An integrated strategy for managing Grapevine leafroll-associated virus 3 in red berry cultivars in New Zealand vineyards

To sustain growth and revenue projections, the New Zealand wine sector aims to produce premium quality wine to supply lucrative export markets. In grapevines, however, the presence of virus and virus-like diseases can negatively influence qualitative parameters of wine production. Where such risks are identified, sustainable remediation protocols should be developed. One risk factor is Grapevine leafroll-associated virus 3 (GLRaV-3), an economically important virus of Vitis. In this thesis, I develop components of an integrated management plan with the aim of reducing and sustaining GLRaV-3 incidence at <1%. In Hawke’s Bay vineyard study blocks, three aspects related to GLRaV-3 management were explored between 2008 and 2013: Firstly, herbicide-treated vines and/or land left fallow after removing infected vines may mitigate the effects of GLRaV-3. Historically though, vine root removal was not well implemented, meaning persistent roots may be long term reservoirs for GLRaV-3. I tested the virus reservoir hypothesis in vineyard blocks where virus incidence of ≥95% necessitated removing all vines. Enzyme-linked immunosorbent assay (ELISA) and/or real-time polymerase chain reaction (real-time PCR) detected GLRaV-3 in most remnant root samples tested, independent of the herbicide active ingredient applied (glyphosate, triclopyr, or metsulfuron) or the fallow duration (6 months to 4 years). On some virus-positive root samples, the GLRaV-3 mealybug vector, Pseudococcus calceolariae, was found, and after real-time PCR testing, virus was detected in some mealybugs. Thus, without effective vine removal, unmanaged sources of virus inoculum and viruliferous vectors could pose a risk to the health of replacement vines. Secondly, in most red berry cultivars, GLRaV-3 is characterised by dark red downward curling leaves with green veins. With visual diagnostics predicted to be a reliable identifier of GLRaV-3-symptomatic red berry vines, early identification could support a cost-effective and sustainable virus management plan. In blocks planted in Merlot, Cabernet Sauvignon, Syrah, and Malbec vines, the reliability of visual symptom identification was compared with ELISA. In terms of sensitivity (binomial generalised linear model, 0.966) and specificity (0.998), late-season visual diagnostics reliably predicted virus infection. Moreover, accuracy appeared unaffected by the genetically divergent GLRaV-3 populations detected in Hawke’s Bay. Thirdly, by acting to visually identify and remove (rogue) symptomatic vines when GLRaV-3 incidence is low (<20%), an epidemic may be averted. In this ongoing study, an integrated approach to virus management was adopted in 13 well established Hawke’s Bay vineyard study blocks. All were planted in vines from one of five red berry cultivars. When monitoring commenced in 2009, all symptomatic vines visually identified (n=2,544 or 12%) were rogued. Thereafter, integrating visual diagnostics with roguing reduced virus incidence so that by 2013, just 434 (2.0%) vines were identified with virus symptoms. Annual monitoring revealed within-row vines immediately either side of an infected vine were most at risk of vector mediated virus transmission, although by 2013, just 4% of these vines had virus symptoms. Hence, roguing symptomatic vines only was recommended. In individual study blocks in 2013, virus management was tracking positively in four blocks; while in another four, results were inconclusive. In the remaining five blocks, contrasting but definitive results were evident. In three of those blocks, mean virus incidence of 10% in 2009 was sustained at ≤0.3% within 2-3 years of roguing commencing; in the other two blocks, mean incidence was 12% but cumulative vine losses of 37% (2011) and 46% (2013) culminated in roguing being replaced with whole block removal. In all five blocks, roguing protocols were standardised but in those with effective virus control, mealybug numbers were significantly lower in all years (mean: <0.2 per vine leaf; p≤0.036) relative to those where all vines were removed (mean: 0.4-2.3 per vine leaf). Overall, the results of this research suggest that rather than adopting a single management tactic in isolation, effective GLRaV-3 control instead requires an integrated plan to be implemented annually.