Improving fungicide-based management of Cercospora leaf spot in table beet in New York, USA

Abstract

Cercospora leaf spot (CLS), caused by Cercospora beticola Sacc., is the dominant foliar disease affecting the profitability of table beet production in New York, USA. The disease is controlled primarily with fungicides, but sustainability is critically threatened by strobilurin resistance within the C. beticola population and a limited suite of alternatives. This study evaluated the temporal dynamics of strobilurin resistance within the C. beticola population over three years, and quantified sensitivity to selected demethylation inhibitors (DMIs; propiconazole and difenoconazole) and succinate dehydrogenase inhibitors (SDHIs; fluopyram, penthiopyrad, boscalid and benzovindiflupyr). Resistance to azoxystrobin (EC₅₀ > 0.2 µg mL⁻¹) was prevalent and detected in all fields sampled in 2015 and 2016, and six of 10 fields in 2017. Within fields, the incidence of azoxystrobin-resistant isolates ranged between 4.8% and 70%. No significant differences were detected in the frequency of azoxystrobin-resistant isolates within populations from fields with conventional and organic practices in 2016 and 2017. Reduced sensitivity to propiconazole and difenoconazole was detected in less than 1% of isolates. The EC₅₀ values for the SDHIs were higher than the maximum concentrations (≥ 20 µg mL⁻¹) evaluated. Two small-plot replicated field trials were also conducted over two years to evaluate alternatives to the commercial standards, propiconazole and azoxystrobin. Two applications of benzovindiflupyr + difenoconazole and pydiflumetofen + difenoconazole significantly reduced CLS severity and epidemic progress compared with plots that received propiconazole and non-treated controls.

Résumé

La cerscosporose, causée par Cercospora beticola Sacc., est la principale maladie foliaire qui compromet la rentabilité de la production de la betterave dans l’État de New York, aux États-Unis. La maladie se gère principalement avec des fongicides, mais la viabilité de la méthode est gravement compromise par la résistance de la population de C. beticola aux strobilurines ainsi que par une gamme limitée de solutions de rechange. Cette étude a évalué, pendant trois ans, la dynamique spatio-temporelle de la résistance aux strobilurines au sein de la population de C. beticola et a quantifié la sensibilité à des inhibiteurs choisis de la déméthylation (IDM: propiconazole et difénoconazole) et de la succinate-déshydrogénase (ISDH: fluopyram, penthiopyrade, boscalide et benzovindiflupyr). La résistance à l’azosxystrobine (EC₅₀ > 0.2 µg mL⁻¹) dominait et a été détectée dans tous les champs échantillonnés en 2015 et 2016 et, en 2017, dans 6 des 10 champs. Dans les champs, l’incidence d’isolats résistants à l’azosxystrobine variait de 4.8% à 70%. Aucune différence significative n’a été détectée quant à la fréquence des isolats résistants à l’azosxystrobine dans les populations des champs cultivés traditionnellement ou biologiquement en 2016 et 2017. Une sensibilité réduite au propiconazole et au difénoconazole a été détectée dans moins de 1% des isolats. Les valeurs d’EC₅₀ pour les ISDH étaient plus élevées que les concentrations maximales (≥ 20 µg mL⁻¹) évaluées. Deux essais répétés dans de petites parcelles ont également été menés pendant deux ans afin d’évaluer des solutions de rechange aux normes commerciales, soit au propiconazole et à l’azosxystrobine. Deux applications de benzovindiflupyr + difénoconazole et pydiflumetofen + difénoconazole ont significativement réduit la gravité de la cerscosporose et ralenti la progression de l’épidémie, et ce, comparativement aux parcelles traitées au propiconazole et aux parcelles témoins non traitées.

Keywords:

• Beta vulgaris
• carboxamides
• Cercospora beticola
• demethylation inhibitors
• fungicide resistance
• strobilurins
• succinate dehydrogenase inhibitors
• table beet

Mots clés:

• Beta vulgaris
• betterave
• carboxamides
• Cercospora beticola
• inhibiteurs de la déméthylation
• inhibiteurs de la succinate-déshydrogénase
• résistance aux fongicides
• strobilurines

Introduction

Cercospora leaf spot (CLS), caused by Cercospora beticola Sacc., is the most significant foliar disease affecting table beet (Beta vulgaris L. ssp. vulgaris) production in New York (NY), USA (Jacobsen and Franc 2009; Pethybridge et al. 2018). The disease causes crop loss by reducing the quality of the foliage making it unsuitable for fresh market sales, and if epidemics occur early, the reduced green leaf area may impair the plant’s growth and ability to achieve adequate root sizes for sale (Pethybridge et al. 2018). If CLS is not controlled, defoliation also deleteriously affects mechanized harvest, which relies upon healthy foliage to pull roots from the ground (Pethybridge et al. 2017a). In NY, the inability to top-pull harvest may result in complete crop loss as digging is hampered by heavy and rocky mineral soils, and often wet conditions in autumn. CLS is also a substantial constraint worldwide to production of the related crop sugar beet, and causes losses through reduced photosynthetic area deleteriously affecting root growth and sugar content (Smith and Ruppel 1973; Smith and Martin 1978; Shane and Teng 1992; Weiland and Koch 2004; Jacobsen and Franc 2009; Tedford et al. 2019).

Symptoms of CLS on table beet begin as distinctive discrete light brown or tan spots that become grey with age and coalesce rapidly when conditions are conducive to infection and disease development (Koike et al. 2010). In red table beet cultivars, which are the most popular for broad-acre production for the processing market, CLS lesions are surrounded by a red halo. Diagnostic for CLS is the presence of signs of C. beticola within lesions including small, black pseudostromata and conidia (McKay and Pool 1918; Weiland and Koch 2004). The conidia are disseminated by wind and/or rain (Koike et al. 2010) and the latent period may range from 5 to 14 days dependent upon cultivar (Pethybridge, unpublished data).

Although there is a broad range of susceptibility to CLS in table beet cultivars (Pethybridge et al. 2017b), the ability to mitigate CLS risk through the selection of less susceptible cultivars is limited for broad-acre table beet production considering the processing market’s strict product standards for canning (USDA 2017). Cultivars grown for processing into value-added products such as juices and snack packs must also have desirable quality characteristics such as high sugar content and colour stability and these typically are highly susceptible to CLS (Pethybridge et al. 2017b).

The rapid spread of CLS from multiple infection cycles within the season results in a polycyclic epidemic (Pool and McKay 1916; Pethybridge et al. 2017a). The dominant tactic used for CLS management is three to four fungicide applications with the aim of reducing the rate of disease progress (Madden et al. 2007; Pethybridge et al. 2017a, 2018). However, the durability of this approach is critically threatened by strobilurin resistance within the local C. beticola population, reported in 41% of isolates collected from broad-acre table beet fields in 2014 (Vaghefi et al. 2016). Strobilurin resistance in the local C. beticola population was associated with G143A and F129L mutations within the cytochrome b gene (Vaghefi et al. 2016). This report followed field control failures despite repeated applications of azoxystrobin (Abawi et al. 2014). Resistance to strobilurin fungicides has also been reported in C. beticola populations associated with CLS in sugar beet across the USA (Kirk et al. 2012; Bolton et al. 2012a, 2013; Rosenzweig et al. 2015) and Ontario, Canada (Trueman et al. 2013). Early season disease control programmes also continue to rely upon azoxystrobin applied in-furrow and post-emergence for Rhizoctonia root rot caused by Rhizoctonia solani (Abawi et al. 1986; Gugino et al. 2008; Pethybridge et al. 2018). However, no information is available on the temporal dynamics of strobilurin resistance within the C. beticola population in NY.

The design of durable management programmes for CLS in NY table beet production based on rotation of fungicides with different site-specific modes of action is problematic due to the restricted number of registered products. Currently, azoxystrobin (as Quadris; Syngenta Corporation) and propiconazole (Tilt; Syngenta Corporation) are the only site-specific fungicides registered for CLS control in table beet in NY (Reiners et al. 2019). Propiconazole is a demethylation inhibitor that interrupts sterol 14α-demethylase activity and the C14-demethylation step during sterol formation (Köller and Scheinpflug 1987). Resistance to DMIs is also prevalent in C. beticola populations from sugar beet across Europe (Karaoglanidis et al. 2000, 2001, 2002, 2003; Karaoglanidis and Bardas 2006), the USA (Secor et al. 2010) and Canada (Trueman et al. 2017). Other products, such as triphenyltin hydroxide (Giannopolitis 1978; Giannopolitis and Chrysayi-Tokousbalides 1980) and benzimidazoles (Georgopoulos and Dovas 1973; Percich et al. 1986; Campbell et al. 1998; Weiland and Halloin 2001; Karaoglanidis and Ioannidis 2010; Vogel et al. 2018) that have been used for CLS management in sugar beet production are not available to table beet growers in NY. Moderate disease control has been reported from selected succinate dehydrogenase inhibitors (SDHIs) alone or in combination with members of other FRAC groups (Pethybridge et al. 2017a). SDHI fungicides act on ubiquinone binding sites within the mitochondrial succinate dehydrogenase leading to respiration inhibition (Sierotzki and Scalliet 2013) and belong to FRAC group 7 (FRAC 2017). A pre-mixed formulation of benzovindiflupyr (FRAC group 7) and difenoconazole (FRAC group 3) (as Aprovia Top; Syngenta Corporation) provided excellent disease control compared with the current commercial standards (Pethybridge et al. 2017a). However, neither of the active ingredients within this pre-mixed formulation were tested individually so it is difficult to determine whether the CLS control resulted from one or both. The only multi-site contact fungicides registered for CLS control in table beet in NY are copper-based products (FRAC group M5) which have moderate efficacy and are best used prior to the first appearance of symptoms (Pethybridge et al. 2017b, 2018). The current commercial recommendation for CLS control is to apply the first fungicide when an average of one CLS lesion per leaf is observed, with subsequent applications repeated based on the reapplication recommendations of specific products from the manufacturer. The presence of only two site-specific FRAC groups registered for CLS control means that fungicide resistance management through rotation of modes of action cannot be achieved.

The objective of this study was to quantify temporal changes in strobilurin resistance and evaluate the in vitro sensitivity to selected DMI (propiconazole) and SDHI fungicides within the C. beticola population in NY. A complementary objective was to evaluate the in-field performance of difenoconazole and selected SDHIs as potential additions to the limited suite of currently available products for CLS control. The status of sensitivity within C. beticola to difenoconazole and selected SDHI fungicides is important baseline information to monitor temporal changes that may result following introduction.

Materials and methods

Isolate collection, culture and storage

Cercospora beticola isolates were obtained from symptomatic table beet leaves collected arbitrarily along linear transects within rows in selected fields in western New York in 2015, 2016 and 2017. Leaves were stored at 4°C for 5–7 days before isolations were conducted. Samples were first rinsed in sterile distilled water and placed in glass Petri dishes in humid boxes at room temperature to promote sporulation. Isolations from plant tissue and single spore transfer were conducted according to the protocol described by Vaghefi et al. (2016), maintained on potato dextrose agar (PDA; Hardy Diagnostics, Santa Maria, CA) plates, and stored at −20°C as colonized mycelial plugs.

In vitro fungicide sensitivity assays

Azoxystrobin. Sensitivity to azoxystrobin was assessed in a conidial germination assay using the methods described by Vaghefi et al. (2016). Sensitivity to azoxystrobin of 452 C. beticola isolates from across NY collected in 2015 (n = 162), 2016 (n = 190) and 2017 (n = 100) was tested. Isolates collected in 2015 were from four table beet fields and within-field population sample sizes ranged between 38 and 42 isolates. Two of the fields were managed using conventional practices and hence for foliar disease management received at least two applications of conventional fungicides. The remainder of the fields were organic and did not receive any fungicides. In 2016, C. beticola populations were collected from three conventional and organic table beet fields (total of six fields), and in-field population sample sizes ranged between 29 and 35. In 2017, C. beticola populations were collected from five conventional and organic table beet fields (total of 10 fields) and 10 isolates were used from each field.

In brief, technical grade azoxystrobin (100 mg; Sigma-Aldrich, St. Louis, MO) was dissolved in warm acetone and made up to 5 mL to make a stock suspension (20 000 µg mL⁻¹) and used to make a dilution series (10⁻¹, 10⁻², 10⁻³ and 10⁻⁴) in sterile distilled water resulting in four concentrations. Suitable aliquots from the dilution series were added to 3% (wt:vol) water agar (WA; Hardy Diagnostics) at 60°C, to achieve concentrations of 0.001, 0.01, 0.1, 1, 10, 25 and 100 µg mL⁻¹. Salicylhydroxyamic acid (SHAM; Sigma-Aldrich) was added to all fungicide-amended media to inhibit the alternative respiration pathway at 60 µg mL⁻¹. This concentration was the maximum to suppress the alternative respiration pathway and not affect conidial germination (Vaghefi et al. 2016). A SHAM only and non-amended control was used in each assay.

Conidia for azoxystrobin sensitivity assays were generated by the technique described in Vaghefi et al. (2016). In brief, this involved culturing C. beticola isolates on clarified V8 agar (Miller 1955; Secor and Rivera 2012) for 14 days under 118 volt/60 hz white fluorescent lights with a 12 h photoperiod at 20°C. Plates were flooded with sterile distilled water, and conidia detached from mycelia using a sterile spatula. The resultant conidial suspension was then passed through four layers of sterile muslin cloth and a 50 µl aliquot placed onto 9 mm² blocks of 3% WA mounted on microscope slides with three replicates for each fungicide concentration, SHAM only, and non-amended control per isolate. Blocks were covered by a glass coverslip and placed in a sealed container in the dark for 24 h at room temperature prior to evaluation of conidial germination at 200 to 400× magnification. Fifty conidia were evaluated from each slide and the number of germinated or non-germinated conidia was recorded. Conidia were considered germinated if any single germ tube was at least half the length of the conidium. The effective concentration required to cause 50% inhibition (EC₅₀) of conidial germination in comparison to the non-amended control was calculated by probit analysis (PROC PROBIT; Statistical Analysis System (SAS) Version 9.4). The effect of SHAM only on conidial germination was quantified by analysis of variance (Genstat Version 18; Hemel Hempstead, UK).

A discriminatory dose associated with azoxystrobin-resistant and -sensitive phenotypes in the NY C. beticola population has been defined (Vaghefi et al. 2016). The presence of the G143A mutation was associated with a less sensitive phenotype (EC₅₀ > 0.2 µg mL⁻¹) and classified as resistant. Isolates with a sensitive phenotype (EC₅₀ < 0.1 µg mL⁻¹) did not have the G143A mutation and were classified as sensitive. A low frequency of isolates with an intermediate phenotype (0.1 < EC₅₀ < 0.2 µg mL⁻¹) had the F129L mutation (Vaghefi et al. 2016). For the 2016 and 2017 populations, the effect of conventional and organic production practices on the frequency of C. beticola isolates with azoxystrobin-resistant phenotypes was tested using an independent t-test assuming unequal sample variances (Genstat Version 18).

Demethylation inhibitors

Propiconazole. Mycelial growth assays were done to quantify the sensitivity of select C. beticola isolates to propiconazole to establish a discriminatory dose. Twenty-seven C. beticola isolates collected in 2017 (one per field) were used to estimate the effective concentration of technical grade propiconazole (Sigma-Aldrich) to reduce mycelial growth by 50% (EC₅₀). Concentrations of 0.1, 1, 10 and 40 µg mL⁻¹ were prepared within the fungicide-amended media. To evaluate the sensitivity of a broader range of C. beticola populations (n = 252) to propiconazole collected in 2016 (n = 152) and 2017 (n = 100), a discriminatory concentration of 20 µg mL⁻¹ was selected. In this assay, 6 mm mycelial plugs from a 14-day-old culture of each isolate were placed in the centre of Petri dishes amended with the discriminatory dose of propiconazole.

Difenoconazole. Twenty-three of the same isolates from 2017 used for propiconazole sensitivity testing were also evaluated for sensitivity to technical grade difenoconazole (Sigma-Aldrich) using the same concentrations used for propiconazole (0.1, 1, 10 and 40 µg ml⁻¹) within the fungicide-amended media, and a non-amended control, using methods described below. The base media for mycelial growth assays was PDA.

For propiconazole and difenoconazole sensitivity testing, the experimental design was completely randomized and for each isolate, three replicates of fungicide-amended and non-amended media were used. Colony diameter of each isolate was measured in two perpendicular directions and averaged. The per cent of mycelial growth inhibition was calculated for each fungicide as: [1−  (colony diameter in fungicide-amended/colony diameter in non-amended control)] × 100. Isolates were insensitive if the per cent inhibition induced by each discriminatory concentration was less than 50% relative to the control, or otherwise sensitive. The association between EC₅₀ values for propiconazole and difenoconazole across isolates was tested using Spearman’s rank correlation coefficient (Spearman 1904; Genstat Version 18).

Succinate dehydrogenase inhibitors. The sensitivity of 28 C. beticola isolates (7 from 2014 and 21 from 2016) to selected succinate dehydrogenase inhibitors (SDHIs) was also tested using mycelial growth assays. Technical grade aliquots of each fungicide (Sigma-Aldrich) were first dissolved in acetone. Stock concentrations of fluopyram (6400 µg mL⁻¹), boscalid (6300 µg mL⁻¹), penthiopyrad (2000 µg mL⁻¹) and benzovindiflupyr (5000 µg mL⁻¹) were used to achieve concentrations of 0.05, 0.5 and 5 µg mL⁻¹ for all fungicides and a maximum concentration of 20 µg mL⁻¹ for all fungicides, except fluopyram for which the maximum was 25 µg mL⁻¹. Non-amended media including an equivalent amount of acetone was used as the control.

Mycelial plugs (6 mm diameter) were removed from a 14-day-old culture of each isolate and placed in the centre of Petri dishes containing each of the fungicide concentrations and non-amended controls. The colony diameter of each isolate was measured in two perpendicular directions after 5 days of incubation at 22°C in the dark. The experimental design was completely randomized and included three replicates of each concentration of fungicide and non-amended control. An experimental unit was defined as a Petri dish with a mycelial plug. The effective concentration required to cause 50% inhibition (EC₅₀) of mycelial growth in comparison to the non-amended control was calculated by probit analysis (PROC PROBIT; Statistical Analysis System (SAS) Version 9.4).

Field trials

One small-plot, replicated trial was conducted in each cropping season of 2017 and 2018 at the Research North facility of Cornell AgriTech at The New York State Agricultural Experiment Station, Geneva, New York. Both trials were established by planting seed of the table beet cultivar Ruby Queen (Seneca Foods, Dayton, WA) which is the commercial standard used for processing crops in NY. The crops were planted on 9 June 2017 and 18 May 2018 with a precision Monosem planter (Monosem, Lenexa, KS) at an in-row plant density of 54 seeds m⁻¹ and 0.76 m between rows. Sufficient nutrients for optimal plant growth was ensured by incorporating 392 kg ha⁻¹ two days prior to planting by broadcast application followed by incorporation with a Coulter mulcher and 337 kg ha⁻¹ in a band at planting. On both occasions, the fertilizer was 10 N: 10 P: 10 K + 2.2 kg ha⁻¹ Boron (Phelps Supply Inc., Phelps, NY). Weeds were managed using the herbicides S-metolachlor (as Dual Magnum (0.78 L ha⁻¹); Syngenta Corporation) + ethofumesate (as Nortron (2.2 L ha⁻¹); Bayer CropScience) within 24 h of planting as a pre-emergence application. A post-emergence application of the herbicides, clopyralid (as Stinger (280.2 g ha⁻¹); Dow AgroSciences) and phenmedipham (as Spin-Aid (1.1 L ha⁻¹); Bayer CropScience) was applied 27–29 days after planting (DAP) when table beet plants had reached two to four true leaves. Trials received supplementary irrigation at 7- to 10-day intervals using overhead sprinklers for optimal plant growth and disease development.

Each plot was inoculated with a suspension of eight C. beticola isolates representative of the genotypes within the NY population in 2014 and used in other trials for comparative purposes (Tb14-047, Tb14-070, Tb14-081, Tb14-113, Tb14-116, Tb14-118, Tb14-146 and Tb14-154; Vaghefi et al. 2016) at 41 and 50 DAP (2017) or54 and 68 DAP (2018). Inoculum suspensions were prepared from liquid clarified V8 (CV8) juice broth (Pethybridge et al. 2017a). The concentration of the inoculum varied between 2.3 × 10³ colony forming units (cfu) mL⁻¹ and 1.2 × 10⁵ cfu mL⁻¹ (+ 0.01% (v/v) polysorbate-20) with a backpack sprayer at 130 L ha⁻¹. Colony forming units were estimated by making serial dilutions of the final inoculum suspension to 10⁻⁵ in sterile distilled water and spreading 50 µL aliquots onto WA plates. Three to four replicate plates were used for each concentration and incubated at room temperatures under white fluorescent light with a 12 h photoperiod. The number of colonies on each plate was counted after four days and averaged between replicates within concentrations.

In the 2017 trial, treatments were Aprovia Top (950 mL ha⁻¹; benzovindiflupyr + difenoconazole), Inspire (365.4 mL ha⁻¹; difenoconazole only), Aprovia (511.5 mL ha⁻¹; benzovindiflupyr only) and Tilt (292.3 mL ha⁻¹; propiconazole conventional standard). Each fungicide was applied either one, two or three times at regular intervals (49, 65 and 72 DAP, respectively). In the 2018 trial, treatments were Aprovia Top (950 mL ha⁻¹; benzovindiflupyr + difenoconazole), Inspire (365.4 mL ha⁻¹; difenoconazole only), Miravis Top (1001.2 mL ha⁻¹; pydiflumetofen + difenoconazole) and Tilt (292.3 mL ha⁻¹; propiconazole conventional standard). Each fungicide was also applied either once, twice or three times (53, 69 and 83 DAP).

Treatments were arranged in a randomized complete block design with four replications, including a non-treated control. Individual plots were 3.1 m sections of two adjacent rows, with a non-treated buffer of 1.5 m between plots within rows. One non-treated and non-inoculated row separated adjacent plots. Fungicides were applied using a CO₂-pressurized backpack sprayer with output volumes of 193 and 233 L ha⁻¹ in 2017 and 2018, respectively. The sprayer was fitted with three TJ 8002VS flat fan nozzles spaced 48.2 cm apart along a 96.5 cm long boom.

The homogeneity of the crop stand was evaluated by counting in-row plant populations in each of two arbitrarily placed 0.5 m sections (one per row in each plot) following full emergence and prior to treatment application at 29 DAP (2017) and 39 DAP (2018). Disease severity was quantified by counting lesions on each of 20 arbitrarily selected leaves within each plot (10 per row) at regular intervals (43, 52, 55, 63, 69 and 76 DAP in 2017; and 51, 75, 83, 89 and 93 DAP in 2018). These data was used to calculate epidemic progress in each plot (area under the disease progress stairs; AUDPS; Simko and Piepho 2012).

Canopy reflectance in the near infrared range (830 nm) was measured using a hand-held multispectral radiometer (MSR5; CropScan Inc., Rochester, MN) at 77 and 93 DAP in 2017 and 2018, respectively. Canopy reflectance at 830 nm has been demonstrated to be a good non-invasive and objective measurement of green leaf area and hence plant health, inversely correlated with disease intensity (Nilsson 1995; Pethybridge et al. 2017a). Measurements were taken between 1200 and 1400 h (Eastern Time) when incoming radiation was at least 400 W m⁻². The radiometer was placed within the middle of each plot 2 m above the soil surface. Canopy reflectance was calculated as a percentage of the voltage for the reflected radiation divided by the voltage for the incident radiation. Two readings were averaged from each plot.

At harvest (80 and 94 DAP in 2017 and 2018, respectively), the efficacy of fungicides was assessed on above and belowground yield components. Foliar biomass was quantified by clipping foliage from the roots within a single 1 m section of each plot and recording weight after drying at 60ºC for 24 h. Roots were dug by hand and the number of roots within the 1 m section was counted and fresh weight recorded. The shoulder diameter of 20 arbitrarily selected roots from each plot was also measured using digital calipers (Mitutoyo 500-196-30; Mitutoyo USA, Aurora, IL) and averaged per plot.

The effect of fungicides on disease severity measured as the average number of CLS lesions per leaf at the final evaluation and AUDPS, canopy reflectance at 830 nm, and yield components (root number at harvest, dry weight of foliage, fresh root weight and average root shoulder diameter) was analysed using a generalized linear model. Data from each trial were analysed separately due to differences in treatments. Treatment and replication were evaluated as fixed and random effects, respectively. Means were separated by a Fisher’s protected least significant difference test at P = 0.05. Linear regression analysis quantified the relationships between table beet yield components within trials (Genstat Version 18).

Results

In vitro azoxystrobin sensitivity

In 2015, 3 of 42 (7.1%), 2 of 42 (7.1%), 24 of 40 (60%) and 14 of 38 (36.8%) C. beticola isolates within each of the table beet fields were classified as resistant (EC₅₀ > 0.2 µg azoxystrobin mL⁻¹ (Fig. 1a). In 2016, the frequency of azoxystrobin-resistant isolates obtained from conventional fields was 12 of 30 (40%), 16 of 35 (45.8%) and 13 of 32 (40.6%); and in organic fields was 21 of 30 (70%), 22 of 34 (64.7%) and 2 of 29 (6.9%) (Fig. 1b). In 2017, in one of the five conventional fields, all C. beticola isolates were sensitive to azoxystrobin but in the remaining four fields, the frequency of resistant isolates ranged between five and six of 10 isolates (Fig. 1c). In three of the organic production fields, all isolates were sensitive to azoxystrobin, but in the remaining two, the frequency of resistant isolates was 1 and 6 of 10 isolates, respectively. In both of the latter years, there was no significant difference (P = 0.785 in 2016; P = 0.101 in 2017) in the proportion of azoxystrobin-resistant isolates from table beet fields grown using organic and conventional production practices.

Fig. 1 Frequency of azoxystrobin-sensitive and -resistant Cercospora beticola isolates collected from conventional and organic table beet fields in 2015 (a), 2016 (b), and 2017 (c) based on in vitro conidial germination assays.

In vitro propiconazole and difenoconazole sensitivity

For the sub-set of C. beticola isolates used to establish a propiconazole discriminatory dose, EC₅₀ values ranged between 5.29 to 19.43 µg mL⁻¹. For two isolates collected in 2016 from broad-acre, conventional fields, the EC₅₀ value was greater than the highest concentration tested (40 µg propiconazole mL⁻¹). Using a discriminatory dose of 20 µg propiconazole mL⁻¹ to evaluate the populations collected in 2016 and 2017, all additional isolates tested were classified as sensitive. EC₅₀ values for difenoconazole for 21 of the 23 isolates collected in 2016 were less than the lowest concentration tested (0.1 µg mL⁻¹) and for the remaining two isolates was greater than the highest concentration tested (40 µg mL⁻¹). The two isolates that were less sensitive to propiconazole were also less sensitive to difenoconazole. There was a significant association between the EC₅₀ values calculated for difenoconazole and propiconazole across isolates (ρ = 0.645; P < 0.001).

Succinate dehydrogenase inhibitor sensitivity

The EC₅₀ values of all C. beticola isolates tested against the SDHI fungicides were above the maximum concentrations tested (20 µg mL⁻¹ for boscalid, penthiopyrad and benzovindiflupyr; and 25 µg mL⁻¹ for fluopyram).

Field trials – 2017

No significant differences (P = 0.559) in plant density were detected between plots prior to application of treatments. Fungicides and application frequency had a significant effect on disease severity and AUDPS (Table 1; Fig. 2). Benzovindiflupyr, difenoconazole, propiconazole and benzovindiflupyr + difenoconazole significantly reduced CLS severity and epidemic progress compared with the non-treated control. Benzovindiflupyr + difenoconazole provided the best control and there was no significant benefit in disease control from repeating the application two or three times compared with a single application. Benzovindiflupyr only provided moderate disease control irrespective of application number, and AUDPS values were significantly higher than in plots receiving difenoconazole only. Two or three applications of difenoconazole provided disease control that was not significantly different than plots receiving benzovindiflupyr + difenoconazole. Propiconazole provided moderate disease control and AUDPS was significantly reduced by 21–27.6% from two applications compared with a single application, but was not significantly improved by a third application (Table 1; Fig. 2). The effects of fungicides on disease control were also observed in canopy reflectance at 830 nm at 77 DAP, with the exception of no significant differences between benzovindiflupyr applications, a single propiconazole application, and non-treated plots. There was also a significant 16.1–25.6% increase in canopy reflectance at 830 nm in plots receiving repeated applications of benzovindiflupyr + difenoconazole compared with a single application (Table 1). The improvements in CLS control from fungicides were also reflected in effects on the dry weight of foliage (Table 1). For example, a single and three applications of benzovindiflupyr + difenoconazole increased the dry weight of foliage compared with non-treated plots by 168.8% and 265%, respectively. Propiconazole also significantly increased the dry weight of foliage compared with non-treated plots by an average of 120.3% and no further benefit was gained from increasing application number. Treatment had no significant effect on the number of roots per metre and average root shoulder diameter, however several treatments that resulted in a significant increase in the dry weight of foliage also increased the fresh weight of roots (Table 1; Fig. S1).

Table 1. Effect of fungicides (benzovindiflupyr, difenoconazole, benzovindiflupyr + difenoconazole, and propiconazole [conventional fungicide standard]) and frequency of applications on Cercospora leaf spot (CLS) disease severity and epidemic progress, canopy reflectance at 830 nm, and table beet yield components in a small-plot, replicated trial at Geneva, New York in 2017.

Fungicide (number of applications)	CLS lesions per leaf (76 DAP) ^a	AUDPS ^b	Canopy reflectance at 830 nm (77 DAP)	Dry weight of foliage (g m^−1; 80 DAP)	Roots m^−1 (80 DAP)	Fresh weight of roots (g m^−1; 80 DAP)	Average root shoulder diameter (cm; 80 DAP)
Benzovindiflupyr (1)	50.3 bc	871 bc	22.6 d	206.1 e	46.0	1,640.0 cd	4.32
Benzovindiflupyr (2)	52.1 bc	798 c	23.9 d	243.5 cde	46.8	1,640.1 cd	4.24
Benzovindiflupyr (3)	48 bc	836 c	26.7 cd	245.2 cde	38.2	1,710.0 bcd	4.57
Difenoconazole (1)	42.3 c	493 d	29.8 bc	222.0 d	44.0	1,630.1 cd	4.42
Difenoconazole (2)	18 d	272 e	36.2 a	308.7 bcd	48.0	2,040.0 ab	4.57
Difenoconazole (3)	4.9 e	122 ef	39.8 a	343.9 ab	45.2	2,150.1 a	4.59
Benzovindiflupyr + Difenoconazole (1)	14.9 de	241 ef	31.6 b	300.2 bcd	44.5	1,810.4 abc	4.42
Benzovindiflupyr + Difenoconazole (2)	9.2 de	181 ef	37.0 a	316.4 abc	43.8	1,890.1 abc	4.49
Benzovindiflupyr + Difenoconazole (3)	3.6 e	110 f	39.7 a	407.7 a	58.2	2,157.9 a	4.34
Propiconazole (1)	58.9 b	997 b	26.3 cd	309.9 bcd	48.2	1,719.9 bc	4.34
Propiconazole (2)	49.7 bc	788 c	30.1 bc	250.6 bcde	56.5	1,759.9 bc	4.42
Propiconazole (3)	41.3 c	722 c	29.8 bc	251.2 bcde	49.0	1,710.0 bcd	4.37
Non-treated	78.9 a	1583 a	22.6 d	111.7 f	48.8	1,340.1 d	4.01
LSD ^c	12.4	149.9	4.4	93.8	-	369.9	-
P	<0.001	<0.001	<0.001	<0.001	0.206 (ns) ^e	0.004	0.162 (ns)
CV ^d (%)	13.0	17.0	2.5	24.2	18.6	14.4	5.9

^a DAP = Days After Planting.

^b AUDPS = Area Under the Disease Progress Stairs (measure of epidemic progress). Larger values indicative of higher disease severity and faster spread.

^c LSD = Least Significant Difference.

^d CV = Coefficient of Variation.

^e ns = not significant.

Fig. 2 Temporal changes in the average number of cercospora leaf spot (CLS) lesions per leaf from 43 to 76 days after planting (DAP) in plots receiving up to three applications (number of applications in parentheses) of benzovindiflupyr (a), benzovindiflupyr + difenoconazole (b), difenoconazole (c), and propiconazole (d) in a small-plot, replicated trial conducted at Geneva, New York in 2017.

Field trials – 2018

No significant differences (P = 0.5) in plant density were detected between plots prior to treatments. Fungicide and application frequency had a significant effect on CLS severity and epidemic progress (Table 2; Fig. 3). When receiving only one application, pydiflumetofen + difenoconazole was the most efficacious product and reduced the number of CLS lesions per leaf and AUDPS by 73.6% and 77.4%, respectively. Additional significant reductions in CLS lesions per leaf and AUDPS resulted from a second application of pydiflumetofen + difenoconazole but no further reduction in disease severity resulted from a third application, resulting in an average reduction of CLS lesions per leaf of 96.4%. For difenoconazole, three applications provided disease control that was not significantly different from two applications of pydiflumetofen + difenoconazole. Although one application of benzovindiflupyr + difenoconazole resulted in a significantly higher number of CLS lesions per leaf and AUDPS compared with the equivalent of pydiflumetofen +difenoconazole, repeated applications resulted in significantly less disease severity that was not significantly different from two applications of pydiflumetofen + difenoconazole. Moreover, there was no benefit in disease control from a third application of benzovindiflupyr + difenoconazole compared with two. For propiconazole (the conventional fungicide standard), one application provided relatively poor disease control and reduced the number of CLS lesions per leaf only by 15.5% compared with non-treated plots. Significant reductions in the number of CLS lesions per leaf and AUDPS were obtained from another application of propiconazole but no additional benefit resulted from a third application (Table 2; Fig. 3). Canopy reflectance at 830 nm also showed the improvements in disease control obtained from fungicide application and hence differences in green leaf area. For example, canopy reflectance at 830 nm was significantly increased by 59.8% in plots receiving two applications of pydiflumetofen + difenoconazole compared with non-treated plots, and by 15.9% over plots receiving only one application (Table 2). All treatments significantly increased the dry weight of foliage compared with non-treated plots. However, increasing application frequency did not increase the dry weight of foliage for any of the fungicides. Treatment had no significant effect on the number of roots, fresh weight of roots, and average root shoulder diameter (Table 2).

Table 2. Effect of fungicides (benzovindiflupyr + difenoconazole, difenoconazole, pydiflumetofen + difenoconazole, and propiconazole) and the number of applications on cercospora leaf spot (CLS) disease severity and epidemic progress, canopy reflectance at 830 nm, and table beet yield components in a small-plot, replicated trial at Geneva, New York in 2018.

Treatment (Application Frequency)	CLS lesions per leaf (93 DAP) ^a	AUDPS ^b	Canopy reflectance at 830 nm (93 DAP)	Dry weight of foliage (g m^−1; 94 DAP)	Roots m^−1 (94 DAP)	Fresh weight of roots (g m^−1 94 DAP)	Average root shoulder diameter (cm; 94 DAP)
Pydiflumetofen + Difenoconazole (1)	20.3 ef	306 f	34.56 d	170.7 ab	39.1	4,010	5.51
Pydiflumetofen + Difenoconazole (2)	3.9 g	41 g	40.06 ab	189.2 ab	39.5	4,250	5.83
Pydiflumetofen + Difenoconazole (3)	1.7 g	21 g	40.17 ab	173.4 ab	30.5	3,921	6.09
Difenoconazole (1)	17.4 f	304 f	33.56 de	184.9 ab	30.5	3,874	5.85
Difenoconazole (2)	26.9 e	409 e	36.56 cd	152.0 b	38.2	3,762	5.82
Difenoconazole (3)	5.0 g	83 g	38.32 bc	162.5 b	35.2	3,688	5.71
Benzovindiflupyr + Difenoconazole (1)	37.6 d	536 d	30.53 e	148.4 b	31.8	4,055	6.01
Benzovindiflupyr + Difenoconazole (2)	6.9 g	83 g	40.69 ab	210.3 a	43.2	4,201	5.77
Benzovindiflupyr + Difenoconazole (3)	5.3 g	82 g	42.81 a	172.5 ab	39.5	3,834	5.49
Propiconazole (1)	65.0 b	1113 b	24.69 f	151.2 b	33.2	3,833	6.78
Propiconazole (2)	39.3 cd	714 c	34.16 d	173.7 ab	46.8	4,192	5.46
Propiconazole (3)	46.3 c	741 c	34.69 cd	168.2 ab	34.2	3,723	6.10
Non-treated	76.9 a	1573 a	25.07 f	71.2 c	33.2	3,674	5.80
LSD ^c	8.1	118	3.36	47.5	-	-	-
P	< 0.001	< 0.001	< 0.001	< 0.001	0.243 (ns) ^e	0.826 (ns)	0.342 (ns)
CV ^d (%)	6.3	4.8	2.6	8.3	14.5	7.5	6.3

^aDAP = Days After Planting.

^bAUDPS = Area Under the Disease Progress Stairs (measure of epidemic progress). Larger values indicative of higher disease severity and faster spread.

^cLSD = Least Significant Difference.

^dCV = Coefficient of Variation.

^ens = not significant.

Fig. 3 Temporal changes in the average number of cercospora leaf spot (CLS) lesions per leaf from 51 to 93 days after planting (DAP) in plots receiving up to three applications (number of applications in parentheses) of benzovindiflupyr + difenoconazole (a), pydiflumetofen + difenoconazole (b), difenoconazole (c), and propiconazole (d) in a small-plot, replicated trial conducted at Geneva, New York in 2018.

Discussion

The high prevalence and incidence of azoxystrobin-resistant C. beticola isolates associated with CLS on table beet in NY found in each of three years is suggestive of persistence of this trait in the population or the annual introduction of isolates with this phenotype. Evidence supporting the relative contributions of a broad range of potential primary inoculum sources for CLS epidemics in table beet is mixed. For example, Knight et al. (2018) did not find evidence for shared multilocus genotypes in C. beticola populations associated with CLS epidemics in three fields planted with table beet between seasons in successive years. The most likely dominant inoculum sources were external to the field of interest and may include infested seed and an unknown sexual morph producing spores able to be disseminated long distances between fields (Knight et al. 2018). In contrast, alternative hosts that could be responsible for local inoculum sources while acting as green bridges between seasons have also been documented (Knight et al. 2019a). Cercospora beticola isolates from NY are pathogenic to several weed species, including Brassica kaber, Chenopodium album, Carthamus tinctorius, Rumex obtusifolius and Spinacia oleacera but surveys of CLS-like lesions on weeds in table beet fields have found C. beticola only at low frequencies. For example, of 71 CLS-like lesions observed on field populations of Ch. album, 51 were associated with C. chenopodii and only two with C. beticola. However, this study also found C. beticola could sporulate on necrotic and senescent tissue, which suggests saprophytic survival on non-hosts may also contribute to the perpetuation of inoculum between seasons (Knight et al. 2019a) and hence contribute to the perpetuation of specific phenotypes within the population, such as azoxystrobin resistance.

The high frequency of azoxystrobin-resistant isolates in C. beticola populations also highlights the critical need to reduce the reliance of disease management programmes on strobilurins. Azoxystrobin is currently the only registered product for early-season and root disease control primarily targeting Rhizoctonia root rot (Abawi et al. 1986; Gugino et al. 2008; Pethybridge et al. 2018), so there is little potential to eliminate its use from table beet production to prevent continued exposure to both R. solani and C. beticola populations. The lack of alternatives to strobilurins for CLS control means that growers have little choice and their use persists in rotation with propiconazole, even in the wake of this knowledge.

In these trials, propiconazole provided only moderate CLS control. Moreover, a tendency to apply propiconazole several times within a season because of the lack of alternatives imposes additional selection pressure on a pathogen population that is highly genotypically diverse (Vaghefi et al. 2016, 2017a, 2017b; Knight et al. 2019b) and at a high risk of resistance development to additional single site mode of action fungicides. Resistance to DMIs is prevalent in C. beticola populations affecting sugar beet (e.g. Karaoglanidis et al. 2000, 2001, 2002, 2003; Bolton et al. 2012b). In this study, the low frequency (0.7%) of C. beticola isolates with reduced sensitivity to propiconazole and difenoconazole is encouraging because the majority of isolates are still highly sensitive to FRAC group 3. This information is an important reference point for the evaluation of temporal changes in DMI sensitivity in the future. If this phenotype becomes more prevalent, further investigation into the molecular basis of resistance within the local C. beticola population affecting table beet will become important, and a basis for the development of assays for detection and monitoring. Point mutations within the CYP51A1 gene have been associated with resistance to another DMI, tetraconazole, within C. beticola populations affecting sugar beet in the USA (Bolton et al. 2012b), and have led to reduced binding affinity in other plant-pathogenic fungi (Asai et al. 1999; Stammler et al. 2009). Overexpression of ABC and MFS transporters (Hayashi et al. 2002; Hulvey et al. 2012), and overexpression of the CYP51A1 gene leading to increased copy number (Hamamoto et al. 2000; Schnabel and Jones 2001; Villani et al. 2016) have also been reported as the molecular basis of DMI resistance.

The superior CLS control in each of the small-plot, replicated trials from benzovindiflupyr + difenoconazole, and pydiflumetofen + difenoconazole in 2018 suggests the potential for improved in-field performance of fungicides for foliar disease control compared with propiconazole and azoxystrobin. However, registration of either of these products for CLS control on table beet still does not offer the ability to improve the best management practices for strobilurin resistance management by rotation to other site-specific modes of action due to the pre-mixed formulations both containing another FRAC group 3 (difenoconazole) (Reddy 2012). In both trials, difenoconazole provided improved disease control compared with propiconazole irrespective of the number of applications. The range of EC₅₀ values for difenoconazole was also several magnitudes lower than for propiconazole which suggests the improvements in CLS control may be caused by a high toxicity which resulted in differential sensitivity to the DMIs within the C. beticola population. Differential sensitivity to the DMIs has also been reported in C. beticola isolates on sugar beet in Ontario, Canada (Trueman et al. 2017). In the 2017 trial, difenoconazole significantly reduced CLS severity compared with the equivalent number of benzovindiflupyr applications when applied two or three times, but when applied in the pre-mixed formulation (benzovindiflupyr + difenoconazole) significantly greater CLS control was realized. In the 2018 trial, the two pre-mixed formulations were included because pydiflumetofen + difenoconazole was more likely than benzovindiflupyr + difenoconazole to be registered for use in speciality crops, and registration of difenoconazole as a sole active ingredient (as Inspire) appeared unlikely. The significant reduction in CLS severity from pydiflumetofen + difenoconazole compared with propiconazole remains encouraging as an alternative to the conventional standard. For pydiflumetofen + difenoconazole, two applications provided significant reduction in CLS severity compared with a single application, but there was no significant benefit of a third. Two applications of pydiflumetofen + difenoconazole significantly reduced CLS severity by 94.9% and increased the dry weight of the foliage by 165.7% at 93 and 94 DAP, compared with non-treated control plots. For difenoconazole only, there were significant reductions in CLS severity with each additional application.

The moderate disease control obtained when benzovindiflupyr was applied alone in the 2017 field trial is similar to discouraging findings towards insensitivity of the selected SDHIs, including benzovindiflupyr, within the in vitro assays. SDHIs act on ubiquinone binding sites within the mitochondrial succinate dehydrogenase and lead to respiration inhibition (Kuhn 1984; Sierotzki and Scalliet 2013). The EC₅₀ values for each of these SDHIs was estimated to be above the highest concentrations tested for fluopyram (25 µg mL⁻¹), and boscalid, benzovindiflupyr and penthiopyrad (all 20 µg mL⁻¹). Resistance to this fungicide group in other plant-pathogenic fungi is relatively complex and diverse, including amino acid exchanges within the SdhB, SdhC or SdhD proteins (e.g. Avenot et al. 2008, 2012; Avenot and Michailides 2010; Mallik et al. 2014). Relatively poor CLS control in sugar beet has been reported with penthiopyrad in Canada (Trueman 2014) and in an earlier trial on table beet (Pethybridge et al. 2017a). Further research is needed to investigate whether EC₅₀ value estimation in C. beticola is sensitive to assay factors such as the selection of base medium. Significant differences in SDHI sensitivity have been noted within Monilinia fructicola isolates when mycelial growth was evaluated on media of varying nutrient status (Hu et al. 2011). Differences in SDHI sensitivity estimation within isolates have also been described between mycelial growth and conidial germination assays (Schnabel et al. 2010).

Given the recalcitrant nature of azoxystrobin resistance within the NY C. beticola population, care should be taken to reduce the selection pressure from other single site mode of action FRAC groups which threatens the sustainability of a fungicide-based management strategy for CLS in table beet. Further research should evaluate the potential to schedule fungicide applications in periods of highest risk. In the CLS-sugar beet pathosystem, fungicides are applied based on risk predictions from a forecasting model that integrates the effects of temperature and relative humidity (Windels et al. 1998; Khan et al. 2007; Kaiser et al. 2010). A forecasting system for CLS has not been optimized for table beet production in the north-eastern USA, and hence prophylactic applications are made in practice because growers are risk averse as a direct function of the high crop losses and absence of information on cost-effective action thresholds. Informed fungicide timing and initiation, together with improved foundational knowledge about the dominant inoculum sources will enhance the sustainability of the current fungicide-based management for CLS.

Acknowledgements

We thank Carol Bowden, Audrey Klein, Sean Murphy, Nicolette Nault and Alex Silva (listed alphabetically by surname) for excellent technical support.

Supplementary material

Supplemental data for this article can be accessed online here: https://doi.org/10.1080/07060661.2019.1690048.

Additional information

Funding

This research was supported by the United States Department of Agriculture National Institute of Food and Agriculture (USDA-NIFA) Hatch project NYG-625424, USDA-NIFA Specialty Crop Block project SCG 15 002 managed by the New York Farm Viability Institute, and USDA-NIFA Crop Protection and Pest Management Applied Research and Development Program project number 2016-07645.