More than 257 million people worldwide live with chronic hepatitis B virus (HBV) infection, facing elevated risks of liver cirrhosis and hepatocellular carcinoma that current drugs can suppress but almost never eliminate. A study published on 30 March 2026 in the Journal of Biomedical Science now reveals an unexpected human defense aimed directly at the virus’s genetic command center. A team led by Nayeon Kim and corresponding author Kyun-Hwan Kim of Sungkyunkwan University School of Medicine, working with proteomics specialists at Kyung Hee University, identified interferon-inducible protein 35 (IFI35) as a potent host restriction factor that shuts down HBV transcription from its covalently closed circular DNA (cccDNA) minichromosome. The mechanism is strikingly direct: IFI35 recruits the E3 ubiquitin ligase TRIM21 to tag hepatocyte nuclear factor 4α (HNF4α)—a liver transcription factor the virus depends on to switch on its genes—for destruction by the proteasome.

HBV belongs to the hepadnavirus family and carries one of the smallest genomes of any human virus: a 3.2-kilobase, partially double-stranded DNA molecule known as relaxed circular DNA, or rcDNA. After the virus enters a hepatocyte, this genome travels to the nucleus, where cellular machinery converts it into cccDNA—a minichromosome that persists for the life of the infected cell and serves as the exclusive template for every viral RNA, including the 3.5-kilobase pregenomic RNA from which new rcDNA is generated by reverse transcription. Nucleos(t)ide analogues silence the viral polymerase but leave cccDNA untouched, which is why patients must take them indefinitely, and pegylated interferon-alpha helps only a subset of patients at the cost of significant side effects. Cytokines such as tumor necrosis factor-alpha (TNF-α) and interferon-gamma (IFN-γ), delivered to the liver by T cells, are known to reduce cccDNA without killing hepatocytes—a celebrated noncytolytic clearance route—yet the intracellular mediators that execute this order have remained only partly identified.

To find the missing executors, the researchers treated Huh7 hepatoma cells with TNF-α and IFN-γ and profiled the proteome using tandem mass tag 6plex labeling coupled to liquid chromatography–tandem mass spectrometry on an Orbitrap Exploris 480 instrument, quantifying a total of 5,598 proteins. Single-sample gene set enrichment analysis against the Hallmark database ranked the “Interferon gamma response” as the most enriched pathway, ahead of TNF-α signaling and inflammatory response, and a protein–protein interaction network assembled from the STRING database placed IFI35 among the most connected nodes alongside STAT1, the signature transcription factor of IFN-γ signaling. Among differentially expressed proteins—defined by an absolute log₂ fold change above 0.5 and an adjusted p-value below 0.05—IFI35 stood out as one of the most dramatically induced, with a log₂ fold change exceeding 1.5 and an adjusted p-value under 0.001. The 35-kDa leucine zipper protein, originally cloned from IFN-γ-treated HeLa cells in 1994, had never before been functionally linked to HBV.

Public clinical datasets reinforced the suspicion that IFI35 matters during genuine infection. Mining Gene Expression Omnibus datasets GSE83148, GSE96851 and GSE159413, the team found IFI35 transcripts significantly elevated in HBV-positive livers compared with uninfected controls. In GSE159413, patients on long-term nucleos(t)ide analogue therapy showed IFI35 levels falling back toward the healthy baseline—exactly what one would expect if the protein tracks active viral replication and the cytokine drive that accompanies it. IFI35 was also markedly increased in liver tissue from patients with HBV-associated acute liver failure, and in a biopsy dataset covering the natural phases of chronic infection it peaked in the immune-clearance phase, when cytotoxic lymphocytes flood the liver with IFN-γ and TNF-α, and sank during the quiescent immune-tolerance phase.

The functional experiments were unambiguous. When HepG2 or Huh7 cells were co-transfected with a replication-competent HBV 1.2mer plasmid and increasing doses of myc-tagged IFI35, Southern blotting of core-associated DNA revealed dose-dependent collapse of viral replication intermediates, while ELISA measurements recorded parallel drops in secreted hepatitis B surface and e antigens. Conversely, small interfering RNA knockdown of IFI35 raised replication and antigen output—and, critically, partially rescued viral replication from suppression by exogenous cytokines, an effect most visible under IFN-γ alone, establishing IFI35 as a working arm of the gamma-interferon response. IFN-γ induced IFI35 in hepatocytes in a dose-dependent manner, whereas IFN-α had only a modest effect. The team then moved into authentic infection systems: primary human hepatocytes isolated by two-step collagenase perfusion from four donor livers, and HepG2-NTCP cells expressing the HBV entry receptor, both challenged with 1,000 genome equivalents per cell of genotype D virions. IFI35 overexpression again blunted replication and antigen secretion, yet cccDNA abundance itself barely moved—indicating that the protein paralyzes the template’s transcriptional output rather than destroying the template.

Northern blotting localized the block with precision: both precore/pregenomic and preS/S RNAs fell dose-dependently as IFI35 rose, yet an actinomycin D chase demonstrated that viral RNA decay kinetics were identical with or without the protein, ruling out effects on RNA stability and pinning IFI35 at the transcriptional step itself. Luciferase reporters for HBV enhancer I and the enhancer II/core promoter lost activity in an IFI35-dose-dependent fashion, and the protein suppressed an HBx-deficient viral genome just as effectively as the wild type, excluding the viral X protein as an intermediary. The target emerged from a screen of candidate hepatocyte nuclear factors: among HNF1α, HNF4α and HNF3β, only HNF4α protein declined as IFI35 accumulated. Real-time PCR showed HNF4α mRNA untouched, pointing to post-transcriptional control. The proteasome inhibitor MG132 rescued HNF4α levels, whereas bafilomycin A1, which blocks lysosomal and autophagic degradation, did not. Ubiquitination assays with HA-tagged ubiquitin constructs revealed that IFI35 drives K48-linked polyubiquitination of HNF4α—the canonical proteasome-targeting mark—while K48R mutant or K63-specific chains failed to do so, and co-immunoprecipitation confirmed a physical IFI35–HNF4α complex.

Because IFI35 carries no HECT, RING or F-box motifs, it cannot ligate ubiquitin itself; the team reasoned that it must act as an adaptor. Their attention turned to TRIM21, an E3 ligase previously implicated by the same group in IFN-γ-mediated HBV suppression and known to ubiquitinate Nmi, the protein that partners with IFI35 in other contexts. Co-immunoprecipitation confirmed an IFI35–TRIM21 interaction in hepatocytes, and the decisive test came with a CRISPR-generated TRIM21 knockout HepG2 line: in these cells, IFI35 lost most of its power to reduce HBV DNA, viral RNA, HNF4α protein and secreted antigens. A cycloheximide chase quantified the difference starkly—the half-life of HNF4α fell below 60 minutes in wild-type cells expressing IFI35, yet barely budged over two hours in the knockout. Enhancer reporters engineered with HNF4α binding sites deleted lost their sensitivity to IFI35, and removing both sites from enhancer II nearly abolished reporter activity altogether. Domain mapping completed the picture: IFI35’s N-terminal leucine zipper is dispensable, at least one of its two tandem N-myc-interacting domains (NID1 and NID2) is essential for antiviral activity, and NID2 in particular is required for TRIM21 engagement—consistent with a scaffold in which IFI35 grips HNF4α through its NIDs while presenting the substrate to TRIM21 for K48-linked destruction. Supplying excess HNF4α partially restored replication suppressed by IFI35, and conversely, when HNF4α was knocked down, IFI35 could no longer push antigen secretion any lower.

The mechanism survived the leap into living animals. In a hydrodynamic injection model, C57BL/6 mice received the HBV 1.2mer plasmid together with either empty vector (n = 4) or an IFI35 expression plasmid (n = 7). Southern and Northern blots of liver tissue revealed dramatic reductions in intrahepatic HBV DNA and viral transcripts in every IFI35-treated animal, with serum HBeAg and HBsAg falling in parallel. Most tellingly, hepatic HNF4α protein declined in exact inverse proportion to IFI35, a correlation significant at p < 0.0001. Correlation analyses across three independent HBV patient datasets—GSE83148, GSE96851 and GSE14668—confirmed the same statistically significant inverse relationship between IFI35 and HNF4α expression in human livers, lending clinical weight to the mouse result.

The discovery slots into a rapidly growing catalogue of interferon-stimulated genes that attack HBV from different angles: APOBEC3A and APOBEC3B deaminate cccDNA toward its destruction, ISG20 selectively degrades N6-methyladenosine-modified viral transcripts, and indoleamine-2,3-dioxygenase starves infected cells of substrates needed for viral protein synthesis. IFI35’s own track record is double-edged—it supports vesicular stomatitis virus replication by dampening RIG-I signaling yet restrains foamy viruses by sequestering their Tas transactivator—but against HBV it is decisively antiviral, and notably it exploits the same TRIM21 partnership documented in earlier work, this time directing K48-linked ubiquitination at a host transcription factor rather than a signaling adaptor. HNF4α has surfaced repeatedly as the virus’s Achilles heel: interleukin-32γ and cleaved c-FLIP lower it through MAPK kinase signaling, steroid receptor coactivator 3 blocks its nuclear import via Akt, and estrogen receptor-alpha physically prevents it from binding the viral enhancers. The authors caution that because HNF4α supplementation did not fully restore replication, IFI35 may harbor additional antiviral targets still to be defined.

The practical implication is a new druggable node in a disease that desperately needs one. Agents that stabilize IFI35, strengthen its recruitment of TRIM21, or mimic its substrate-handling could push cccDNA transcription below the threshold required to sustain infection, complementing polymerase inhibitors and potentially shortening the lifelong treatment courses they currently demand. As the authors conclude, the IFI35–TRIM21–HNF4α axis appears to play a crucial role in the TNF-α- and IFN-γ-mediated suppression of HBV and holds clear value for the development of alternative anti-HBV drugs. For a virus that has evaded eradication for half a century by hiding a circular DNA template inside the chromosomes of the liver’s own cells, the disclosure of a human protein capable of disarming that template’s master regulator marks a genuinely consequential advance.