To understand why this finding matters, it helps to begin with the biology of protein turnover. Inside every cell, proteins that are no longer needed, damaged, or dangerous are tagged with small marker proteins called ubiquitins. Enzymes assemble these tags into chains, and the way the chains are linked determines the tagged protein’s fate. K48-linked polyubiquitin chains, in which ubiquitin molecules are connected through a specific lysine residue, act as a molecular kiss of death: they flag the protein for delivery to the proteasome, the cell’s garbage disposal, where it is chopped into fragments. Deubiquitinating enzymes, or DUBs, do the opposite. They snip ubiquitin tags off proteins, rescuing them from degradation. Among the largest DUB family is the ubiquitin-specific proteases, or USPs, and when one of these enzymes removes death marks from a protein that promotes cancer, the effect can be dramatic: the oncoprotein accumulates, and the cell drifts further toward uncontrolled growth.

CIAP2, also known as baculoviral IAP repeat containing protein 2 and a member of the inhibitor of apoptosis protein family, was already known to be overexpressed in hepatocellular carcinoma and to function there as an oncogene. What remained incompletely understood was how tumor cells manage to keep CIAP2 levels so high. The research team approached this question systematically. Rather than guessing which of the many ubiquitin-specific proteases might be responsible, they screened USPs that could plausibly regulate CIAP2 expression and identified USP45 as a key regulator that enhances CIAP2 levels in hepatocellular carcinoma cells. That screening result set the stage for a deeper interrogation of what USP45 is doing in these tumors and whether its activity has real consequences for cancer behavior.

Bioinformatic analysis of clinical data provided the first hint that USP45 is more than a laboratory curiosity. The researchers found that USP45 is overexpressed in hepatocellular carcinoma and, critically, that elevated USP45 is associated with worse patient survival. In other words, the enzyme is not merely present in these tumors; its abundance tracks with the outcome that matters most to patients. This kind of correlation does not by itself prove causation, but it supplied the rationale for the functional experiments that followed. If USP45 truly helps drive the disease, then removing it from cancer cells should blunt their malignant behavior in measurable ways.

That is precisely what the team observed. When the researchers knocked down USP45 in two hepatocellular carcinoma cell lines, SNU449 and SK-Hep-1, the consequences were broad and consistent. The cells lost viability, and their ability to form colonies in culture was impaired. The cell cycle was disrupted, with cells accumulating in an arrested state rather than progressing smoothly through division. Apoptosis, the programmed self-destruction that cancer cells typically resist, increased. Beyond growth and survival, the cells’ metastatic potential also weakened: migration and invasion, the two behaviors that allow a tumor to spread beyond its original site, were attenuated. Finally, sphere formation, an in vitro assay that reflects the self-renewal capacity associated with cancer stem cells, was compromised, suggesting that USP45 supports not just bulk tumor growth but also the stem-like subpopulation that often fuels relapse and treatment resistance.

The downstream signaling changes reinforced this picture. USP45 deficiency suppressed the NF-κB signaling pathway, a major pro-survival and pro-inflammatory cascade that many cancers exploit to keep growing under hostile conditions. At the same time, loss of USP45 reduced the expression of key transcription factors involved in cancer cell stemness, the molecular program that endows a subset of tumor cells with the self-renewal and differentiation plasticity of stem cells. Taken together, these results indicate that USP45 sits upstream of at least two malignant programs in hepatocellular carcinoma cells: the inflammatory survival signaling that keeps cells alive and the stemness program that maintains a reservoir of aggressive, therapy-resistant cells.

The mechanistic core of the study lies in the physical and functional relationship between USP45 and CIAP2. Using co-immunoprecipitation, the researchers showed that USP45 interacts with CIAP2 in hepatocellular carcinoma cells, and endogenous co-immunoprecipitation experiments, which detect interactions between the proteins as they naturally occur in cells rather than in artificial overexpression systems, confirmed that this partnership is physiological. Functionally, USP45 removes ubiquitination from CIAP2 and enhances its stability, meaning the enzyme directly extends the lifespan of the oncoprotein inside the cell. The team then narrowed the mechanism further with ubiquitin mutant assays. By testing cells co-transfected with Flag-tagged CIAP2, HA-tagged USP45, and either wild-type ubiquitin or mutant ubiquitin lacking the K48 or K63 linkage sites, they demonstrated that USP45 specifically targets K48-linked polyubiquitin chains on CIAP2. In plain terms, USP45 strips off precisely the ubiquitin marks that would otherwise condemn CIAP2 to the proteasome, sparing the K63-linked chains that serve other signaling roles.

A crucial rescue experiment established that CIAP2 is not merely one target among many but the functionally relevant mediator of USP45’s oncogenic activity. When the researchers overexpressed CIAP2 in USP45-deficient hepatocellular carcinoma cells, the malignant traits that had been lost came back: the cells regained the aggressive behaviors that USP45 depletion had taken away. This epistasis-style logic, in which restoring a downstream factor rescues a phenotype caused by losing an upstream factor, is a strong indication that the USP45–CIAP2 axis is the operative pathway rather than a side effect of some other change. It also suggests that the amount of CIAP2 available to the tumor cell, which USP45 controls by editing its ubiquitin tags, is a limiting factor for the malignant state.

The final piece of evidence came from an animal model. In a hepatocellular carcinoma xenograft mouse model, in which human cancer cells are grown as tumors in immunodeficient mice, depletion of USP45 inhibited tumor growth and disrupted various malignant characteristics of the tumors. This in vivo result matters because cell culture experiments, however thorough, cannot fully capture the complexity of a growing tumor with its nutrient gradients, stromal interactions and selective pressures. The xenograft data indicate that the USP45–CIAP2 mechanism is not an artifact of the culture dish but operates in the context of an actual expanding tumor, strengthening the case that the pathway is relevant to real disease progression.

Collectively, the findings position USP45 as an oncogene in hepatocellular carcinoma and define its mechanism of action with unusual precision: USP45 binds CIAP2, removes K48-linked polyubiquitin chains from it, stabilizes the protein, and thereby sustains the survival, proliferation, motility and stem-like behavior of liver cancer cells, with the axis confirmed both in cultured cells and in xenograft tumors. The study was approved by the Ethics Committee of The Affiliated Hospital of Putian University and was supported by the Fujian Provincial Natural Science Foundation and the Joint Funds for the Innovation of Science and Technology of Fujian Province. For the field, the work adds USP45 to the growing list of deubiquitinating enzymes implicated in cancer and nominates the USP45–CIAP2 pair as a potential therapeutic axis. Because the enzyme–substrate relationship is defined at the level of a specific ubiquitin linkage, it offers a concrete biochemical handle for future drug discovery efforts aimed at destabilizing CIAP2 in liver tumors, although translating such mechanistic insight into clinical interventions will require the usual long road of validation and development.

Subject of Research: The role of the deubiquitinating enzyme USP45 in stabilizing CIAP2 and promoting hepatocellular carcinoma progression

Article Title: USP45 Promotes hepatocellular carcinoma progression by regulating the ubiquitination and stability of CIAP2

Article References: Jiang, X., Chen, J., Lin, W., Huang, K., Wang, L., & Lin, J. (2026). USP45 Promotes hepatocellular carcinoma progression by regulating the ubiquitination and stability of CIAP2. Cellular and Molecular Life Sciences. https://doi.org/10.1007/s00018-026-06409-5

Image Credits: AI Generated

DOI: 10.1007/s00018-026-06409-5

Keywords: hepatocellular carcinoma, USP45, CIAP2, deubiquitinating enzymes, ubiquitination, K48-linked polyubiquitin, proteasomal degradation, NF-kB signaling, cancer stem cells, xenograft mouse model, liver cancer, oncoprotein stability

The centerpiece of the discovery is the UT3 domain of Ufd1. Using a combination of X-ray crystallography, AlphaFold3-based structural modeling, chemical synthesis of ubiquitin conjugates, hydrogen-deuterium exchange mass spectrometry, dye-accessibility assays, and functional tests in yeast and human cells, the researchers demonstrated that UT3 binds simultaneously to two K48-linked ubiquitins and actively destabilizes one of them. The mechanism is elegant and, in structural terms, surprisingly simple. The C-terminal beta-strand of a ubiquitin molecule is pried loose and tucked into a conserved hydrophobic cleft on the surface of UT3. This beta-strand augmentation—an energetic trick also seen in other ubiquitin-binding systems—effectively rips the strand out of ubiquitin’s own beta-sheet, collapsing the fold and creating an unstructured initiation region.

Crucially, binding to a single ubiquitin is not enough. The team showed that UT3 presents two distinct interaction surfaces: a ridge site and the hydrophobic cleft. Both sites must be occupied at once for unfolding to occur. When UT3 grips two K48-linked ubiquitins—one on the ridge and one in the cleft—the cooperative binding provides the energetic leverage needed to overcome the substantial thermodynamic barrier that keeps ubiquitin folded. Computational predictions with AlphaFold3, guided by specified isopeptide linkages, supported a model in which the geometry of K48-linked diubiquitin is uniquely suited to span the two sites, explaining why Ufd1 preferentially handles K48 chains, the canonical proteasomal degradation signal.

To probe this mechanism directly, the researchers turned to chemical biology. They synthesized K48-linked ubiquitin conjugates carrying a modified C-terminal peptide and fused a 19-residue peptide corresponding to the ubiquitin C-terminus to UT3, crystallizing the resulting construct. The structure deposited in the Protein Data Bank (PDB 22ID) revealed the cleft-bound C-terminal strand in atomic detail, showing how conserved hydrophobic residues of UT3 cradle the peptide while the rest of the ubiquitin fold peels away. Dye-accessibility experiments, in which a buried fluorophore in ubiquitin becomes fluorescent only when the protein unfolds, provided direct biochemical evidence that UT3 unfolds K48-linked diubiquitin and triubiquitin in the complete absence of ATP, nucleotides, or any other energy source.

The team then traced the handoff of the unfolded tag. Once UT3 has destabilized the initiator ubiquitin, the unfolded polypeptide is captured by Npl4 and the Cdc48 hexamer. Earlier cryo-EM structures had shown the initiation complex poised with an unfolded ubiquitin at the pore of the ATPase, but they could not explain how the ubiquitin got there. The new work supplies the missing step: Npl4, together with a second region of Ufd1 called the UT6 domain, facilitates the transfer of the UT3-unfolded ubiquitin to the central pore of Cdc48, allowing the ATPase’s pore loops to grip the chain and begin ATP-powered translocation of the entire substrate.

The functional importance of the mechanism was confirmed with mutants. When the researchers mutated conserved residues in either the ridge or the cleft site of UT3, ubiquitin unfolding was impaired both in vitro and in cells. These unfolding-defective Ufd1 mutants compromised the ability of the Cdc48/p97 machinery to extract and process polyubiquitinated model substrates, including fluorescently labeled sfGFP and Dendra reporters. Complementation experiments in human cells further showed that the mutants fail to support p97-dependent protein quality control. Notably, conventional ubiquitin-binding domains such as UBA domains could not substitute for UT3, underscoring that Ufd1’s domain is not merely a tether but a true ATP-independent unfoldase.

The finding reframes a long-standing question in protein degradation. The proteasome and its associated unfoldases typically need an unstructured initiation region to engage a substrate, and ubiquitin chains themselves serve as that initiation point for Cdc48/p97. But ubiquitin’s exceptional stability posed a paradox: how does a system that runs on ATP hand off a substrate that is still fully folded? The answer, according to the new study, is that the cofactor solves the problem before the ATPase is even engaged. Simple protein-protein interactions—the burial of a beta-strand in a hydrophobic groove, multiplied across two cooperatively bound ubiquitins—are sufficient to pay the energetic cost of unfolding. The ATP of the machine is then spent exclusively on pulling the substrate through the pore.

The implications reach well beyond basic biochemistry. Cdc48/p97 is a major drug target: inhibitors such as CB-5083 disrupt protein homeostasis in multiple myeloma and other cancers, and the disulfiram metabolite has been shown to act through the p97 adaptor NPL4. Mutations in p97 cause multisystem proteinopathy, a degenerative disease affecting muscle, bone, and brain, and p97 dysfunction is implicated in a range of neurodegenerative conditions. By pinpointing the UT3 cleft and ridge as obligate, druggable interaction surfaces that the machine cannot function without, the study offers a new structural blueprint for molecules that could selectively tune or sabotage ubiquitin-chain engagement—potentially complementing existing ATPase-pocket inhibitors with cofactor-targeted strategies.

The work also highlights a recurring design principle in cell biology: binding energy can substitute for chemical energy. Beta-strand augmentation, the same mechanism UT3 exploits, is used by HECT E3 ligases during ubiquitin ligation, and ubiquitylation itself is known to destabilize the folds of modified proteins. The Westlake team’s demonstration that two simultaneous ubiquitin-binding events can push a remarkably stable protein over its unfolding barrier shows how cells choreograph degradation with remarkable economy—recruiting a chain, unzipping one ubiquitin with nothing more than a hydrophobic pocket, and only then firing the ATP-driven motor. As the authors note, the results provide a reasonable and satisfying explanation of how simple protein-protein interactions cause the unfolding of one of nature’s most stable small proteins, resolving the initiating step of an essential quality-control pathway that every cell relies upon, and that medicine increasingly seeks to control.

Ubiquitin’s reputation for stability has deep historical roots. Its structure was solved at high resolution as early as 1987, revealing the compact beta-grasp architecture that has since made it a textbook model of a small, well-folded protein. Decades of biophysical work, including mechanical unfolding simulations and single-molecule pulling experiments, established that ubiquitin withstands considerable force before its beta-sheet gives way. That resilience is precisely what makes it a durable molecular tag in the crowded cellular environment, but it also created the puzzle that the Cdc48 system had to solve: a degradation signal that must itself be destroyed before the degradation machinery can begin.

The concept of an initiation region has long shaped thinking about ATP-dependent proteolysis. Studies of the proteasome showed that efficiently degraded substrates generally require a loosely folded or unstructured segment long enough to be grasped by the pore loops of the AAA+ motor and pulled through. For many substrates, intrinsically disordered tails serve this purpose. For the vast population of proteins marked only by polyubiquitin, the chain itself must double as the initiation region, which means the cell needs a way to convert a folded, stable ubiquitin module into an unstructured peptide strand. The new findings assign that conversion to a specific moment in the pathway, upstream of both Npl4 engagement and ATP hydrolysis by the hexamer.

The link-type specificity of the system also gains a mechanistic foundation. K48 linkages are the canonical destruction signal, and cellular surveys of ubiquitin linkage abundance have shown that the Cdc48 axis contributes disproportionately to the processing of K48-marked proteins. The requirement that two K48-linked ubiquitins simultaneously occupy the ridge and cleft surfaces of UT3 provides a structural rationale for this preference: the geometry of the linkage positions the two ubiquitin units at exactly the spacing needed for cooperative binding, whereas other linkage types would fail to bridge the two sites productively.

There is also a broader systems context worth noting. Cdc48/p97 participates in pathways ranging from ER-associated degradation to the extraction of proteins from chromatin and ribosome assembly intermediates, and recent work has described bidirectional shuttling of substrates between the ATPase and the proteasome. In all of these settings, the same bottleneck applies: a polyubiquitinated substrate must be committed to translocation before ATP consumption becomes productive. An ATP-independent unfolding step at the front end therefore has economy implications for the entire network, ensuring that the motor’s nucleotide cycle is not wasted on substrates that cannot be engaged.

Methodologically, the study illustrates the growing power of combining classical crystallography with modern prediction tools. Guided structural modeling of defined isopeptide linkages, validated against a deposited crystal structure and orthogonal biochemical assays, allowed the team to test a cooperative two-site binding model that would have been difficult to capture in a single static snapshot. The convergence of such evidence, from dye-accessibility kinetics to cellular complementation, sets a high bar for the claim that a small cofactor domain can perform genuine mechanical work on a protein fold using binding energy alone.

Subject of Research: ATP-independent ubiquitin unfolding by the Ufd1 UT3 domain during Cdc48/p97-mediated substrate processing

Article Title: ATP-independent unfolding of ubiquitin by Ufd1 initiates Cdc48/p97-mediated substrate processing

Article References: Wang, Y., Zhang, Z., He, W., Wang, P., Du, J., Pan, J., Feng, S., Huang, J., & Ji, Z. (2026). ATP-independent unfolding of ubiquitin by Ufd1 initiates Cdc48/p97-mediated substrate processing. Nature Structural & Molecular Biology. https://doi.org/10.1038/s41594-026-01884-7

Image Credits: AI Generated

DOI: 10.1038/s41594-026-01884-7

Keywords: Ufd1, Cdc48, p97, VCP, Npl4, ubiquitin, K48-linked polyubiquitin, protein degradation, proteasome, ATP-independent unfolding, protein quality control, UT3 domain