The Hepatitis B Virus Nucleocapsid-Dynamic Compartment for Infectious Virus Production and New Antiviral Target

Abstract

Hepatitis B virus (HBV) is a small enveloped DNA virus which replicates its tiny 3.2 kb genome by reverse transcription inside an icosahedral nucleocapsid, formed by a single ~180 amino acid capsid, or core, protein (Cp). HBV causes chronic hepatitis B (CHB), a severe liver disease responsible for nearly a million deaths each year. Most of HBV's only seven primary gene products are multifunctional. Though less obvious than for the multi-domain polymerase, P protein, this is equally crucial for Cp with its multiple roles in the viral life-cycle. Cp provides a stable genome container during extracellular phases, allows for directed intracellular genome transport and timely release from the capsid, and subsequent assembly of new nucleocapsids around P protein and the pregenomic (pg) RNA, forming a distinct compartment for reverse transcription. These opposing features are enabled by dynamic post-transcriptional modifications of Cp which result in dynamic structural alterations. Their perturbation by capsid assembly modulators (CAMs) is a promising new antiviral concept. CAMs inappropriately accelerate assembly and/or distort the capsid shell. We summarize the functional, biochemical, and structural dynamics of Cp, and discuss the therapeutic potential of CAMs based on clinical data. Presently, CAMs appear as a valuable addition but not a substitute for existing therapies. However, as part of rational combination therapies CAMs may bring the ambitious goal of a cure for CHB closer to reality.

Keywords: HBV; HBV cure; HBcAg; HBeAg; capsid; capsid assembly; capsid assembly modulator (CAM); capsid protein; chronic hepatitis B (CHB); combination therapy; core protein; core protein allosteric modulator (CpAM).

Figure 1

HBV genome organization, transcription, and translation. (A) Genome organization. The blue and the black line depict rcDNA with the 5′ terminally linked TP domain of P protein on the minus-strand and the RNA primer (red) on the plus-strand; the dashed black line symbolizes the incompletely filled-in 3′end. Colored internal arrows represent the ORFs with their respective preS1, preS2, and preC extensions. The outer red line depicts pgRNA with the 5′ proximal ε signal. Note that the actual transcription template is cccDNA, not rcDNA. (B) Viral transcripts and protein products. RNAs are shown above (subgenomic) and below (greater-than-genome length) a linear representation of the ORFs as present on the terminally redundant pgRNA; DR1, DR1*, and DR2 are direct repeats involved in rcDNA formation. As is common for eukaryotes, the most cap-proximal ORF on each RNA is translated, for precore protein enabled by the inclusion of the preC start codon in the 5′ extension; precore processing at both termini and secretion yield HBeAg (see Section 3.2). Translation of pgRNA yields Cp from the first and P protein from the second ORF; the latter initiation mechanism is still unclear.

Figure 2

The HBV infectious cycle. The virion binds to heparan sulfate proteoglycans (HSPG) and via a high-affinity interaction of the L protein’s PreS1 domain to the hepatocyte-specific bile acid transporter NTCP, triggering entry into the cell. The subsequent steps leading to production of infectious progeny virions are numbered. (1) Endosomal uptake and loss of the envelope; (2) exposure of NLSs to importins; (3) nuclear transport and capsid shell disassembly at the nuclear pore complex (NPC); (4) release of the rcDNA genome into the nucleus; (5) repair to cccDNA; (6) minichromosome formation; (7) transcription by RNA polymerase II; (8) translation of viral proteins, including HBx; (9) stimulation of cccDNA transcription by HBx; (10) packaging of the pgRNA-P protein complex into newly forming nucleocapsids; (11) capsid-internal reverse transcription of pgRNA into new rcDNA; via (11a) to (12) envelopment of mature rcDNA containing nucleocapsids and (13) egress from the cell via the multivesicular body (MVB) compartment; or via (11b) recycling of rcDNA (2) to the nucleus to replenish the cccDNA pool. Infected cells also release empty virions, i.e., enveloped empty capsids and small amounts of enveloped RNA containing capsids (not shown). Further viral particles include non-enveloped “naked” capsids, and spherical and filamentous subviral particles (SVPs) comprising only envelope proteins. Secreted nonparticulate HBeAg arises from processing of the precore protein (not shown).

Figure 3

(A) The HBV capsid as distinct replication compartment. The pgRNA-P protein complex recruits Cp subunits, initiating assembly of nucleocapsids wherein pgRNA is reverse transcribed into rcDNA. P binding to ε also mediates protein-primed initiation of DNA synthesis at the ε bulge, covalently linking the TP domain to the DNA 5′ end. The short DNA oligo is transferred close to the pgRNA 3′ end (first switch) and extended towards the 5′ end, with degradation of the RNA in the nascent DNA-RNA hybrid. A 5′ terminal RNA oligo remains, is transferred to an acceptor not far from the 5′ end of the new minus-strand DNA (second switch), and primes plus-strand DNA synthesis. When reaching the template’s 5′ end a third switch to the 3′ end enables circularization and further extension into mature rcDNA. NUCs inhibit DNA synthesis but not immature nucleocapsid formation. Genome maturation is assumed to induce a structural change on the capsid that signals readiness for envelopment, and also for nuclear transport as part of the recycling pathway; in neither case is the new viral DNA exposed in the cytoplasm, minimizing immune recognition as non-self. Conceivably, the entire genome maturation process is accompanied by as yet unidentified conformational alterations in the capsid. RNA is shown as red wavy line, minus-strand DNA in blue, plus-strand DNA in black. Small amounts of ds linear (dsL) DNA from a failed second template switch are not indicated. (B) Nonproductive assembly paths of recombinant Cps expressed in E. coli. In the absence of the basic C terminal domain (CTD), Cps comprising an intact N-terminal assembly domain such as Cp149 assemble empty CLP shells, essentially via Cp dimer contacts; this is the subject of most current in vitro studies. In wild-type Cp183 electrostatic interactions between the R residues in the CTD and the sugar-phosphate backbone of bacterial RNAs contribute to assembly nucleation and stabilize the CLP against inter-CTD repulsion. In highly CTD-phosphorylated Cp183 the phosphoryl groups neutralize the R residues and block their general RNA binding capacity, yielding empty, phospho-CTD containing CLPs.

Figure 4

Basic structural features of HBV Cp. (A) Primary sequence and its correlation to the fold of the NTD. The bars symbolize the N terminal assembly domain (NTD), the linker (blue), and the basic nucleic acid binding C terminal domain (CTD). Numbers indicate amino acid positions, + signs the clustered R residues in the CTD. The linker and CTD sequences are explicitly shown. S and T residues in the CTD that are potential phosphorylation targets are highlighted; in recombinant Cp183 SRPK1 phosphorylates all but S181 and thereby blocks RNA binding. The models on the right (derived from PDB 6HTX) show two views of one monomer with the all-α-helical fold of the NTD. M1 and L143 are highlighted as spheres, the larger helices α2 to α5 are color-coded as in the linear representation on the left. The CTD is likely disordered and not resolved in current structures. (B) The Cp dimer. One monomer is shown in green, the other in blue. Dimerization relies on formation of a four-helix-bundle by the α3/α4 hairpins from two monomers, generating a prominent spike; in addition, the N terminal sequences embrace the spike base, as is most easily seen in the top view. (C) Cp dimer assembly into icosahedral capsid shells. On the left an icosahedron with its defining 2-fold, 3-fold, and 5-fold symmetry axes at one of the 20 triangular faces is shown. Cartoons in the center show the arrangement of Cp dimers in T = 3 and T = 4 particles. A subunits (green) cluster around 5-fold axes, B (blue) and C (yellow) subunits around 3-folds. The two C subunits are related by 2-fold symmetry and thus identical. In the larger T = 4 particle the two halves of the non-AB dimers are in distinct environments and thus termed C (yellow) and D (red). Counting the dimers per triangular face yields 4.5 for T = 3 and 6 for T = 4, hence totally 90 and 120 dimers, respectively. The model on the right shows the asymmetric unit, i.e., neighboring AB and CD dimers (residues as spheres), as seen in the context of a real T = 4 CLP (6HTX).

Figure 5

HBeAg, a distinct structure despite nearly identical primary sequence to the Cp assembly domain. (A) Processing pathway of the precore protein. Cotranslation of the 29 codons of the preC region with the core ORF generates the precore protein. The first 19 PreC aa act as signal peptide directing the protein into the ER where N terminal cleavage occurs; during transit through the secretory pathway furin-like proteases clip-off most of the basic CTD, and in the oxidizing environment a specific disulfide bridge is formed between C-7 and C61 which determines the distinct structure of HBeAg. Failed signal peptide cleavage can create an only C terminally processed recently termed PreC protein (not shown) that contributes serologically to the “hepatitis B core-related antigen” (HBcrAg). (B) Crystal structure of one monomer in the HBeAg dimer (PDB 3V6Z). The preC-derived extra sequence is shown in yellow; C-7 and C61 are depicted as spheres. The second monomer is omitted for clarity. (C) The HBeAg dimer. The second monomer is shown in grey. Note the completely different dimerization mode compared to Cp which creates new epitopes, exposes epitopes that are hidden in the assembled capsid, and prevents HBeAg multimerization.

Figure 6

A model correlating HBV Cp phosphorylation and replication dynamics. By default, Cp produced in hepatocytes (1) is a substrate for various protein kinases (PKs), causing high-level phosphorylation, suppression of nucleic binding capacity and eventually empty virion formation; this prevents packaging of non-relevant viral and cellular RNAs. For infectious virion production (2) the model predicts that specific encapsidation of the pgRNA-P complex is enabled by the localized dephosphorylation of nearby Cp dimers by (a) protein phosphatase(s) associated with that complex. This restores the previously suppressed RNA binding capacity of Cp, with dephosphorylation of further Cp subunits enabling interactions that result in progressive and, eventually, complete packaging of pgRNA into newly forming nucleocapsids, and, subsequently, infectious virions in which Cp is largely unphosphorylated. The envelope blocks access of ATP and dNTPs. Upon infection of a new cell, removal of the envelope (de-envelopment) allows influx of dNTPs and ATP into the particle, enabling further fill-in of the plus-strand DNA, and re-phosphorylation of previously dephosphorylated Cp subunits. In effect, this increases conformational stress from the increasing DNA stiffness, and reduces Cp’s nucleic acid binding capacity, weakening the nucleocapsid structure. Eventually, this destabilization mediates disintegration of the capsid shell and release of the rcDNA into the nucleus to start of a new cycle.

Figure 7

The HBV Cp inter-dimer interaction. (A) The Cp dimer is tetravalent. The “hand region” comprising helix α5, the P-rich turn and the downstream sequence to the end of the assembly domain at L140 (highlighted in gold for each monomer) is the main module for inter-dimer contacts. Each hand region provides two interfaces, hence the dimer is tetravalent, as indicated by the broad yellow arrows in the top view. Only helices α3, α4, and α5 are shown as ribbons. (B) Close-up of the B–C interdimer contact. For chain B the entire NTD is represented, for chain C only the descending helix α4 and the hand region. Properly oriented chain B residues from F23 through helix α2 (P25 to Y38), from α4b (W102 to F110), plus W125 and residues from P138 to S141 form a largely hydrophobic cavity, against which α5 from chain C can snuggly pack, in particular via V120, V124, W125, R127, T128, Y132, R133, P134, and P138. In the capsid, the respective residues undergo analogous interaction with their other neighbors. The pocket formed by the dimer–dimer interaction is the target for all currently known CAMs (“HAP pocket”) and the main site for CAM resistance confering mutations.

Figure 8

Solid-state NMR can detect conformational alterations with high resolution at near-native conditions. Once all resonances are assigned in a reference structure other, similar samples can be compared against the reference. In the example recombinant pgRNA loaded CLPs from wild-type Cp183 and its variant F97L were compared; in cell culture this mutation relaxes the restriction on envelopment of nucleocapsid carrying immature genomes (adapted from Lecoq et al.). Residues whose conformational environment changes by the mutation experience chemical shift differences (CSDs) compared to the reference; larger differences indicate larger alterations. The linear representation at the bottom assigns such residues to the primary sequence, the model above to the structure of the Cp dimer. Note that numerous residues not only in immediate vicinity to the mutation are affected. The impact of bound ligands on the structure can similarly be determined, even in irregularly assembled multimers. Figure courtesy of Lauriane Lecoq and Anja Böckmann.

Figure 9

Typical impact of a CAM-A compound on in vitro assembly of Cp149. E. coli derived Cp149 CLPs were disassembled and reassembled in the absence (left) or presence (right) of an experimental CAM-A compound, JNJ 890, at a molar Cp:compound ratio of 2:1. Samples were analyzed by negative staining EM and are shown at identical magnification. Formation of large flat aggregates is a hallmark of CAM-A compounds; CAM-E compounds would not produce a particle phenotype that can be distinguished by this technique. Figure courtesy of Lauriane Lecoq and Anja Böckmann.

Figure 10

Different CAMs share a common binding pocket yet display non-identical binding modes. All currently known CAMs bind in the HAP pocket formed at the Cp inter-dimer interface with one Cp subunit (in blue) providing most contacts, and the second subunit (in gold) “capping” the bound compound. The compounds represent the classical chemotypes of HAP (HAP18), phenylpropenamide (AT-130) and sulfamoylbenzamidine (SBA_R01). Shown from left to right are the name and chemical formula of the respective CAM; a close-up of the HAP pocket with the most relevant protein residues as spheres and the CAM as ball-and-stick model; and the same view but some relevant protein residues as sticks and the CAM atoms as spheres. The corresponding PDB entries are given on the right. Note that despite their different chemical structure all three CAMs fit snuggly into the binding pocket yet engage both common and distinct Cp residues. This correlates with overlapping but non-identical resistance profiles and addresses different pressure points in the capsid shell which, via allostery, results in different overall responses.