Research Projects
Replication machinery of segmented negative strand RNA viruses
Viruses from the order of Bunyavirales with a segmented negative strand RNA genome produce only a handful of own proteins, some of them get by with only four gene products. Nonetheless several of the most deadly and newly emerging pathogens like the Lassa virus and the Crimean-Congo hemorrhagic fever virus belong to this group. Another virus, the Rift Valley fever virus, has been known for approx. 100 years, but large outbreaks in humans have been reported more and more often during the last years. Thus in addition to the newly emerged viruses, some of the long-known viruses gained a more important role for public health. How these small viruses are able to use their 4-6 proteins in a way that allows them to have such a dramatic effect on their host is fascinating and not yet understood.

Two proteins are mainly in the focus of our research: The L protein and the N protein that together with the RNA genome of the virus form the ribonucleoparticle, which is sufficient for replication and transcription of the genome. We apply a variety of methods from structural biology, biochemistry and virology fields to understand bunyavirus genome replication and transcription processes in tiny detail.

We produce the viral proteins in a suitable expression system, e.g. bacteria, insect cells or mammalian cells. In an interdisciplinary approach combining structural biology techniques (protein crystallography, Small angle X-ray scattering, cryo-electronmicroscopy and mass spectrometry) with functional studies from in vitro enzyme activity assays and cell-based replication and transcription experiments as well as advanced bioinformatics we aim at understanding catalytic mechanisms and enzyme dynamics.
Besides the central polymerase domain, the viral L protein contains an endoribonuclease in its N-terminus. Hantaviruses, which are another group of important pathogens within the order of Bunyavirales, contain an endonuclease, which is too active to be recombinantly produced: it degrades all RNA making it toxic for the expressing cell. We studied the endonuclease in detail by introducing a set of mutations, which result in an attenuated enyzme, allowing for expression in E. coli. By determination of the crystal structure we could propose a role of all mutated amino acids and gain new insights into an enzyme, which is essential for the viral life cycle and therefore an attractive drug target.
Furthermore, we solved the structures of the N- and C-terminal domains of the California Academy of Sciences virus (reptarenavirus) L protein. The N-terminal domain of this L protein contains an endoribonuclease, which is structurally homologous to already published arenavirus endonucleases. The C terminus of the L protein contains a putative cap-binding site very similar to influenza virus cap-binding protein PB2. Both of these functions, cap-binding and endonuclease, are required for the cap-snatching mechanism, by which segmented negative strand RNA viruses presumably initiate transcription. In our structure however, the residues potentially involved in cap-binding did not show the expected conformation and also functional studies could not proof a cap-binding activity leaving it unclear how cap-snatching actually works in bunyaviruses.

The endonuclease activity in the N terminus of the hantaviral L protein. A: Expression level of full L protein variants and their mRNA in mammalian cells. B: Endonuclease activity of recombinantly produced isolated endonuclease domain variants of the L protein. Radioactively labelled RNA is separated by denatured PAGE. C: Crystal structure of the hantaviral endonuclease domain. All residues mutated in A are shown and grouped by their proposed role.

Left: The structure of the N-terminus of reptarenavirus CASV L protein (A) is similar to other endonucleases of mammarenavirus L proteins (C). The active site comformation is almost identical (B).
Right: The structure of the C-terminus of CASV L protein contains two separate domains (A) and is very similar to parts of influenza virus PB2, even concerning the domain topology (B). This is exemplarily shown for domain 2 of CASV L protein C-terminus and PB2 cap-binding domain (C).
With more recenty solved structures of Rift Valley fever virus and Severe fever with thrombocytopenia syndrome virus L protein C-terminal domains we were able to demonstrate the presence of a functional cap-binding domain within the bunyavirus L protein. Comparison with influenza virus cap-binding domain revealed commonalities and differences in the binding mode for the cap. Yet, many details about bunyavirus cap-snatching remain unclear.


A schematic overview about bunyavirus transcription. Influenza virus transcribes its genome using cap snatching in the nucleus, the viral mRNAs resemble cellular mRNAs and are exported and translated just as any host cell mRNA. By contrast, bunyaviruses perform cap snatching in the cytoplasm. Most probably, the viral L protein is responsible for cap binding and cleavage of host mRNA a few nucleotides downstream of the 5′ cap. This short, capped RNA fragment is subsequently used as a primer for viral transcription. This results in a chimeric mRNA, which is then translated at host ribosomes into viral proteins. To date, several steps of this process are not understood: (1) Which cellular capped RNAs are targeted by bunyavirus cap snatching? (2) With which high-affinity cellular cap-binding proteins do bunyaviruses compete and how? (3) Where in the cytoplasm do bunyaviruses perform cap snatching? (4) And finally, how do bunyaviruses assure the viral mRNAs are translated at the ribosomes? All these unknown aspects are indicated by question marks. (Olschewski-Pavlita et al. Trends Microbiol.2020)
The endonuclease activity in the N terminus of the hantaviral L protein. A: Expression level of full L protein variants and their mRNA in mammalian cells. B: Endonuclease activity of recombinantly produced isolated endonuclease domain variants of the L protein. Radioactively labelled RNA is separated by denatured PAGE. C: Crystal structure of the hantaviral endonuclease domain. All residues mutated in A are shown and grouped by their proposed role.
Recently, we investigated the conformational changes associated with promoter binding and polymerase activity in Lassa virus L protein using cryoEM (presented as a preprint on BioRxiv). We were able to visualize the L protein in functional states of pre-initiation and early elongation at high resolution and could also observe changes in conformation upon 3' RNA binding to a so-called secondary binding site (outside the polymerase active site). Functional in vitro and cell-based studies were the basis for this work and complemented the structural insights.
This figure shows the ~250 kDa L protein in an early elongation state with template and product RNA in the active site of the RNA-dependent RNA polymerase. The different domain names and their location both in the 3D model and the linear protein sequence are indicated.
