Molecular Analysis of Vanilla Mosaic Virus From the Cook Islands
Vanilla was first introduced to French Polynesia in 1848 and from 1899-1966 was a major export for French Polynesia who then produced an average of 158 tonnes of cured Vanilla tahitensis beans annually. In 1967, vanilla production declined rapidly to a low of 0.6 tonnes by 1981, which prompted a nation-wide investigation with the aim of restoring vanilla production to its former levels. As a result, a mosaic-inducing virus was discovered infecting V. tahitensis that was distinct from Cymbidium mosaic virus (CyMV) and Odontoglossum ringspot virus (ORSV) but serologically related to dasheen mosaic virus (DsMV). The potyvirus was subsequently named vanilla mosaic virus (VanMV) and was later reported to infect V. tahitensis in the Cook Islands and V. planifolia in Fiji and Vanuatu. Attempts were made to mechanically inoculate VanMV to a number of plants that are susceptible to DsMV, but with no success. Based on a partial sequence analysis, VanMV-FP (French Polynesian isolate) and VanMV-CI (Cook Islands isolate) were later characterised as strains of DsMV exclusively infecting vanilla. Since its discovery, little information is known about how VanMV-CI acquired the ability to exclusively infect vanilla and lose its ability to infect natural hosts of DsMV or vice versa.
The aims of this research were to characterise the VanMV genome and attempt to determine the molecular basis for host range specificity of VanMV-CI. VanMV-CI has a typical potyvirus genome structure encoding a single polyprotein which can potentially be cleaved into ten proteins, flanked by 5’ and 3’ untranslated regions (UTR). Sequence comparisons of individual genes and complete genomes from members of the bean common mosaic virus (BCMV) group to which DsMV belongs revealed VanMV-CI is most closely related to DsMV. Sequence analyses revealed VanMV-CI is 75.1-77.5% and 81.8-84.1% identical to DsMV at the nucleotide (nt) and amino acid (aa) levels, respectively in the coat protein (CP) and 75.8-77.1% and 81.9-86.7% identical to DsMV at the nt and aa levels, respectively over the entire genome. These data are consistent with the findings of Farreyrol et al. (2005) who first suggested that VanMV-CI is a strain of DsMV. The origin of VanMV-CI is unclear. Did it evolve from DsMV or was it the progenitor of DsMV? To try answer this, a recombination and phylogenetic analysis was carried out. A recombination analysis between DsMV and VanMV-CI genome sequences was carried out against all available potyvirus genomes to determine whether VanMV-CI emerged as a result of a recombination event between DsMV and another virus. This analysis was also to determine whether the N-terminal P1 gene of VanMV-CI arose as a result of a recombination event giving VanMV the ability to exclusively infect vanilla spp. The reasoning for focussing on this genome region is outlined below. This analysis found two recombination events in the P1 gene of DsMV, one in its N-terminal region and another in its C-terminus. Sequence identities between DsMV and its major parent however, were lower than 75% and the recombination event was only detected by one detection method producing insufficient evidence for VanMV arising as a result of one or more recombination events. A phylogenetic analysis based on the complete genome showed that the emergence of VanMV-CI preceded the publicly available DsMV genomes. However, a maximum likelihood analysis of the entire CP of VanMV-CI and all available DsMV CP sequences showed that VanMV-CI was most closely related to DsMV in the South Pacific and the origin of DsMV in this region preceded VanMV-CI. In addition, this analysis showed that DsMV likely originated in the Asia region (China, Japan and India) where its natural host (Colocasia esculenta and Xanthosoma species) originated- it is possible that DsMV spread to the Pacific with human migration. Thus, it is most likely that VanMV arose from DsMV, by an as yet unknown mechanism- possibly by random mutation and natural selection or by a recombination event between DsMV and an as yet to be discovered virus. Upon sequence comparison with DsMV, the P1-HC-Pro region of VanMV-CI became a region of interest for determining a molecular basis for the difference in host range, as this was the most divergent area between the two genomes. This difference was largely due to an insertion/deletion (indel) found in the N-terminal region of the VanMV-CI P1. According to a number of studies, the N-terminal region of the potyviral P1 gene must be compatible with a still unknown host factor before its C-terminus can effectively cleave between itself and HC-Pro. Without successful separation from P1, HC-Pro is unable to counterattack RNA silencing by plants and the virus ceases to infect its host. This suggests that the VanMV-CI P1 may be the region responsible for its ability to infect vanilla but not aroids. An agroinfiltration experiment was designed to test the functionality of the P1 region as a viral suppressor of RNA interference (VSR) to provide support for its role as a host range determinant. A series of gene constructs were designed to test the function of the VanMV-CI and DsMV P1 genes in combination with the HC-Pro region to determine if the HC-Pro activity of each virus is affected by the origin of the P1. The HC-Pro and P1-HC-Pro regions of VanMV-CI and a New Zealand isolate of DsMV were amplified to generate a series of wild type and hybrid constructs for future use in determining their ability to counterattack RNA silencing by plants (silencing suppressor activity). The wild type constructs (PKP2 and PKP4) and a hybrid PKP6 construct containing DsMV P1/VanMV-CI HC-Pro were generated. Generating the alternative hybrid PKP5 construct (VanMV-CI P1/ DsMV HC-Pro) was unsuccessful, most probably due to a lack of sequence complementarity between the fusion templates (VanMV-CI P1 and DsMV HC-Pro fusions). Attempts were made to transform the constructs PKP1-PKP4 (VanMV HC-Pro, VanMV P1/VanMV HC-Pro, DsMV HC-Pro and DsMV P1/DsMV HC-Pro respectively) into a pHEX2 expression vector for agroinfiltration into Nicotiana benthamiana to examine the difference in silencing suppressor activity in light of the difference in P1. Cloning of the PKP6 construct into the entry vector (pCR8) was unsuccessful and transformation of expression vectors containing PKP constructs with Agrobacterium tumefaciens GV3101 also encountered a number of problems. This was most likely due to the use of the inefficient freeze-thaw method. Although the agroinfiltration experiment could not be completed, PKP constructs including HC-Pro and P1/HC-Pro of VanMV and DsMV were amplified and incorporated into the destination expression vector pHEX2 and their sequences confirmed.