Mutational and structural analyses of tomato curly stunt virus pathogenicity determinants

Abstract

Geminiviruses infect a wide variety of important crop plants causing severe disease resulting in significant yield losses worldwide. Tomato production in the subtropical growing regions of South Africa (SA) is threatened by the emergence of Tomato curly stunt virus (ToCSV), a monopartite ssDNA geminivirus belonging to the genus Begomovirus. The broad aim of this study was to examine the coding capacity, sequence features, and the role of identified sequence differences on the infectivity in vivo of two variant group isolates (V30 and V22) identified in a previous nationwide survey of tomato (Esterhuizen, 2012). An additional study aim was to explore the structure and investigate the function of ToCSV replication-associated protein (Rep) using in silico and lab-based techniques. Both V30 and V22 DNA-A genomes were predicted to contain three putative complementary-sense (cs) ORFs (C5, C6, C7) and two virion-sense (vs) ORFs (V3 and V4). Sequence alignment of putative ORFs identified in related African mono- and bipartite begomoviruses revealed that the C5/AC5 and C6/AC6 ORFs were the most frequently identified putative ORFs encoding divergent proteins. Two novel putative protein domains were described: a C5/AC5 N-terminal long helix region (NLHr) containing a predicted transmembrane domain, and a C6/AC6 conserved core domain (C6cd). This project was the first to conduct a comprehensive assessment of putative ORFs in African begomoviruses. Additional experiments are necessary to confirm their expression and determine their role in infection. Previously, an agroinfectious clone of V30 was shown to induce severe disease symptoms, whilst an agroinfectious clone of V22 induced significantly milder disease symptoms in tomato cv Rooi Khaki (Esterhuizen, 2012). Infectivity assays of V30 and V22 were conducted for the first time in Nicotiana benthamiana in the current study. The symptom phenotype induced by V30 was characterised by severe swollen veins (SV) and leaf rugosity with no recovery in disease symptoms. V22 induced moderate to severe upward leaf roll (ULR) with a significant recovery in both ULR and SV disease symptoms being observed between 24 and 36 days post inoculation (dpi). Viral load was quantified using qPCR and determined to be significantly higher in V30- infected plants at peak (20 dpi) and late (36 dpi) infection timepoints. Using RT-PCR with strand-specific primers, RNA transcripts derived from both vs and cs strands were xxxiii identified in V30-infected plants. The identified transcripts spanned multiple canonical and putative ORFs, putative V2 and C1 5’ ORF extension regions, as well as the intergenic region (IR) crossing the origin of replication (Ori). The presence of these transcripts suggests that putative ORFs are transcribed, and the presence of long transcripts (~1.2 kb) may be evidence of bi-directional readthrough transcription to generate dsRNA for downstream RNA silencing pathways, although additional experiments are required to confirm this. The role of sequence differences in the 3’ IR and V2 ORF regions on infectivity of V30 and V22 in Nicotiana benthamiana was investigated using V30 partial 3’ IR swap mutant chimeras, as well as V30 and V22 canonical V2 ORF swap mutant chimeras. Plants infected with the V30ΔIR-s mutant containing a 78 nt segment of the V22 3’ IR developed ULR, a symptom characteristic of V22 infection. A significant increase in viral load at 28 dpi was also observed. Altered symptom phenotypes induced by shorter IR swap mutants suggested that the unique arrangement of a novel cis-acting TATA-associated composite element (TACE) situated to the right of the conserved TATA box promoter contributes to the observed phenotype differences. N. benthamiana infected with V30ΔV2-s displayed a significant reduction in SV symptom severity at late infection and significantly reduced viral titre was measured at this timepoint compared to that measured in V30-infected plants. The V22ΔV2-s mutant induced significantly increased disease symptom severity which was associated with a significant increase in viral load at peak and late infection. A loss of symptom recovery phenotype was reported with symptom severity remaining high at 36 dpi. Taken together, these findings indicate that the ToCSV 3’ IR region is a symptom phenotype determinant in N. benthamiana whilst sequence differences present in the canonical V2 ORF are responsible for the modulating viral titre, disease symptom severity, and V22-specific disease recovery phenotype. The precise mechanisms for these observed changes warrant further investigation and additional infectivity assays should be conducted in tomato. Sequence analysis of the Rep proteins of ToCSV and related African begomoviruses was conducted and the presence of key HUH endonuclease motifs (motif I, II, III), and ATPase/superfamily 3 (SF3) helicase motifs (Walker A, Walker B, B’ motif, C motif) were confirmed. Homology modelling of ToCSV Rep showed that the N-terminal endonuclease domain (monomer) was predicted to contain an antiparallel β-sheet xxxiv (five β-strands), whilst the C-terminal ATPase/helicase domain (homohexamer) contained a parallel β-sheet (four β-strands). The arrangement of important functional motifs was similar to that present in the solved structures of other DNA virus Rep proteins. Agroinoculation of N. benthamiana with V30 C1 deletion mutants indicated that deletion of the 14 C-terminal residues of Rep (aa 346 – 359) resulted in significantly reduced viral load and greatly delayed symptom development. Removal of the 22 C-terminal residues of Rep (aa 338 – 359) abolished infectivity entirely. From these results it was concluded that the 14 C-terminal residues (predicted to be disordered) are not essential for Rep enzymatic activity which is a prerequisite for successful viral replication in vivo. The ToCSV RepC protein (aa 122 – 359) was successfully expressed in Escherichia coli as a recombinant fusion. Additional optimisation of expression and purification is required to obtain sufficient purified protein for future structural studies. Miniaturised expression trials of RepC truncated by 14 C-terminal residues appeared to yield higher quantities of protein compared to full length RepC as evidenced by SDS-PAGE. Targeted removal of disordered C terminal residues is therefore suggested as a practical approach to improve yield of recombinantly expressed RepC whilst retaining the enzymatic activity of the domain of interest. The present study contributed to increasing the knowledge of putative begomovirus ORFs as well as providing insight into the possible functions of selected sequence elements of interest present in the ToCSV genome. It is anticipated that the broad scope of this study will generate several intriguing avenues for future research of ToCSV and related African begomoviruses which remain largely understudied. It is hoped that this research will contribute to improving the understanding of these pathogens with the aim of developing measures aimed at protecting the expanding agriculture industry in Southern Africa