key: cord-0710286-bsddq3ex
authors: Ueda, Chie; Langton, Michelle; Chen, Jiahua; Pandelia, Maria-Eirini
title: The HBx protein from hepatitis B virus coordinates a redox-active Fe-S cluster
date: 2022-02-08
journal: J Biol Chem
DOI: 10.1016/j.jbc.2022.101698
sha: d73f79b8e95d1f70d8e3cbb4783c05270c46f273
doc_id: 710286
cord_uid: bsddq3ex

The viral protein HBx is the key regulatory factor of the hepatitis B virus (HBV) and the main etiology for HBV-associated liver diseases, such as cirrhosis and hepatocellular carcinoma. Historically, HBx has defied biochemical and structural characterization, deterring efforts to understand its molecular mechanisms. Here we show that soluble HBx fused to solubility tags copurifies with either a [2Fe-2S] or a [4Fe-4S] cluster, a feature that is shared among five HBV genotypes. We show that the O(2)-stable [2Fe-2S] cluster form converts to an O(2)-sensitive [4Fe-4S] state when reacted with chemical reductants, a transformation that is best described by a reductive coupling mechanism reminiscent of Fe-S cluster scaffold proteins. In addition, the Fe-S cluster conversions are partially reversible in successive reduction–oxidation cycles, with cluster loss mainly occurring during (re)oxidation. The considerably negative reduction potential of the [4Fe-4S](2+/1+) couple (−520 mV) suggests that electron transfer may not be likely in the cell. Collectively, our findings identify HBx as an Fe-S protein with striking similarities to Fe-S scaffold proteins both in cluster type and reductive transformation. An Fe-S cluster in HBx offers new insights into its previously unknown molecular properties and sets the stage for deciphering the roles of HBx-associated iron (mis)regulation and reactive oxygen species in the context of liver tumorigenesis.

The viral protein HBx is the key regulatory factor of the hepatitis B virus (HBV) and the main etiology for HBVassociated liver diseases, such as cirrhosis and hepatocellular carcinoma. Historically, HBx has defied biochemical and structural characterization, deterring efforts to understand its molecular mechanisms. Here we show that soluble HBx fused to solubility tags copurifies with either a [2Fe-2S] or a [4Fe-4S] cluster, a feature that is shared among five HBV genotypes. We show that the O 2 -stable [2Fe-2S] cluster form converts to an O 2 -sensitive [4Fe-4S] state when reacted with chemical reductants, a transformation that is best described by a reductive coupling mechanism reminiscent of Fe-S cluster scaffold proteins. In addition, the Fe-S cluster conversions are partially reversible in successive reduction-oxidation cycles, with cluster loss mainly occurring during (re)oxidation. The considerably negative reduction potential of the [4Fe-4S] 2+/1+ couple (−520 mV) suggests that electron transfer may not be likely in the cell. Collectively, our findings identify HBx as an Fe-S protein with striking similarities to Fe-S scaffold proteins both in cluster type and reductive transformation. An Fe-S cluster in HBx offers new insights into its previously unknown molecular properties and sets the stage for deciphering the roles of HBx-associated iron (mis)regulation and reactive oxygen species in the context of liver tumorigenesis.

Chronic hepatitis B virus (HBV) infection affects 300 million people globally and is a major risk factor for the development of hepatocellular carcinoma (1, 2) . HBV is classified into 10 genotypes that appear in distinct geographical locations and are associated with different clinical outcomes or disease severity (3, 4) . All HBV genotypes share a common genomic architecture featuring four overlapping genes (P, S, C, and X) encoding for seven viral proteins (5) . Protein X (or HBx) is required for viral replication and is a key factor in chronic HBV-related hepatocarcinogenesis (6) (7) (8) . HBx promotes transcription of the extrachromosomal HBV genome by hijacking the DNA-damage binding protein 1 (DDB1)-containing E3 ubiquitin ligase to target the "structural maintenance of chromosomes" (Smc) Smc5/6 complex for degradation (9) . This mechanism stimulates transcription of episomal DNA templates as well as the HBV genome, giving HBx the title of a transcriptional transactivator (10) . HBx has also been shown to promote viral replication by recruiting Parvulin 14 and 17 to the HBV covalently closed circular DNA to enhance transcriptional activation (11) . In addition, HBx is reported to affect many oncogenic-related cellular processes, such as cell proliferation, DNA damage repair, and Fe regulation (12) . For example, HBx has been linked to altered iron metabolism and iron regulatory protein 1 (IRP1) expression in hepatic cells, and is suggested to inhibit the tumor suppressor PTEN via the production of reactive oxygen species (ROS) (13, 14) . Most of our current understanding of HBx is inferred from cellular experiments (reviewed in detail in the following refs: (12, 15, 16) ), and thus the molecular mechanisms that underlie HBx activity remain obscure.

HBx is a small 154 amino acid protein with a cysteine-rich sequence that is highly conserved across the various HBV genotypes. Naturally occurring C-terminally truncated (30-40 aa) or N-terminally elongated (52 aa) forms of HBx are associated with different disease outcomes (17, 18) . Although both canonical and noncanonical HBx proteins share pronounced sequence identity, the lack of sequence similarity to any known proteins hinders inference of molecular structure and function. The historically poor solution behavior of HBx has confined in vitro studies to samples that are nonnative, highly truncated, or contain extensive amino acid substitutions, often leading to conflicting outcomes and inadequate biophysical and biochemical characterization (19) (20) (21) . Solubility enhancement tags have successfully provided a first insight into the molecular properties of HBx, especially in its metal binding potential. HBx fused to the Maltose Binding Protein (MBP-HBx) copurifies with significant amounts of Fe and Zn (20) and was recently assigned as a Zn-binding protein based on the copurification of Zn when coexpressed or fused to DDB1 (22) . C61, C69, C137, and H139 were identified as potential ligands to the Zn cofactor, and the residues of this CCCH motif were shown to be critical for HBx viral replication function (i.e., targeting the Smc5/6 complex for degradation). However, there was no evidence that Zn binding facilitates protein-protein interactions with DDB1, leaving the functional role of Zn binding in HBx unresolved.

In the present study, we focus on a soluble construct of HBx fused to disulfide bond isomerase C (DsbC-HBx) (23) to delve into the potential of HBx to coordinate transition metal ions.

We find that HBx copurifies with Fe assembled into an Fe-S cluster, establishing HBx as an unrecognized Fe-S protein.

HBx binds two different types of Fe-S clusters: a [2Fe-2S] or [4Fe-4S] cluster depending on the presence or absence of oxygen, respectively. Through multiple spectroscopic approaches, we characterize the chemical nature and redox attributes of the newfound Fe-S cluster in HBx. Interestingly, the [2Fe-2S] cluster form undergoes a reductive rearrangement that affords formation of [4Fe-4S] clusters. This transformation is best described by a reductive coupling mechanism that is a well-studied and characteristic feature of Fe-S cluster scaffold proteins, such as IscU and IscA. This parallel is intriguing and invokes the notion that HBx may potentially hijack cellular pathways by affecting Fe-S cluster biogenesis (24) . Despite this similarity, however, there are multiple ways by which the Fe-S cluster in HBx may be important either at a structural or a regulatory level. Fe-S clusters are functionally diverse and ubiquitous cofactors that are becoming increasingly recognized as key components in viral proteins from rotaviruses, polyomaviruses, and more recently coronaviruses (25) (26) (27) (28) (29) . The roles of Fe-S cofactors in viral proteins are proposed to be structural with critical outcomes in the host protein activity. In our study, the presence of the newfound Fe-S cluster sets the stage for interrogating the molecular and structural underpinnings of HBx, whereas its redox-dependent behavior outlines possible roles of the Fe-S cluster-bound HBx in Fe homeostasis.

HBx is an Fe-S cluster-binding protein Aerobically isolated DsbC-or MBP-HBx copurifies with 2.4 to 2.5 mol Fe per mol protein (Table S1 ). Both proteins exhibit a reddish brown color and optical features at 325, 415, and 460 nm, suggestive of the presence of a [2Fe-2S] cluster (30) (Fig. 1A) . 57 Fe-enriched samples were examined by Mössbauer spectroscopy to confirm the exact chemical nature of the bound Fe species. The 4.2 K Mössbauer spectrum of the aerobically isolated DsbC-HBx was collected in a small magnetic field (78 mT) applied parallel to the direction of the γ irradiation (Fig. 1B) and shows a quadrupole doublet with parameters characteristic of antiferromagnetically coupled allferric [2Fe-2S] 2+ clusters (δ = 0.28 mm/s, ΔE Q = 0.51 mm/s, 84% of the total Fe absorption) (Table S2) (31, 32) . A broad baseline spectrum makes up the rest of the signal intensity (16%) and is attributed to adventitiously bound high-spin (S = 5/2) Fe 3+ . The Mössbauer spectrum of the aerobically isolated MBP-HBx is essentially identical to that of DsbC-HBx (Fig. 1C) . Elemental analyses in tandem with Mössbauer spectroscopy indicate binding of one [2Fe-2S] cluster per polypeptide. Exact Fe-S cluster stoichiometry cannot be unequivocally determined because both fusion constructs purify as poorly defined higher-order oligomers as previously reported (20) . The observed aggregation is partially due to intermolecular disulfides (Fig. S1 ) and unknown interactions that are salt and concentration independent.

The HBx fusion proteins were also isolated under O 2 -free conditions to examine whether the [2Fe-2S] cluster arises from oxidative degradation of a more O 2 -sensitive [4Fe-4S] cluster form. The UV-visible spectrum of the anaerobically purified DsbC-HBx exhibits a band at 420 nm, consistent with the presence of [4Fe-4S] clusters ( Fig. 1A ) (30) . The 4.2 K/78 mT Mössbauer spectrum of the anaerobically purified DsbC-HBx is dominated by a single quadrupole doublet with δ = 0.45 mm/s and ΔE Q = 1.11 mm/s (Fig. 1B) , confirming the presence of [4Fe-4S] 2+ clusters and supporting the hypothesis that the [2Fe-2S] clusters detected in the aerobically isolated protein are oxidative products. The anaerobically purified MBP-HBx has a similar spectrum with the [4Fe-4S] 2+ cluster as the major component (60% of total Fe absorption). The remaining absorption can be ascribed to a subspectrum of [2Fe-2S] 2+ clusters (17%), and a broad paramagnetic subspectrum assigned to Fe polysulfides formed during the purification process (23%) (Fig. 1C) . Parallel experiments on whole Escherichia (E.) coli cells monitored by Mössbauer spectroscopy showed that the [2Fe-2S] cluster form is also present in whole cells (Fig. S2) , suggesting that [2Fe-2S] clusters in HBx may be both an oxidative outcome and a consequence of incomplete Fe-S cluster insertion due to overexpression in E. coli. There is no evidence for formation of [3Fe-4S] 1+ clusters in the aerobically purified protein, as demonstrated by electron paramagnetic resonance (EPR) (Fig. S3) ; therefore, the [2Fe-2S] 2+ form appears to be the only stable oxidative by-product of the [4Fe-4S] cluster form.

Dithionite reduction of the aerobically isolated DsbC-and MBP-HBx yields EPR spectra with identical parameters and lineshape irrespective of the solubility tag (Fig. 1D) . The observed signals have axial symmetry, principal g-values of 2.04, 1.94, and 1.94, and relaxation properties that are best attributed to S = 1/2 [4Fe-4S] 1+ clusters rather than [2Fe-2S] 1+ clusters (Figs. 1D and S4 ). These data suggest that reductive treatment generates [4Fe-4S] 1+ clusters, possibly as a result of reductive coupling (vide infra). This process, which has been observed in Fe-S cluster chaperone proteins and proteins that act as Fe-S cluster donors, involves the one-electron reduction of [2Fe-2S] 2+ to [2Fe-2S] 1+ clusters, followed by the coupling of two [2Fe-2S] 1+ clusters to form one [4Fe-4S] 2+ cluster. Cleavage of the DsbC tag by tobacco etch virus protease does not result in Fe-S cluster loss (as seen by EPR, Fig. S5 ); however, we were unable to separate out the DsbC tag.

Bioinformatic analyses identified 6800 HBx sequences that share a high degree of sequence identity (>80%). HBx sequences are rich in cysteines, seven of which are invariant ( Fig. 2A) . Because Fe-S cluster binding is primarily cysteine dependent, we suspect that some of these conserved cysteines are involved in cofactor binding in HBx. Other residues, such as histidine, aspartate, and arginine can also contribute to Fe-S cluster coordination (33) . In this respect, the strictly conserved H139 may also serve as a ligand. Single substitution of each of the conserved cysteines ( Fig. 2A) with alanine did not affect Fe-S incorporation, and only a variant lacking all cysteines was unable to coordinate an Fe-S cluster (Fig. S6 ). We also identified a highly conserved leucine-tyrosine-arginine (LYR)-like motif in the HBx primary sequence ( Fig. 2A) . Such a motif has been shown to be a recognition sequence for direct interaction with the Fe-S cochaperone HSC20 (34) , which may further support the fact that HBx can receive an Fe-S cluster from the biosynthetic machinery.

Although the sequence similarity network (SSN) maps little segregation due to high sequence identity, it is overall clustered by the extant HBV genotype. HBx from the HBV genotypes A, C, and D (shown in red, orange, and green, respectively, Fig. 2B ) exhibit the highest sequence homology to each other. Genotypes A and D are distributed pandemically in populations worldwide, whereas genotype C is found primarily in East Asia (3, 4) . From genotype C we selected a sequence designated "whole-HBx" or wHBx, which includes an N-terminal 52-aa extension, to increase sequence variability. The two most sequence-diverse genotypes that cluster separately in the SSN (i.e., B and F1) are mainly found in East Asia and South America, respectively (3, 4) . On the basis of their distribution and clinical importance, DsbC fusions were generated with select sequences from the A2, B, C, D, and F1 HBV genotypes. All five proteins were 57 Fe-enriched for characterization by Mössbauer spectroscopy and isolated under O 2 -free conditions. The isolated DsbC-HBx fusions all coordinate both [2Fe-2S] 2+ and [4Fe-4S] 2+ clusters (Fig. 2C ), albeit at different relative ratios (Table S3 ), demonstrating that Fe-S cluster binding is a common feature across diverse HBx sequences.

DsbC-HBx was expressed in LB media supplemented with Fe and Zn salts to examine the extent of cofactor incorporation on the basis of metal availability. Cells overexpressing DsbC-HBx exhibit a light brown color, indicating enrichment in Fe-S clusters, whereas cells overexpressing DsbC alone lack color (Fig. 3A) . Therefore, cell pellet color is an initial indicator for Fe-S cluster incorporation under different metal supplementation conditions. Supplementation with 250 μM ZnSO 4 causes a lighter-colored pellet, whereas, supplementation with 250 μM Fe(NH 4 ) 2 (SO 4 ) 2 results in a dark brown pellet (Fig. 3A ). DsbC-HBx expression with simultaneous Zn and Fe supplementation leads to brown cell pellets that are lighter colored than those obtained from cells supplemented with Fe alone, suggesting that Zn may partially inhibit Fe incorporation when added in equimolar amounts. SDS-PAGE analysis of normalized cell pellets shows a similar expression level of the 42-kDa DsbC-HBx in the LB, LB(Fe), LB(Zn), and LB(Fe,Zn) cells (Fig. 3B ). The optical spectra of the aerobically purified proteins show formation of a [2Fe-2S] 2+ cluster under all conditions, but signal intensity is decreased when Zn is added to the media, consistent with compromised Fe-S incorporation (Fig. 3 , C and D). These results highlight that (i) Zn competes with the Fe-S cluster for binding to HBx, suggestive of a common coordination site, and (ii) Fe-S cofactor assembly is favored over Zn incorporation supporting that the Fe-S cluster may be the preferred cofactor in cellulo.

The aerobically purified DsbC-HBx coordinates 1.0 [2Fe-2S] cluster per protein (Figs. 4 and S7) and some adventitiously bound high-spin Fe 3+ , which is inaccessible to chelating agents. The observed formation of [4Fe-4S] 1+ clusters by dithionite reduction of DsbC-HBx:[2Fe-2S] in EPR (Fig. 1D ) prompted us to examine this process in more detail by 57 Fe Mössbauer spectroscopy. The [2Fe-2S] to [4Fe-4S] cluster transformation requires a strong reducing agent (i.e., sodium dithionite) and is not induced by weaker thiol-based reductants such as DTT or by the Azotobacter (A.) vinelandii iscFdx (Fig. S8 ). This implies that the reduction potential of DsbC-HBx:[2Fe-2S] is fairly negative when compared with canonical [2Fe-2S] clusters (30, 33) . In a first experiment we mimicked the conditions employed previously for the bacterial IscU, in which reductive transformation of [2Fe-2S] clusters was afforded by stoichiometric addition of reducing equiv (24 (Fig. 4C ). The nonstoichiometric conversion may be rationalized by loss of reducing equiv toward the one-electron reduction of adventitiously bound Fe 3+ and [2Fe-2S] 2+ clusters (the latter verified by EPR by appearance of the axial signal with principal values g II = 2.00, g ? =1.93) (Fig. 4D) demonstrating that reductive degradation of Fe-S clusters is negligible (Table S4) . Despite the unavoidable consumption of electrons by adventitious free Fe and further reduction of the nascent [4Fe-4S] 2+ clusters (Fig. 4D ), our findings are consistent with a reductive coupling mechanism like that observed for IscU.

The time-dependent reduction of the [2Fe-2S] to [4Fe-4S] cluster conversion was also monitored by Mössbauer and EPR spectroscopies in an analogous experiment to the one performed for the Nif IscA chaperone (35) . The starting sample contained 0.8 [2Fe-2S] clusters per protein (Fig. S9) and was reacted with excess sodium dithionite for 1, 5, and 30 min prior to freezing (Fig. 5A) (Fig. 6A) . Treatment of the [4Fe-4S] 2+ state with excess dithionite, therefore, does not lead to significant cluster disassembly, unlike the IscU chaperone for which complete cluster decomposition is observed (36) . The lack of reductive degradation of the [4Fe-4S] cluster enabled us to measure the midpoint potential of the [4Fe-4S] 2+/1+ couple by EPR potentiometric titrations on a [4Fe-4S] cluster-enriched sample (Fig. S10) . The fraction of [4Fe-4S] 1+ clusters was followed as a function of the measured potential ( Fig. 6B) . At the lowest attainable potential (−525 mV), only 55% of the [4Fe-4S] 2+ clusters present in the sample were reduced to the 1+ state, precluding measurement of the complete titration curve. The estimated midpoint potential (E m ) for the [4Fe-4S] 2+/1+ couple in HBx is −520 mV (Fig. 6B) .

We exposed DsbC-HBx:[4Fe-4S] to molecular O 2 to monitor the extent of [2Fe-2S] cluster formation and other oxidative products by Mössbauer spectroscopy (Fig. 7A) (Table S6 ). The possibility of the small amount of paramagnetic species observed to be assigned to [3Fe-4S] 1+ clusters cannot be excluded; however, the resonances are too broad to be consistent with formation of a trinuclear cluster, and also, we have no evidence for accumulation of [3Fe-4S] 1+ clusters in the aerobically exposed DsbC-HBx (Figs. 1B, S3, and S7) . Therefore, we consider the formation of [3Fe-4S] 1+ clusters unlikely.

To examine the extent of reversibility of the [2Fe-2S] to [4Fe-4S] transformation, we subjected DsbC-HBx:[2Fe-2S] to consecutive reduction-oxidation cycles while monitoring Fe-S cluster species by UV-visible and EPR spectroscopy. Dithionite reduction of the as-purified DsbC-HBx leads to the disappearance of the [2Fe-2S] 2+ cluster optical features (i.e., 415 and 460 nm) and appearance of a single band at 420 nm, consistent with the formation of [4Fe-4S] clusters (Fig. 7B) . After removing the reductant, the air-exposed sample recovers the [2Fe-2S] 2+ cluster optical spectrum, but to a lower extent. The optical features progressively bleach upon multiple rounds of oxidation and reduction, but cycle between the [2Fe-2S] and [4Fe-4S] cluster forms, showing that they may interconvert multiple times. Cognate EPR experiments were conducted using the [4Fe-4S] 1+ state as a quantitative probe of the Fe-S clusters remaining in the sample after each reduction step (Fig. 7C ). If the first reduction corresponds to the maximal attainable [4Fe-4S] 1+ clusters, the signal is reduced to 32% and 11% after the second and third reduction steps, respectively. Overall, these results show that the oxidatively formed [2Fe-2S] clusters can undergo successive rounds of reductive coupling; however, there is significant Fe-S cluster degradation upon each reduction-oxidation cycle (Table S6) .

To precisely determine the fate of the Fe-S species during reduction-oxidation, analogous experiments were performed by Mössbauer spectroscopy (Fig. 7D) 2+ clusters. This sample was exposed to air for 4 h in the Mössbauer cup and refrozen. The resultant spectrum is assigned to mononuclear Fe 2+ originating from oxidative cluster disassembly, as estimated by low-and hightemperature spectra (Fig. S11) . Considering that there is minimal degradation during the reductive coupling process (Fig. 4) , our findings demonstrate that most of the observed Fe-S cluster degradation during oxidation-reduction occurs during the oxidation process.

Despite multidecade studies on HBx, little is known about the biophysical and biochemical mechanisms by which HBx affects cellular processes during HBV infection. Here we present evidence that HBx is an Fe-S protein and establish the nuclearity and redox properties of the metallocofactor, providing a framework for future studies and development of therapeutics. HBx binds a [2Fe-2S] or [4Fe-4S] cluster depending on the presence or absence of O 2, respectively, and we have observed that a small amount of [2Fe-2S] clusters is present in whole E. coli cells. The [2Fe-2S] cluster appears to be an oxidative outcome, but its persistence in whole cells also suggests an inability of the E. coli biosynthetic machinery to insert intact Fe-S clusters. This compromised Fe-S cluster insertion may be alleviated if HBx is overexpressed in mammalian or insect cells. However, the low solubility of HBx when isolated from insect cells (37) and the variability of HBx overexpression in different mammalian lines pose a significant challenge toward these studies, which nevertheless should be pursued in the near future.

The ability of HBx to coordinate both [2Fe-2S] and [4Fe-4S] clusters is a common feature among sequences from the five HBV genotypes examined (A2, B, C, D, F1), highlighting the potential importance of the Fe-S cofactor in HBx function. HBx is a cysteine-rich polypeptide, offering seven strictly conserved cysteines as likely ligand candidates for the Fe-S cluster. Many of the conserved cysteines have been shown to hold important functional roles, such as C61, C69, and C137, which are crucial for HBx transactivation function (38) . We found that single-point cysteine to alanine substitutions allow for wildtype-like Fe-S cluster incorporation and that complete abolition of Fe-S cluster binding is only afforded in a variant lacking all cysteines. Thus, in vitro cofactor binding is cysteine dependent but not easily abrogated, suggesting that HBx may reorganize itself to accommodate for Fe-S cluster binding using other available cysteines, protein, or nonprotein residues.

Among the conserved HBx residues, C61, C69, C137, and H139 constitute a CCCH motif reported to coordinate a Zn ion (22) . The identified CCCH motif was shown to be critical for HBV replication via degradation of the Smc5/6 complex, but a functional role for Zn binding in this respect has not been identified (22) . In our study, expression of DsbC-HBx in media with excess Zn results in a decrease in Fe-S cluster enrichment, reflecting competitive ligation of these two cofactors that may occupy the same binding site. Many Fe-S cluster proteins have historically been misidentified as Zn proteins owing to similar ligand requirements (39). The mitochondrial membrane protein mitoNEET was first annotated as a Zn-finger protein based on primary sequence but was found to instead bind a [2Fe-2S] cluster (39) . The catalytic subunit of RNA polymerase (nsp12) from the severe acute respiratory syndrome coronavirus 2 (SARS-cov-2) was first assigned to be Zn dependent based on cryo-EM structures (29) . Subsequent studies revealed that nsp12 harbors instead two [4Fe-4S] clusters, one of which is critical for polymerase activity. Zn has been known to inhibit Fe-S biosynthesis in vivo by competitively binding to Fe-S scaffold proteins, including IscU and IscA (40) . In biochemical studies and during aerobic purification, Zn can supplant an Fe-S cluster that has been oxidatively degraded, leading to the misassignment of proteins as Zn dependent. In some cases, such as with the SARS-cov-2 nsp12, Zn supports limited protein activity by binding to the same site as the Fe-S cluster (29) and partially fulfilling a structural role. Thus, from our present data, we cannot exclude that both Zn and the Fe-S cluster may support HBx cellular activity.

The ability of the Fe-S cluster of HBx to transform from a [2Fe-2S] to a [4Fe-4S] cluster under reducing conditions is especially intriguing, since this is a distinctive behavior observed in Fe-S scaffold proteins. Our spectroscopic studies of the DsbC-HBx [2Fe-2S]→[4Fe-4S] transformation are consistent with a reductive coupling mechanism that was demonstrated for the bacterial IscU and IscA scaffold proteins (24, 35) . This mechanism is thought to be the final step in the de novo biosynthesis of [4Fe-4S] clusters, allowing scaffold proteins to donate both [2Fe-2S] and [4Fe-4S] clusters to appropriate acceptor proteins (41) . Indirect evidence of reductive coupling is also implied in cluster transfer studies of glutaredoxin-type and other A-type scaffold proteins (e.g., GLRX5, SufA, ISCA1/ISCA2) that require reducing agents to transfer clusters to [4Fe-4S] acceptor proteins (42) (43) (44) (45) . Reductive coupling in HBx requires the presence of a strong reducing agent, which comes in contrast to A-type scaffolds and glutaredoxins, in which a milder reductant (DTT) can drive the reaction. A. vinelandii iscFdx (E m = −344 mV) can also serve as the electron donor for reductive coupling in IscU, Fe 2+ ) . A, top, anaerobically isolated DsbC-HBx, (middle) exposed to 1 equiv of O 2 , (bottom) exposed to 2 equiv of O 2 . B, redox cycling monitored by UV-visible spectroscopy. The initial sample (black solid line) was reduced with dithionite (dashed black line), exposed to air for 1 h (solid orange line), reduced again with dithionite (dashed orange line), exposed to air for 1 h (solid blue line), and finally reduced again with dithionite (dashed blue line). C, redox cycling monitored by X-band cw EPR spectroscopy (microwave frequency = 9.38 GHz, T = 10 K, microwave power = 0.64 mW, modulation amplitude = 1 mT). The experiment analogous to (B) was performed: one reduction (black), two rounds of reduction (orange), three rounds of reduction (blue). *A small sharp signal present in all samples is not Fe-S cluster related and arises from subtraction of a paramagnetic impurity in the resonator. D, top, anaerobically isolated DsbC-HBx, (middle) exposed to air for 4 h, (bottom) rereduced with excess sodium dithionite.

but not DsbC-HBx. This observation may not be surprising considering that reduction by ferredoxins is often dependent on protein-protein interactions to facilitate efficient electron transfer (46) . The physiological reductant of HBx is presently unknown, but human ferredoxins (i.e., FDX1, FDX2) may serve as better candidates for the reductive coupling in HBx.

There is also some variability in the reversibility of the [2Fe-2S] to [4Fe-4S] conversion among Fe-S scaffold proteins. The majority of [4Fe-4S] clusters in IscU experience O 2 -mediated degradation into mononuclear Fe 3+ , and only a small fraction converts back to [2Fe-2S] clusters (24) . In IscA, the process is highly reversible with minimal Fe-S cluster degradation, which can be rationalized considering its proposed role in the maturation of [4Fe-4S] proteins during oxidative stress (35, 47) . DsbC-HBx seems to fall in between these two extremes: the Fe-S cluster persists after multiple rounds of oxidation-reduction, albeit with increasing concomitant degradation. Reaction of the reduced Fe-S cluster with O 2 and concomitant release of Fe may act as sources of ROS via Febased Fenton chemistry (48) . ROS can cause irreversible oxidative damage by nonspecifically reacting with proteins, lipids, and DNA, leading to carcinogenic mutations (49, 50) . There is evidence that HBx is associated with an increase in ROS, thereby causing low-level liver inflammation that promotes carcinogenesis (51, 52) . In addition, HBx is proposed to associate with the mitochondria and colocalize with membrane proteins such as cytochrome c oxidase, which is a major source of ROS generation in the cell (53, 54) . Therefore, partial oxidative degradation of the Fe-S cluster in HBx can offer an alternative route by which HBx can modulate mitochondrial reduction potential or protein activities by ROS formation.

The measured reduction potential of the [4Fe-4S] 2+/1+ transition in DsbC-HBx is appreciably negative (−520 mV), which may argue against this state's physiological relevance. The estimated reduction potential of the [4Fe-4S] 2+/1+ couple in IscU is similarly low at −570 mV (24) . The observed low potential of the IscU [4Fe-4S] 2+/1+ couple may prevent unwanted electron transfer, which may also be true for HBx. However, at present we cannot rule out the possibility that binding of a partner protein or small molecule may cause a shift to a more positive reduction potential. Taken collectively, our results suggest that the [2Fe-2S] 2+ and [4Fe-4S] 2+ forms are likely the relevant cofactor states. The redox behavior of the Fe-S cluster in HBx exhibits similarities to Fe-S donating proteins, suggesting that HBx may be involved in Fe regulation and modulate activity of Fe-S containing proteins, such as the iron regulatory protein 1 (IRP1) (55) . The cluster sensitivity to molecular oxygen draws parallels to sensor proteins, such as the oxygen-sensing Fe-S containing transcription factor Fumarate nitrate reductase or the redox-sensitive transcriptional activator SoxR (56, 57) . The discovery of an Fe-S cluster in HBx adds to a growing body of work identifying Fe-S clusters in viral proteins and opens new exciting directions to provide insight into the mechanisms that lead to carcinogenic outcomes during chronic HBV infection.

All chemicals were obtained from Fisher Scientific (unless specified otherwise) and were of high purity grade. NCBI accession codes for HBx genotypes: A2, P69713; B, Q8B8Q5; C, Q9YZS7; D, O11883; F1, Q05499.

The plasmids encoding MBP-HBx or DsbC-HBx were transformed into Rosetta (DE3) competent cells (New England Biolabs) with the pDB1282 plasmid (kindly gifted by Dr Squire Booker, Pennsylvania State University). Cells were grown in standard Lennox Broth (LB) or M9 minimal media for 57 Fe enrichment at 37 C with shaking (200 rpm) until an A 600 of 0.3 was reached, at which time solid L-Arabinose (0.2% (w/v) final concentration) and solid L-cysteine (200 μM final concentration) were added. At A 600 of 0.7, the cells were cold shocked at 4 C for 1 h. Protein expression was induced by addition of 0.5 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) and supplemented with 250 μM 56 Fe or 125 μM 57 Fe. Cells were then incubated at 18 C for 18 to 20 h and harvested by centrifugation at 7000 rpm for 15 min at 4 C. Cell pellets were flash frozen in liquid N 2 and stored at −80 C prior to purification.

DsbC-HBx cell pellets were resuspended in Lysis Buffer (50 mM Hepes, 300 mM NaCl, and 10 mM imidazole, 1 mM TCEP, pH 8.0). Phenylmethylsulfonyl fluoride (PMSF) was added to a final concentration of 45 μg/ml. Cells were broken by sonication, and the clarified lysate was applied onto a gravity Ni-NTA column (McLab) equilibrated in Lysis Buffer. The column was washed with Lysis Buffer, followed by an additional washing with Wash Buffer (50 mM Hepes, 300 mM NaCl, 25 mM imidazole, 1 mM TCEP, pH 8.0). The bound protein was eluted with Elution Buffer (50 mM Hepes, 300 mM NaCl, 250 mM imidazole, 1 mM TCEP, pH 8.0), and fractions containing DsbC-HBx were combined and concentrated using a 30 kDa MWCO Amicon Centrifugal Filter (Millipore, Sigma-Aldrich). The concentrated protein was loaded onto a size exclusion S200 column (GE Healthcare), which was equilibrated with Storage Buffer (50 mM Hepes, 150 mM NaCl, 10% glycerol, pH 8.0). Fractions containing the pure protein were pooled and further concentrated.

DsbC-HBx was purified under strictly O 2 -free conditions in an anaerobic chamber (Coy Labs) using degassed buffer solutions. DsbC-HBx cell pellets were resuspended in Lysis Buffer, followed by addition of 45 μg/ml PMSF, 0.5 mM 57 Fe, 5 mM sodium dithionite, and 2 mM TCEP. Cells were lysed by sonication for a total time of 30 min (15 s pulse, 59 s pause, 60% amplification), during which Na 2 S was gradually added to a final concentration of 0.5 mM. The clarified lysate was applied onto Ni-NTA column equilibrated in Lysis Buffer. The column was washed with Lysis Buffer, followed by an additional washing with Wash Buffer. The bound protein was eluted with Elution Buffer, and fractions containing DsbC-HBx were combined and concentrated using a 30 kDa MWCO Amicon Centrifugal Filter (Millipore, Sigma-Aldrich). Following the purification steps described above, the obtained DsbC-HBx was stored under anaerobic conditions until further use.

MBP-HBx was purified with a MBPtrap column (5 ml, Cytiva) following the same protocols for DsbC-HBx, apart from small changes in buffer composition: Lysis and Wash Buffer (50 mM Hepes, 300 mM NaCl, 1 mM TCEP, pH 8.0), Elution Buffer (50 mM Hepes, 300 mM NaCl, 10 mM Maltose, 1 mM TCEP, pH 8.0).

All protein concentrations were determined by the Bradford assay, and protein purity was assessed by SDS-PAGE with Coomassie staining (Thermo Fisher Scientific). Metal analysis of the as-purified proteins was determined by inductively coupled atomic emission spectrometry (ICP-AES). Fe content was determined colorimetrically as described (58) .

The SSN was generated with the web-based Enzyme Function Initiative-Enzyme Similarity Tool (EFI-EST) using the Transactivation protein X InterPro family IPR000236 (13,756 sequences) as an input, which was trimmed to contain only full-length HBx (6792 sequences) (59) . For the generation of the SSN, an alignment score of 80 was selected, and the final network was further refined by removing edges with a percent sequence identity that was less than 90.6%.

Mössbauer spectra were recorded on a WEB Research spectrometer equipped with a Janis SVT-400 variable-temperature cryostat. The external magnetic field (0.078 T) was applied parallel to the γ-beam. Zero velocity was calibrated as the centroid of the spectrum of α-Fe recorded at room temperature. WMOSS spectral analysis software (www.wmoss.org, WEB Research) was used to simulate and analyze Mössbauer spectra.

EPR spectra were recorded on a Bruker E500 Elexsys continuous wave (cw) X-band spectrometer (operating at approximately 9.38 GHz) equipped with a rectangular resonator (TE102) and a continuous-flow cryostat (Oxford 910) with a temperature controller (Oxford ITC 503). All EPR samples were prepared in storage buffer under O 2 -free conditions in an anaerobic glovebox (CoyLab). Samples were reduced with excess sodium dithionite for 30 min at 22 C prior to being frozen in liquid N 2 , unless stated otherwise. Spin concentration was quantified by double integration of signals using 100 μM CuSO 4 as a standard.

Stock sodium dithionite solutions were prepared fresh prior to every reaction under strictly anaerobic conditions in Storage Buffer and quantified by measuring the absorbance at 315 nm (ε 315 = 8043 M −1 cm −1 ). For stoichiometric reduction experiments, samples were reacted with sodium dithionite for 30 min at room temperature (22 C) prior to freezing into Mössbauer cups or EPR tubes. For the time-dependent reduction experiments, DsbC-HBx was reacted with excess sodium dithionite (10 mM) at room temperature, and aliquots of the reaction mixture were frozen into Mössbauer cups or EPR tubes at the appropriate timepoints.

DsbC-HBx was first enriched in [4Fe-4S] by reducing the aerobically purified protein with excess sodium dithionite (10 mM), which was further removed by a size exclusion HiLoadTM 16/600 Superdex 200 pg column (Cytiva) housed in the anaerobic chamber. For the titration experiment, 4.4 ml of 972.9 μM protein containing 0.61 Fe/protein was maintained under O 2 -free conditions in the titration cell by continuous flushing with hydrated argon gas. Reduction or oxidation was afforded by dropwise addition of sodium dithionite or potassium ferricyanide, respectively. The potentials were measured with a pH/redox meter (GPHR 1400, Greisinger) using a combination Pt/Ag/AgCl microelectrode (3 M KCl, Mettler Toledo) and are quoted relative to the standard hydrogen electrode. The electrode was calibrated with a saturated quinhydrone solution at pH 8 (1 M Hepes buffer). The following mediators were added and allowed to equilibrate for 15 min with the protein prior to the actual experiment: 50 μM phenazine methosulfate and 100 μM of the following: 2,6-dichlorophenolindophenol, 1,2-napthoquinone, 1,4-naptho quinone, 5-hydroxy-1,4-napthoquinone, methylene blue, 1,4-dihydroxynapthoquinone, indigo carmine, 2-hydroxyl-1,4napthoquinone, anthraquinone-1,5-disulfonate, phenosafranine, safranine T, methyl viologen, benzyl viologen. The reduction potential was estimated by simulating the fraction of reduced [4Fe-4S] 1+ for a one-electron reduction process using the Nernst equation.

For the stoichiometric oxidation experiments, DsbC-HBx:[4Fe-4S] samples were reacted with O 2 by adding appropriate amounts of O 2 -saturated Storage Buffer (1.8 mM O 2 ) in a gastight vial for 30 min at room temperature. For time-dependent oxidation, samples were exposed to air at 4 C before freezing in Mössbauer cups at the respective timepoints. For the redox (reduction-oxidation) cycling, samples were reduced with excess sodium dithionite and exposed to air. Excess dithionite was removed in between redox cycles with Zeba spin desalting columns (ThermoFisher).

All data generated or analyzed during this study are included in this published article (and its supplemental information).

Supporting information-This article contains supporting information.

Viral hepatitis and hepatocellular carcinoma: Etiology and management

Estimations of worldwide prevalence of chronic hepatitis B virus infection: A systematic review of data published between

The global hepatitis B virus genotype distribution approximated from available genotyping data

Hepatitis B virus genotypes: Global distribution and clinical importance

Hepatitis B: The virus and disease

HBx gene of hepatitis B virus induces liver cancer in transgenic mice

Hepatitis B virus X protein is essential to initiate and maintain virus replication after infection

Hepatitis B virus X gene and hepatocarcinogenesis

Hepatitis B virus X protein identifies the Smc5/6 complex as a host restriction factor

Hepatitis B virus X protein stimulates gene expression selectively from extrachromosomal DNA templates

Parvulin 14 and Parvulin 17 bind to HBx and cccDNA and upregulate hepatitis B virus replication from cccDNA to virion in an HBx-dependent manner

The enigmatic X gene of hepatitis B virus

HBx-induced reactive oxygen species activates hepatocellular carcinogenesis via dysregulation of PTEN/Akt pathway

HBx modulates iron regulatory protein 1-mediated iron metabolism via reactive oxygen species

Molecular mechanism of hepatitis B virus X protein function in hepatocarcinogenesis

Molecular functions and biological roles of hepatitis B virus x protein

Cterminal truncated hepatitis B virus x protein is associated with metastasis and enhances invasiveness by C-Jun/matrix metalloproteinase protein 10 activation in hepatocellular carcinoma

Hepatitis B virus whole-X and X protein play distinct roles in HBVrelated hepatocellular carcinoma progression

The X protein of hepatitis B virus has a ribo/deoxy ATPase activity

Structural and biochemical analysis of Bcl-2 interaction with the hepatitis B virus protein HBx

Structural characterization of an intrinsically unfolded mini-HBX protein from hepatitis B virus

Hepatitis B virus X protein function requires zinc binding

Hepatitis B virus oncoprotein HBx is not an ATPase

Formation and properties of

Iron-sulfur clusters: Nature's modular, multipurpose structures

Iron-Sulfur Clusters in Chemistry and Biology

The rotavirus nonstructural protein NSP5 coordinates a [2Fe-2S] iron-sulfur cluster that modulates interaction to RNA

The oncogenic small tumor antigen of merkel cell polyomavirus is an iron-sulfur cluster protein that enhances viral DNA replication

Fe-S cofactors in the SARS-CoV-2 RNA-dependent RNA polymerase are potential antiviral targets

Principles of Bioinorganic Chemistry

Characterization of Fe-S clusters in proteins by mössbauer spectroscopy

Alternative FeS cluster ligands: Tuning redox potentials and chemistry

Cochaperone binding to LYR motifs confers specificity of iron sulfur cluster delivery

Spectroscopic and functional characterization of iron-sulfur cluster-bound forms of azotobacter vinelandii (Nif)IscA

IscU as a scaffold for iron-sulfur cluster biosynthesis: Sequential assembly of

Isolation and molecular characterization of hepatitis B virus X-protein from a baculovirus expression system

A truncated mutant (residues 58-140) of the hepatitis B virus X protein retains transactivation function

Iron-sulfur clusters in zinc finger proteins

Zinc toxicity and iron-sulfur cluster biogenesis in Escherichia coli

Mechanisms of mitochondrial ironsulfur protein biogenesis

ISCA1 is essential for mitochondrial Fe4S4 biogenesis in vivo

A pathway for assembling [4Fe-4S](2+) clusters in mitochondrial iron-sulfur protein biogenesis

Native Escherichia coli SufA, coexpressed with SufBCDSE, purifies as a [2Fe-2S] protein and acts as an Fe-S transporter to Fe-S target enzymes

Formation of [4Fe-4S] clusters in the mitochondrial iron-sulfur cluster assembly machinery

Mitochondrial [4Fe-4S] protein assembly involves reductive [2Fe-2S] cluster fusion on ISCA1-ISCA2 by electron flow from ferredoxin FDX2

Assembly of Fe/S proteins in bacterial systems: Biochemistry of the bacterial ISC system

Fenton chemistry: An introduction

Iron and dioxygen chemistry is an important route to initiation of biological free radical oxidations: An electron paramagnetic resonance spin trapping study. Free Radic

Role of iron in carcinogenesis: Cancer as a ferrotoxic disease

Effect of HBx on inflammation and mitochondrial oxidative stress in mouse hepatocytes

HBx targeting to mitochondria and ROS generation are necessary but insufficient for HBV-induced cyclooxygenase-2 expression

Accurately mapping the location of the binding site for the interaction between hepatitis B virus X protein and cytochrome c oxidase III

The co-localization of HBx and COXIII upregulates COX-2 promoting HepG2 cell growth

Regulation of iron homeostasis by the p53-ISCU pathway

Iron-sulfur cluster disassembly in the FNR protein of Escherichia coli by O2

Binuclear [2Fe-2S] clusters in the Escherichia coli SoxR protein and role of the metal centers in transcription

Ferrozine-a new spectrophotometric reagent for iron

Enzyme function initiative-enzyme similarity tool (EFI-EST): A web tool for generating protein sequence similarity networks

Acknowledgments-The authors thank Paul S. Ralifo and Andrew Weitz for providing us access to the EPR spectrometer at the Chemical Instrumentation Center of Boston University (Boston, MA). We thank Laura J. Liermann for the ICP-AES analysis at the Laboratory for Isotopes and Metals in the Environment at the Pennsylvania State University (University Park, PA). The authors are additionally grateful to Amy Milne and Trent Quist for their contributions and helpful discussions. The authors are also thankful to Mehmet Berkhmen (New England Biolabs) for donating the backbone of the DsbC vector. Conflict of interest-The authors declare that they have no conflicts of interest with the contents of this article.Abbreviations-The abbreviations used are: DDB1, DNA-damage binding protein 1; DsbC, disulfide bond isomerase C; EPR, electron paramagnetic resonance; HBV, hepatitis B virus; ROS, reactive oxygen species; Smc, structural maintenance of chromosomes; SSN, sequence similarity network.