Thesis (Ph.D.) - University of St Andrews, June 2008. / Pagination differs from that of the electronic version in the Digital Research Repository.

Thesis (Ph.D.) - University of St Andrews, November 2007.

Lipopolysaccharide (LPS) is an essential component of the outer membrane (OM) of Gram-negative bacteria and plays a fundamental role in protecting the bacteria from harsh environments and toxic compounds. The LPS transport system is responsible for transporting LPS from the periplasmic side of the inner membrane (IM) to the OM, in a process involving seven LptA-LptG proteins. The current model for lipopolysaccharide transport (Lpt) suggests that LPS is initially extracted by a four-protein complex, LptBCFG, from the inner membrane to the periplasm, where LptA mediates further transport to the OM. Another two protein complex, LptD/E, catalyses the assembly of LPS at the OM cell surface. However, the details of this transport mechanism have remained unknown, mainly due to a lack of structural information. In chapter 1 and 2 of this thesis, I report materials and methods for all LptD/E, and Schmallenberg virus (SBV) nucleoprotein (NP) experiments and the theories and software that were used in determining structures of LptD/E, SBV NP and the SBV NP/RNA complex. In chapter 3 of this thesis, I report the first crystal structure of the outer membrane protein LptD/E complex. LptD forms a 26-strand ß-barrel in a closed form and LptE is a roll-like structure located inside LptD to form “barrel and plug” architecture. Through structural analysis, function assay and molecular dynamics simulation, we proposed a mechanism in which the hydrophilic head of LPS molecule, including the oligosaccharide core and the O antigen, directly penetrates through the hydrophilic ß- barrel whilst the hydrophobic lipid A tail is inserted into an intramembrane hole, with a lateral opening between strand ß1 and ß26 of the LptD. LptE may assist this process. In chapter 4, I report the crystal structure of the SBV NP in two conformations: tetrameric when the protein was purified under native conditions, and trimeric when denatured and refolded during purification. The SBV NP has a novel fold and we have also identified that the N-terminal arm is crucial for RNA binding, and the N- and the C-terminal arm is essential for RNA multimerisation with adjacent protomers and for viral RNA encapsidation. Chapter 5 describes the crystal structure of SBV NP in complex with a 42 nucleotide long RNA (polyU). This ribonucleoprotein (RNP) complex was crystallized as a ring-like tetramer with each protomer bound to 11 ribonucleotides. Eight of these nucleotides are bound in a positively charged cleft between N- and C- terminal domains and three are bound in the N-terminal arm. I also compared the structure to that of other NPs from negative-sense RNA viruses, and found that SBV NP sequesters RNA using a different mechanism. Furthermore, the structure suggests that when RNA binds the protein, there are conformational changes in the RNA-binding cleft, and in the N- and C-terminal arms. Thus our results reveal a novel mechanism of RNA encapsidation by orthobunyaviruses NP.

Viruses in the genus Orthobunyavirus (family Bunyaviridae) are classified serologically into 18 serogroups. The viruses have a tripartite genome of negative sense RNA composed of large (L), medium (M) and small (S) segments. The L segment encodes the polymerase protein, the M segment encodes two glycoproteins, Gc and Gn, and a non-structural protein (NSm), and the S segment encodes nucleocapsid (N) and NSs proteins, in overlapping reading frames (ORF). The NSs proteins of Bunyamwera and California serogroup viruses have been shown to play a role in inhibiting host cell protein synthesis and preventing induction of interferon in infected cells. To-date, viruses in only 4 serogroups: Bunyamwera, California, Group C and Simbu, have been studied intensively. Therefore, this study was conducted with the aim to sequence the S RNA segments of representative viruses in the other 14 orthobunyavirus serogroups, to analyse virus-encoded proteins synthesised in infected cells, and to investigate their ability to cause shutoff of host protein synthesis. S RNA segment sequences were obtained from cloned RT-PCR products. They were compared with the available sequences and each other. Complete S RNA sequences of Anopheles A (ANAV) and Tacaiuma virus (TCMV) [Anopheles A serogroup], Anopheles B (ANBV) and Boraceia virus (BORV) [Anopheles B serogroup], Eretmapodites (E147V) and Nyando virus (NDV)[Nyando serogroup], Bwamba virus (BWAV) [Bwamba serogroup], M’Poko virus (MPOV) [Turlock serogroup], Tete (TETEV) and Batama virus (BMAV) [Tete serogroup], and Gamboa (GAMV) and San Juan 2441 virus (SJ244V) [Gamboa serogroup], and partial sequences of Patois virus (PATV) [Patois serogroup], Guama (GMAV) and Bertioga virus (BERV) [Guama serogroup], Capim virus (CAPV) [Capim serogroup] and Palestina virus (PLSV) [Minatitlan serogroup] were obtained. Complete S segment sequences revealed that viruses in the same serogroup have same length of N and NSs proteins, except for the viruses in Gamboa serogroup which were found to have two lengths of NSs protein. Viruses in 4 serogroups (Anopheles A, Anopheles B, Tete and Capim) were found not to encode an NSs ORF, presenting the first report of naturally isolated orthobunyaviruses without an NSs protein. Most of these viruses were found to have longer N proteins compared to those with NSs protein, with the largest N protein observed to date in TETEV and BMAV (258 amino acids). Other viruses 3 (EREV, NDV, GAMV, SJ2441V, BWAV and MPOV) were found to encode both N and NSs proteins in their S segment with the largest and smallest NSs protein detected to date in SJ2441V (137 amino acids) and MPOV (70 amino acids) respectively. The conserved CA rich motif in 5’ non coding region (NCR) of Bunyamwera and California serogroups viruses was absent in BWAV and MPOV, while ANBV and BORV were found to have two copies of this motif. Repeated sequences, as observed previously in the 5’ NCR of genomic-sense RNA of Lumbo virus (LUMV), were also detected in BWAV and TCMV S RNA segments. Sequence comparisons and phylogenetic analyses of the sequences determined in this study were in agreement with previous serological classification of the viruses, except for BERV and TCMV. BERV, in the Guama serogroup, was found to have a closer relationship with CAPV compared to GMAV. However high sequence identities (>70%) were observed between these 3 viruses, suggesting that they are derived from the same ancestor. N protein and nucleotide sequence identities of TCMV with ANAV were only 53% and 59% respectively. However, Neighbour-Joining (NJ) plot based on complete N amino acid sequence and Maximum Parsimony (MP) plot based on partial N sequence supported previous serological classification which placed this virus in the same clade as ANAV. This study first reports on the proteins synthesised by Bakau, Bwamba, Koongol, Gamboa, Minatitlan, Olifantsvlei and Tete serogroup viruses. Analysis of radio-labelled cell extracts revealed similar protein migration patterns for all the studied viruses compared with other viruses in the genus Orthobunyavirus. Shutoff of host cell protein synthesis, similar to that seen in Bunyamwera virus (BUNV)-infected cells was only observed in ACAV, BAKV, BWAV, CAPV, PAHV, PATV and WONV-infected cells. However, this shutoff was found not related to the presence of NSs protein. In general, viruses in the same serogroup were found to have almost same size of plaque and plaque-size did not correlate with the presence of NSs protein and the virulence of the virus in the mice. In vitro transcription and translation (TnT) using rabbit reticulocyte and wheat germ lysate expression systems further confirmed the sequencing results that no NSs protein was expressed from S cDNA clones of ANAV, TCMV, ANBV, BORV, BMAV and TETEV. S RNA segments shutoff almost similar to BUNV-infected cells was observed in A549 cells infected with TCMV, suggesting that TCMV might use a different mechanism to induce shutoff. No significant shutoff was observed in Hep2, Hep2/V and C6/36 cells infected with any of the viruses. RT-PCR specific for IFN- ß mRNA in 293 infected cells and IFN reporter gene assays revealed that TCMV was capable of counteracting IFN production similar to wt BUNV, whereas the other NSs minus viruses (ANAV, ANBV, BORV, TETEV and BMAV) were found to be capable of inducing IFN in infected cells. However, only low level of IFN- ß mRNA and weak activation of the IFN- ß promoter was detected in ANAV and BMAV- infected cells.

The La Crosse (LAC) virus is a member of the California encephalitis group of bunyaviridae (Porterfield et al., 1975 and 1976). It is one of an envelope virus and this virus under acidic conditions (below pH 6.3) has been demonstrated to result in cell-to-cell fusion (Gonzalez-Scarano, 1984). The LAC virus is also capable of forming virus-to-virus fusion particles. The focus of this thesis is the analysis of the structure and function of this virus-to-virus fusion by cryo-electron microscopy at different pH and temperatures. The results of this study provide the basis for further study of the structure and function of the LAC virus. The virus-to-virus fusion event shows a dependence on both pH and temperature. The frequency of the fusion event increases with an elevation in temperature (in the range 4 to 37°C) and with a decrease in pH from 7.3 to 5.4. The process of virus-to-virus fusion gives rise to the formation budding to a chain of fused viruses.

The family Bunyaviridae is the largest known taxonomic group of arboviruses. Four of the five genera possess members which are responsible for serious human and livestock disease. The worldwide distribution of these viruses justify studies which will allow understanding of the replication and transcription cycles within permissive cells. The bunyaviruses have been shown to possess a tripartite single strand RNA genome of negative polarity. Replication is confined to the cytoplasm and the virion envelope is acquired when the genome ribonucleoproteins bud into the golgi. Virus release is presumed to be through exocytosis and ultimately cell lysis. The messenger RNA species of all five genera do not possess a poly-A tail of sufficient length to bind to an oligo(dT) cellulose column. This has made separation of viral transcripts from replicating RNAs difficult. In an effort to achieve this separation, infected cell extracts were centrifuged over 20-40% CsCl gradients which permitted replicating RNA structures to band at a density of 1.32 while cellular and viral mRNAs pellet. Recovery of viral transcripts from the CsCl pelleted RNA required synthesis of a cDNA copy of the virus genome to use as a probe. This was done by an unusual method which employs both genome and antigenomic RNA as templates for reverse transcriptase in a first strand synthesis reaction. Recombinant viral clones were then used in a hybrid selection scheme to recover virus mRNA from pelleted material. After recovery, the messages were visualized on acid urea agarose gels pH 3.5, or used to program an in vitro translation reaction. Using these methods, it was established that each genome segment codes for a single messenger RNA which is most likely capped, and that for at least the mid sized segment, proteins with molecular weights which exceed the coding capacity of the genome are translated from the single message.

Bunyamwera virus (BUNV) is the prototype of the family Bunyaviridae. It has a tripartite genome consisting of negative sense RNA segments called large (L), medium (M) and small (S). The S segment encodes the nucleocapsid protein (N) of 233 amino acids. The N protein encapsidates all three segments to form transcriptionally active ribonucleoproteins (RNPs). The aim of this project was to determine the domain map of BUNV N protein. To investigate residues in BUNV N crucial for its functionality, random and site- specific mutagenesis were performed on a cDNA clone encoding the BUNV N protein. In total, 102 single amino acid substitutions were generated in the BUNV N protein sequence. All mutant N proteins were used in a BUNV minigenome system to compare their activity to wt BUNV N. The mutant proteins displayed a wide-range of activity, from parental-like to essentially inactive. The most disruptive mutations were R94A, I118N, W134A, Y141C, L177A, K179I and W193A. Sixty-four clones carrying single substitutions in the BUNV N protein were used in the BUNV rescue system in an attempt to recover viable mutant viruses. Fifty recombinant mutant viruses were rescued and 14 N genes were nonrescuable. The 50 mutant viruses were characterized by: titration, protein labelling, western blotting, temperature sensitivity and host-restriction. Mutant viruses displayed a wide range of titers between 10³ -10⁸ pfu/ml, and three different plaque sizes large, medium and small. Protein labelling and western blotting showed that mutations in the N gene did not affect expression of the other viral genes as much as affecting N protein expression. It was demonstrated that single amino acid substitutions could alter N protein electrophoretic mobility in SDS- PAGE (e.g. P19Q and L53F). Temperature sensitivity tests showed that recombinant viruses N74S, S96S, K228T and G230R were ts, growing at 33˚C but not at 37˚C or 38˚C, while the parental virus grew at all temperatures. Using the northern blotting technique, mutant viruses N74S and S96G were shown to have a ts defect in genome-synthesis (late replication step), while mutant viruses K228T and G230R had a ts defect in antigenome- synthesis (early replication step). Host-restriction experiments were performed using 5 different cell lines (Vero-E6, BHK-21, 2FTGH-V, A549-V and 293-V). Overall, the parental virus grew similarly in all cell lines. Likewise, the majority of mutant viruses follow this pattern except mutant virus Y23A. It showed a 100-fold reduction in titer in 2FTGH-V cells. Comparing the ratios of intracellular and extracellular particles revealed that only 15% of the total virus particles of mutant Y23A was released as extracellular particles compared to 30% of the parental virus. Fourteen N genes were nonrescuable. They were characterized by (i) their activity in the BUNV minigenome system, (ii) their activity in BUNV packaging assay, (iii) their ability to form multimers, (iv) their ability to interact with L protein, and (v) their impact on RNA synthesis. In summary, BUNV N protein was shown to be multi-functional and involved in the regulation of virus transcription and replication, RNA synthesis and assembly, via interactions with the viral L polymerase, RNA backbone, itself or the viral glycoproteins.

The Bunyaviridae family is one of the largest among RNA viruses, comprising more than 350 serologically distinct viruses. The family is classified into five genera, Orthobunyavirus, Hantavirus, Nairovirus, Phlebovirus, and Tospovirus. Orthobunyaviruses, nairoviruses and phleboviruses are maintained in nature by a propagative cycle involving blood-feeding arthropods and susceptible vertebrate hosts. Like most arthropod-borne viruses, bunyavirus replication causes little damage to the vector, whereas infection of the mammalian host may lead to death. This situation is mimicked in the laboratory: in cultured mosquito cells no cytopathology is observed and a persistent infection is established, whereas in cultured mammalian cells orthobunyavirus infection is lytic and leads to cell death. Bunyaviruses encode four common structural proteins: an RNA-dependent RNA polymerase, two glycoproteins (Gc and Gn), and a nucleoprotein N. Some viruses also code for nonstructural proteins called NSm and NSs. The NSs protein of the prototype bunyavirus, Bunyamwera virus, seems to be one of the factors responsible for the different outcomes of infection in mammalian and mosquito cell lines. However, only limited information is available on the growth of bunyaviruses in cultured mosquito cell lines other than Aedes albopictus C6/36 cells. Here, I compared the replication of Bunyamwera virus in two additional Aedes albopictus cell clones, C7-10 and U4.4, and two Aedes aegypti cell clones, Ae and A20, and investigated the impact of virus replication on cell function. In addition, whereas the vertebrate innate immune response to arbovirus infection is well studied, relatively little is known about mosquitoes’ reaction to these infections. I investigated the immune responses of the different mosquito cells to Bunyamwera virus infection, in particular antimicrobial signaling pathways (Toll and IMD) and RNA interference (RNAi). The data obtained in U4.4 cells suggest that NSs plays an important role in the infection of mosquitoes. Moreover infection of U4.4 cells more closely resembles infection in Ae and A20 cells and live Aedes aegypti mosquitoes. My data showed that the investigated cell lines have various properties, and therefore they can be used to study different aspects of mosquito-virus interactions.

Oropouche virus (OROV) is a medically important orthobunyavirus, which causes frequent outbreaks of a febrile illness in the northern parts of Brazil. However, despite being the cause of an estimated half a million human infections since its first isolation in Trinidad in 1955, details of the molecular biology of this tripartite, negative-sense RNA virus remain limited. The work presented in this thesis has re-determined the nucleotide sequences of OROV strain BeAn19991 (GenBank accession numbers: L, KP052850; M, KP052851 and S, KP052852), and demonstrates that the S segment is significantly longer than the published sequence with an additional 204 nucleotides at the 3' end. Data analysis revealed that there is a critical nucleotide mismatch at position 9 within the base-paired terminal panhandle structure of each genomic segment. Using a combination of deep sequencing and Sanger sequencing the complete genome sequences of 10 field isolates of OROV were also determined for the first time, and led to the identification of a novel OROV reassortant virus. Phylogenetic analysis of these sequences and of published sequences showed that there are two genotypes of OROV, rather than the four genotypes previously proposed. Further work led to the development of a T7-RNA polymerase-driven minigenome and virus-like particle (VLP) production systems for OROV; the information from these was subsequently used to develop a reverse genetics system for OROV. Using reverse genetics, OROV mutants that lack either the non-structural proteins NSm or NSs were generated. In vitro growth properties of the OROV mutant lacking NSm were indistinguishable from the wild-type virus, but the NSs mutant was attenuated in growth, particularly in interferon (IFN) competent cells. Further work demonstrated NSs as a viral IFN antagonist and that it's C-terminus is required for this activity. Interestingly, OROV is more resistant to IFN-α treatment than Bunyamwera virus, but this is not related to its NSs protein. The development of a reverse genetics system for OROV, which is the main human pathogen within the Simbu serogroup of orthobunyaviruses, will prove invaluable for future studies designed to further investigate the molecular pathogenesis of this virus and in the development of attenuated vaccine strains.

Bunyaviridae is one of the biggest known viral families, and includes many viruses of clinical and economic importance. The major virulence factor of most bunyaviruses is the non-structural protein (NSs). NSs is expressed early in infection and inhibits the innate immune response of the host by blocking several steps in the interferon induction and signalling pathways. Hence, NSs significantly contributes to the establishment of a successful viral infection and replication, persistent infection and the zoonotic capacity of bunyaviruses. Although functions and structures of many viral interferon antagonists are known, no structure of a bunyavirus NSs protein has been solved to date. This strongly limits our understanding of the role and the mechanism of interferon antagonism in this large virus family. In this work the first structure for a bunyavirus interferon antagonist, the core domain crystal structure of NSs from the Rift Valley fever virus (RVFV) is presented. RVFV is one of the most clinically significant members of the Bunyaviridae family, causing recurrent epidemics in Africa and Arabia, often featuring high-mortality haemorrhagic fevers. The structure shows a novel all-helical fold. The unique molecular packing of NSs in the crystal creates stable fibrillar networks, which could correspond to the characteristic fibrillation of NSs observed in vivo in the nuclei of RVFV infected cells. This first NSs structure might be a useful template for future structure-aided design of drugs that target the RVFV interferon antagonism. Attempts at characterising other bunyavirus NSs proteins of other genera were made, but were hampered by problems with obtaining sufficient amounts of soluble and folded protein. The approaches that proved unsuccessful for the solubilisation of these NSs proteins, however, should inform future experiments aimed at obtaining recombinant NSs for structural studies.