APS Online Publications
Skip main navigation
Previous
Next
RESEARCHFree Access icon

The Change in Winter Wheat Response to Deoxynivalenol and Fusarium Head Blight Through Technological and Agronomic Progress

    Affiliations
    Authors and Affiliations
    • Ruoxi Xia1
    • Arthur W. Schaafsma1
    • Felicia Wu2
    • David C. Hooker1 †
    1. 1Department of Plant Agriculture, University of Guelph, Ridgetown Campus, Ridgetown, ON N0P 2C0, Canada
    2. 2Department of Agricultural, Food, and Resource Economics, Michigan State University, East Lansing, MI 48824, U.S.A.
    Published Online:1 Mar 2021https://doi.org/10.1094/PDIS-06-20-1310-RE

    Abstract

    Fusarium head blight (FHB) in wheat causes yield loss, quality reduction, and mycotoxin contamination in temperate wheat production areas worldwide. The objective of this study was to quantify the progress of agronomic and FHB management strategies over the past two decades in FHB suppression and agronomic performance of winter wheat in environments favorable for FHB. Field experiments were conducted in environments typical of FHB epidemics to compare common agronomic and FHB management practices used in the 1996 era compared with those used in 2016. The experiments included a comparison of three different nitrogen (N) fertilizer application rates and six old (1996-era) and new (modern-era) winter wheat cultivars representing combinations of susceptibility and era to FHB, with and without a fungicide applied at flowering (pydiflumetofen + propiconazole). To mimic environments favorable for infection (similar to 1996 in Ontario, Canada), plots were challenged at 50% anthesis with F. graminearum macroconidia suspension followed by mist irrigation. The modern management strategy of using moderately resistant cultivars, a fungicide applied at flowering, and a high rate of N fertilizer reduced total deoxynivalenol by 67%, reduced Fusarium-damaged kernels by 49%, reduced FHB index by 86%, increased grain test weight by 11%, and increased grain yield by 31% compared with the standard management practice of seeding highly susceptible cultivars with no fungicide and a lower rate of N fertilizer recommended in the 1996 era. This study enabled a published economic assessment of the return on investment for the improvements in cultivars, fungicide, and N fertilizer applications since 1996.

    Fusarium head blight (FHB) in wheat, caused mainly by Fusarium graminearum (Schwabe) in North America, has caused yield loss (McMullen et al. 1997; Paul et al. 2010), grade reductions (Dexter et al. 1996; McMullen et al. 1997; Schaafsma et al. 2001), and most importantly, mycotoxin-contaminated grain (Champeil et al. 2004; Hooker et al. 2002; Miller et al. 2014). Southern Ontario, among other North American regions, experienced a severe FHB epidemic in 1996; effective management strategies were not available when conditions were highly favorable for fungal infection and disease development for winter wheat. It was estimated that the epidemic caused more than CAD$100 million loss to wheat growers in the province of Ontario, through yield losses and grade reductions (Schaafsma 2002). Since 1996, epidemics have occurred in at least localized regions across Ontario (Schaafsma 1999; Tamburic-Ilincic et al. 2011; Canadian Grain Commission 2019a).

    It has been postulated that since 1996, technological innovations have reduced the impact of FHB epidemics, despite frequent occurrences of environmental conditions that favor epidemics (McMullen et al. 2012). Currently, the most effective management strategy to reduce the risk of FHB and the mycotoxin deoxynivalenol (DON) is to plant modern wheat cultivars with some resistance to FHB and to apply an effective fungicide (Gilbert and Haber 2013; Paul et al. 2019).

    The occurrence of FHB concerns more than just DON contamination. Fusarium-damaged kernels (FDKs) are usually bleached, shriveled, and less dense than noninfected kernels. This leads to lower test weight and protein content, which results in lower grain grade (Canadian Grain Commission 2019b; Dill-Macky 2010). Up to 30% yield loss has been reported due to floret sterility and poor seed filling in infected wheat heads (Dill-Macky 2010). Nevertheless, the major concern of FHB is the accumulation of mycotoxins in harvested grain. The major mycotoxins produced by F. graminearum include DON, zearalenone, and DON derivatives (Limay-Rios and Schaafsma 2018). DON, the predominant mycotoxin, can cause vomiting, diarrhea, weight loss, and immunosuppression if animals consume excessive amounts (da Rocha et al. 2014; Pestka 2003).

    F. graminearum overwinters as mycelia or spores on crop residues (Schmale and Bergstrom 2010). Mycelia develop into perithecia, the sexual reproduction structure, with moderate temperature, rain events, or high relative humidity (RH) and the presence of light (Champeil et al. 2004). Once perithecia mature, rain events or high RH trigger the primary inoculum, ascospores, to be released from perithecia (Trail et al. 2002). Ascospores carried by wind or rain-splash land on wheat spikes (McMullen et al. 1997). Wheat heads are most susceptible to infection during anthesis (Zadoks growth stage [ZGS] 60 to 69; Zadoks et al. 1974) (Champeil et al. 2004).

    Genetic resistance against FHB is a quantitative trait involving complex mechanisms that can be categorized into three main types (Bai and Shaner 2004; Mesterházy 1995; Miller et al. 1985; Schroeder and Christensen 1963). Among them, the major resistance mechanisms used in current commercial cultivars are type I, resistance to initial infection, and type II, resistance to spread of the disease within the spike by the formation of a protective layer between infected and healthy tissue (Schroeder and Christensen 1963). Type II resistance in wheat has been found to be more stable and less affected by nongenetic factors than type I resistance (Bai and Shaner 1994). Resistance in cultivars currently in North America was obtained mostly from two sources: the Chinese cultivar Sumai 3 (Bai and Shaner 2004; Steiner et al. 2017) and native cultivars possessing some resistance (Benson et al. 2012; Castro Aviles et al. 2020). The latter has been used more extensively in breeding programs (Islam et al. 2016) or combined with resistance quantitative trait locus (QTL) Fhb1 from Sumai 3 (Bai et al. 2018).

    One of the challenges in FHB resistance breeding is linkage drag, by which undesirable yield or quality traits are introduced when resistance QTLs are incorporated (Salameh et al. 2011). Linkage drag has been cited as a reason for the lack of cultivars that are both high-yielding and resistant to FHB. More recently, molecular methods such as marker-assisted backcrossing aid in introducing FHB-resistant QTLs into cultivars while maintaining grain yield, kernel weight, test weight, and protein content (Salameh et al. 2011). Because of the quantitative nature of FHB resistance, the highest FHB resistance level achieved in commercial cultivars is moderately resistant (MR) (OCCC 2018). In Ontario, between 1999 and 2015, the proportion of available cultivars that were highly susceptible (HS) dropped from 40% to zero, and the proportion of moderately susceptible (MS) and MR cultivars increased from 28 to 56% (OCCC 2018). Genetic resistance has been proven to be an effective tactic in FHB and DON suppression, with some remaining challenges and room to improve (Salameh et al. 2011).

    A fungicide application at 50% anthesis (ZGS 65) is another key FHB management tool developed in the past 2 decades. The fungicide products currently available for reducing DON include Folicur (tebuconazole), Proline (prothioconazole), Prosaro (prothioconazole and tebuconazole) from Bayer CropScience (2016a, b, c, d), and Caramba (metconazole) from BASF (2015). These fungicides are all triazoles, belonging to group 3, the demethylation inhibitors (DMIs) that inhibit biosynthesis of sterols in fungi (FRAC 2018). More recently, a new fungicide formulation has been developed and marketed as Miravis Ace (Health Canada 2019; Syngenta 2018), which contains the active ingredient pydiflumetofen (trademark ADEPIDYN), a carboxamide (group 7; FRAC 2018), plus propiconazole. Pydiflumetofen is a succinate dehydrogenase inhibitor that inhibits cellular respiration of fungi (FRAC 2018). Propiconazole is a DMI, belonging to group 3 (FRAC 2018).

    According to a meta-analysis of fungicide trials conducted by the U.S. Wheat and Barley Scab Initiative, all commercial triazole fungicides suppress FHB and reduce DON concentration (Paul et al. 2005a) and increase yield and test weight when applied according to their labels. In the meta-analysis, Prosaro, Caramba, Proline, and Folicur reduced the FHB index (FHBI) by 40 to 52%, and most of them reduced DON by 42 to 45%, except that Folicur reduced DON by 23% (Paul et al. 2008). In a similar meta-analysis excluding Folicur, the average grain yield increase across fungicides was 14%, and test weight increased by 2.6% (Paul et al. 2010). In other experiments, fungicides were most effective in reducing FHB and DON, with better coverage of wheat heads. Hooker et al. (2005) used Turbo TeeJet nozzles arranged backward–forward pointing, or alternating Turbo FloodJet nozzles for best spray coverage.

    It is suspected that much progress has been made in managing FHB and DON in winter wheat since the 1996 epidemic. The current recommendation is an integrated strategy to manage FHB (Beres et al. 2018; Paul et al. 2018), but no strategy achieves complete control of FHB and elimination of DON. Genetic resistance against FHB is still improving; however, high yields and quality challenges remain an obstacle because undesirable yield and quality traits are usually introduced along with FHB resistance (Bai and Shaner 2004; Salameh et al. 2011). Moreover, because triazole fungicides have the same mode of action, there is concern about the pathogen developing resistance (FRAC 2018; Salgado et al. 2018) or about DMI fungicides declining in efficacy (Salgado et al. 2018). The pydiflumetofen-containing product introduces a different mode of action that may reduce some of these concerns, especially because preliminary work has shown comparable levels of FHB and DON reduction compared with the older fungicides (Salgado et al. 2018). Although progress has been made in FHB management since 1996, no direct comparison has been made for FHB suppression and grain productivity between practices used in the 1996 era and those available in the modern era. Therefore, the primary objective of the present study was to quantify the progress of FHB management strategies for reducing FHB symptoms and DON accumulation and increasing agronomic performance by using one or two major winter wheat cultivars from the 1996 era compared with four major cultivars from the modern era in environments favorable for FHB.

    Materials and Methods

    Experimental design.

    The study was conducted over 2 years on separate fields at Ridgetown, Ontario, Canada in 2016–2017 and 2017–2018. Plots were inoculated with F. graminearum and mist-irrigated to favor FHB infection and development, to mimic environments that produced similar DON levels in typical epidemics in Ontario, such as the one that occurred in 1996. In each experiment, treatments were arranged in a split-plot design with three levels of nitrogen (N) as the whole plot and 12 subplots consisting of combinations of six cultivars and two fungicide levels in each replication. Plot size was 4 m in length by 1.15 m in width. There were four replications. Different N rates were separated by a guard plot or a 1-m-wide path. Levels of all factors were selected as an attempt to represent and compare typical FHB and agronomic management in the eras of 1996 (referred to as the old practice) and 2016 (referred to as the modern practice). The winter wheat cultivars and their susceptibility to FHB are presented in Table 1. The susceptibility ratings of all cultivars were sourced from the Ontario Cereal Crop Committee (www.gocereals.ca). Given the spatial size of the experiment under a misting system and the availability of seed, it is important to note that only a few cultivars were chosen and included in this study to represent each era and susceptibility to FHB. For the 1996 era, the cultivar ‘AC Ron’ was preferred because it was the most popular soft white wheat cultivar grown in Ontario in 1996, mainly for its high yield potential (Teich et al. 1992), as evidenced in field surveys (Schaafsma et al. 2001); however, within a few years after 1996, it was discontinued because of its high susceptibility to FHB. The archived seed of ‘AC Ron’ was in short supply at the start of this experiment, so we multiplied ‘AC Ron’ during the first year of the study elsewhere to be used in the second year of the experiment; the cultivar ‘AC Mountain’ was used in both years of the study, which was comparable to ‘AC Ron’ in both yield and susceptibility to FHB in provincial tests (OCCC 2001). The modern cultivars used in this study were divided into two susceptibility groups: ones that were susceptible (S) to FHB and DON accumulation (Pioneer 25R40, and Pioneer 25W31) and ones that were MR (‘Ava’ and Pioneer 25R46) (OCCC 2016). Cultivars 25R40 and 25W31 were used by the Ontario Cereal Crop Committee as FHB-susceptible checks for the registration of new cultivars. These four cultivars were chosen because they were produced on most of the winter wheat acres during the modern era across Ontario. In the modern era, MR cultivars are available to growers; similar cultivars were not available in 1996. Overall, popular cultivars were chosen to estimate the progress of plant breeding on productivity, FHB, and DON, across three era–FHB susceptibility classes (1996 S, 2016 S, and 2016 MR). Of course, more cultivars to represent each era would have been better, but the size of the experiment under a misting system was at a maximum with six cultivars, two fungicide treatments, three N rates, and four replications.

    Table 1. Factors included in experimental design, representing Fusarium head blight and agronomic management practices in the 1996 era and 2016 era. All combinations of factors are represented, with the exception of the 0 kg N/ha control.

    Nitrogen fertilizer recommendations have also changed from 1996 to the modern era; therefore, the N rate of 100 kg N/ha represented the 1996 standard practice, and 170 kg N/ha represented the modern practice. The recommended N rates in Ontario changed from 1996 to 2016 with evidence that modern cultivars respond more to N when a fungicide is applied at ZGS 65 (Brinkman et al. 2014). Some studies indicate that low amounts of N fertilizer (≤80 kg N/ha) increased FHB severity and DON compared with zero N fertilizer, but the majority of N studies indicate that the rate of N fertilizer is not a management factor for FHB (Lemmens et al. 2004; Schaafsma et al. 2001); nevertheless, N rate was investigated as a factor affecting mainly agronomic performance. A zero N check treatment was included mainly to examine the wheat responsiveness to N on the field sites.

    Fungicide treatments included no fungicide (untreated), which represented the normal practice in the 1996 era when no fungicides were used or available, and the pydiflumetofen-containing product Miravis Ace (150 g a.i./liter of pydiflumetofen + 125 g a.i./liter of propiconazole) was applied according to the label rate of 1 liter of product/ha at ZGS 65 (dates and weather conditions at time of spray are presented in Table 2). Miravis Ace was a new fungicide in the 2016 era with an efficacy similar to other fungicides used for suppressing FHB and reducing DON (Salgado et al. 2018). The fungicide was applied with a CO2 backpack sprayer equipped with two nozzle assemblies spaced 50 cm apart. Each nozzle assembly consisted of two Turbo TeeJet TT11001 nozzles (TeeJet Technologies, Springfield, IL) on a double-swivel Quick TeeJet nozzle body (QJ8600-2-1/4-NYB, TeeJet Technologies); one nozzle in the nozzle body was adjusted to point forward in the direction of travel and angled 45° down from horizontal, and the other nozzle was adjusted to point backward and angled 45° down from horizontal. The nozzles were positioned approximately 35 cm above the canopy of each cultivar during spraying. The sprayer was calibrated to deliver 200 liters/ha with 240 kPa pressure at a traveling speed of 3.6 km/h.

    Table 2. Weather conditions at all fungicide application timings

    Plot management.

    In both years, the experimental areas were in a crop rotation of winter wheat–corn–soybean, which is a typical crop rotation in southern Ontario (Brinkman et al. 2014). The fertilizer 6-24-24 was broadcast at a rate of 336 kg/ha and incorporated before planting, according to soil test recommendations. The 2016–2017 and 2017–2018 sites were planted on 24 October 2016 and 26 October 2017, respectively, with a Wintersteiger TRIM planter (Laval, QC, Canada) equipped with six row units spaced 19 cm apart. The target seeding rate was 450 seeds/m2. Seed for individual plots were counted with Seedburo 801 Count-A-Pak (Seedburo Equipment Company, Des Plaines, IL). N fertilizer treatments were split-applied on the 170 kg N/ha treatments at rates of 100 and 70 kg N/ha, with calcium ammonium nitrate (27-0-0), on 21 April and 9 May 2017 and 19 April and 15 May 2018. Sulfur (11 kg/ha) was broadcast with a spinner spreader to all plots in the form of calcium sulfate (gypsum; 0-0-0 to 17% S) on 10 May 2017 and 15 May 2018. Broadleaf weeds were controlled with Refine SG herbicide (33.35% thifensulfuron methyl and 16.65% tribenuron methyl; FMC Corporation, Philadelphia, PA), applied on 3 May 2017, and with Refine SG again tankmixed with Buctril M herbicide (280 g/liter bromoxynil and 280 g a.i./liter 2-methyl-4-chlorophenoxyacetic acid) (Bayer CropScience, Guelph, ON, Canada) on 23 May 2018.

    Foliar fungicide application, inoculation, and mist irrigation.

    In 2018, all plots were sprayed with 0.3 liters/ha of the foliar fungicide Headline EC (250 g/liter pyraclostrobin; BASF, Mississauga, ON, Canada) at ZGS 39 (flag leaf fully emerged) for controlling stripe rust (Puccinia striiformis f. sp. tritici) that would otherwise probably confound the experimental objectives. The fungicide was applied with a CO2 sprayer equipped with two air bubble jet nozzles, NNB003002 (BFS, Brandon, MB, Canada) spaced 50 cm apart. It was calibrated to deliver 200 liters/ha with 240 kPa pressure. In 2017, no fungicide was applied at ZGS 39 (flag leaf fully emerged) because the potential for foliar disease was low.

    The primary goal of the inoculated-misting procedure used in this study was to modify an environment to produce FHB and DON effects similar to what occurred in typical epidemics in Ontario; the goal was achieved because the DON values in 1996-era cultivars were similar to those from farms in 1996 (Schaafsma et al. 2001). In both years of the current study, plots were inoculated with F. graminearum from two sources: infected corn stalks collected from a commercial corn field, consisting of five infected stalks scattered on the ground in each plot around ZGS 31, and a spore suspension prepared in liquid culture according to Hamilton et al. (1997). Briefly, the spore suspension was a mixture of multiple isolates containing both 3ADON and 15ADON chemotypes. Suspensions of each isolate were filtered separately with cheesecloth. Before inoculation, suspensions of all the isolates were mixed, and spore concentration adjusted to 5 × 104 macroconidia/ml with a hemocytometer. In 2017, five isolates, DAOM 251952, 251953, 251954, 251955, and 251956 (Canadian National Mycological Herbarium [DAOM], Ottawa, ON), representing both 3ADON (DAOM 251952, 251953, and 251956) and 15ADON (DAOM 251954 and 251955), chemotypes were used. Each plot was spray-inoculated twice with 225 ml of the spore suspension (Tamburic-Ilincic et al. 2011). The first spray inoculation was within 24 h after the fungicide application at ZGS 65, and the second was 48 h after the first. In 2017, plots were mist-irrigated with an overhead mist irrigation system every 8 min from 10:00 to 17:00 in 1-min durations, delivering about 7.5 mm water per day, from 2 June (after the first spray inoculation) to 26 June (Tamburic-Ilincic et al. 2007). In 2018, plots were spray-inoculated with only three F. graminearum isolates, DAOM 251953 (3ADON), 251955 (15ADON), and 251956 (3ADON), because DAOM 251952 and 251954 lacked vigor in producing macroconidia. Also in 2018, the mist irrigation system was turned on from 4 June, after the first inoculation, to 14 June, when first FHB symptom appeared.

    FHB and agronomic wheat performance assessments.

    Because the cultivars used in this study were the ones most widely grown in their respective eras, we can provide a reasonable estimate of performance in the old era and the modern era. In both years, FHB incidence and severity assessments were made 21 ± 2 days after the fungicide treatment at ZGS 65 from 20 heads per plot, as described by Stack and McMullen (1998). In each plot, the FHB incidence and severity estimates were used to calculate FHBI (Paul et al. 2005a). Wheat agronomic performance was assessed by three measurements: grain yield; test weight, which is one indicator of grain quality; and the normalized difference vegetation index (NDVI), which reflects the photosynthetic potential or green leaf area during grain filling (Thomas and Howarth 2000; Thomas and Smart 1993). The NDVI values range from 0 to 1, with the higher value indicating a denser and greener canopy. NDVI has been used extensively to assess crop condition and yield (Mkhabela et al. 2011). In both years, NDVI was measured with a handheld GreenSeeker optical sensor (Trimble Inc., Sunnyvale, CA) at approximately ZGS 61 to 65 to indicate the amount of green in the canopy of the various cultivar and N rate combinations. In 2018, at the onset of canopy senescence, differences in canopy greenness were noted between plots, and an additional measurement of NDVI was taken at approximately ZGS 85. The change of NDVI values from ZGS 65 to ZGS 85 (ΔNDVI) was calculated to quantify the level of canopy senesce to the end of the grain filling stage close to physiological maturity.

    Plots were harvested on 20 July in both years with a small plot combine. Grain moisture was measured with a moisture meter. Grain yields were adjusted to Mg/ha at 14% moisture. Test weight was measured with a standard 0.5-liter test weight cup and a filling hopper approved by the Canadian Grain Commission. The percentage of FDKs (wt/wt %) was estimated with near-infrared spectroscopy according to the method developed by Limay-Rios et al. (2012). The FDK values reported were an average of three subsamples.

    Mycotoxin analysis.

    To collect a representative subsample for mycotoxin analysis, 500 g of grain was subsampled from each plot. This sample was coarsely ground (10-mesh sieves) through a Romer grinding mill (Romer Labs Inc., Newark, DE) and divided into three more subsamples of 150 to 175 g each. One of these was finely ground with a Stein Laboratory Mill (The Steinlite Corporation, Atchison, KS) until 90% of the flour material passed through a 20-mesh sieve. Then 40 ml of mycotoxin extraction solvent (79% acetonitrile, 20% water, and 1% acetic acid) was added to 10 g of the flour sample and shaken on an Eberbach shaker for 90 min. The mixture was centrifuged at 3,000 rpm for 3 min. Then 200 µl of supernatant was transferred into a test tube and diluted with 1.8 ml of mycotoxin extraction solvent. Then 200 µl of the diluted sample was transferred to a new test tube, together with 100 µl of internal standard dilution solvent. The mixture was dried under a nitrogen stream and reconstituted with 400 µl of mycotoxin dilution solvent (45.5% H2O, 53.5% MeOH, 1% acetic acid, and 5 mM ammonium acetate). The 400-µl sample was transferred to vials with inserts and stored at 4°C until analyzed (Limay-Rios and Schaafsma 2018).

    The procedure used for mycotoxin liquid chromatography–tandem mass spectrometry quantification was as described by Limay-Rios and Schaafsma (2018) with minor modifications. In 2017, the limits of detection (LODs) for DON and DON derivatives deoxynivalenol-3-glucoside (D3G), 15-acetyl-deoxynivalenol (15ADON), and 3-acetyl-deoxynivalenol (3ADON) were 23, 6, 76, and 12 ppb, respectively. In 2018, the LODs were 3, 0.03, 2, and 4 ppb, respectively. The LODs were lower in the second year because of improvements in analysis and quantitation methods (Dr. V. Limay-Rios, University of Guelph, personal communication). Total DON concentration (TDON) was calculated as the sum of DON, D3G, 15ADON, and 3ADON.

    Statistical analysis.

    Data analyses were conducted in SAS version 9.4 (SAS Institute, Cary, NC). Data from the 2017 and 2018 agronomy experiments were pooled across years, because the experiments followed the same protocol, year × treatment interactions were not statistically significant (P > 0.05), and the effect variances between years were homogeneous (Bowley 2015). The PROC UNIVARIATE procedure was used to test the assumptions of normality of the residuals. Outliers were identified by distribution and probability plot for studentized residuals. If the absolute value of a studentized residual was >3.4, the data point was identified as a putative outlier and removed if it did not affect the results (Bowley 2015).

    PROC GLIMMIX was used to conduct variance analysis for a split-plot design pooled across years. The cultivars in the experiments were grouped into three cultivar susceptibility–era combinations (HS, highly susceptible; S/MS, susceptible or moderately susceptible; MR, moderately resistant according to Ontario Cereal Crop Committee ratings from 2001 and 2018): 1996-HS, ‘AC Ron’ and ‘AC Mountain’; 2016-S/MS, 25R40 and 25W31; and 2016-MR, 25R46 and ‘Ava’ (Table 1). Fixed effects included the three cultivar susceptibility–era combinations, two or three N rates, and two fungicide treatments at ZGS 65; random effects included year, replication block nested within year, and N × replication block nested within year to reflect the split-plot design.

    The responsiveness of wheat to N, measured by FHB (TDON, FDKs, FHBI) and wheat performance (yield, test weight, and NDVI at ZGS 65), was subjected to variance analysis. Least squares means were separated via Tukey’s test (P = 0.05). The impact of N deficiency on FHB and wheat performance was investigated by comparing the least squares means of 0 N with those of 100 and 170 kg N/ha. If the least squares means of 0 N were significantly different from those of 100 and 170 kg N/ha, wheat in this study was confirmed to be responsive to fertilizer N applied.

    After the initial variance analysis, the 0-N control treatments were removed from the dataset, so the changes in agronomic and FHB management practices between the eras of 1996 and 2016 could be compared independently from the 0-N treatment. The statistical significance of the fixed effects was tested on TDON, percentage FDKs, FHBI, yield, test weight, and NDVI at ZGS 65 and 85 and the ΔNDVI from the two growth stages. Least squares means were separated via Tukey’s test. The level of statistical significance used for all analyses was P = 0.05.

    Results

    The addition of 100 kg N/ha of fertilizer increased grain yield from 2.09 to 2.73 Mg/ha compared with the 0 N, depending on the cultivar susceptibility–era combination. The inclusion of a 0-N control treatment showed that both site-years were responsive to N fertilizer (Supplementary Table S1), which is the expectation on farm fields. Inclusion of the 0-N treatment enables some inferences about wheat agronomic performance, FHB, and DON in N-deficient systems (i.e., organic), but any discussion of these effects is beyond the scope of this article. Nonetheless, when the 0 N rate was included, cultivar era interacted with N rate on TDON content, grain yield, and canopy NDVI at ZGS 65 and 85, whereas fungicide interacted with N rate on grain yield, test weight, and canopy NDVI at ZGS 85 (Supplementary Table S1). These interactions were caused primarily by differences between 0 N and the two N rates (100 and 170 kg N/ha) used in the study. Because the focus of this study was not on the effect of 0 N, differences between treatments with 0 N (or in N-deficient environments) are presented in Supplementary Tables S2 and S3.

    Treatment effects on DON concentration.

    Variance analyses for addressing the main objectives of this study were performed on a reduced dataset that excluded the 0-N rate data. The TDON, FDKs, and FHBI were affected by the three cultivar susceptibility–era combinations and the fungicide treatment at ZGS 65 (P < 0.0001; Table 3); however, the magnitude of TDON content and FHBI response to cultivar susceptibility–era combinations depended on whether a fungicide was applied at ZGS 65, which produced cultivar susceptibility–era combination × fungicide interactions (P = 0.0100 for TDON, P = 0.0344 for FHBI; Table 3). The application of a fungicide at flowering on old HS cultivars was more effective at reducing the concentration of TDON than the same application on modern MR cultivars. A fungicide applied at flowering reduced TDON content from 19.0 to 10.6 ppm (reduction of 8.4 ppm, or 44%; P < 0.0001) in HS old cultivars and from 10.2 to 6.3 ppm (reduction of 3.9 ppm; P = 0.0004) in modern MR cultivars (Table 4). When compared with the practice of using old HS cultivars and no fungicide, the application of a fungicide on MR cultivars was most effective at reducing TDON content, from 19.0 to 6.3 ppm (reduction of 12.7 ppm, or 67%; Table 4). Interestingly, modern S/MS cultivars had similar TDON content compared with old HS cultivars with and without a fungicide.

    Table 3. Fixed effects in mixed model variance analysis of the effects of cultivar susceptibility–era, N rate, and fungicide to total DON concentration, FDKs, and FHBI

    Table 4. Least squares means of total DON concentration, FDKs, and FHBI as affected by cultivar susceptibility–era and fungicide interactiont

    Treatment effects on FDKs.

    The percentage FDKs in the grain was affected by cultivar susceptibility–era category, fungicide, and N treatments; no interactions were detected (P > 0.2930; Table 3). When averaged across the three cultivar susceptibility–era combinations, the application of a fungicide reduced FDKs from 7.9 to 5.1% (P < 0.0001; Tables 3 and 4). When averaged across the two levels of the fungicide treatment, FDKs were the highest among old HS and modern S/MS cultivars (7 to 7.1%) compared with those in modern MR cultivars (5.4%; Table 4). The use of modern MR cultivars combined with a fungicide application at flowering reduced FDKs from 8.3 to 4.2% compared with old HS cultivars with no fungicide (Table 4). An increase in N rate from 100 to 170 kg N/ha decreased FDKs statistically from 6.8 to 6.1% (P = 0.0361) across the three cultivar susceptibility–era combinations and fungicide treatments (data not shown), but the difference caused by N rate was small compared with the effect of cultivar susceptibility–era combination and fungicide on FDKs.

    Treatment effects on FHBI.

    The use of a fungicide applied at flowering on old HS cultivars reduced FHBI from 17.8 to 4.6%, a reduction of 74% (P < 0.0001; Table 4). On modern MR cultivars, fungicide application reduced FHBI from 5.8 to 1.5%. In other words, the use of a fungicide at flowering on modern MR cultivars reduced FHBI from 17.8 to 1.5% compared with the old HS cultivar with no fungicide (Table 4). The FHBI of modern S/MS cultivars, with and without fungicide, was 7.3 and 1.5%, respectively, similar to the modern MR cultivars (Table 4). The reduction of FHBI was similar proportionally between old HS and modern S/MS cultivars (74 to 80%), in contrast to the reduction in modern MR cultivars (53%; Table 4).

    Treatment effects on wheat agronomic performance.

    Grain yield was affected by cultivar susceptibility–era combination and fungicide treatments (P < 0.0001; Table 5). The grain yield response to the fungicide was not consistent with N rate, so the N rate by fungicide interaction was sliced across N rate (Table 6; Littell et al. 2006). Across the three cultivar susceptibility–era combinations, grain yield responded to an increase in fertilizer N from 100 to 170 kg N/ha but only when a fungicide was applied at flowering; the yield increased by 0.48 Mg/ha, or 8.9% (P = 0.0240; Table 6) as the N rate increased from 100 to 170 kg N/ha. The yield response to fungicide varied across N rates. At the 100 kg N/ha N rate, fungicide applied at anthesis increased yield by 10 to 18%, depending on the cultivar susceptibility–era combination, or 13% (0.63 Mg/ha) across the three cultivar susceptibility–era combinations (P < 0.0001). At the 170 kg N/ha N rate, fungicide increased yield by 19 to 26% depending on the cultivar susceptibility–era combination, or 21% (0.99 Mg/ha) across cultivar susceptibility–era combinations (P < 0.0001; Tables 6 and 7). It is noteworthy that old HS cultivars and modern MR cultivars had similar yields, both with (5.21 to 5.39 Mg/ha) and without fungicide (4.28 to 4.67 Mg/ha). In contrast, modern S/MS cultivars yielded 5.33 Mg/ha without and 6.12 Mg/ha with a fungicide; S/MS cultivars yielded more than either old HS or modern MR cultivars regardless of whether a fungicide was applied (Table 7).

    Table 5. Variance analysis on the effects of cultivar susceptibility–era, N rate (100 versus 170 kg/ha), and fungicide on grain yield (Mg/ha), test weight (kg/hl), and canopy NDVI in 2017 and 2018

    Table 6. Least squares means of yield, test weight, canopy NDVI at ZGS 65 and ZGS 85, and their difference as affected by N and fungicide interaction

    Table 7. Least squares means of grain yield, test weight, canopy NDVI at ZGS 65 and ZGS 85, and their difference as affected by cultivar susceptibility–era and fungicide interaction

    Test weight was affected by the simple effects of cultivar susceptibility–era combination and fungicide treatment (P < 0.0001), with no interaction at P = 0.05 (Table 5). The application of a fungicide at flowering increased the average test weight from 72.0 to 75.6 kg/hl (P < 0.0001) across cultivar susceptibility–era combinations (Table 6). When a modern MR cultivar was used with a fungicide, test weight increased 7.7 kg/hl compared with the old HS cultivars and no fungicide (Table 7; P < 0.0001, contrast not shown). Modern S/MS cultivars had test weights of 76.4 and 72.8 kg/hl, with and without fungicide, respectively, both equivalent to that observed in modern MR cultivars (76.9 vs. 73.9 kg/hl) (Table 7).

    The progression of canopy senescence during grain filling was estimated by the decrease in NDVI (ΔNDVI) between full anthesis (ZGS 65) and soft dough (ZGS 85). The ΔNDVI was affected by cultivar susceptibility–era combination (P < 0.0001), fungicide (P < 0.0001), and N rate (P = 0.0017; Table 5). However, the ΔNDVI response to cultivar susceptibility–era combination depended on whether a fungicide was applied, which produced a cultivar susceptibility–era combination × fungicide interaction (P = 0.0155; Table 5). Within old HS cultivars, a fungicide application at flowering reduced the ΔNDVI from 0.32 to 0.21 (P < 0.0001; Table 7), which indicates that the fungicide may have caused the wheat canopy to stay green longer into the grain fill phase. Similarly, the canopies of modern MR cultivars tended to stay green longer than old HS cultivars, as the ΔNDVI was reduced from 0.32 in old HS cultivars to 0.19 in modern MR cultivars (P < 0.0001; Table 7). When the fungicide was applied to modern MR cultivars, the ΔNDVI from flowering to soft dough stage was reduced from 0.32 to 0.13 (Table 7). The modern S/MS cultivars had similar ΔNDVI as MR cultivars from the same era.

    The effect of N rate on canopy senescence (magnitude of ΔNDVI response) depended on cultivar susceptibility–era combination and produced a cultivar susceptibility–era combination × N rate interaction (P = 0.0109; Table 5). The average ΔNDVI in old HS cultivars decreased from 0.29 to 0.23 as the N rate increased from 100 to 170 kg N/ha (P < 0.0001 Table 7), implying that the higher N rate delayed canopy senescence. Similarly, in modern MR cultivars, the increase in N rate reduced ΔNDVI from 0.19 to 0.13 (P < 0.0001; Table 7). In this case, the ΔNDVI indicates that the N rate effect was more pronounced in modern MR and S/MS cultivars, which probably caused the N rate × fungicide interaction (Table 5).

    Discussion

    In the Great Lakes Region, which includes southern Ontario, Canada, three major improvements in FHB and agronomic management have been made in winter wheat since the FHB epidemic in 1996: Genetic resistance to FHB in commercial cultivars has increased, most growers now apply fungicides at ZGS 65 to protect wheat against FHB (and leaf diseases), and application rates of N have increased from 100 to 170 kg N/ha. The purpose of this study was to measure the impact of these improvements on FHB, DON content, and wheat agronomic performance in environments typical of FHB epidemics and to provide the background and framework for estimating the economic impact of these improvements (see Xia et al. 2020). Our goal to simulate favorable environments for FHB was achieved because DON concentrations and FDKs in this study were similar to those reported across farms in 1996 (Schaafsma et al. 2001).

    To our knowledge, the present study is the first to compare the progression of common agronomic and FHB management practices from an old (1996) era to a modern era in side-by-side trials with uniform environments favorable for FHB infection. However, the results should be interpreted with caution, because one or two cultivars represented each FHB susceptibility level within era, at least a portion of the test weight and grain yield responses to fungicide application at flowering were probably physiological or caused by foliar disease control, which probably differed between the cultivars, nitrogen and fungicide responses may have been altered by the misting system, and the study involved only two site-years or environments, which interacted with grain yield and test weight responses. Cultivar resistance to both FHB and DON accumulation improved from old to modern eras, and when combined with a modern fungicide applied at flowering, the two management strategies had an additive effect on reducing FHB and DON content and increasing agronomic performance under environments favorable for FHB development and DON accumulation. A modern fungicide applied at flowering, along with higher N rates, resulted in higher grain yield and test weight in the common modern era practice, compared with the old era practice that consisted typically of lower N and no fungicide. Higher agronomic responses may be associated with improved FHB control, but they are also associated or confounded with physiological responses in the absence of disease or the control of fungal leaf diseases provided by the fungicide (Parry et al. 1995; Paul et al. 2010). Modern MR cultivars resulted in 44% lower total DON accumulation and 22% lower FDKs compared with old HS cultivars, whereas modern S/MS cultivars still had similar total DON content and FDKs as old HS cultivars. Our results agreed with those of Tamburic-Ilincic et al. (2011), who found that under a mist-irrigated environment, the average DON concentration in MR cultivars was 9.2 ppm, or 50% lower than that of HS and S cultivars combined (18.4 ppm). McMullen et al. (2008) also reported that DON concentrations in MR cultivars under natural infection were reduced to 0.6 ppm compared with 1.2 ppm in S/HS cultivars. The difference in DON concentrations between studies was probably caused by the environment and the presence of the pathogen (i.e., natural vs. our inoculated, misted). Inconsistent results have been found previously during FHB resistance testing because of genotype and environment interactions and heterogeneous sources of resistance parents (Kolb et al. 2001). However, the present study showed that MR cultivars can reduce DON concentrations by at least 45% under environments highly favorable to FHB and DON accumulation.

    An unexpected observation from this study was that modern cultivars, regardless of FHB susceptibility, had 60 to 70% lower FHB visual symptoms than old HS cultivars. The reason for this observation may be an error related to the awned versus awnless trait (the presence or absence of awns) in the cultivars. The modern S/MS cultivars (25R40, 25W31) are both awned, whereas the old HS cultivars (‘AC Ron,’ ‘AC Mountain’) are awnless. Campbell and Lipps (1998) described that awned cultivars may obscure visible FHB on spikelets, or awns may provide physical protection from infection. The latter may be categorized as passive type I resistance, which is the resistance to initial infection (Champeil et al. 2004; Schroeder and Christensen 1963). Despite having lower FHBI, the modern S/MS cultivars used in this study still had TDON concentrations similar to those of the old HS cultivars. Considering the significance of mycotoxins to the wheat-related industry, the finding in the present study suggests room for improvement in other types of resistance, including: active type I, which is the resistance to initial infection through pathogenesis-related proteins (Champeil et al. 2004); type II, which is the resistance to propagation of the pathogenic agent in the tissues (Schroeder and Christensen 1963); and type III, the resistance to DON accumulation by degrading DON (Miller et al. 1985).

    In the environment favorable to FHB, modern S/MS and MR cultivars had similar test weights, which were higher than that of old HS cultivars (Table 7). This observation coincided with that of FHBI, where modern cultivars had lower FHBI compared with old HS cultivars (Table 4). Salgado et al. (2015) also reported a negative correlation between test weight and FHBI; with each unit increase of FHBI, the test weight was reduced by 0.32 to 0.39%. In the present study, with no fungicide application, old HS cultivars had an FHBI 10.5% higher than that of modern S/MS cultivars. Given the FHBI–test weight relationship reported by Salgado et al. (2015), the 10.5% higher FHBI with old HS cultivars in the current study translates to a test weight reduction of 3.4 to 4.1%, which is similar to the 5.4% difference observed in the present study. Therefore, our study showed that modern cultivars had lower FHBI than old cultivars, which may be one reason for the improved test weight under the environments tested. Other factors that may explain test weight responses include cultivar differences (OCCC 2018), differences in foliar diseases such as leaf rust and Septoria leaf blotch, and environmental conditions (Milus 1994). Because foliar diseases were at low levels in this study, we assume cautiously that test weight was associated with the level of FHB infection.

    The modern S/MS cultivars used in this study had higher grain yields compared with both old HS cultivars and modern MR cultivars. Madden and Paul (2009) reported that grain yield was negatively correlated with FHBI, with a 0.038 Mg/ha decrease in grain yield for each unit increase in FHBI. In our study, FHBI of old HS cultivars was 10.5% higher than that of modern S/MS cultivars, which would translate to 0.4 Mg/ha yield difference under the FHBI–grain yield relationship reported by Madden and Paul (2009). Another reason for the grain yield difference between susceptible cultivars in old and new eras may be that the canopy stayed green longer into grain fill in modern cultivars compared with the old ones tested in this study. Between anthesis (ZGS 65) and soft dough (ZGS 85), canopy NDVI decreased by 0.32 and 0.21 in old HS and modern S/MS cultivars, respectively (Table 7). The effect of staying green increases grain yield by maintaining photosynthetic capacity during late grain fill (Thomas and Howarth 2000; Thomas and Smart 1993). Modern MR cultivars had similar FHBI and canopy senescence during grain fill compared with the modern S/MS cultivars. The reason for grain yield differences between modern MR and S/MS cultivars might be a combination of: limited representation of cultivars within each susceptibility–era combination in this study; linkage drag, where undesirable yield traits are introduced when resistance QTLs are incorporated (Salameh et al. 2011); and selection for high-yielding cultivars associated with decreased tolerance of plant diseases (e.g., Foulkes et al. 2006). Because only two cultivars were included in each cultivar susceptibility–era combination in our study, more cultivars are needed to support the notion of a linkage drag associated with MR cultivars. Compared with old HS cultivars, modern MR cultivars had lower FHB symptoms and DON accumulation, higher test weight, delayed canopy senescence, and higher grain yields under favorable conditions for FHB infection. More effort is needed to reduce the potential linkage drag for MR cultivars, because high-yield cultivars are favored highly by growers.

    The fungicide used in this study contained pydiflumetofen + propiconazole, which has an additional mode of action (group 7) compared with the triazole fungicides (group 3) (FRAC 2018). Compared with the common practice of no fungicide in the 1996 era, pydiflumetofen + propiconazole fungicide reduced total DON by 39 to 47%, FDKs by 35%, and FHBI by 52 to 80% depending on the cultivar. According to a meta-analysis by Paul et al. (2008), the main triazole fungicides in that analysis (i.e., metconazole, prothioconazole, and tebuconazole + prothioconazole) achieved 42 to 45% lower DON content and 48 to 52% lower FHBI compared with no-fungicide controls. Thus, the new fungicide pydiflumetofen + propiconazole provided similar efficacy on FHB and DON content reduction to the triazole fungicides in that study. The pydiflumetofen-containing fungicide increased test weight and grain yields compared with the no-fungicide control (i.e., 1996 era). The application of the fungicide increased test weight by 5%, or 3.6 kg/hl, in the FHB-favorable environments, which was slightly higher than 2.5 to 2.8% found by Paul et al. (2010) using triazole fungicides, probably through a combination of cultivar, environment, FHB suppression, or differences in leaf disease control between studies. Across N rates, the pydiflumetofen-containing fungicide increased grain yield by 15 to 22%, depending on the cultivar, which is slightly higher than that reported by Paul et al. (2010), who reported 14 to 15% grain yield increases where only triazole fungicides were used. Therefore, it appears that the new pydiflumetofen-containing fungicide had similar efficacy to the triazole fungicides in reducing FHB and DON content and showed slightly better agronomic performance in test weight and grain yield responses compared with triazoles reported in other studies. Because test weight and grain yield are affected by other genetic and environmental factors (Ma et al. 2004; Milus 1994), side-by-side comparisons between the new pydiflumetofen-containing fungicide and the conventional triazoles are needed for comparing FHB and the control of various leaf diseases; however, the data from this study and unpublished studies support the use of the pydiflumetofen-containing fungicide as an alternative to triazoles for FHB management in a broader scheme of fungicide resistance management of FHB and DON content.

    The development and regulatory approval of pydiflumetofen + propiconazole (Miravis Ace) is advantageous from a fungicide resistance management view because it has a new mode of action. The current fungicides for managing FHB are all triazoles, which are DMIs belonging to group 3 (FRAC 2018). These have a single-site mode of action, which has a moderate risk of being overcome by the pathogen (OMAFRA 2013). Although fungicide-resistant F. graminearum has not been found in North America, in vitro studies have detected strains with higher half maximal effective concentration values and produce higher levels of nivalenol (Becher et al. 2010; Talas and McDonald 2015). Pydiflumetofen, a succinate dehydrogenase inhibitor belonging to group 7 (FRAC 2018), provides an alternative mode of action; alternating fungicides with different modes of action may be important for resistance management, where fungicides can be rotated or tank mixed (Brent and Hollomon 2007).

    Similar to the findings by Krnjaja et al. (2015), the increase in N rate from 100 (the common rate in the 1996 era) to 170 kg N/ha (the common rate in the modern era) had no detectable effect on FHB, TDON content, or test weight (Tables 3 and 5). However, the N rates and fungicide had a synergistic effect on grain yield; the highest yield response occurred where the high N rate was applied followed by a fungicide application at flowering, which was similar to responses across multiple cultivars and nine environments in a study reported by Brinkman et al. (2014). However, in the Brinkman et al. study, most of the yield responses were attributed to leaf disease control in larger canopies produced with high N rates in modern cultivars, because there was low incidence of FHB. In the current study, grain yields were 17 to 26% higher where the pydiflumetofen-containing fungicide was applied in plots that received 170 kg N/ha, whereas the yield response to fungicide was only 10 to 18% at the 100 kg N/ha rate, depending on the cultivar susceptibility–era combination. The synergy between a high N rate and fungicide in grain yield responses was documented in Brinkman et al. (2014), where high-yielding modern cultivars were hypothesized as factor that contributed to the response. In the current study, it was surprising that the old 1996 cultivars had a higher numerical response where a fungicide was applied to high-N plots compared with the use of the low N rate alone (Table 7). In old cultivars, the application of a fungicide at flowering to plots that received 170 kg N/ha increased grain yield by 1.17 Mg/ha, compared with the old standard of using 100 kg N/ha with no fungicide (4.25 to 5.42 Mg/ha; Table 7). In the modern cultivars, grain yields with both fungicide and high N increased by 1.13 and 0.99 Mg/ha for modern S/MS cultivars (5.30 to 6.43 Mg/ha; Table 7) and modern MR cultivars (4.58 to 5.57 Mg/ha; Table 7), respectively. The 100 kg N/ha rate was the common N rate used in the 1996 era. Another reason for the synergy between fungicide and N may be that both N and fungicide delayed canopy senescence or perhaps extended the grain fill period. Across cultivars, canopy NDVI decreased by 0.27 where 100 kg N/ha and no fungicide were applied; when either the pydiflumetofen-containing fungicide, 170 kg N/ha alone, or both fungicide and high-rate N were applied, the ΔNDVI from ZGS 65 to 85 decreased from 0.27 to 0.20, 0.20, and 0.12, respectively (Table 6). The improvements in cultivars, fungicide, and the use of higher N rates all positively influenced the stay-green effect, which contributed to grain yield responses. The use of a modern MR cultivar, high N rate, and a fungicide at flowering increased grain yield by a maximum of 31% in this study (from 4.25 to 5.57 Mg/ha, Table 7); grain yields have similarly increased by 32% in provincial 5-year average yield from 1992 and 1996 to 2012 and 2016 (OMAFRA 2019).

    In conclusion, the improvements in cultivar genetic resistance and fungicide, as represented by the cultivars and the fungicide in this study, were effective at suppressing FHB and reducing TDON contents. Compared with the use of an HS cultivar with no fungicide in the 1996 era, the use of modern MR cultivars and a fungicide reduced TDON content by 67% (from 19.0 to 6.3 ppm), FDKs by 49% (from 8.3 to 4.2%), and FHBI by 86% (from 17.8 to 2.6%; Table 4) and increased test weight by 11% (from 69.2 to 76.9 kg/hl; Table 7) in environments favorable for the development of FHB. The use of modern MR cultivars, a fungicide, and a higher N rate increased yield by 31% compared with the use of old HS cultivars, no fungicide, and a lower N rate in FHB environments. The magnitude of FHB suppression, yield, and test weight increase might vary given different cultivars, fungicides, and environmental conditions. As noted, the data must be interpreted with caution, because TDON, grain yield, and test weight may be confounded by the limited number of cultivars chosen to represent each susceptibility–era combination and through physiological impacts of nitrogen and fungicide effects on canopy development and leaf disease control. Nevertheless, this study has demonstrated progress in managing FHB and agronomic performance from the 1996 era to the modern era of wheat production.

    Acknowledgment

    We thank technicians Todd Phibbs, Jonathan Brinkman, Ken Van Raay, Darrell Galbraith, and co-op students Blake Morey, Alex Derewianko, Emma Langlois, Xuetao Hong, Julia Lupton, and Hyerin Kim for technical support. Special thanks to research associate Dr. Victor Limay-Rios for the guidance and support in mycotoxin analysis.

    The author(s) declare no conflict of interest.

    Literature Cited

    The author(s) declare no conflict of interest.

    Funding: This study was sponsored by Ontario Ministry of Agriculture, Food, and Rural Affairs; Agriculture and Agri-Food Canada through the Growing Forward 2 Program; University of Guelph, Ridgetown Campus, Research Support Lab; Grain Farmers of Ontario; Food & Consumer Products of Canada; Canadian National Millers Association; Pepsico Foods; and Nestlé.