Structural dynamics of the nucleocapsid of hepatitis B viruses
Final Report Abstract
Hepatitis B virus (HBV), the prototypic hepadnavirus and an important human pathogen, replicates its tiny DNA genome by capsid-internal reverse transcription of a pregenomic (pg) RNA. DNA formation enables envelopment of the icosahedral nucleocapsid and egress of stable infectious virions; upon infection of a new cell the particles must disintegrate to release the packaged DNA for conversion into a distinct nuclear DNA form, the template for new viral RNAs. Progress through these stages is long thought to go along with structural alterations in the capsid which themselves are accompanied by changes in phosphorylation state of the capsid (or core) protein (HBc). However, neither the structural changes as such nor the number, position or occupancy of phosphorylation sites during individual replication steps are well established. This is largely owed to >500 phosphotransferases in human cells and the lack of techniques to discriminate differentially phosphorylated HBc forms. Our previous work had suggested the larger core protein of duck HBV (DHBc) as a model in which structural alterations are more profound and thus better detectable. Hence one of two major objectives proposed was to analyze the structures of recombinant DHBc capsid-like particles (CLPs) and eventually nucleocapsids from replicating DHBV by electron cryo microscopy (cryoEM), in collaboration with Prof. Bettina Böttcher, a renowned cryoEM expert. The project was foccused on the second objective, i.e. the biochemical characterization of phosphorylation of the human virus HBc. To cut down the (de)phosphorylation complexity in human cells we exploited the lack in E. coli bacteria of eukaryote-like kinases and developed a recombinant co-expression system for HBc plus individual mammalian kinases, with an initial focus on a major candidate HBc kinase, SRPK1. Next, we established that a gel electrophoresis technique based on the phosphoryl group binding Phos-tag chelator can efficiently discriminate HBc species carrying different numbers of phosphoryl groups. By mass spectrometry (MS) we were able to determine that SRPK1 phosphorylates seven sites in HBc. Combining MS with site-specific mutagenesis we could map them to seven of the eight hydroxy amino acid residues in the Arg-rich C terminal nucleic acid binding domain (CTD) of HBc. This allowed for the first time to evaluate how actual seven-fold phosphorylation (rather than Asp or Glu carboxylates as mutational mimics) affects basic properties of HBc such as RNA binding (which was almost completely suppressed) and capsid stability against detergent or proteases. Replacing SRPK1 by other candidate HBc kinases such as PKA or PKC caused clearly detectable but lower level phosphorylation. Importantly, the bulk of HBc in capsids from human cells, replicating HBV or not, was similarly highly phosphorylated as the recombinantly SRPK1-phosphorylated CLPs. The largely electrostatic nature of the HBc-RNA interaction implies that high-level CTD phosphorylation suppresses the RNA binding capacity of HBc in human cells similarly as in E. coli. Hence high-level phosphorylation provides a means how the virus avoids packaging of irrelevant RNAs whose huge excess could easily outcompete encapsidation of the proper pgRNA/polymerase complex. Conversely, preventing phosphorylation of just one of the seven sites restored substantial RNA binding. Hence, a phosphatase activity associated with the pgRNA/polymerase could locally unleash HBc´s RNA binding ability, promoting specific encapsidation of the close-by pgRNA. While these data provided novel insights into the nature and principal consequences of HBc phosphorylation direct cryoEM analyses confirmed the virtual absence of internal RNA in the recombinant phospho-CLPs but at around 8 Å resolution revealed only minor structural differences in the particles´ protein shells. In a new collaboration with solid-state nuclear magnetic resonance (ssNMR) experts Dr. A. Böckmann (Lyon) and Prof. B. Meier (Zurich) the improved methods for expression and enrichment of HBc CLPs were used to generate 13C/15N-labeled preparations of HBc (and recently also DHBc) particles in which by now the NMR-signals of 140 residues could be assigned; in a first application this enabled the identification of individual residues acting as flexible hinges in the HBc structure. Similarly the binding sites for new capsid-targeting drug candidates could preliminarily be defined. In addition, technical hurdles in the way of analyzing enveloped HBV particles have been cleared by demonstrating the cell-free expression of DHBV envelopes in an ssNMR suitable state. Future combination of ssNMR with cryoEM will open new avenues to truly define the dynamics of the hepadnaviral capsid which are emerging as a new target for therapeutic intervention. Virologically the project data contribute to a new understanding of the long prevailing notion that HBc phosphorylation is necessary for pgRNA encapsidation; it certainly is but not in a direct sense (which assumed non-phosphorylated HBc as the default state) but rather indirectly by preventing encapsidation of improper RNAs. Conversely, with phosphorylated HBc as default, dephosphorylation appears as important for specific pgRNA packaging and subsequent genome replication.
