Hepadnaviruses (hepatotropic DNA viruses) are a family of small enveloped retrotranscribing DNA viruses. In humans, chronic infections with HBV cause ~800,000 deaths per year (Report, 2017). Current therapies can control but rarely cure infection (Fanning et al., 2019; Martinez et al., 2019; Nassal, 2015). Most fundamental aspects of the hepadnaviral lifecycle such as replication of the tiny ~3 kb DNA via reverse transcription of a pregenomic (pg) RNA (Summers and Mason, 1982), or the use of a covalently closed circular (ccc) DNA form (Tuttleman et al., 1986) as episomal persistence reservoir (Nassal, 2015), were uncovered using more tractable model viruses, mostly duck HBV (DHBV) (Schultz et al., 2004), prototype of the genus avihepadnaviridae.

Despite basal similarities in genome size, genetic organization and replication strategy the avian viruses exhibit characteristic differences to human and the other mammalian HBVs (genus orthohepadnaviridae). They lack an X protein, a regulator of cccDNA transcription (Livingston et al., 2017), and a medium sized (M) envelope or surface protein besides the small (S) and large (L) forms; in addition, the avian virus surface proteins are smaller, with only three predicted transmembrane helices versus four in the orthohepadnaviruses. Particularly striking is the much larger size (~260 amino acids) of their single capsid or core proteins (CPs) which in mammalian HBVs comprise only ~180 aa. Recently discovered HBV-like viruses in distant hosts (Lauber et al., 2017) maintain this division. Most employ the small type CP but amphibian and reptilian viruses, e.g. Tibetan frog HBV (Dill et al., 2016) and lizard HBVs (Lauber et al., 2017), encode the large type CP. As the small CPs can obviously perform all essential CP functions (see below, and Venkatakrishnan and Zlotnick, 2016) the biology behind the large type CPs is enigmatic. A key step would be knowledge of their structure but, ironically, here DHBV lags much behind HBV for which various structures are available for recombinant capsid-like particles (CLPs) (Böttcher and Nassal, 2018; Böttcher et al., 1997; Conway et al., 1997; Crowther et al., 1994; Wynne et al., 1999), serum-derived nucleocapsids (Roseman et al., 2005) and virions (Dryden et al., 2006; Seitz et al., 2007).

CPs are multifunctional (Zlotnick et al., 2015; Diab et al., 2018), making the HBV CP (HBc) a prime novel antiviral target (Lahlali et al., 2018; Schlicksup et al., 2018). Most obvious is their role as the building block of the icosahedral viral capsid. Production of progeny virions during infection involves transcription from nuclear cccDNA, translation in the cytoplasm of CP, polymerase and the other viral proteins, CP phosphorylation-regulated encapsidation of a complex of pgRNA and polymerase (Heger-Stevic et al., 2018b; Zhao et al., 2018), and reverse transcription of pgRNA into relaxed circular (RC plus some double-stranded linear (dsL)) DNA. Such nucleocapsids can interact with the surface proteins to be secreted as enveloped virions. Alternatively, they may interact with nuclear import receptors to be redirected into the nuclear pore where they disintegrate (Gallucci and Kann, 2017), releasing the rcDNA for conversion into cccDNA (Nassal, 2015; Schreiner and Nassal, 2017).

How hepadnaviral CPs integrate the different interaction cues to follow one path or the other is poorly understood but must somehow relate to their three-dimensional (3D) structure. Typical small type CPs comprise an N-terminal ~140 aa assembly domain (Birnbaum and Nassal, 1990) and an arginine-rich ~34 aa C-terminal domain (CTD) whose nucleic acid binding capacity is modulated by phosphorylation (Heger-Stevic et al., 2018b; Zhao et al., 2018). Both domains are connected by an ~10 aa linker (Watts et al., 2002).

The assembly domains feature five α-helices (Figure 1—figure supplement 1). α3 and α4 form an antiparallel hairpin providing the interface for dimerization with a second subunit. The resulting four-helix bundles protrude as spikes from the capsid surface. The helix α5 plus the downstream sequence (the ‘hand region’) provide the inter-dimer contacts, mostly arranged in T = 4 icosahedral symmetry which comprises 120 Cp dimers (the minor T = 3 class encompasses 90 dimers). The asymmetric unit is composed of two dimers in spatially different surroundings with four quasi-equivalent monomer conformations, A to D (Figure 1—figure supplement 1). The first residues of the CTD are capsid-internal (Zlotnick et al., 1997) but residues past position 150 usually lack sufficient order for visualization (Böttcher and Nassal, 2018).

Early cryo-EM reconstructions of DHBc suggested a T = 4 arrangement similar to HBc CLPs but with laterally widened spikes (Kenney et al., 1995). Extensive mutagenesis led to a DHBc model (Nassal et al., 2007; Vorreiter et al., 2007) predicting an HBc-like body of dimeric assembly domains comprising the N-terminal ~180 aa, an ~80 aa CTD with a more profound morphogenetic impact than in HBc, and a much extended ~45 aa proline-rich loop (residues 78–122) connecting helices α3 and α4. However, all attempts to directly verify this model by cryo-EM reconstructions fell short of sufficient resolution. The broader distribution of DHBc vs. HBc CLPs during sedimentation in sucrose gradients and native agarose gel electrophoresis (NAGE) suggested that morphological heterogeneity obscured the data. Some DHBc variants such as R124E produced more homogeneous CLPs (termed class2 Nassal et al., 2007) but did not yield better resolved cryo-EM reconstructions. In particular, the predicted loop at the spike tips remained invisible, in line with complete disorder.

Here we discovered that over several months of storage this loop, now termed extension domain, adopted a defined fold. This enabled us to determine the structure of whole DHBc CLPs at a resolution of 3.7 Å enabling de novo model building. Functional studies revealed that the extension domain´s ability to fold is dispensable for recombinant CLP assembly but correlates with capsid stability in hepatoma cells. We propose that plasticity of the extension domain represents a large-type CP-specific evolutionary adaptation.

DHBV CP (DHBc) expressed in E. coli self-assembles into CLPs which package bacterial RNA (Kenney et al., 1995; Nassal et al., 2007). Here, improved expression was achieved using vectors (Heger-Stevic et al., 2018b) with an E. coli optimized DHBc gene encoding the DHBV16 CP sequence (Uniprot accession: P0C6J7.1). CLPs for cryo-EM analysis were enriched by two sequential ammonium sulfphate precipitations (Birnbaum and Nassal, 1990) followed by sucrose gradient sedimentation (Heger-Stevic et al., 2018a; Figure 1—figure supplement 2).

After two weeks of storage at 4°C we exchanged the sucrose-containing buffer against a cryo-EM compatible buffer with low solute concentration and vitrified the samples. Micrographs showed distinct particles of spherical shape, aggregates of such particles and some proteo-liposomes (Figure 1—figure supplement 2). 2D-class averages revealed particles with protruding spikes but also particles that lacked distinct spikes or were deformed (Figure 1—figure supplement 2), indicating that only a fraction of particles had assembled into regular icosahedral capsids. The resulting 3D-map had a nominal resolution of 6.1 Å and showed capsids with an arrangement of spikes and holes resembling that of HBc-capsids with T = 4 triangulation (Crowther et al., 1994) but with 15 Å broader and 5 Å longer spikes (Figure 1—figure supplement 3). The characteristic broad spikes of the DHBc CLPs were only visible at lower density thresholds but disappeared at higher density thresholds, suggesting a markedly higher structural variability in the extensions than in the core of the spikes.

To test this hypothesis, we analysed the individual asymmetric units by 3D-classification. The class averages (Figure 1—figure supplement 3) showed 42% of asymmetric units without proper spikes and lower average density consistent with dimers either being misfolded, largely displaced or completely missing. Another third of the asymmetric units grouped into classes in which both spikes were broad, suggesting that all extension domains in the asymmetric unit were folded. The remaining 25% of asymmetric units grouped into classes with one narrow and one broad spike, consistent with folded extension domains in the broad spikes and still unfolded or flexible ones in the narrow spikes. Furthermore, the radial position of the spikes in the CLPs varied by up to 7 Å, indicating local deformations of the capsid shell.

The analysis of the individual asymmetric units required accurate determination of the orientations of an ordered asymmetric unit of only some 80 kDa, which is at the lower size limit that can be confidently analysed. We speculated that our analysis might have been limited by the imaging conditions (integrating mode, exposure <40 e/Ų), which were chosen for entire capsids (Song et al., 2019) rather than for the individual asymmetric units. Therefore, we acquired another data set at a higher dose in counting mode (Table 1). For this we prepared new grids of the CLPs that had now been stored in sucrose at 4°C for 10 months.

Much to our surprise 2D-class averages showed crisper projection averages (Figure 1—figure supplement 2) and the 3D-maps of the CLPs were better resolved than before. The 3D-map of the CLPs had a resolution of 3.7 Å (Figure 1, Figure 1—figure supplement 4), but more importantly, the extension domains were clearly visible at the same threshold as the core of the spikes, indicating that during long-term storage most extension domains had become folded. To investigate the folding of individual extension domains, we analysed the asymmetric units as described above. We observed three well resolved classes with broad spikes comprising 74% of the asymmetric units (Figure 1, Figure 1—figure supplement 4), one class with broad spikes and low resolution (9% of the asymmetric units), and one class with a very low average density (18% of the asymmetric units), in line with a still significant number of missing or grossly displaced asymmetric units in the capsids. The three well-resolved classes varied in their radial position by only 2 Å with no major structural differences. Local refinement of the orientations of these particles further improved the overall resolution of the asymmetric unit to a nominal resolution of 3 Å (Figure 1—figure supplement 4) with the extension domains still less well resolved than the rest of the asymmetric unit.

Figure 1 with 5 supplements see all

Download asset Open asset

We were able to build a de novo model of DHBc’s assembly domain into the 3D-map of the asymmetric unit of DHBc (Figure 1, Figure 1—figure supplement 4, Table 2). All four subunits of the asymmetric unit had the same overall fold with adaptations to the different spatial surroundings close to the N-termini and in the C-terminal hand regions (Figure 2a–c). The core of the spikes was formed by four helices, with each DHBc monomer contributing an ascending helix (α3; residues 44–78, 35 aa) and a descending helix (α4; residues 120–157, 38 aa). The ascending helix α3 was kinked in the center between F60 and W61 and tilted away from the dimer axis in the upper part of the spike. The helices were considerably longer than in HBc (Figure 3 and Figure 3—figure supplement 1), where α3 is straight (25 aa) and α4 is kinked (31 aa). The four helices of the spike of HBc and DHBc had a very similar packing below the kink (Figure 3) but diverged above the kink with a different twist of the helices around each other (Figure 3). At the tips of the spikes the positions of the helices were dramatically different, basically swapping the positions of DHBc α3 with HBc α4 in the same subunit and of DHBc α4 with HBc α3 between different subunits of the dimer. Moreover, DHBc helices α3 and α4 were connected through a 41 aa long extension domain (residues 79–119, Figure 2g) whereas in HBc this connection is just a short loop of 3–4 aa. The extension with two short helical segments (αe1: 92–99; αe2: 103–117), was positioned at the side of the spike close to α4 of the same subunit and in contact with α3 of the other subunit (Figure 2e). As independently confirmed below, the packing of the extension domain to the flank of the spike was stabilized by a salt bridge between R124 in α4 of the core spike and E109 in αe2 of the extension domain (Figure 2—figure supplement 1).

Figure 2 with 2 supplements see all

Download asset Open asset

Figure 3 with 2 supplements see all

Download asset Open asset

Analysis with Intersurf (Ray et al., 2005) showed that the extension domain increased the inter-dimer surface area from ~1,900 Ų to ~3,400 Ų, highlighting its important contribution to the intra-dimer contact. Nonetheless, the absence of folded extension domains from many spikes in freshly prepared CLPs implied that its contribution is not strictly required for capsid formation and stability, as directly addressed below.

In HBc which lacks an extension domain a disulphide-bridge between Cys61 on opposite monomers (Figure 1—figure supplement 1) can add to dimer stabilization (Nassal et al., 1992). DHBc lacks a homologous Cys residue; instead at a similar position in the spike Phe60 made a π-stacking interaction with Phe60 on the other monomer, and so did His52 with His52 (Figure 2—figure supplement 2). Another difference between HBc and DHBc was the position of the N-termini. In HBc the N-terminus is located at the dimer interface at the base of the spikes and contributes to the intra-dimer contact (Wynne et al., 1999). In DHBc the seven N-terminal residues pointed towards the lumen of the capsid, without contributing to the intra-dimer contact (Figure 2c). Interestingly, the two N-termini within one dimer had different orientations, with those of chains B and C directed towards the base of the spike and those of chains A and D pointing away (Figure 2).

The pockets at the base of the spikes close to the N-termini of chains A and D contained density that was not accounted for by the so far described model (Figure 1e,f magenta). This density resembled a short peptide and was modelled with the 13 very C-terminal residues 250–262. In the other two pockets similar albeit weaker density accounted for the 6 C-terminal residues (257-262). As the connecting residues to the assembly domain were unresolved the C-termini could not be assigned to a specific subunit. Overall, the C-termini appear to take a similar structural role in DHBc as the N-termini do in HBc (Figure 3).

While the intra-dimer contacts differed significantly between HBc and DHBc, the inter-dimer packing in the C-terminal hand-region was largely conserved. The last resolved residues in our model agreed well with the biochemically derived C-terminal border of the DHBc assembly domain around position 187 (Nassal et al., 2007). At the given resolution, the fold of the hand-region was similar to that of HBc (Figure 2d) as it maintained the length of the α5 helix and the distance to the adjacent extended stretch. Overall, the RMSD between HBc and DHBc of the hand-region and the part of the spike below the kink and the fulcrum is 1.9 Å, further highlighting the preservation of the whole shell-region of the capsids (Figure 3—figure supplement 1). Consistent with the preserved inter-dimer contacts, the radius of inner and outer surface of the protein shell and the distance to a diffuse density shell in the interior of the capsid that is attributed to RNA and unresolved parts of the CTDs (Figure 1b) was the same in DHBc and HBc.

The DHBc extension domain is dispensable for recombinant CLP assembly

The low folding propensity of the extension domain questioned a direct role in capsid assembly. We therefore replaced residues 78–122 by a short G₂SG₂ linker and evaluated the consequences on assembly and particle morphology. The corresponding protein, DHBc_Δ78–122, was virtually insoluble; however, solubility was drastically improved by the additional mutation R124E (DHBcR124EΔ). Sedimentation, NAGE and negative staining EM data all indicated the formation of orderly CLPs, prompting a closer look by cryo-EM image (Figure 1—figure supplement 2) reconstruction. The resulting density maps at 3 Å resolution (Figure 4, Figure 4—figure supplement 1, Figure 1—figure supplements 4 and 5) revealed the same architecture as in the wt DHBc particles, except that all spikes remained narrow as also confirmed by the asymmetric analysis (Figure 4—figure supplement 1). Also, the fold of the subunits in the asymmetric unit was nearly superimposable on that of the full length protein subunits with an RMSD of 0.6 Å, except for the missing sideward pointing density. This missing density was accounted for by the 45 amino acids of the deleted extension domain whereas the flexible G₂SG₂ linker was not resolved. Difference mapping confirmed that all density differences between full length and internal deletion variant localized to the extension domain which therefore behaves much like an independent internal appendix.

Figure 4 with 1 supplement see all

Download asset Open asset

Folding competence of the DHBc extension domain is critical for capsid stability in eukaryotic cells and hence viral replication

To reveal a potential virological relevance of the extension domain we examined the impact of extension domain-affecting mutations on capsid expression, viral nucleic acid packaging and reverse transcription in transfected LMH cells, an avian hepatoma line supporting formation of infectious DHBV (Condreay et al., 1990; Dallmeier et al., 2008). Most variants, including the additional mutant R124Q, were tested by trans-complementation of a cloned CP-deficient DHBV16 genome (pCD16_core^-) with the respective DHBc protein expressed from a separate plasmid vector. This excludes inadvertent effects on the virus genome. Mutant E109R_R124E was also tested in the cis-context. Controls included the non-complemented core^- genome and a variant with a mutated YMDD motif in the polymerase active site (pCD16_YMhD), allowing pgRNA encapsidation but not reverse transcription (Radziwill et al., 1990). By NAGE assay of cytoplasmic lysates and subsequent anti-DHBc immunoblotting (Figure 5a) the pCD16 constructs wt, YMhD and E109R_R124E, but not the core-deficient vector, produced similarly intense capsid signals whose mobilities matched those of the recombinant CLPs (lanes 16–19). Comparable signals were generated in the trans-format yet only with wt DHBc and mutants R124K, E109R_R124E and E110R. Variants R124Q, E109R, R124E and i95 gave very weak or no signals at all. Intriguingly, all these variants had exhibited the more homogeneous class2 phenotype when expressed in bacteria. Of the extension-less variants only Δ78–122_R124E produced specific though weak signals (Figure 5—figure supplement 1). To get more insight into the low or lacking capsid signals we repeated the pCD16_core^- cotransfections with variants R124E, Δ78–122_R124E, and E109R_R124E (including at a five-fold reduced plasmid amount) versus wt-DHBc, which reproduced the previously seen NAGE blot data for capsids and capsid-borne DNA (Figure 5—figure supplement 2a,b). This time, however, we additionally addressed non-assembled CP subunits, employing immunoprecipitation (IP) with a polyclonal anti-DHBc antiserum (12/99, Vorreiter et al., 2007) to enrich DHBc-related proteins from lysates of the transfected cells, followed by SDS-PAGE separation of the immunopellets and immunoblotting, again with the polyclonal antiserum; lysate from cells transfected with a GFP expression vector served as control. All DHBc transfected samples produced specific signals with the same mobility as the recombinant full-length proteins (Figure 5—figure supplement 2c); the Δ78–122_R124E band was weak, and thus consistent with the weak NAGE blot capsid signal. Strikingly, however, the R124E sample presented with a whole ladder of smaller products which in sum by far exceeded the left-over full-length band. Hence, the 124E mutation markedly enhances proteolytic susceptibility of the variant and concomitantly impedes its ability to form stable capsids in avian hepatoma cells.

Figure 5 with 2 supplements see all

Download asset Open asset

To reveal capsid-borne viral nucleic acids a parallel NAGE blot was hybridized with a ³²P-labeled DHBV DNA probe, after breaking up the blotted capsids by brief exposure to 0.1 N NaOH; this leaves both RNA and DNA intact. All capsid-positive samples, except the recombinant non-viral RNA containing CLPs, scored positive in this assay (Figure 5b, top). Longer NaOH treatment ablated the signals for sample pCD16_YMhD and reduced it for DHBc_E110R (Figure 5b, bottom), in line with an exclusive and partial RNA content, respectively. Persistence of the other signals indicated they were DNA. A more sensitive Southern blot assay revealed in all these variants a wt-typical pattern of rc-DNA and dsL-DNA (Figure 5c). Such signals were also faintly visible for R124Q and the triple variant E109,110R_R124E but not for variant Δ78–122_R124E. In sum, only variant R124K preserving the positive charge, and the charge-polarity reversing variant E109R_R124E maintained an in all aspects wt-like replication phenotype (Figure 5c and Figure 5—figure supplement 1). Hence opposite charges at positions 109 and 124 and the extension domain´s ability to fold are critical for virus replication.

Despite DHBV's long history as hepadnaviral model virus, knowledge of its structural biology has lagged behind that of HBV. Our EM-map of the DHBc CLP at 3.7 Å and of the DHBc_Δ78–122_R124E CLP at 3 Å resolution allowed us for the first time to build de novo models of DHBc’s assembly domain. Confirming selective aspects of evolutionary conservation between avi- and orthohepadnaviruses (Kenney et al., 1995), the new data also reveal unique structural modules, including a direct contribution of the very C-terminal residues to the capsid shell and a centrally inserted extension of ~45 aa that behaves like an independent folding entity.

A temporarily mobile spike extension - defining feature of hepadnaviruses employing large type CPs?

The most distinctive feature in DHBc vs. HBc are the large extensions at the tips of the spikes that comprise residues 79–119, in line with the biochemical prediction (Nassal et al., 2007) and directly supported by the structure of the extension-less Δ78–122_R124E CLPs (Figure 4). The extension sequence is unusually rich in P (8 of 45 aa, Figure 2) and E residues (6 of 45 aa), the two most disorder-promoting amino acids (Theillet et al., 2013; Uversky, 2013). Notably, P79/P80 and P117/P120 at the borders of the extension domain are ideally located to insulate it from the spike-forming four-helix bundle (Figure 2). These features are preserved in the CPs of diverse other avian hosts (Figure 1—figure supplement 1). Non-avian hosts with large-type CPs such as the 266 aa sequence of the Tibetan frog HBV CP diverge at many positions. Yet, the sequence aligning to the DHBc extension domain also has a very high proline content (10 of 48 aa; Figure 1—figure supplement 1), indicating evolutionary conservation of this feature over millions of years. We therefore expect that the fundamental structural aspects of the DHBV capsid revealed in this study are prototypical for all hepadnaviruses employing the large-type CPs. The plasticity of the extension domain is also reflected by local resolution analysis and high local B-factors in the model (Figure 1—figure supplement 4). Also the spike tips in HBc have high local B-factors (Böttcher and Nassal, 2018), perhaps in both cases facilitating induced-fit interactions, e.g. with the envelope proteins, which for DHBV might be more akin to a folding-by-binding process (Weng and Wang, 2020). Notably, the structure of one HBc spike tip fits well into the DHBc extension domain emerging from the core-spike which provides two such sites per spike (Figure 3—figure supplement 2). Thus, large-type CPs could have evolved two surface protein binding sites per spike in the extension domains, compared to only one in small-type CPs. This coincides with a generally smaller membrane part of the surface proteins in hepadnaviridae with large-type CPs allowing for more surface proteins per area of envelope.

Despite its disposition for disorder, the extension domain in wt DHBc can adopt a stably folded state, albeit very slowly in vitro. The time-scale in live cells is unknown but in sum our data favour that assembly of the capsid shell precedes folding of the extension domains. Assembly can occur in the complete absence of the extension domain, and CLPs with permanently disordered extension domains, such as R124E, present with more homogeneous morphologies than those from wt DHBc. Hence the large contact area to the intra-dimer interface that forms upon folding of the extension domain may even disfavour regular icosahedral assembly; conversely, the requirement for a folding helper would increase the time-span for undisturbed assembly. Thereafter, extension domain folding provides extra stability to the assembled structure. As a resolution of around 7 Å is sufficient to distinguish folded from unfolded extension domains, defining their states in various stages of the viral life-cycle, including in enveloped virions, appears worthwhile and feasible.

Based on the high density of prolines in the extension domain we expect the folding helper to be a peptidyl-prolyl cis-trans isomerase, as is supported by the higher frequency of DHBc CLPs with folded extension domains upon coexpression of E. coli FkpA. From such activity of a bacterial enzyme it appears unlikely that a specialized isomerase is required in DHBV host cells.

In contrast to wt DHBc CLPs, the extension domains in R124E CLPs remained invisible (Figure 4—figure supplement 1), suggesting the homogeneous appearance of class2 CLPs relates to mostly disordered extension domains, as supported by their similarity to the extension-less variant (Figure 4, Figure 4—figure supplement 1).

Intriguingly, despite their efficient expression in bacteria DHBc class2 variants were barely if at all detectable in hepatoma cells (Figure 5); also Δ78–122_R124E gave only faint signals in the NAGE immunoblot (Figure 5—figure supplement 1). As the same eukaryotic vectors led to efficient production of wt DHBc and other class1 particles, the class2 mutations probably interfered with some aspect of protein stability. In fact, decreased CP stability and lack of capsid formation have independently been reported for a fortuitously identified variant lacking H107 (Guo et al., 2006) which we now know to reside within the center of helix αe2. Deletion of H107 thus puts the remaining residues of the helix out of register and, specifically, ablates a hydrophobic interaction of H107 with I99 in αe1 (Figure 2—figure supplement 1) which contributes to the stability of the folded extension domain.Likely, long-term unfolded extension domains enhance accessibility of instability determinants, e.g. target sites for proteases, or for ubiquitylation which initiates proteasomal degradation. This view is strongly supported by the multiple proteolysis products seen with DHBc _R124E (Figure 5—figure supplement 2), and the simultaneous absence of stable capsids and/or nucleocapsids from this variant (Figure 5;.Figure 5—figure supplements 1 and 2). In addition, without the extra contribution of the folded extension domains to the intra-dimer interface the capsid shell may become too labile to withstand the replication process, in particular double-stranded DNA formation. While these are testable hypotheses the lack of stable capsids from these variants precluded statements on a specific role of the extension domain in DHBV replication. Most striking, however, is the complete rescue of wt DHBV-like replication by combining two per se detrimental mutations in the double mutant E109R_R124E (Figure 5), fully supporting our structural model as well as the importance of a salt-bridge between these two specific positions for the ordered state of the extension domain (Figure 6). Conversely, protonation of the glutamic acid carboxylate, realistically achievable at pH values of certain cell compartments (Huotari and Helenius, 2011), would be sufficient to break this bridge and thus create a reversible pH-responsive conformational switch for capsid stabilization and capsid disintegration.

Figure 6

Download asset Open asset

EM-maps are deposited in the EMDB. Where applicable models were deposited in the pdb DHBc capsid: 10800 (EMDB) 6ygh (pdb) DHBC co expressed with FkpA: 10801 (EMDB) DHBC R124E (mutant): 10802 (EMDB) DHBCR124E_del (deletion-mutant): 10803 (EMDB) 6ygi (pdb).

1. 1. Makbul C
   2. Nassal M
   3. Bottcher B

   (2020) RCSB Protein Data Bank

   ID 6ygh. DHBc capsid.

   https://www.rcsb.org/structure/6ygh
2. 1. Makbul C
   2. Nassal M
   3. Bottcher B

   (2020) RCSB Protein Data Bank

   ID 6ygi. DHBCR124E_del (deletion-mutant).

   https://www.rcsb.org/structure/6ygi
3. 1. Makbul C
   2. Nassal M
   3. Bottcher B

   (2020) Electron Microscopy Data Bank

   ID EMD-10800. DHBc capsid.

   https://www.ebi.ac.uk/pdbe/entry/emdb/EMD-10800
4. 1. Makbul C
   2. Nassal M
   3. Bottcher B

   (2020) Electron Microscopy Data Bank

   ID EMD-10801. DHBC co expressed with FkpA.

   https://www.ebi.ac.uk/pdbe/entry/emdb/EMD-10801
5. 1. Makbul C
   2. Nassal M
   3. Bottcher B

   (2020) Electron Microscopy Data Bank

   ID EMD-10802. DHBC R124E (mutant).

   https://www.ebi.ac.uk/pdbe/entry/emdb/EMD-10802
6. 1. Makbul C
   2. Nassal M
   3. Bottcher B

   (2020) Electron Microscopy Data Bank

   ID EMD-10803. DHBCR124E_del (deletion-mutant).

   https://www.ebi.ac.uk/pdbe/entry/emdb/EMD-10803

1. 1. Mandart E
   2. Kay A
   3. Galibert F

   (1984) Uniprot

   ID P0C6J7. Nucleotide sequence of a cloned duck hepatitis B virus genome: comparison with woodchuck and human hepatitis B virus sequences.

   https://www.uniprot.org/uniprot/P0C6J7
2. 1. Mandart E
   2. Kay A
   3. Galibert F

   (1984) NCBI Nucleotide

   ID K01834.1. Nucleotide sequence of a cloned duck hepatitis B virus genome: comparison with woodchuck and human hepatitis B virus sequences.

   https://www.ncbi.nlm.nih.gov/nuccore/K01834
3. 1. Horne SM
   2. Young KD

   (1995) NCBI Protein

   ID AAC41459.1. Escherichia coli and other species of the Enterobacteriaceae encode a protein similar to the family of Mip-like FK506-binding proteins.

   https://www.ncbi.nlm.nih.gov/protein/AAC41459.1
4. 1. Böttcher B
   2. Nassal M

   (2018) RCSB Protein Data Bank

   ID 6htx. WT Hepatitis B core protein capsid.

   https://www.rcsb.org/structure/6HTX
5. 1. Leslie AGW
   2. Wynne SA
   3. Crowther RA

   (1999) RCSB Protein Data Bank

   ID 1qgt. HUMAN HEPATITIS B VIRAL CAPSID (HBCAG).

   https://www.rcsb.org/structure/1QGT
6. 1. Galibert F
   2. Mandart E
   3. Fitoussi F
   4. Tiollais P
   5. Charnay P

   (1979) Uniprot

   ID P03146. Nucleotide sequence of the hepatitis B virus genome (subtype ayw) cloned in E. coli.

   https://www.uniprot.org/uniprot/P03146
7. 1. El-Gebali S
   2. Mistry J
   3. Bateman A
   4. Eddy SR
   5. Luciani A
   6. Potter SC
   7. Qureshi M
   8. Richardson LJ
   9. Salazar GA
   10. Smart A
   11. Sonnhammer ELL
   12. Hirsh L
   13. Paladin L
   14. Piovesan D
   15. Tosatto SCE
   16. Finn RD

   (2019) Pfam

   ID PF00906. Hepatitis core antigen.

   http://pfam.xfam.org/family/PF00906

1. 1. Birnbaum F
   2. Nassal M

   (1990) Hepatitis B virus nucleocapsid assembly: primary structure requirements in the core protein

   Journal of Virology 64:3319–3330.

   https://doi.org/10.1128/JVI.64.7.3319-3330.1990

   • PubMed
   • Google Scholar
2. 1. Böttcher B
   2. Wynne SA
   3. Crowther RA

   (1997) Determination of the fold of the core protein of hepatitis B virus by electron cryomicroscopy

   Nature 386:88–91.

   https://doi.org/10.1038/386088a0

   • PubMed
   • Google Scholar
3. 1. Böttcher B
   2. Nassal M

   (2018) Structure of mutant hepatitis B core protein capsids with premature secretion phenotype

   Journal of Molecular Biology 430:4941–4954.

   https://doi.org/10.1016/j.jmb.2018.10.018

   • PubMed
   • Google Scholar
4. 1. Condreay LD
   2. Aldrich CE
   3. Coates L
   4. Mason WS
   5. Wu TT

   (1990) Efficient duck hepatitis B virus production by an avian liver tumor cell line

   Journal of Virology 64:3249–3258.

   https://doi.org/10.1128/JVI.64.7.3249-3258.1990

   • PubMed
   • Google Scholar
5. 1. Conway JF
   2. Cheng N
   3. Zlotnick A
   4. Wingfield PT
   5. Stahl SJ
   6. Steven AC

   (1997) Visualization of a 4-helix bundle in the hepatitis B virus capsid by cryo-electron microscopy

   Nature 386:91–94.

   https://doi.org/10.1038/386091a0

   • PubMed
   • Google Scholar
6. 1. Crowther RA
   2. Kiselev NA
   3. Böttcher B
   4. Berriman JA
   5. Borisova GP
   6. Ose V
   7. Pumpens P

   (1994) Three-dimensional structure of hepatitis B virus core particles determined by electron cryomicroscopy

   Cell 77:943–950.

   https://doi.org/10.1016/0092-8674(94)90142-2

   • PubMed
   • Google Scholar
7. 1. Dallmeier K
   2. Schultz U
   3. Nassal M

   (2008) Heterologous replacement of the supposed host determining region of avihepadnaviruses: high in vivo infectivity despite low infectivity for hepatocytes

   PLOS Pathogens 4:e1000230.

   https://doi.org/10.1371/journal.ppat.1000230

   • PubMed
   • Google Scholar
8. 1. Diab A
   2. Foca A
   3. Zoulim F
   4. Durantel D
   5. Andrisani O

   (2018) The diverse functions of the hepatitis B core/capsid protein (HBc) in the viral life cycle: implications for the development of HBc-targeting antivirals

   Antiviral Research 149:211–220.

   https://doi.org/10.1016/j.antiviral.2017.11.015

   • PubMed
   • Google Scholar
9. 1. Dill JA
   2. Camus AC
   3. Leary JH
   4. Di Giallonardo F
   5. Holmes EC
   6. Ng TF

   (2016) Distinct viral lineages from fish and amphibians reveal the complex evolutionary history of hepadnaviruses

   Journal of Virology 90:7920–7933.

   https://doi.org/10.1128/JVI.00832-16

   • PubMed
   • Google Scholar
10. 1. Dryden KA
    2. Wieland SF
    3. Whitten-Bauer C
    4. Gerin JL
    5. Chisari FV
    6. Yeager M

    (2006) Native hepatitis B virions and capsids visualized by electron cryomicroscopy

    Molecular Cell 22:843–850.

    https://doi.org/10.1016/j.molcel.2006.04.025

    • PubMed
    • Google Scholar
11. 1. Dunyak BM
    2. Gestwicki JE

    (2016) Peptidyl-Proline isomerases (PPIases): Targets for natural products and natural Product-Inspired compounds

    Journal of Medicinal Chemistry 59:9622–9644.

    https://doi.org/10.1021/acs.jmedchem.6b00411

    • PubMed
    • Google Scholar
12. 1. Fanning GC
    2. Zoulim F
    3. Hou J
    4. Bertoletti A

    (2019) Therapeutic strategies for hepatitis B virus infection: towards a cure

    Nature Reviews Drug Discovery 18:827–844.

    https://doi.org/10.1038/s41573-019-0037-0

    • PubMed
    • Google Scholar
13. 1. Gallucci L
    2. Kann M

    (2017) Nuclear import of hepatitis B virus capsids and genome

    Viruses 9:21.

    https://doi.org/10.3390/v9010021

    • Google Scholar
14. 1. Guo H
    2. Aldrich CE
    3. Saputelli J
    4. Xu C
    5. Mason WS

    (2006) The insertion domain of the duck hepatitis B virus core protein plays a role in nucleocapsid assembly

    Virology 353:443–450.

    https://doi.org/10.1016/j.virol.2006.06.004

    • PubMed
    • Google Scholar
15. 1. Heger-Stevic J
    2. Kolb P
    3. Walker A
    4. Nassal M

    (2018a) Displaying Whole-Chain proteins on hepatitis B virus Capsid-Like particles

    Methods in Molecular Biology 1776:503–531.

    https://doi.org/10.1007/978-1-4939-7808-3_33

    • PubMed
    • Google Scholar
16. 1. Heger-Stevic J
    2. Zimmermann P
    3. Lecoq L
    4. Böttcher B
    5. Nassal M

    (2018b) Hepatitis B virus core protein phosphorylation: identification of the SRPK1 target sites and impact of their occupancy on RNA binding and capsid structure

    PLOS Pathogens 14:e1007488.

    https://doi.org/10.1371/journal.ppat.1007488

    • PubMed
    • Google Scholar
17. 1. Huotari J
    2. Helenius A

    (2011) Endosome maturation

    The EMBO Journal 30:3481–3500.

    https://doi.org/10.1038/emboj.2011.286

    • PubMed
    • Google Scholar
18. 1. Kenney JM
    2. von Bonsdorff CH
    3. Nassal M
    4. Fuller SD

    (1995) Evolutionary conservation in the hepatitis B virus core structure: comparison of human and duck cores

    Structure 3:1009–1019.

    https://doi.org/10.1016/S0969-2126(01)00237-4

    • PubMed
    • Google Scholar
19. 1. Köck J
    2. Rösler C
    3. Zhang JJ
    4. Blum HE
    5. Nassal M
    6. Thoma C

    (2010) Generation of covalently closed circular DNA of hepatitis B viruses via intracellular recycling is regulated in a virus specific manner

    PLOS Pathogens 6:e1001082.

    https://doi.org/10.1371/journal.ppat.1001082

    • PubMed
    • Google Scholar
20. 1. Lahlali T
    2. Berke JM
    3. Vergauwen K
    4. Foca A
    5. Vandyck K
    6. Pauwels F
    7. Zoulim F
    8. Durantel D

    (2018) Novel potent capsid assembly modulators regulate multiple steps of the hepatitis B virus life cycle

    Antimicrobial Agents and Chemotherapy 62:00835-18.

    https://doi.org/10.1128/AAC.00835-18

    • Google Scholar
21. 1. Lauber C
    2. Seitz S
    3. Mattei S
    4. Suh A
    5. Beck J
    6. Herstein J
    7. Börold J
    8. Salzburger W
    9. Kaderali L
    10. Briggs JAG
    11. Bartenschlager R

    (2017) Deciphering the origin and evolution of hepatitis B viruses by means of a family of Non-enveloped fish viruses

    Cell Host & Microbe 22:387–399.

    https://doi.org/10.1016/j.chom.2017.07.019

    • PubMed
    • Google Scholar
22. 1. Levy R
    2. Ahluwalia K
    3. Bohmann DJ
    4. Giang HM
    5. Schwimmer LJ
    6. Issafras H
    7. Reddy NB
    8. Chan C
    9. Horwitz AH
    10. Takeuchi T

    (2013) Enhancement of antibody fragment secretion into the Escherichia coli periplasm by co-expression with the peptidyl prolyl isomerase, FkpA, in the cytoplasm

    Journal of Immunological Methods 394:10–21.

    https://doi.org/10.1016/j.jim.2013.04.010

    • PubMed
    • Google Scholar
23. 1. Livingston C
    2. Ramakrishnan D
    3. Strubin M
    4. Fletcher S
    5. Beran R

    (2017) Identifying and characterizing interplay between hepatitis B virus X protein and Smc5/6

    Viruses 9:69.

    https://doi.org/10.3390/v9040069

    • Google Scholar
24. 1. Martinez MG
    2. Testoni B
    3. Zoulim F

    (2019) Biological basis for functional cure of chronic hepatitis B

    Journal of Viral Hepatitis 26:786–794.

    https://doi.org/10.1111/jvh.13090

    • PubMed
    • Google Scholar
25. 1. Nassal M
    2. Rieger A
    3. Steinau O

    (1992) Topological analysis of the hepatitis B virus core particle by cysteine-cysteine cross-linking

    Journal of Molecular Biology 225:1013–1025.

    https://doi.org/10.1016/0022-2836(92)90101-O

    • PubMed
    • Google Scholar
26. 1. Nassal M
    2. Leifer I
    3. Wingert I
    4. Dallmeier K
    5. Prinz S
    6. Vorreiter J

    (2007) A structural model for duck hepatitis B virus core protein derived by extensive mutagenesis

    Journal of Virology 81:13218–13229.

    https://doi.org/10.1128/JVI.00846-07

    • PubMed
    • Google Scholar
27. 1. Nassal M

    (2015) HBV cccDNA: viral persistence reservoir and key obstacle for a cure of chronic hepatitis B

    Gut 64:1972–1984.

    https://doi.org/10.1136/gutjnl-2015-309809

    • PubMed
    • Google Scholar
28. 1. Pettersen EF
    2. Goddard TD
    3. Huang CC
    4. Couch GS
    5. Greenblatt DM
    6. Meng EC
    7. Ferrin TE

    (2004) UCSF chimera--a visualization system for exploratory research and analysis

    Journal of Computational Chemistry 25:1605–1612.

    https://doi.org/10.1002/jcc.20084

    • PubMed
    • Google Scholar
29. 1. Radziwill G
    2. Tucker W
    3. Schaller H

    (1990) Mutational analysis of the hepatitis B virus P gene product: domain structure and RNase H activity

    Journal of Virology 64:613–620.

    https://doi.org/10.1128/JVI.64.2.613-620.1990

    • PubMed
    • Google Scholar
30. 1. Ray N
    2. Cavin X
    3. Paul JC
    4. Maigret B

    (2005) Intersurf: dynamic interface between proteins

    Journal of Molecular Graphics and Modelling 23:347–354.

    https://doi.org/10.1016/j.jmgm.2004.11.004

    • PubMed
    • Google Scholar
31. Software

    1. Report GH

    (2017) Geneva: World Health Organization; 2017, version 3.0 IGO

    World Health Organization.

    https://www.who.int/gho/publications/world_health_statistics/2017/EN_WHS2017_TOC.pdf?ua=1
32. 1. Rohou A
    2. Grigorieff N

    (2015) CTFFIND4: fast and accurate defocus estimation from electron micrographs

    Journal of Structural Biology 192:216–221.

    https://doi.org/10.1016/j.jsb.2015.08.008

    • PubMed
    • Google Scholar
33. 1. Roseman AM
    2. Berriman JA
    3. Wynne SA
    4. Butler PJ
    5. Crowther RA

    (2005) A structural model for maturation of the hepatitis B virus core

    PNAS 102:15821–15826.

    https://doi.org/10.1073/pnas.0504874102

    • PubMed
    • Google Scholar
34. 1. Schlicht HJ
    2. Bartenschlager R
    3. Schaller H

    (1989) The duck hepatitis B virus core protein contains a highly phosphorylated C terminus that is essential for replication but not for RNA packaging

    Journal of Virology 63:2995–3000.

    https://doi.org/10.1128/JVI.63.7.2995-3000.1989

    • PubMed
    • Google Scholar
35. 1. Schlicksup CJ
    2. Wang JC
    3. Francis S
    4. Venkatakrishnan B
    5. Turner WW
    6. VanNieuwenhze M
    7. Zlotnick A

    (2018) Hepatitis B virus core protein allosteric modulators can distort and disrupt intact capsids

    eLife 7:e31473.

    https://doi.org/10.7554/eLife.31473

    • PubMed
    • Google Scholar
36. 1. Schmid B
    2. Rösler C
    3. Nassal M

    (2011) A high level of mutation tolerance in the multifunctional sequence encoding the RNA encapsidation signal of an avian hepatitis B virus and slow evolution rate revealed by in vivo infection

    Journal of Virology 85:9300–9313.

    https://doi.org/10.1128/JVI.05005-11

    • PubMed
    • Google Scholar
37. 1. Schreiner S
    2. Nassal M

    (2017) A role for the host DNA damage response in hepatitis B virus cccDNA Formation-and beyond?

    Viruses 9:125.

    https://doi.org/10.3390/v9050125

    • PubMed
    • Google Scholar
38. 1. Schultz U
    2. Grgacic E
    3. Nassal M

    (2004) Duck hepatitis B virus: an invaluable model system for HBV infection

    Advances in Virus Research 63:1–70.

    https://doi.org/10.1016/S0065-3527(04)63001-6

    • PubMed
    • Google Scholar
39. 1. Seitz S
    2. Urban S
    3. Antoni C
    4. Böttcher B

    (2007) Cryo-electron microscopy of hepatitis B virions reveals variability in envelope capsid interactions

    The EMBO Journal 26:4160–4167.

    https://doi.org/10.1038/sj.emboj.7601841

    • PubMed
    • Google Scholar
40. 1. Song B
    2. Lenhart J
    3. Flegler VJ
    4. Makbul C
    5. Rasmussen T
    6. Böttcher B

    (2019) Capabilities of the falcon III detector for single-particle structure determination

    Ultramicroscopy 203:145–154.

    https://doi.org/10.1016/j.ultramic.2019.01.002

    • PubMed
    • Google Scholar
41. 1. Summers J
    2. Mason WS

    (1982) Replication of the genome of a hepatitis B--like virus by reverse transcription of an RNA intermediate

    Cell 29:403–415.

    https://doi.org/10.1016/0092-8674(82)90157-X

    • PubMed
    • Google Scholar
42. 1. Theillet FX
    2. Kalmar L
    3. Tompa P
    4. Han KH
    5. Selenko P
    6. Dunker AK
    7. Daughdrill GW
    8. Uversky VN

    (2013) The alphabet of intrinsic disorder: I. act like a pro: on the abundance and roles of proline residues in intrinsically disordered proteins

    Intrinsically Disordered Proteins 1:e24360.

    https://doi.org/10.4161/idp.24360

    • PubMed
    • Google Scholar
43. 1. Tuttleman JS
    2. Pourcel C
    3. Summers J

    (1986) Formation of the pool of covalently closed circular viral DNA in hepadnavirus-infected cells

    Cell 47:451–460.

    https://doi.org/10.1016/0092-8674(86)90602-1

    • PubMed
    • Google Scholar
44. 1. Uphoff CC
    2. Drexler HG

    (2013) Detection of mycoplasma contaminations

    Methods in Molecular Biology 946:1–13.

    https://doi.org/10.1007/978-1-62703-128-8_1

    • PubMed
    • Google Scholar
45. 1. Uversky VN

    (2013) The alphabet of intrinsic disorder: ii. various roles of glutamic acid in ordered and intrinsically disordered proteins

    Intrinsically Disordered Proteins 1:e24684.

    https://doi.org/10.4161/idp.24684

    • PubMed
    • Google Scholar
46. 1. Venkatakrishnan B
    2. Zlotnick A

    (2016) The structural biology of hepatitis B virus: form and function

    Annual Review of Virology 3:429–451.

    https://doi.org/10.1146/annurev-virology-110615-042238

    • PubMed
    • Google Scholar
47. 1. Vorreiter J
    2. Leifer I
    3. Rösler C
    4. Jackevica L
    5. Pumpens P
    6. Nassal M

    (2007) Monoclonal antibodies providing topological information on the duck hepatitis B virus core protein and avihepadnaviral nucleocapsid structure

    Journal of Virology 81:13230–13234.

    https://doi.org/10.1128/JVI.00847-07

    • PubMed
    • Google Scholar
48. Preprint

    1. Wagner T
    2. Merino F
    3. Stabrin M
    4. Moriya T
    5. Antoni C
    6. Apelbaum A
    7. Hagel P
    8. Sitsel O
    9. Raisch T
    10. Prumbaum D
    11. Quentin D
    12. Roderer D
    13. Tacke S
    14. Siebolds B
    15. Schubert E
    16. Shaikh TR
    17. Lill P
    18. Gatsogiannis C
    19. Raunser S

    (2019) SPHIRE-crYOLO: a fast and accurate fully automated particle picker for cryo-EM

    bioRxiv.

    https://doi.org/10.1101/356584

    • Google Scholar
49. 1. Watts NR
    2. Conway JF
    3. Cheng N
    4. Stahl SJ
    5. Belnap DM
    6. Steven AC
    7. Wingfield PT

    (2002) The morphogenic Linker peptide of HBV capsid protein forms a mobile array on the interior surface

    The EMBO Journal 21:876–884.

    https://doi.org/10.1093/emboj/21.5.876

    • PubMed
    • Google Scholar
50. 1. Weng J
    2. Wang W

    (2020) Dynamic multivalent interactions of intrinsically disordered proteins

    Current Opinion in Structural Biology 62:9–13.

    https://doi.org/10.1016/j.sbi.2019.11.001

    • PubMed
    • Google Scholar
51. 1. Wynne SA
    2. Crowther RA
    3. Leslie AG

    (1999) The crystal structure of the human hepatitis B virus capsid

    Molecular Cell 3:771–780.

    https://doi.org/10.1016/S1097-2765(01)80009-5

    • PubMed
    • Google Scholar
52. 1. Yang Z
    2. Lasker K
    3. Schneidman-Duhovny D
    4. Webb B
    5. Huang CC
    6. Pettersen EF
    7. Goddard TD
    8. Meng EC
    9. Sali A
    10. Ferrin TE

    (2012) UCSF chimera, MODELLER, and IMP: an integrated modeling system

    Journal of Structural Biology 179:269–278.

    https://doi.org/10.1016/j.jsb.2011.09.006

    • PubMed
    • Google Scholar
53. 1. Zhao Q
    2. Hu Z
    3. Cheng J
    4. Wu S
    5. Luo Y
    6. Chang J
    7. Hu J
    8. Guo JT

    (2018) Hepatitis B virus core protein dephosphorylation occurs during pregenomic RNA encapsidation

    Journal of Virology 92:e02139-17.

    https://doi.org/10.1128/JVI.02139-17

    • PubMed
    • Google Scholar
54. 1. Zheng SQ
    2. Palovcak E
    3. Armache JP
    4. Verba KA
    5. Cheng Y
    6. Agard DA

    (2017) MotionCor2: anisotropic correction of beam-induced motion for improved cryo-electron microscopy

    Nature Methods 14:331–332.

    https://doi.org/10.1038/nmeth.4193

    • PubMed
    • Google Scholar
55. 1. Zivanov J
    2. Nakane T
    3. Forsberg BO
    4. Kimanius D
    5. Hagen WJ
    6. Lindahl E
    7. Scheres SH

    (2018) New tools for automated high-resolution cryo-EM structure determination in RELION-3

    eLife 7:e42166.

    https://doi.org/10.7554/eLife.42166

    • PubMed
    • Google Scholar
56. 1. Zlotnick A
    2. Cheng N
    3. Stahl SJ
    4. Conway JF
    5. Steven AC
    6. Wingfield PT

    (1997) Localization of the C terminus of the assembly domain of hepatitis B virus capsid protein: implications for morphogenesis and organization of encapsidated RNA

    PNAS 94:9556–9561.

    https://doi.org/10.1073/pnas.94.18.9556

    • PubMed
    • Google Scholar
57. 1. Zlotnick A
    2. Venkatakrishnan B
    3. Tan Z
    4. Lewellyn E
    5. Turner W
    6. Francis S

    (2015) Core protein: a pleiotropic keystone in the HBV lifecycle

    Antiviral Research 121:82–93.

    https://doi.org/10.1016/j.antiviral.2015.06.020

    • PubMed
    • Google Scholar
58. 1. Zolotukhin S
    2. Potter M
    3. Hauswirth WW
    4. Guy J
    5. Muzyczka N

    (1996) A "humanized" green fluorescent protein cDNA adapted for high-level expression in mammalian cells

    Journal of Virology 70:4646–4654.

    https://doi.org/10.1128/JVI.70.7.4646-4654.1996

    • PubMed
    • Google Scholar

Acceptance summary:

Duck hepatitis B virus (DHBV) has been used widely as a model to study the hepatitis B virus (HBV) – a human virus infects over 250 million people worldwide. This work revealed the structure of DHBV core protein, which is substantially larger than that of HBV. The study generates important new information regarding the assembly and replication of HBV-like viruses (hepadnaviruses).

Decision letter after peer review:

Thank you for submitting your article "Slowly folding surface extension in the prototypic avian hepatitis B virus capsid governs stability" for consideration by eLife. Your article has been reviewed by two peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Cynthia Wolberger as the Senior Editor. The reviewers have opted to remain anonymous.

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

We would like to draw your attention to changes in our revision policy that we have made in response to COVID-19 (https://elifesciences.org/articles/57162). Specifically, we are asking editors to accept without delay manuscripts, like yours, that they judge can stand as eLife papers without additional data, even if they feel that they would make the manuscript stronger. Thus the revisions requested below mainly address clarity and presentation.

Summary:

The duck HBV (DHBV) has been used widely as a model to study the human pathogen, the hepatitis B virus (HBV). A striking difference between HBV and DHBV is that the DHBV capsid protein is substantially larger than that of HBV (ca. 260 vs. ca. 180 amino acid residues). The three significant accomplishments of this paper are (i) the observation that DHBV capsid protein is partially unfolded and yet still assembles, (ii) the protein will slowly refold and that can be catalyzed by the presence of proline isomerase, and (iii) the actual structure of the protruding domain.

Essential revisions:

There are some points to be addressed or discussed to further strengthen the manuscript, for example:

It is well known that HBc lacks the extension domain and yet is fully functional in human and avian cells. Conversely, DHBc with the extra domain is also functional in both human and avian cells. If there were indeed a host-cell specific determinant for the folding/functions of the extension domain in DHBV, as implied by the authors, it would be interesting to test the various extension domain deletion/substitution mutants in human cells, in addition to avian cells. As it stands now, little could be learnt regarding the exact roles of this extension in viral replication from the phenotype analysis of these mutants in avian cells, due to the apparently severe deleterious effect on capsid assembly.

Figure 5. The authors should check levels of the mutant DHBV capsid protein subunits using SDS-PAGE and western blotting. Did the subunits still accumulate? Was the defect in subunit stability, assembly, or the stability of the assembled capsids?

The question of which chaperone is responsible for capsid protein refolding in vivo is not answered in this paper. The effect of proline isomerase indicates that refolding can be catalyzed. However the limited effect of proline isomerase suggests that some other chaperone(s) may be involved. Also, the interaction of the capsid protein with the surface protein may also contribute to folding. The possibilities should be discussed with an open mind.

The evolutionary connections from avian to other HBVs is important to understanding virus evolution. The authors should also relate the DHBV structure to other insert containing sequences instead of just the limited sequence alignment of Figure 1—figure supplement 1D.

Following please find more comments from the two reviewers.

Reviewer #1:

The duck HBV (DHBV) has been used widely as a model to study the human pathogen, the hepatitis B virus (HBV). A striking difference between HBV and DHBV is that the DHBV capsid protein is substantially larger than that of HBV (ca. 260 vs. ca. 180 amino acid residues). The authors report high-resolution structures of several DHBV capsid-like particles (CLPs) determined by electron cryo-microscopy. These structures show that DHBV CLPs are generally similar to those of HBV, consisting of a dimeric α-helical backbone with protruding spikes at the dimer interface. The ca. 45 amino acid proline-rich extension unique to the DHBV capsid forms an extra domain near the spike as compared to HBV. The authors noticed that folding of the extension domain seemed to take months in vitro, which could accelerated by prolyl peptidyl isomerase, consistent with a catalyzed process in vivo. DHBc variants lacking a folding-proficient extension domain were competent to assemble capsids in bacteria but failed to form stable nucleocapsids in avian cells. The authors propose that the extension domain acts as a conformational switch with differential response options during viral infection.

This is an interesting study that has a strong rationale and generates important new information regarding the assembly and replication of HBV-like viruses (hepadnaviruses). The authors should consider the following points in revising the manuscript.

Reviewer #2:

In this paper the authors show the structure of Duck Hepatitic B Virus Core Protein and variants. DHBc is a dimeric protein that assembles into T=4 capsids. While closely related to the human homolog, DHBc has a bulky insert at the dimer interface. The authors show:

1) The insert is disordered in freshly purified protein

2) Can slowly fold, over several months, to an ordered form

3) Co-expression in E. coli with a prolyl isomerase yields capsids that are somewhat better folded.

4) Compensating charged residue mutations support a stable insert. Non-compensated mutations do not. The resulting proteins with a misfolded insert do not support assembly of functional capsid.

5) Deletion of the disordered residues results DHBcR124E∆ which has a well-ordered assembly domain and shows how the helical dimer interface is preserved

The most striking feature of the results is the folded-unfolded transition of the insert as discussed by the authors. This insert is present in lizard and fish variants of HBV but not in mammalian variants. There is an evolutionary argument. The authors do not discuss whether the insert may also interact with receptors or other host-protein.

The authors include a peculiar discussion of a gap at the dimer interface which may be "pocket factor" binding site. Has a pocket factor ever been observed or is this a speculation? The authors should clarify if this discussion is entirely speculative.

The quality of the reconstructions is high. However many of the figures are too small and busy. The text is clear with a few exceptions. The structure of the insert will facilitate future experimentation and will raise interest in non-mammalian HBV viruses. The puzzle of the folded to unfolded transition is striking. Co-expression of DHBc with proline isomerase yields a marginally better reconstruction. In the model structures, are prolines in the cis conformation? Would any chaperone do? The conclusion that the delay in folding is due to proline isomerization is ambiguous.

Essential revisions:

There are some points to be addressed or discussed to further strengthen the manuscript, for example:

It is well known that HBc lacks the extension domain and yet is fully functional in human and avian cells. Conversely, DHBc with the extra domain is also functional in both human and avian cells. If there were indeed a host-cell specific determinant for the folding/functions of the extension domain in DHBV, as implied by the authors, it would be interesting to test the various extension domain deletion/substitution mutants in human cells, in addition to avian cells. As it stands now, little could be learnt regarding the exact roles of this extension in viral replication from the phenotype analysis of these mutants in avian cells, due to the apparently severe deleterious effect on capsid assembly.

Figure 5. The authors should check levels of the mutant DHBV capsid protein subunits using SDS-PAGE and western blotting. Did the subunits still accumulate? Was the defect in subunit stability, assembly, or the stability of the assembled capsids?

We agree that the biological function of the extension domain in the large-type Cp remains an intriguing open question, and that investigating the impact of a different, e.g. human, cell environment is a potentially promising approach for new insight. An example is our demonstration, some time ago that the avian DHBV produces much more cccDNA in human hepatoma cells than human HBV. These phenotypic differences might indeed relate to the presence vs. absence of the extension domain in the respective Cps. Hence these and further host-factor focussed investigations are certainly worthwhile but in our view are beyond the scope of the current structure-focussed study.

However, we experimentally followed up the reviewer suggestion to address the fate of the unassembled subunits of the key mutants DHBcR124E and DHBcR124E∆ which failed to generate stable, replication-competent nucleo-capsids in chicken LMH cells – whereas DHBc WT as well as the charge-reversal mutant DHBc_E109R_R124E did. This was initially shown by DHBc capsid and DHBV DNA detection after NAGE (Figure 5). As direct SDS-PAGE immunoblotting of cell lysates did not yield conclusive data, we adopted an immunoprecipitation (IP) protocol to enrich potentially low amounts and/or partially degraded Cp subunits (about half of the lysate from a 10 cm dish as IP input). To cover as many DHBc epitopes as possible we used a polyclonal rabbit antiserum (12/99; Vorreiter et al., 2007) both for the IP and the subsequent immunoblot (IB) detection step. To minimize signals from the IP antibody chains on the blot we then used an Fc-part specific anti-rabbit secondary antibody-peroxidase conjugate in the final visualization step (which does not react strongly with the antibody light chains in the relevant 25 kDa range of the gel), producing the data now shown in Figure 5—figure supplement 2C. To corroborate the previously seen nucleocapsid phenotypes in the very same source samples as used for the IP/IB analysis we also subjected about 5% of each lysate to the NAGE-based DHBV capsid and capsid-associated DNA assays (Figure 5—figure supplement 2A, B).

As before the extension-less variant DHBcR124EΔ gave a very weak NAGE capsid signal (and no detectable capsid DNA signal) and this correlated with a weak signal in the SDS-PAGE IB, comparable to the E109R_R124E lo sample that had received one fifth the amount of DHBc plasmid DNA of the others. Hence the lower steady-state level of DHBcR124EΔ subunits (be it from less efficient production or more efficient degradation) might explain the lower levels of capsids detected in the NAGE IB; however, as previously surmised, the complete lack of detectable DHBV DNA supports a replicative defect of the mutant (pgRNA packaging, and/or reverse transcription). This assumption is now further strengthened by the detectability of viral DNA in the DHBcR124EΔ lo sample, which produced a similarly low capsid signal.

The most striking new data concern variant DHBc_R124E which, as before, gave no capsid and no capsid-DNA signal in the NAGE blots (although the E. coli derived faster-than-wt migrating R124E capsids were clearly detectable on the same blot). The IP/SDS-PAGE/IB protocol revealed the presence of substantial amounts of DHBc-related signals, however, besides little full-length protein the faster mobility of the other bands indicates heavy, variant-specific proteolytic processing, as similar products are completely absent from the DHBc WT and E109R_R124E samples.

One can still argue whether a lack of assembly competence of R124E in eukaryotic cells (vs. clear capsid formation in bacteria) causes higher sensitivity to proteolysis, or vice versa. However, although these new data do not answer all of the reviewer´s respective questions they directly support the previously only assumed defect in subunit stability when no salt bridge can form between positions 109 and 124 and when the extension domain does not fold.

We have added these new data as Figure 5—figure supplement 2, and added statements on the increased proteolysis susceptibility of DHBc_R124E in eukaryotic cells in the Results (subsection “Folding competence of the DHBc extension domain is critical for capsid stability in eukaryotic cells and hence viral replication”) and the Discussion section (subsection “A temporarily mobile spike extension – defining feature of hepadnaviruses employing large type CPs?”) of the main text. We also added additional information to the Materials and methods (subsection “Detection of viral capsids, capsid proteins, and capsid-borne DNA”).

The question of which chaperone is responsible for capsid protein refolding in vivo is not answered in this paper. The effect of proline isomerase indicates that refolding can be catalyzed. However the limited effect of proline isomerase suggests that some other chaperone(s) may be involved. Also, the interaction of the capsid protein with the surface protein may also contribute to folding. The possibilities should be discussed with an open mind.

Yes, it is true that we cannot say which chaperone is (or which chaperones are) acting in vivo. However, we respectfully disagree with the reviewer about the "limited effect" of the PPiase. While after two weeks and 2 months (not shown) without co-expression of a PPiase, most DHBc subunits in capsids have still unfolded extension domains this has changed to the majority of extension domains being folded within 2 weeks when co-expressed with a PPiases. A general PPiase is sufficient for the catalysis as it works with the E. coli FkpA PPiase. Typically, PPiases are among the most abundant proteins. Therefore, we would expect that the extension domain folds catalysed by a general PPiase of the host cells.

Another effect on the folding propensity of the extension domain is the formation of salt bridges (R124-E109) and hydrophobic interactions (H109-I99). These interactions are unlikely to require additional chaperones for formation and might happen instantaneously after the prolines are in the correct isomerization state. More importantly the significance of these stabilizing interactions can be interrogated by mutations: Mutants with a destabilized extension domain (DHBc R124E) behave very similar in cell culture to mutants with an entirely missing extension domain (DHBc R124E∆) as they do not form stable capsids that support replication. Considering that the folded extension domain almost doubles the intra-dimer interface it is conceivable that both the mutant capsids are less stable. In line with a reduced stability of capsids harbouring CPs with disordered extension domains, we observed that DHBcR124E∆ and DHBcR124E CLPs burst in thin, vitrified buffer whereas DHBc CLPs remains stable. Moreover, the new cell culture data (see above) illustrate a clearly enhanced proteolysis susceptibility of the R124E variant in cell culture.

Thus proline isomerization probably acts by delaying the folding of the extension domain, whereas interactions within the extension domain are likely to fine-tune capsid stability during the viral life cycle. We have broadened the Discussion along these lines (subsection “A temporarily mobile spike extension – defining feature of hepadnaviruses employing large type CPs?”)

The evolutionary connections from avian to other HBVs is important to understanding virus evolution. The authors should also relate the DHBV structure to other insert containing sequences instead of just the limited sequence alignment of Figure 1—figure supplement 1D.

While avian large-type CPs have a well preserved extension domain, the sequence conservation of the extension domain to large-type CPs in other species is less well preserved. However, a large number of prolines in the extension domains of other host species is maintained. This is what we wanted to show with the alignment. Detailed, sequence based discussions of evolutionary aspects have been made by Lauber et al., 2017, and Dill et al., 2016, which is given in the Introduction.

Based on the weak sequence conservation we would not necessarily expect that amphibian or lizard extension domains have a very similar structure – although structural data in Lauber et al., 2017, suggest that small-type CPs of nackedna viruses adopt an HBV CP-like fold despite highly divergent primary sequence. We have modelled the structure of Tibetan Frog CPs based on our DHBc structure. This suggests indeed a similar fold of the extension domain with maintenance of αe2 but not of αe1 and the positioning of the prolines at similar strategic positions as in avian viruses (Author response image 1). However, a large none-helical insert of some 20 residues with 2 prolines in α3 at the base of the core-spike raises doubts about details of the model. Thus a proper structure should be more revealing as to which extent salt bridges, charge distribution and hydrophobic interactions are maintained in the extension domains of other large-type CPs. Considering the low conservation and the lack of structure of other large-type CPs, we have chosen not to include a structure-based discussion on evolution of the large-type CPs.

Author response image 1

Download asset Open asset

However, large-type CPs go together with small-type surface proteins. We think that the large type CPs have evolved two potential surface protein binding sites per spike compared to only one in small type CPs. Support for this idea comes from the observation that the tip of the small-type HBc CPs matches in its overall fold and arrangement of loops the upper part of one extension domain. We have added the information on the size of the surface proteins in the Introduction (second paragraph) and have discussed the aspect of two potential binding sites per spike in the Discussion (subsection “A temporarily mobile spike extension – defining feature of hepadnaviruses employing large type CPs?”). We also added a figure to illustrate the point (Figure 3—figure supplement 2).

Author details

1. Cihan Makbul

   Julius Maximilian University of Würzburg, Department of Biochemistry and Rudolf Virchow Centre, Würzburg, Germany

   Contribution

   Data curation, Formal analysis, Validation, Investigation, Visualization, Methodology, Writing - original draft, Writing - review and editing

   Competing interests

   No competing interests declared

2. Michael Nassal

   University Hospital Freiburg, Internal Medicine 2/Molecular Biology, Freiburg, Germany

   Contribution

   Conceptualization, Resources, Funding acquisition, Validation, Investigation, Visualization, Methodology, Writing - original draft, Project administration, Writing - review and editing

   Contributed equally with

   Bettina Böttcher

   For correspondence

   nassal2@ukl.uni-freiburg.de

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-2204-9158

3. Bettina Böttcher

   Julius Maximilian University of Würzburg, Department of Biochemistry and Rudolf Virchow Centre, Würzburg, Germany

   Contribution

   Conceptualization, Data curation, Formal analysis, Funding acquisition, Validation, Investigation, Visualization, Methodology, Writing - original draft, Project administration, Writing - review and editing

   Contributed equally with

   Michael Nassal

   For correspondence

   bettina.boettcher@uni-wuerzburg.de

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-7962-4849

• 1,154

  Page views

• 156

  Downloads

• 10

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.