Folded DNA Structures in the Hepatitis B Genome Steer Viral Gene Expression

Hepatitis B virus remains one of the world’s most stubborn pathogens, chronically infecting hundreds of millions of people and driving liver cirrhosis and hepatocellular carcinoma despite the availability of an effective preventive vaccine. Current therapies suppress viral replication but rarely eliminate the virus, largely because the viral covalently closed circular DNA, or cccDNA, persists in hepatocytes as a minichromosome that continues to direct transcription of viral genes. A new study published in Virology Journal now adds an unexpected layer to the understanding of how this small DNA virus controls its own gene expression, showing that specific folded structures within the viral genome act as docking platforms for host proteins that fine-tune viral protein production.

The structures in question are G-quadruplexes, or G4s, four-stranded nucleic acid architectures that form in guanine-rich sequences. In a G4, four guanine bases pair through Hoogsteen hydrogen bonding to create a planar G-quartet, and the stacking of at least two such quartets, stabilized by pi-pi interactions and by monovalent cations such as potassium, produces a compact and remarkably stable fold. Depending on the orientation of the four guanine tracts that form the core, G4s can adopt parallel, antiparallel or hybrid topologies. In human cells, G4s are known to regulate telomere maintenance, DNA replication, transcription, mRNA processing and chromatin remodeling, and conserved potential G-quadruplex sequences have been documented in the genomes of Epstein-Barr virus, hepatitis C virus, HIV and SARS-CoV-2, where they influence replication and infection.

Researchers led by a team at Anhui Medical University systematically searched full-length hepatitis B virus genomes from genotypes A through H for sequences capable of forming G4s, using the QGRS Mapper algorithm to score the likelihood of quadruplex formation. High-scoring motifs clustered at positions 1204, 1732, 1886 and 3021 in a representative genotype B genome. Multiple sequence alignment and WebLogo analysis revealed that three of these motifs, at positions 1204, 1732 and 1886, are highly conserved across genotypes, with the exception of genotype G, while the 3021 motif is not evolutionarily conserved. Crucially, the 1204 motif lies within the region overlapping the X promoter and Enhancer I, the 1732 motif sits in the overlap between the core promoter and Enhancer II, and the 1886 motif falls within the epsilon RNA element of the precore and core gene.

To confirm that these predicted sequences actually fold, the team synthesized the corresponding oligonucleotides and subjected them to a battery of biophysical tests. Native polyacrylamide gel electrophoresis showed that the wild-type sequences migrated faster than guanine-mutated counterparts, consistent with the formation of compact intramolecular quadruplexes. Proton nuclear magnetic resonance spectroscopy detected the characteristic imino proton signals of Hoogsteen base pairing in the 10 to 12.5 parts per million range for all three sequences, signals that shifted toward Watson-Crick pairing signatures when guanines were mutated. Circular dichroism spectroscopy then resolved the topologies: the 1732 sequence adopts a parallel G4, while 1204 and 1886 fold into antiparallel conformations. Potassium ions proved preferred for stabilization, and the G4-stabilizing ligand BRACO-19 raised the melting temperatures of the structures, with 1732’s melting temperature exceeding the detectable range after ligand addition, underscoring the exceptional stability of these folds under physiological conditions.

Functional relevance was established with reporter systems. In a luciferase construct in which expression is driven by the hepatitis B virus core promoter together with Enhancer I and Enhancer II, disrupting the G4 at 1204, 1732 or 1886 significantly reduced reporter activity, with the 1732 mutation producing the strongest effect. In a more complete HBV 1.3-mer plasmid system, destroying the 1732 G4 markedly lowered the levels of hepatitis B surface antigen, e antigen and core antigen. Intriguingly, total viral RNA and the 3.5 kilobase pregenomic RNA were not significantly affected, pointing to a post-transcriptional role for the 1732 structure, possibly through modulation of translation efficiency, RNA conformation or transcript processing, rather than a direct effect on transcription itself. The 1204 mutation, constrained by the overlapping polymerase reading frame to a single G-to-A substitution that only partially weakened the fold, produced no significant changes in the 1.3-mer model, consistent with residual G4 signatures still visible in its circular dichroism spectra.

The next question was which host proteins these viral structures recruit. Using biotinylated 1204 and 1732 oligonucleotides as bait in pull-down experiments from HepG2.2.15 cell lysates, followed by data-independent acquisition mass spectrometry, the researchers identified 619 proteins enriched more than 1.5-fold over a non-G4 control, of which 34 participate in DNA repair, transcriptional regulation, RNA processing or G4 binding. The cellular nucleic acid-binding protein CNBP, a CCHC-type zinc finger protein, emerged as the most highly enriched G4-binding partner, while the hepatocyte nuclear factor 4 alpha, HNF4A, a master transcriptional activator of the hepatitis B virus core promoter, had never previously been reported to bind G4 structures. RNA interference against five candidate proteins, including CNBP and HNF4A, significantly reduced viral antigen and RNA levels, confirming their positive contribution to viral gene expression.

Surface plasmon resonance quantified the physical interactions. Both the 1204 and 1732 quadruplexes bound HNF4A directly, with equilibrium dissociation constants of 2.552 times ten to the minus five molar and 1.991 times ten to the minus six molar respectively, meaning the 1732 structure binds roughly an order of magnitude more tightly. Neither the mutated sequences nor a non-G4 control showed any detectable binding, and replacing potassium with lithium, which disrupts G4 folding, abolished the interaction entirely, demonstrating that the folded structure itself, not the primary sequence, is what HNF4A recognizes. BRACO-19, by further stabilizing the quadruplexes, actively enhanced HNF4A binding in pull-down assays. AlphaFold3-based structural predictions supported these findings, revealing hydrogen bonds between both G4 elements and specific HNF4A residues, with partially overlapping binding interfaces.

Functional assays then separated the two host factors mechanistically. Overexpression of HNF4A markedly increased activity of the core promoter-enhancer reporter, but when the 1732 G4 was destroyed, HNF4A could no longer fully restore reporter output, indicating that an intact 1732 quadruplex is required for optimal HNF4A-dependent activation. The authors propose a dual regulatory model in which HNF4A activates the core promoter-enhancer region primarily through its classical DNA binding sites, while a secondary, G4-dependent mechanism involving direct engagement of the 1732 structure potentiates maximal promoter output, likely facilitated by the genomic proximity of the two elements. CNBP, by contrast, boosted viral antigen and RNA levels when overexpressed and reduced them when knocked down, yet it did not enhance, and even slightly decreased, core promoter-enhancer reporter activity, indicating that CNBP promotes hepatitis B virus expression through a mechanism independent of this promoter region, perhaps by stabilizing or unwinding RNA G4s within viral transcripts, as it has been shown to do with SARS-CoV-2 genomic G4s.

The study’s implications extend beyond basic virology. Because properly folded G4s in the core promoter-enhancer region support, rather than repress, viral gene expression, these structures and the proteins that bind them represent candidate targets for new anti-hepatitis B strategies, complementing G4-directed approaches already explored against SARS-CoV-2 and other viruses. The authors caution that their experiments relied largely on plasmid-based reporter and 1.3-mer systems, and that future work using authentic infection models and fully chromatinized cccDNA templates will be needed to confirm whether HNF4A associates directly with viral cccDNA in living cells. Quantifying effects on nascent RNA synthesis, RNA polymerase II recruitment and liquid-liquid phase separation, a process recently implicated in G4-driven cccDNA transcription, could further clarify how these unusual DNA folds choreograph the life of a virus that has evaded elimination for millennia.

Subject of Research: Conserved G-quadruplex structures in the hepatitis B virus core promoter and enhancer regions and their host protein partners regulate viral gene expression

Article Title: Regulatory roles of G-quadruplexes and G-quadruplex-binding proteins across the enhancer and promoter of the HBV genome

Article References: He, L., Huang, B., Ma, H., Wang, L., Wu, Q., Zhang, J., Yu, L., & Lv, S. (2026). Regulatory roles of G-quadruplexes and G-quadruplex-binding proteins across the enhancer and promoter of the HBV genome. Virology Journal, 23(1), Article 218. https://doi.org/10.1186/s12985-026-03132-2

Image Credits: AI Generated

DOI: 10.1186/s12985-026-03132-2

Keywords: hepatitis B virus, G-quadruplex, core promoter, enhancer, G4-binding proteins, HNF4A, CNBP, BRACO-19, viral gene expression, cccDNA, surface plasmon resonance, antiviral therapy