In vitro Rekonstitution der Replikationsinitiation des humanen Hepatitis B Virus

• Main Author: Dörnbrack, Katharina
• Format: Article in Journal/Newspaper
• Language: German
• Published: Universität Freiburg 2014
• Subjects: Arctic
• Online Access: https://dx.doi.org/10.6094/unifr/10404; https://www.freidok.uni-freiburg.de/data/10404

Hepatitis B virus (HBV), the prototypic hepadnavirus and causative agent of B-type hepatitis, is one of the relevant human pathogens. Hepadnaviruses are small hepatotropic DNA viruses that replicate through reverse transcription of an RNA intermediate, the pregenomic (pg) RNA. Replication is initiated by an unusual protein-priming mechanism, mediated by the binding of the viral polymerase to a highly conserved stem-loop structure, ε, close to the pgRNA´s 5´ end. This triggers encapsidation of the RNP complex yet furthermore, a bulged region within ε becomes the template for a short DNA oligonucleotide that is covalently linked to the polymerase via the Terminal Protein domain that is unique to hepadnaviral reverse transcriptases. Following translocation to a 3´proximal acceptor site on the pgRNA, the oligonucleotide is extended into complete minus-strand DNA and eventually converted into the partially double-stranded relaxed circular DNA found in infectious virions. For duck HBV (DHBV), already years ago the protein-priming reaction has successfully been reconstituted in cell-free systems, initially in rabbit reticulocyte lysate (RL), then with recombinant DHBV polymerase derivatives produced in E. coli. These in vitro systems yielded invaluable mechanistic insights, including the strict dependence of the reaction on cellular chaperones, and enabled the mapping of functionally important regions in both the polymerase and the DHBV ε (Dε) RNA. In addition, they indicated that after initial binding to the polymerase the Dε RNA must undergo a major conformational change in order to function as template for the oligonucleotide Primer. For HBV, however, all attempts to establish analogous in vitro systems have thus far failed. While HBV ε RNA can bind to HBV polymerase in vitro, the resulting RNP complexes never showed any priming activity. Suspected explanations were the inability of recombinant HBV polymerase to properly fold, or the lack of required cell factors. The focus of this thesis was an alternative hypothesis, namely that the higher stability of the HBV ε versus Dε RNA fold would prevent the ε RNA in the RNP complex from adopting a new, priming-active conformation. If true, destabilizing the ε RNA fold might allow in vitro priming. In the first part of this thesis, an "in-cell-SELEX" approach was employed to identify structurally destabilized yet functional ε variants. To this end, pools of HBV expression vectors carrying randomized sequences in the upper ε stem were transfected into human hepatoma cells. From progeny virus DNA the ε-containing genome regions were PCRamplified and used to generate new HBV vector pools for the next selection round. From four rounds, 23 replication-competent individual clones carrying from 1 to 8 nucleotide exchanges compared to wild-type ε were obtained. As determined by enzymatic and chemical (SHAPE) secondary structure analyses, all adopted the ε-typical structure consisting of a lower stem, a bulge, an upper stem, and an apical loop. However, due to the mismatches caused by the mutations, basepairing in, and therefore stability of, the upper stem was reduced, as desired. In the second part various HBV polymerase constructs were evaluated for expression in different E. coli strains. The best results were obtained using fusion constructs consisting of solubility-enhancing partner proteins and C terminally truncated polymerase ("miniPol"); analogous DHBV polymerase constructs are known to have chaperone-independent priming activity. Purification was either achieved by renaturing the protein isolated from inclusion bodies, or by a native procedure; this was only possible using Arctic Express cells which overexpress a GroEL/ES chaperonine. In the third part, the in vitro priming competence of the variant ε RNAs versus wild-type ε RNA was addressed, using HBV polymerase in vitro translated in RL, or the renatured and natively purified HBV miniPol protein from E. coli. Using conditions allowing well detectable priming activity for DHBV, in particular the presence of Mn2+, the destabilized ε variants, but not the wild-type ε RNA, showed unambiguous priming activity in all three systems. Activity required at least two nucleotide exchanges compared to the wild-type ε sequence. In a first application addressing the initiation site on the RNA template, in addition to the genetically derived 3´ terminal bulge nucleotide and the adjacent first nucleotide of the upper stem, also the following second upper stem nucleotide was found to be able to serve as template for the first nucleotide of the DNA oligonucleotide Primer. The natively purified HBV miniPol from E. coli allowed, for the first time, to separately investigate the impact of bivalent metal ions on initial polymerase/ε RNA complex formation, and the subsequent priming reaction. Most surprisingly, Mg2+ during complex formation, and Mn2+ during priming drastically enhanced priming-activity of the wild-type ε RNA. For the destabilized RNA variants, in contrast, the initial presence of Mg2+ had neither a positive nor a negative impact. Preliminary SHAPE analyses of the RNA structures in the complex revealed a dramatic, Mg2+ independent rearrangement in the upper stem of the variant RNAs whereas for the wild-type a similar change occurred exclusively when prior complex formation was performed in the presence of Mg2+. These data firstly indicate that also for HBV a structural change in the RNA is necessary for its function as priming template. Secondly, the structure of wild-type ε RNA appears to contain a Mg2+ modulatable energy barrier against rearrangement which in the destabilized RNA variants is reduced by the mutations. Lastly, the starting hypothesis was further corroborated by a novel "split ε" system in which this barrier is decreased by physical interruption of the RNA backbone. Even without prior Mg2+ activation, a split ε based on the wild-type sequence showed a distinct priming activity that strongly exceeded that of the contigous chain RNA. In sum, the studies performed in this thesis achieved, for the first time, the complete in vitro reconstitution of the protein-priming reaction for human HBV. Compared to another recently reported cell-free system which relies on the cotransfection of mammalian cells with expression vectors for HBV polymerase and ε RNA, followed by affinity purification of the RNP complexes, our new system offers an extremely simple composition (just recombinant HBV polymerase plus e RNA), as well as another decisive advantage: In the mammalian cellderived complexes, the bound εRNA can neither be exchanged, nor can polymerase expressed without ε RNA be loaded with ε RNA in vitro. Hence extensive mutagenesis analyses of the polymerase and/or ε RNA are not feasible. Our simple system, in contrast, should be adaptable to high-throughput applications, including screening for small molecule inhibitors of the highly specific HBV protein-priming reaction. Furthermore, the split ε system is not dependent on a natural ribonucleotide backbone anymore and thus opens completely new options for mechanistic analyses of the HBV priming reaction.