-dependent amplification shifts are consistent with cytidine deamination-induced DNA destabilization, as C>U/T mismatches reduce the duplex thermal stability. Together, these data provide clinical evidence linking elevated hnRNPA2B1 levels to cytidine deamination-associated editing to cccDNA decay.

Figure 4.

hnRNPA2B1 promotes cccDNA decay. Protein, RNA, and Hirt DNA were extracted from para-tumor tissues of HBV-related HCC patients (n = 28) for further detection of hnRNPA2B1 protein (A), preC/pgRNA (B), and cccDNA (B, C). (A) hnRNPA2B1 expression was detected by western blot and normalized to β-actin. (B) Correlation analysis of hnRNPA2B1 expression with cccDNA and preC/pgRNA levels in patients with HCC. (C and D) 3D-PCR analysis of cccDNA stability in HCC tissues using a denaturation temperature gradient, followed by agarose gel electrophoresis. The numbers on the left show the relative expression levels of hnRNPA2B1 (A2B1/β-actin) in each patient (C). Correlation analysis of hnRNPA2B1 expression with the ratio of cccDNA amplification products at 84.2°C versus 85°C denaturation temperatures (D). P-values were determined using two-tailed Pearson correlation analysis (B and D). Each dot represents a single patient.

hnRNPA2B1 recruits APOBEC3B to destabilize cccDNA

Given the established role of APOBEC3 cytidine deaminases in catalyzing C-to-U editing and subsequent AP site formation [27], we investigated whether hnRNPA2B1 mediates cccDNA decay by APOBEC3 recruitment. To address this, we cloned human APOBEC3 members and investigated their interaction with hnRNPA2B1. Co-IP assay revealed specific interactions between HA-hnRNPA2B1 and APOBEC3B, but not between other APOBEC3s (Fig. 5A). This interaction was reciprocally confirmed using FLAG-APOBEC3B immunoprecipitation (Supplementary Fig. S4A) and validated endogenously in HLCZ01 cells using Co-IP assay (Fig. 5B). Further ChIP analysis showed that hnRNPA2B1 overexpression significantly enhanced APOBEC3B occupancy of cccDNA in HBV-infected HepG2^NTCP and HLCZ01 cells, whereas hnRNPA2B1 knockdown reduced the accumulation of APOBEC3B in cccDNA (Fig. 5C and Supplementary Fig. S4B). 3D-PCR assay showed that APOBEC3B knockdown largely abolished the reduction in cccDNA thermostability caused by hnRNPA2B1 (Fig. 5D). APOBEC3B silencing abolished the hnRNPA2B1-mediated increase in C>T and G>A mutations in cccDNA (Fig. 5E and Supplementary Fig. S4C). Concurrently, it reversed hnRNPA2B1-mediated suppression of cccDNA and downstream viral replication outputs, including preC/pgRNA transcripts and viral antigens (HBsAg and HBeAg) (Fig. 5F). Furthermore, overexpression of wild-type APOBEC3B (APOBEC3B-WT) resulted in a stronger suppression of cccDNA levels and enhanced its deamination compared to hnRNPA2B1 alone. In contrast, overexpression of the catalytically inactive APOBEC3B mutant (A3B-E255A) [28] did not produce this additional suppressive effect (Supplementary Fig. S4D). Moreover, hnRNPA2B1 knockdown abrogated APOBEC3B-triggered reduction in cccDNA abundance and stability (Supplementary Fig. S4E and F). These data demonstrate the important role of the hnRNPA2B1-APOBEC3B axis in inducing cccDNA destabilization and HBV inhibition.

Figure 5.

hnRNPA2B1 interacts with APOBEC3B to promote its recruitment to cccDNA. (A) HA-A2B1 was co-transfected with FLAG-APOBEC3s (A3A, A3B, A3C, A3D, A3F, or A3G) in Huh7 cells for 48 h. Co-IP assay was performed to analyze their interactions. The arrow indicates the target band. (B) Co-IP assay was performed with anti-A2B1 antibody or anti-A3B antibody in HLCZ01 cells to evaluate the endogenous interaction of hnRNPA2B1 and APOBEC3B. (C) hnRNPA2B1 overexpression was induced by Dox in HBV-infected HepG2^NTCP-A2B1 cells (left panel), while hnRNPA2B1 knockdown was accessed by shA2B1 lentivirus in HBV-infected HepG2^NTCP cells (right panel). The enrichment of A3B on cccDNA was analyzed by ChIP assay, and the immunoprecipitation efficiency was measured by western blot. HBV-infected HLCZ01^A2B1-Teton cells transfected with APOBEC3B-siRNA (siA3B) and treated with Dox for 72 h were subjected to: 3D-PCR with denaturation temperature gradient (85°C–83°C) (D); deep sequencing of 85°C-amplified cccDNA fragments showing C>T and G>A events and mutation frequency (E); and parallel assessments of cccDNA, HBsAg/HBeAg, and preC/pgRNA levels by Southern blot, qPCR, ELISA, and RT-qPCR, respectively (F). (G) Schematic diagram of FLAG-hnRNPA2B1 truncates (left panel). HA-A3B was co-transfected with different FLAG-hnRNPA2B1 truncates in Huh7 cells for 48 h. Co-IP assay was performed to analyze their interactions. (H) FLAG-A2B1 and A2B1ΔPrLD were transfected in HBV-infected HLCZ01 cells for 72 h, the enrichment of A3B on cccDNA was analyzed by ChIP assay, and the efficiency of immunoprecipitation was detected by western blot. Results are representative of two [(A, B) and (G)] or three [(C, D) to (F) and (H)] independent experiments. P-values were determined by one-way ANOVA (F) or two-way ANOVA (C, H); **P < .01, NS: non-significant.

hnRNPA2B1 contains two RNA recognition motifs (RRM1 and RRM2) and a prion-like domain (PrLD) [29]. To delineate the APOBEC3B-binding domain, we constructed a series of truncated hnRNPA2B1 mutants that retained their nuclear localization signal (Fig. 5G, left). Co-IP assay showed that APOBEC3B was immunoprecipitated by hnRNPA2B1-FL, ΔRRM1, and ΔRRM2 variants but not by the ΔPrLD mutant (Fig. 5G, middle). Crucially, the PrLD domain itself was sufficient for its interaction with APOBEC3B (Fig. 5G, right). Deletion of this domain did not affect the binding of hnRNPA2B1 to the G4 structures of cccDNA (Supplementary Fig. S4G and H), but it attenuated the hnRNPA2B1–mediated enhancement of APOBEC3B recruitment to cccDNA (Fig. 5H) and abolished the ability to reduce cccDNA and viral antigen levels (Supplementary Fig. S4I), confirming that PrLD acts as the molecular linker for hnRNPA2B1 function. 3D-PCR analysis showed that overexpression of wild-type A2B1 reduced the PCR denaturation temperature, indicating deamination-induced mismatches, whereas the G4-binding-deficient mutant (A2B1-mG4bd) had no such effect (Supplementary Fig. S4J). This confirms that G4-binding is essential for A2B1 to promote cccDNA deamination.

HBx targets hnRNPA2B1 for CRL4-DDB1-mediated ubiquitination and degradation

It is well established that HBV has developed multiple strategies to circumvent host restriction mechanisms [30]. Therefore, we investigated whether the expression of hnRNPA2B1 is modulated by HBV. As shown in Fig. 6A and Supplementary Fig. S5A, hnRNPA2B1 mRNA levels did not change, but its protein abundance was reduced in MC-HBV-transfected Huh7 cells, HBV-infected HLCZ01 cells, and HBV-infected HepG2^NTCP cells, suggesting post-translational regulation. To determine how HBV regulates hnRNPA2B1 expression, HBV proteins (HBx, HBc, and L-HBsAg) were overexpressed in Huh7 cells. Among them, HBx, rather than other viral proteins, repressed hnRNPA2B1 expression in a dose-dependent manner (Fig. 6B and C and Supplementary Fig. S5B). Furthermore, either mutation of A at 1376 (ATG→TTG) that results in HBx deletion in MC-HBV (MC-HBV∆HBx) or the use of HBx-deficient HBV damaged the HBV-mediated suppression of hnRNPA2B1 (Fig. 6D and E). Consistently, the CHX-chase assay confirmed that ectopic HBx expression reduced the half-life of hnRNPA2B1 protein (Fig. 6F). These results suggested that HBx inhibits hnRNPA2B1 protein expression by accelerating its turnover.

Figure 6.

HBx promotes the ubiquitination and degradation of hnRNPA2B1. (A) HLCZ01 or HepG2^NTCP cells were infected with HBV for 5 days. hnRNPA2B1 mRNA and protein were detected by RT-qPCR and western blot, respectively. Huh7 cells were transfected with MC-HBV and different viral protein constructs (B), or HA-HBx construct at different doses (C), or MC-HBV and MC-HBVΔHBx (D) for 3 days, hnRNPA2B1 expression was detected by western blot. (E) HepG2^NTCP cells were infected with HBV and HBVΔHBx for 5 days and hnRNPA2B1 expression was detected by western blot. (F) Huh7 cells with HBx overexpression were incubated with CHX for indicated times. hnRNPA2B1 turnover was detected by western blot and quantified using ImageJ (band intensity was normalized to β-actin and band intensity at 0 h was defined as 100%). (G) HEK293T cells co-transfected with HA-HBx and FLAG-hnRNPA2B1 plasmids for 48 h were treated with MG-132 for 8 h before harvest, followed by a Co-IP assay to assess their interaction. (H) Co-localization of hnRNPA2B1 (green) with HA-HBx (red) in HepG2 cells detected by confocal microscopy. The intensity profiles correspond to the white lines drawn in the merged images. Scale bars, 10 μm. (I) HA-HBx was co-transfected with different FLAG-hnRNPA2B1 truncates in Huh7 cells for 48 h. Co-IP assay was performed to analyze their interactions. Huh7 cells were transfected with HA-HBx (J) or HA-HBx-R96E (L) for 48 h. HepG2^NTCP cells infected with HBV and HBVΔHBx for 5 days (K). All cells were then treated with MG132 for 8 h before harvest, and endogenous hnRNPA2B1 was immunoprecipitated and immunoblotted with anti-Ub antibody to measure its ubiquitylation. Results are representative of two [(A, B) and (G)] or three [(C–F), (H–L)] independent experiments. P-values were determined by unpaired t-test (A); NS: non-significant.

It is known that HBx binds to the cullin 4 RING E3 ubiquitin ligase complex (CRL4) by interacting with DDB1 to degrade host proteins [31, 32]. We explored whether HBx is involved in modulating hnRNPA2B1 ubiquitination. Co-IP analysis revealed the interaction of hnRNPA2B1 with HBx (Fig. 6G). Immunofluorescence staining further demonstrated the colocalization of HBx and hnRNPA2B1, and a weaker hnRNPA2B1 signal in HBx-expressing cells (Fig. 6H). This interaction was observed with hnRNPA2B1-FL, ΔRRM1, and ΔPrLD variants but not with the ΔRRM2 mutant (Fig. 6I). Ubiquitination analysis showed that HBx overexpression upregulated the polyubiquitination of endogenous hnRNPA2B1 upon pretreatment with the proteasome inhibitor MG132 (Fig. 6J). HBV infection also upregulated the ubiquitination of hnRNPA2B1, and this effect is abolished by HBx deletion (Fig. 6K). Crucially, the DDB1-binding-defective HBx mutant, HBx-R96E [31], can still bind hnRNPA2B1 (Supplementary Fig. S5C) but fails to induce its polyubiquitination (Fig. 6L). Next, we investigated whether cccDNA becomes more susceptible to deamination in the absence of HBx. Given that HBx is essential for cccDNA transcription, we knocked down SMC5 to silence the SMC5/6 complex that was reported to restore the transcriptional activity of cccDNA with HBx deficiency [32] and found that cccDNA in HBVΔHBx-infected HepG2^NTCP cells exhibited increased sensitivity to APOBEC3B-mediated deamination (Supplementary Fig. S5D). Taken together, these results demonstrate that HBV suppresses hnRNPA2B1-mediated antiviral defense via HBx-driven recruitment of the CRL4-DDB1 E3 ligase complex, which targets hnRNPA2B1 for ubiquitination-dependent proteasomal degradation.

Discussion

HBV cccDNA is a viral persistence reservoir and a key obstacle for the treatment of chronic hepatitis B. Understanding the mechanisms and factors involved in the stability of cccDNA could be essential for the development of cccDNA elimination strategies. In the present study, we identified hnRNPA2B1 as a cccDNA-binding factor and host restriction factor for HBV infection. hnRNPA2B1 engages cccDNA and recruits APOBEC3B to induce cccDNA C>T editing and destabilization. To maintain persistence, HBV evades the repression of hnRNPA2B1 via HBx-induced ubiquitin-proteasome degradation. Therefore, this study expands the host defense mechanisms to degrade cccDNA and provides a promising therapeutic target for HBV treatment.

The recognition of viral nucleic acids by pattern recognition receptors that initiate the induction of type I IFNs is a critical mechanism for eliminating viral infection. However, HBV is considered a “stealth” hepatotropic virus that does not invoke strong IFN responses or IFN-stimulated gene induction [33–35]. hnRNPA2B1, an RNA-binding protein and m6A reader, plays a broad role in RNA metabolism, including mRNA/miRNA splicing, transport, stabilization, and translational regulation [29]. In addition, hnRNPA2B1 has been characterized as a nuclear DNA sensor that undergoes cytoplasmic translocation following pathogen-derived DNA detection to activate innate immune cascades [19]. Here, we identified hnRNPA2B1 as a nuclear cofactor that directly engages cccDNA via G4 structural motifs. This interaction facilitates the recruitment of cytidine deaminase APOBEC3B via hnRNPA2B1’s PrLD domain, establishing a targeted antiviral deamination complex on the cccDNA minichromosome. This model finds structural concordance in published work, showing that the hnRNPA2B1 homodimer forms in solution by binding to pre-generated ssDNA or dsDNA with a U-shaped bulge, similar to a G4 structure [15]. G4 usually acts as a switch for the regulation of gene expression. Biswas et al. were the first to report G4 formation in the preS2/S promoter of HBV genotype B, and mutations in G4 resulted in the reduction of preS2/S transcripts and HBsAg levels [36]. Meier-Stephenson et al. discovered another G4 in the pre-core promoter of HBV cccDNA, which was recognized by the known G4-binding protein DHX36 [37]. Recently, Giraud et al. demonstrated that cccDNA contains ten G4 structures, among which two within enhancer I facilitate FUS-mediated phase separation and transcription of cccDNA [20]. The minichromosomal and episomal nature of cccDNA often prevents recognition by DNA damage sensors. Interestingly, functioning as a G4-binding factor and molecular scaffold, hnRNPA2B1 capitalizes on G4 vulnerability that is important for cccDNA activity, thereby coupling cccDNA recognition and APOBEC3B recruitment to achieve cccDNA degradation. G4 also exists in other viral extrachromosomal circular DNA, such as Kaposi’s sarcoma-associated herpesvirus (KSHV), Epstein–Barr virus (EBV), and human papillomavirus (HPV) [38–40]. Our findings imply that the hnRNPA2B1-APOBEC3B axis may serve as an important host defense mechanism against the replication of multiple DNA viruses.

Comprising seven DNA cytidine deaminases (APOBEC3A, B, C, D, F, G, and H), the APOBEC3 protein family plays important roles in innate immunity against viral infections by introducing mutagenic deamination into viral genomes. Cellular localization critically governs the APOBEC3 functionality. Unlike the dispersive nucleocytoplasmic distribution of A3A, A3C, and A3H, A3B is predominantly expressed in the nucleus [41]. However, the cofactor of A3B remains unclear. Interactome analysis of APOBEC3B in multiple myeloma identified several hnRNPs, including hnRNPA1, hnRNPK, hnRNPA3, and hnRNPC [42]. Another protein interaction map of the APOBEC3 family in HEK293T also revealed the binding of A3B to hnRNPL, hnRNPL1, and hnRNPL2 [43]. Recently, Chen et al. reported that disruption of protein interactions between A3 and hnRNPs through A3G and A3B mutagenesis decreased A3 mutational activity, indicating that hnRNPs are closely associated with A3s [12]. Here, we found that hnRNPA2B1 selectively interacts with A3B via its PrLD domain and is of great importance for loading A3B onto cccDNA, which highlights the role of hnRNPA2B1 as an A3B cofactor in regulating its orientation and function.

HBx is a multifunctional regulator that interacts with host factors to modulate protein degradation, transcription, signal transduction, cell cycle progression, and genetic stability [44]. Degradation of the SMC5/6 complex by hijacking the DDB1-CRL4 E3 ubiquitin ligase complex is considered to be the vital mechanism of HBx in maintaining cccDNA transcription activity [32]. However, the role of HBx in regulating cccDNA stability remains unclear. Wang et al. found that cccDNA without HBx expression exhibits low transcriptional activity and an epigenetically inactive pattern, making it more difficult to access by APOBEC3A and exhibiting a significantly longer persistence in mice [45]. In contrast, Gao et al. reported that HBx-elevated male-specific lethal 2 (MSL2) expression facilitates the maintenance of HBV cccDNA stability through the MSL2-mediated degradation of APOBEC3B by ubiquitylation [46]. Zhang et al. discovered that HBx upregulates heat shock protein B1 (HSPB1) to stabilize HBV cccDNA and suppress immune responses [47]. Similarly, our study indicated that HBx-induced degradation of hnRNPA2B1 may enable the virus to bypass the recruitment of APOBEC3B and prevent cccDNA deamination. Correspondingly, under SMC5 knockdown condition that can maintain the transcriptional activity of cccDNA with HBx deficiency [32], cccDNAΔHBx exhibited greater sensitivity to APOBEC3B–induced deamination, highlighting the importance of HBx in maintaining cccDNA stability beyond its transcriptional regulation.

The binding of HBx to hnRNPA2B1, which induces ubiquitination and degradation, raises the question of whether hnRNPA2B1 can be targeted to block its interaction with HBx. In 2022, PAC5, a novel broad-spectrum antiviral compound, was found to target Asp49-contained pocket 4 in the RRM1 domain of hnRNPA2B1. This interaction promotes its translocation to the cytoplasm and activates the TBK1-IRF3 signaling pathway [48]. Recently, Zhang et al. identified 29 hnRNPA2B1-bound endogenous metabolites from a library containing 769 endogenous metabolites using surface plasmon resonance (SPR) screening. Adenine interacts with hnRNPA2B1 depending on its phenylalanine at position 12 and then recruits hnRNPA2B1 to Il1b enhancers [49]. Dicoumarol, a competitive NADPH quinone oxidoreductase (NQO1) inhibitor, and nitazoxanide (NTZ), a thiazolide anti-infective agent, have been reported to promote HBx proteasomal degradation and block HBx-DDB1 protein interaction, respectively [50, 51]. These findings provide valuable clues for future development of targeted intervention strategies aimed at the interaction of HBx with hnRNPA2B1.

In summary, our findings uncover the pivotal role of hnRNPA2B1 as a cccDNA interactor, which recruits APOBEC3B to catalyze the deamination of cytosines and elimination of cccDNA. HBx represses hnRNPA2B1 expression to evade hnRNPA2B1-APOBEC3B axis-mediated inhibition. This work broadens our understanding of host-HBV interactions and provides a novel strategy for HBV therapeutics.

Supplementary data

Supplementary data is available at NAR online.

Conflict of interest

None declared.

Funding

This study was supported by grants from the National Key R&D Program of China (2022YFC2303600), Taishan Scholarship (tsqn202306008), National Science Foundation of China (8 2321002, and 82402604), National Science Foundation of Shandong (ZR2024YQ072), Cutting Edge Development Fund of Advanced Medical Research Institute (GYY2023QY01), and the Major Science and Technology Innovation Project of Shandong Province (2024ZLGX01). Funding to pay the Open Access publication charges for this article was provided by National Key R&D Program of China (2022YFC2303600) and National Science Foundation of Shandong (ZR2024YQ072).

Data availability

All data relevant to the study are included in the article or uploaded as supplemental information.