Thesis--Auburn University, 2005. / Abstract. Vita. Includes bibliographic references (p.24-30)

Assembly of virus capsid protein (CP) into icosahedrally symmetric particles is an intriguing and elegant process. In most cases of virus assembly, a large number of identical protein subunits self-assemble to generate a shell that protects the viral genome. Studies on virus assembly have resulted in a new scientific technique that uses these proteinaceous shells as nano-particles for a variety of biological applications. The current thesis deals with understanding the factors that govern the assembly of the Sesbania mosaic virus (SeMV) and a pleomorphic virus, Tobacco streak virus (TSV). CP of SeMV, a T=3 plant virus, consists of a disordered N-terminal R-domain and an ordered S-domain. The importance of the R-domain in the assembly was probed by replacement with polypeptides such as the B-domain of Staphylococcus aureus protein A and polypeptides P10 and P8 of SeMV. These chimera assembled into T=3 or larger virus like particles (VLPs). Addition of divalent cations resulted in the formation of heterogeneous nucleoprotein complexes that disappeared upon treatment with EDTA/RNAse. One of the chimeras (N∆65-B) purified in a dimeric form by affinity chromatography assembled into T=1 VLPs during crystallization. The three dimensional structure of these VLPs showed that they were devoid of divalent ions and the B-domain was disordered. These studies demonstrate the importance of N-terminal residues, metal ions in virus assembly and robustness of the assembly process. Also, the B-domain was functional in N∆65-B VLPs, suggesting possible biotechnological applications. Tobacco streak virus (TSV) is a polymorphic virus and a major plant pathogen. TSV capsids encapsidate the tri-partite ss-RNA genome of the virus in three spheroidal particles of diameters 27, 30 and 33 nm, respectively. CPs of ilarviruses are also involved in genome activation. The labile nature of ilarviruses has posed difficulties in their structure determination. This thesis describes the first crystal structure of truncated TSV-CP. The core of TSV CP conforms to the canonical β-barrel jelly roll tertiary structure found in other viral coat proteins. Dimers of CP with swapped C-terminal arms (C-arm) were observed in the two crystal structures determined. The C-arm was found to be flexible and responsible for the polymorphic and pleomorphic nature of TSV capsids. Mutations in the hinge region of the C-arm that reduce the flexibility resulted in the formation of more uniform particles. TSV CP was also found to be structurally similar to that of Alfalfa mosaic virus (AMV) accounting for similar mechanism of genome activation in alfamo and ilar viruses.

In the present thesis, two positive sense single-stranded RNA viruses have been used as models to understand the structure and function of viral-encoded proteins. One of them, Pepper Vein Banding Virus (PVBV; genus Potyvirus; family Potyviridae) is a flexuous, rod-shaped virus that encodes for a polyprotein of size ~340 kDa. The polyprotein undergoes proteolytic processing by viral-encoded proteases, of which Nuclear Inclusion-a Protease (NIa-Pro) is the major protease. It is a serine-like cysteine protease which cleaves between a Q/A or Q/S, present in the context of the heptapeptide recognition sequence. The temporal regulation of intermediates and mature proteins released by NIa-Pro cleavage is crucial for a successful infection. In the present study, histidine-tagged NIa-Pro, Viral Protein genome-linked (VPg), and the cleavage site mutant (E191A) VPg-Pro were over-expressed in E. coli and purified. The protease activity of NIa-Pro was monitored using an HPLC-based protease assay developed using a peptide substrate. NIa-Pro protease activity was found to get modulated upon interaction with VPg and upon undergoing phosphorylation. Both these events have been found to involve the face of NIa-Pro which contains the solvent-exposed Trp143. Mutational studies and molecular dynamics analyses provide evidence that this residue is buried upon interaction of NIa-Pro with VPg, and any perturbation of its orientation influences the active site Cys151 via an extensive interaction network. This interaction was found to enhance the velocity of NIa-Pro protease activity, especially if the two domains were present in trans (VPg+Pro). In addition, the main-chain –NH2 group of Trp143 was found to be hydrogen-bonded to the side chain –OH group of Ser129, the residue which was identified to undergo phosphorylation by host plant kinases. Interestingly, when the two domains were present in cis (E191A VPg-Pro), no phosphorylation was observed. Mutations of Ser129 (to phosphorylation-mimic Asp or phosphorylation-deficient Ala residues) which affected this H-bond were found to disturb Trp143 and Cys151 orientation, which drastically reduced the protease activity of NIa-Pro. Within the polyprotein, VPg is present at the N-terminus of NIa-Pro and the cleavage site between them is suboptimal (E/A). In the present study, VPg-Pro was shown to be covalently linked to the genomic RNA present in the virions. Interestingly, during purification, VPg could only be purified from the soluble when it was expressed at the N-terminus of NIa-Pro. A series of bioinformatics and biophysical analysis of VPg showed that PVBV VPg, like other potyviral VPgs, exists as a molten-globule. Moreover, while VPg was shown to harbour the Walker motifs, it was found to exhibit an ATPase activity only when it was present with the NIa-Pro (especially in cis). Lys47 and Asp88:Glu89 were found crucial for optimal activity. Over all the results demonstrated that there is a reciprocal modulation of structure and function of the VPg and NIa-Pro domains. These results can explain the possible significance of an impeded cleavage rate between the two domains of VPg-Pro during PVBV infection. The precursor, VPg-Pro, could offer the advantage of evading the inhibitory phosphorylation of NIa-Pro by the host, as well as drive certain viral processes by virtue of its ATPase activity. And subsequent cleavage of the domains and their trans interaction could offer a higher turnover rate which might assist sufficient CP production required for viral morphogenesis. Another virus, Tobacco Streak Virus (TSV) that belongs to the Ilarvirus genus of the Bromoviridae family is a spherical virus which forms pleiomorphic icosahedral virus particles. It has a tripartite genome and each RNA is encapsidated individually. In the present thesis, TSV was used as a model to understand the properties of its structural protein-the coat protein (CP), with the aim of deciphering TSV assembly process. Thus, the CP gene from TSV RNA 3 was cloned and over-expressed in E. coli. The coat protein thus expressed formed virus-like particles (VLPs), which could be disassembled into dimers using high CaCl2 concentrations. Reassembly of VLPs was possible from dimers even in the absence of any nucleic acid. Mutational analysis of the N-terminal disordered domain showed that 26 amino acid residues from the amino-terminus could be crucial for capsid heterogeneity while, zinc-binding domain was essential for assembly. Overall, the present study shows that the flexible W-C loop of PVBV NIa-Pro, the disordered N-terminal region of PVBV VPg and the disordered N-terminal region of TSV CP harbour residues crucial for regulation of protein function. Such regulatory elements would ultimately allow viruses to maintain a smaller protein number, and thus a smaller genome size.