Objective measures for powdery mildew

Abstract

Powdery mildew compromises wine quality. Visual inspection, currently used to estimate disease severity, is subjective and prone to inaccuracy. Research was undertaken to improve visual assessment and to develop objective measures. A diagrammatic key and smartphone app, with supporting online training resource, were developed, trialled in Australia in 2015 and released globally in 2016. A DNA-based assay (qPCR) was developed for absolute quantification of powdery mildew fungus in grapes. Mid- and near-infrared spectroscopy proved insufficiently sensitive to discriminate levels of powdery mildew severity critical to the sector. Arachidic fatty acid was identified as a potential biomarker for quantifying powdery mildew on grapes.

Summary

Powdery mildew, caused by the fungus Erysiphe necator, occurs in most viticultural regions of Australia and worldwide, and causes loss of yield and quality if not adequately controlled. Many Australian wineries have thresholds for powdery mildew contamination, such that grapes that exceed 3-5% of the surface area affected (disease severity) may be rejected or downgraded. Assessment of disease severity is based on visual inspection in the vineyard and/or at the winery. Visual assessment or estimation of disease severity is subjective and prone to variation depending on the assessor and the conditions of assessment. Uncertainty about visual assessments can lead to confusion and disputes about quality and pricing, so objective, quantitative and reliable measures for powdery mildew are required to allow better-informed decisions. Preliminary research that preceded this project suggested that near infrared spectroscopy and DNA-based methods had potential to provide objective measurement of powdery mildew. The research reported in this document was undertaken to develop (i) tools to improve the accuracy and reliability of visual assessment and (ii) objective measures for powdery mildew.

The development of tools to improve in-field visual assessment is described in Chapter 5 of this report. The concepts, components and features of the tools were discussed and refined at a series of six meetings of the Project Steering Group from January 2013 to September 2016. The main outputs from this work are: a new diagrammatic key, a smartphone application (app) and a supporting online resource to provide training in field assessment of powdery mildew. The diagrammatic key features standard area diagrams of grape bunches with increments in shading to represent powdery mildew in increments of 2% in the range 2-12% severity. The app, PMapp, was made available in Australia in December 2015 and globally in November 2016 as a free download for Apple and Android devices. PMapp has four components; assessment data entry, image browser, self calibration test for visual assessment and the diagrammatic key with 2% increments. The assessment screen allows the user to enter the score for each bunch by tapping the button which represents the best-fit severity category. The emphasis is on categories at the low end of the severity scale. The date, time and location (latitude and longitude) are recorded. The user can review cumulative disease incidence and severity as they go and toggle between data for the row and patch being assessed. On completion of the patch assessment, results can be sent in a csv or xml format to a designated email account for analysis. The image browser contains computer-generated images of bunches of various configurations with pale blue shading to represent powdery mildew, comprising 25 severity categories. The self calibration function allows the user to test their skills in estimating area, presenting 10 or 20 images for assessment followed by results for accuracy and bias. The online resource was developed in response to requests from the Steering Group to provide training in recognising and assessing powdery mildew in the field to support PMapp, and is designed for pre-vintage training for new staff and refreshing skills of experienced assessors. This free-to-access resource (pmassessment.com.au) comprises; a best practice guide for assessing powdery mildew, an exercise to assist recognition of powdery mildew, an area assessment training tool and the diagrammatic key. The vineyard assessment guide is presented as a stepwise dichotomous key that allows the user to identify his/her training needs and link directly to PMapp, the disease recognition exercise and the area assessment training tool. It also provides guidance about sampling strategy for use when a winery protocol is not available. The disease recognition exercise features high definition photographs of bunches of white and black grapes, which can be enlarged, and the user is asked to identify which have powdery mildew symptoms outlined correctly. The area assessment training tool features the images used in PMapp and the user can choose to assess 20 images spanning 0.5-15% severity or 30 images spanning 0.5-90% severity, and provides the user with information about the accuracy, repeatability and speed with which they assessed area shaded to represent powdery mildew.

A specific, sensitive and reliable quantitative polymerase chain reason (qPCR) assay was developed for the absolute quantification of biomass of E. necator on grape berries, based on DNA content, to serve as a means of calibrating a spectroscopic or biochemical assay. The qPCR assay was applied to homogenates of Chardonnay, Riesling, Grenache and Pinot Noir grapes which had been manufactured, by mixing homogenates of healthy and fully-infected berries, to represent various degrees of powdery mildew severity (weight:weight), as well as to homogenates of composite samples (surface area affected) and individual bunches of selected visual disease severity. The qPCR data, expressed as pathogen coefficient (log of ratio of quantification cycle for V. vinifera and E. necator), were used to develop statistical models for quantifying powdery mildew severity in these four varieties and predicting the visual severity by assuming a spherical berry shape and taking into account the decrease in weight between healthy and infected berries. There was no obvious relationship between the amount of E. necator as measured by qPCR and disease severity as estimated by visual assessment of bunches and individual berries using a magnifying lamp and dissecting microscope, respectively. This can be attributed in part to variation in the density of the fungus (mycelium, spores) on the surface of infected berries, and further confirms the subjective nature of visual assessment.

Mid-infrared spectroscopy was applied to homogenised individual berries, individual bunches and composite bunches for which E. necator biomass had been determined by qPCR. Differences attributed to powdery mildew were observed in the spectral region 1800-1185 cm-1, indicative of a mixture of lipid moieties, amide I and II, protein carboxyls, nucleic acids and fatty acid esters. Soluble and insoluble proteins in the spectral region 1695-1300 cm–1 contributed to the separation of some individual bunches and berries of Chardonnay with diverse E. necator biomass measured by qPCR. Principal component analysis followed by partial least squares analyses identified the optimal number of PCs (factors) that covered relevant spectral information related to powdery mildew, leading to development of a good calibration model for Grenache (R2 = 0.83). This model needs to be validated in the future. However, the calibration models established for Chardonnay and Riesling demonstrated poor predictive performance in validation using separate test samples. The calibration model for Pinot Noir was informative only for samples from healthy grapes. MIR spectroscopy was therefore limited in terms of practical implementation. Near infrared spectroscopy was not informative at the levels of discrimination required and calibration models could not be obtained for Chardonnay, Riesling or Pinot Noir.

Fatty acid profiling of spores of E. necator and of powdery mildew-affected Chardonnay berries identified arachidic acid (C20:0) as a potential biomarker for powdery mildew. It was the predominant fatty acid in spores of E. necator but was not detected in Botrytis cinerea (botrytis bunch rot) or Plasmopara viticola (downy mildew). Arachidic acid concentration increased as powdery mildew severity increased from healthy through half-infected to fully-infected berries, and correctly classified 90% of healthy berries. Analysis of arachidic acid and three other medium-long-chain fatty acids (behenic, myristic and pentadecanoic) correctly classified 97% of healthy berries and 75% of half- and fully-infected berries. Fatty acid analysis offers a routine, accurate and potentially rapid means of measuring powdery mildew and can be conducted in a standard analytical laboratory with minimal sample preparation. Further research is required to determine if analysis of fatty acids will allow discrimination of powdery mildew severity levels with sufficient sensitivity.

Fatty acids are known to affect the sensory qualities of food and beverages and the increase in particular fatty acids with increasing disease severity observed here offers insights into the mechanism(s) by which powdery mildew may affect wine quality and, in particular, cause the oily/viscous mouth-feel identified in previous research. Research to identify the concentrations of arachidic and other fatty acids that compromise wine may lead to an objective measure for wine quality that can be adopted for routine use in the wine sector.

Wine Australia and the School of Agriculture, Food and Wine of the University of Adelaide provided financial and in-kind support. Collaborators include RW Emmett Horticultural Pathology Research, the Tasmanian Institute of Agriculture, Accolade Wines, Lemur Software, Arris Pty Ltd, the Australian Wine Research Institute, Wine TQ and Fraunhofer (Germany). This project benefited immensely from an engaged and enthusiastic Project Steering Group that comprised personnel of large and small wine sector companies, research organisations and Wine Australia.