Tomato spotted wilt virus particle assembly : studying the role of the structural proteins in vivo

Members of the Bunyaviridae have spherical, enveloped virus particles that acquire their lipid membrane at the Golgi complex. For the animal-infecting bunyaviruses, virus assembly involves budding of ribonucleoprotein particles (RNPs) into vacuolised lumen of the Golgi complex, after which the enveloped particles are secreted. The maturation of tomato spotted wilt virus (TSWV), a bunyavirus infecting plants, is different in that virions acquire their membrane by wrapping of a Golgi stack around RNPs after which the enveloped particles eventually accumulate in large vesicles in the plant cell. TSWV also multiplies in its insect vector thrips, and here particles are secreted from salivary gland cells into the gland ducts. The latter seems a logic requirement to allow virus passage to healthy host plants.To further study the process of TSWV particle assembly, the interactions between the structural N, Gn and Gc proteins in mature virus particles, as well as their intracellular behaviour invivo havebeen the main target of this Ph. D. thesis.After an introductory chapter on bunyavirus particle assembly (chapter 1), the protein composition of purified TSWV RNPs and enveloped particles was studied in chapter 2.In enveloped virus preparations, the three major structural proteins, i.e. the nucleocapsid protein (N) and the two envelope glycoproteins Gn andGc, were detected in monomeric as well as oligomeric forms. GlycoproteinGcbut not Gn was observed tightly bound to RNPs, suggesting Gc is involved in RNP envelopment. Analysis of cytoplasmic RNPs and mature virus particles for other viral proteins revealed, surprisingly, the presence of the so-called nonstructural protein NSs. Whereas mature virus particles contained only traces of NSs, RNP preparations clearly contained larger amounts of this protein, which could be related to an earlier reported difference in transcriptional/replicational activity between both.To study the process of virus assembly in more detail, fluorescence microscopy methods were employed for the in vivo detection of protein interactions, rendering information concerning the intracellular localisation simultaneously (chapter 3). For this a system was set up in mammalian cells and as a first protein to be studied the cytosolic N protein was selected. This protein was already known to form homo-oligomeric structures in vitro. Using fusions of N with either yellow or cyan fluorescent protein (YFP and CFP, respectively), pairs were created to function as a donor (CFP) and acceptor (YFP) fluorophore for fluorescence resonance energy transfer (FRET). Using acceptor photobleaching and fluorescence lifetime imaging microscopy (FLIM) to measure FRET, N was observed to form homodimers and -multimers throughout the cytoplasm before eventually accumulating in a non-Golgi, perinuclear region ona microtubuli- and actin-dependent manner.In a similar way, potential in vivo interactions between N and the viral glycoproteins were investigated (chapter 4). While no interaction between N and Gn was observed, these studies demonstrated interaction between N andGc, inagreementwith the earlier observation (chapter 2) that some Gc remains tightly bound to purified RNPs. The interactions between N andGclocalised to the non-Golgi perinuclear area, similar to transiently expressed N. These data provided further support for the idea that interactions between N andGcare involved in envelopment of the viral RNPs. While studying the possible formation of heterodimers of Gn andGc, it appeared that with the constructs used, FRET could not be applied for this purpose, as fluorescence from the two fluorophore fusion proteins was never observed in the same cell. This could be due to the fact that interaction between the two glycoproteins interfered with proper folding of (one of) the fluorophores, resulting in greatly reduced and possibly undetectable fluorescence.Another intriguing question-relevant to a broader cell biological field as well-concerns the signal responsible for Golgi localisation of the two glycoproteins during infection. Previous work had shown that Gn carries a Golgi retention signal and is able to rescueGcfrom the ER to the Golgi, suggesting a heterodimerisation of Gn and Gc. For a number of other bunyaviruses, the Golgi retention signal had been mapped to the C-terminal transmembrane domain (TMD) and / or cytoplasmic tail of Gn. Using C-terminal deletion mutants of TSWV Gn and chimeric Gc and vesicular stomatitis virus glycoprotein (VSV-G) constructs (chapter 5), it was shown that both the TMD and the first 30 amino acids of the 60 residues-sized cytoplasmic tail of Gn are necessary for Golgi localisation. The lumenal domain was shown not to be required for Golgi localisation, nor was its presence required for rescuing ofGc. By deletion mapping the 20 most C-terminal residues of the cytoplasmic tail were shown to be crucial for interaction withGc.Large cells containing multiple nuclei were frequently observed whenGcwas expressed. This phenomenon was further investigated in chapter 6, and was shown probably to result from a cell fusion activity of this glycoprotein. Cell fusion is not likely to occur during the plant infection cycle, but may play a role in the infection cycle or the process of virus entry within the thrips vector. The fusion was not pH-dependent and not observed with Gn.Chapter 7 discusses the major findings of this Ph. D. research in a broader perspective, and presents a model for TSWV particle assembly in which all observations have been accomodated.