Fusarium Head Blight of Small Grains in Pennsylvania: Unravelling Species Diversity, Toxin Types, Growth, and Triazole Sensitivity

Fusarium head blight (FHB) is one of the most destructive and widespread diseases of wheat, barley, and other small grain cereals worldwide, causing significant losses to both yield and grain quality (McMullen et al. 2012). During the 1990s, the disease caused a tremendous impact on North American agriculture with economic losses due to FHB estimated at approximately $3 billion in the United States (Windels 2000). Recently, FHB was considered the second most important pest and pathogen in wheat globally (Savary et al. 2019). Direct losses due to FHB occur from failed kernel development, or grains that become shriveled and discolored and with low test weights (Osborne and Stein 2007). In addition, infected grains are usually contaminated with type-B trichothecene mycotoxins, particularly deoxynivalenol (DON), and its acetylated forms 3-acetyl-deoxynivalenol (3ADON) and 15-acetyl-deoxynivalenol (15ADON), as well as nivalenol (NIV) and its acetylated derivative 4-acetyl nivalenol (4ANIV) (Miller et al. 1991). While DON is the predominant mycotoxin contaminating small grains in North America (McMullen et al. 1997), both DON and NIV are common in some parts of Asia, Europe, and South America, likely due to the presence of other toxigenic species (Del Ponte et al. 2015; Ichinoe et al. 1983). These mycotoxins may accumulate in the grains at levels considered unsafe for both human and livestock consumption and pose a serious threat to food security and crop production (Pestka 2010; Rocha et al. 2005).

In North America, an endemic population of the Fusarium graminearum sensu stricto (hereafter F. graminearum) of the 15ADON toxin type has been historically responsible for FHB (Zeller et al. 2003, 2004). However, a displacement of this dominant population by a newly introduced and highly virulent 3ADON population has been documented in the Upper Midwest of the United States and western Canada (Gale et al. 2007; Puri and Zhong 2010; Ward et al. 2008). Similarly, 3ADON genotypes have been reported in surveys conducted from southeastern (North Carolina) to the northeastern (New York [NY]) areas of the United States (Schmale et al. 2011). The NIV toxin type of the F. graminearum species complex (FGSC) is absent or found at very low frequencies in North America (Campbell et al. 2002; Starkey et al. 2007) but was prevalent (79%) in limited small wheat growing areas dominated by rice production such as those in Louisiana (Gale et al. 2011). In addition, a more recently described NX-2 genotype has been found with unusually high frequency in northeastern NY (Lofgren et al. 2018; Varga et al. 2015).

Reports of shifts in toxin profile of FHB populations in the United States raise concerns about the current distribution of these populations, especially in poorly sampled regions such as many areas of Pennsylvania (PA). Previously, Schmale et al. (2011) reported 15ADON as the predominant type (92%) in almost 1,000 F. graminearum isolates obtained from six eastern U.S. states. In that study, 62 isolates were obtained from four commercial winter wheat fields during 2006 in PA, of which 5 represented the 3ADON genotype. Comparisons among the trichothecene genotypes of F. graminearum isolates obtained from different fungal habitats in NY including wheat, corn, aerial populations, and wild spikes and stems, have shown an increased frequency of the 3ADON genotype with a general frequency of 15% (Schmale et al. 2011), 18% (Kuhnem et al. 2015), and 22% (Fulcher et al. 2019). In some of those studies, the 3ADON genotype comprised more than 40% of isolates in some individual fields.

While the reasons associated with significant changes in FHB populations and trichothecene genotype frequency are not entirely clear, several factors have been suggested including migration, host preference, and ecological and environmental conditions (Backhouse 2014; Boutigny et al. 2011; Lee et al. 2009). Data from Ward et al. (2002) previously suggested that trichothecene genotype diversity has been maintained in FHB pathogen populations by balancing selection, indicating that chemotype differences among F. graminearum populations may have fitness consequences. Indeed, some studies showed that 3ADON isolates were more fertile (sexual and asexual) (Ward et al. 2008), more aggressive, and produced more toxin than 15ADON isolates (Foroud et al. 2012; Puri and Zhong 2010; Ward et al. 2008). Conversely, Spolti et al. (2014a) compared 14 different attributes of saprophytic and pathogenic fitness and could not find differences in the 15ADON and 3ADON populations from NY. The reasons for those inconsistencies are not entirely clear but may be likely due to population or isolate-specific toxin profiles that are encountered in some regions.

Continuous surveillance of plant pathogens is critical for assessing risk and defining management strategies, especially in PA. The state currently ranks first in the U.S. for craft beer production, which coincides with an increased demand for locally sourced barley and wheat, further raising concerns about the risk and impact of FHB epidemics and mycotoxin contamination (Schwarz and Horsley 2019). In this context, the main objectives of this study were to (i) determine the composition of FHB-causing species isolated from spikes of wheat and other small-grain cereals as well as overwintering crop residues, in PA, and to assess their mycotoxin profiles, and (ii) compare the trichothecene types of FGSC with regard to several fitness-related traits related to the saprophytic phase of the pathogen, such as mycelium growth, macroconidia production (asexual spores), perithecia and ascospore production (sexual spores), and triazole sensitivity.

MATERIALS AND METHODS

Study area and sampling.

During summer 2018 and 2019, naturally FHB-affected spikes of different small grain hosts (wheat, barley, spelt, rye) were collected from the major cereal-producing regions of PA (defined as South, Central, and North). Samples consisting of symptomatic spikes were collected along 10 arbitrary transects through each field where 10 spikes were collected, totaling 100 spikes per field. Twelve fields (10 wheat and 2 spelt) located in seven counties were surveyed in 2018 and 11 fields (7 wheat, 2 barley, 1 spelt, and 1 rye) in six counties were surveyed in 2019 (Table 1). FHB severity (%), defined as the mean percentage of diseased spikelets per spike, was recorded in 11 and 10 fields of the 2018 and 2019 seasons, respectively. Previous crop information was also obtained from 11 and 7 fields in those respective years (Supplementary Tables S1 and S2).

TABLE 1. Trichothecene genotype frequency of Fusarium graminearum isolates collected from naturally Fusarium head blight-affected spikes of different small grain hosts (wheat, barley, spelt, and rye) across Pennsylvania in 2018 and 2019

Fa	Year	Location	Rb	Host	Nc	Isolates (%)d
						15-ADON	3-ADON	NIV	NX-2
1	2018	Lancaster	South	Wheat	35	100.0	–	–	–
2	2018	York	South	Spelt	34	100.0	–	–	–
3	2018	Armstrong	Central	Wheat	28	89.3	7.1	3.6	0.0
4	2018	Centre	Central	Wheat	140	97.9	2.1	0.0	0.0
5	2018	Lebanon	South	Wheat	3	100.0	–	–	–
6	2018	Centre	Central	Wheat	21	90.5	9.5	–	–
7	2018	Potter	North	Wheat	3	100.0	–	–	–
8	2018	Tioga	North	Wheat	28	92.9	7.1	0.0	0.0
9	2018	Tioga	North	Wheat	3	66.7	33.3	0.0	0.0
10	2018	Tioga	North	Wheat	4	100.0	–	–	–
11	2018	Potter	North	Spelt	4	–	100.0	–	–
12	2018	Potter	North	Wheat	8	50.0	25.0	2.5	0.0
13	2019	Lancaster	South	Barley	25	100.0	–	–	–
14	2019	Centre	Central	Barley	14	100.0	–	–	–
15	2019	Lancaster	South	Rye	6	100.0	–	–	–
16	2019	Lancaster	South	Spelt	6	100.0	–	–	–
17	2019	Lancaster	South	Wheat	1	100.0	–	–	–
18	2019	Potter	North	Wheat	5	100.0	–	–	–
19	2019	Erie	North	Wheat	11	100.0	–	–	–
20	2019	Crawford	North	Wheat	3	100.0	–	–	–
21	2019	Lebanon	South	Wheat	12	100.0	–	–	–
22	2019	Lebanon	South	Wheat	3	100.0	–	–	–
23	2019	Lebanon	South	Wheat	17	88.2	5.9	0.0	5.9

^a Individual fields sampled in each location (F).

^b Region of Pennsylvania from where the isolates were collected.

^c N = total number of isolates.

^d Genotypes 3-acetyldeoxynivalenol (3-ADON), 15-acetyldeoxynivalenol (15-ADON), and nivalenol (NIV).

View as image HTML

Fusarium spp. isolate collections.

In the laboratory, kernels symptomatic of FHB were surface sterilized in 70% ethanol for 1 min, immersed in 5% sodium hypochlorite for 2 min, and rinsed three times in sterile distilled water (Qu et al. 2008). The dried kernels were transferred to potato dextrose agar (PDA) medium supplemented with streptomycin at 10 μg·ml⁻¹. Plates were incubated for 4 days at 25 ± 2°C under continuous darkness. Fragments of mycelia from visibly typical Fusarium spp. colonies were transferred to new PDA plates. Colonies identified both macroscopically and microscopically as Fusarium spp. (Leslie and Summerell 2006) were subcultured on synthetic nutrient agar media (SNA) to obtain pure, single-spore isolates. Isolates were frozen at –80°C in 15% glycerol for long-term storage.

DNA extraction and species identification.

Mycelia for DNA extraction were obtained by growing isolates in potato dextrose broth (PDB) medium and incubating them on an orbital shaker (150 rpm) for approximately 4 days at 25 ± 2°C. Mycelia were filtered through two layers of sterilized cheesecloth, dried on filter papers, and stored at –80°C. DNA was extracted using the MasterPure Yeast DNA Purification Kit (Lucigen Corporation) protocol. The DNA from all isolates was then amplified by PCR using the FGSC-specific primers, Fg16F/Fg16R, which produced polymorphic products (∼450 bp) with DNA from members of the FGSC (Nicholson et al. 1998). Isolates that could not be identified by FGSC-specific primers were identified based on partial sequences of either the translation elongation factor 1α (EF-1α) or RNA polymerase II second largest subunit (RPB2) genes (O’Donnell et al. 1998, 2013). Sequence similarity searches were performed with the BLAST network service of the Fusarium ID database (https://github.com/fusariumid/fusariumid) (Geiser et al. 2004).

Determination of trichothecene genotype.

The trichothecene genotype (3ADON, 15ADON, or NIV) of the isolates previously identified as F. graminearum sensu lato was determined through multiplex PCR assays targeting portions of the TRI3 and TRI12 genes (Starkey et al. 2007). Four separate primers were used to amplify each gene: 3CON, 3NA, 3D15A, and 3D3A for TRI3 and 12CON, 12NF, 12-15F, and 12-3F for TRI12 (Ward et al. 2002). Both multiplex reactions were performed in 25-μl volumes containing 1× PCR buffer, 2 mM MgCl₂, 2 units of Taq DNA polymerase, 0.2 mM concentrations of each deoxynucleotide triphosphate, 0.2 μM concentrations of each primer, and 25 ng of genomic DNA. Cycling conditions consisted of an initial denaturation step at 94°C for 2 min, followed by 25 cycles of 30 s at 94°C, 30 s at 52°C, and 1 min at 72°C (Starkey et al. 2007). The resulting PCR products were separated by electrophoresis in an 1% agarose gel with 1% TBE buffer and visualized under ultraviolet light following staining of the DNA gel with SYBR safe (Life Technologies-Invitrogen). The TRI3 multiplex produced amplicons of approximately 840, 610, and 243 bp, which correspond to NIV, 15ADON, and 3ADON chemotypes, respectively. The TRI12 multiplex produced amplicons of approximately 840, 670, and 410 bp, which correspond to NIV, 15ADON, and 3ADON chemotypes, respectively. The reference isolate PH-1 was used throughout the PCR assays as a positive control for F. graminearum species possessing the 15ADON genotype (Cuomo et al. 2007). Additionally, a TRI1 PCR-RFLP assay (Liang et al. 2014) was used to screen F. graminearum isolates possessing the 3ADON genotype for the NX-2 TRI1 allele, because NX-2 strains also would have a 3ADON trichothecene gene cluster genotype. This assay was performed following Liang et al. (2014).

Trichothecene genotypes from overwintering residues.

The trichothecene genotype (3ADON, 15ADON, or NIV) in a collection of 77 F. graminearum isolates obtained from overwintering residues was also determined. Samples were obtained during winter 2012 from the primary corn-producing regions of PA, with a total of 40 fields sampled across 16 locations (Supplementary Table S3). Only samples showing signs of the pathogen, such as aggregations of light pink/salmon-colored spores (sporodochia) or perithecia, were randomly selected from each field. In the laboratory, the isolation process was conducted as previously described. Isolates were identified based on partial DNA sequencing of the translation elongation factor-1α (EF-1α) gene (O’Donnell et al. 1998). Sequence similarity searches and multiplex PCR assays targeting portions of the TRI3 and TRI12 genes were conducted as previously described.

Mycelial growth.

Thirty-one isolates of F. graminearum were selected from the collection obtained during summer 2018. Fifteen isolates of F. graminearum possessing the 15ADON genotype and 16 possessing the 3ADON genotype were selected to represent the geographic distribution of these two populations across PA. Detailed information for these isolates is shown in Supplementary Table S4. Isolates were grown on PDA for 7 days at 25°C under continuous darkness. Mycelium plugs (6 mm in diameter) of each isolate were then excised from the margins of the colonies and individually placed in the center of PDA plates (90 mm in diameter). Cultures were incubated in growth chambers at 23°C under continuous darkness. To obtain mycelial growth rate (cm²) measurements, plates were photographed on the third and fourth days of incubation by opening plates in a laminar flow cabinet and taking a photo using an iPhone XR camera (12MP) at a standard distance between the mobile device and the colonies. Photographs were analyzed in ImageJ (https://imagej.nih.gov/ij), where the perimeter of the colony was manually annotated on photographs within the ImageJ software and the enclosed area was recorded in square centimeters (Newbery et al. 2020). The mycelial growth rate of each isolate was calculated as the difference in the area covered with mycelia between the fourth and third days, and the area of the agar plug (0.283 cm²) was subtracted from this value. Two replicates (plates) were used per isolate. The experiment was performed twice.

Macroconidia production.

Isolates were grown on SNA media for 7 days under a cycle of 12 h of light and 12 h of darkness at 25°C (Leslie and Summerell 2006). After this period, three discs (6 mm in diameter) were removed from the edges of the developing colonies and immersed in 5 ml of sterile water containing 0.1 ml of 0.001% Tween in a test tube and shaken for 20 s (Nicolli et al. 2018). The spore concentration was quantified by using a hemocytometer and expressed as the number of macroconidia per ml (spores/ml). Two replicates (plates) were used per isolate. The experiment was performed twice.

Perithecia production on carrot agar.

This experiment was conducted following a standard protocol (Cavinder et al. 2012) with some adaptations. Briefly, F. graminearum isolates were grown on PDA for 7 days under continuous darkness at 25°C. Mycelium plugs (6 mm in diameter) of each isolate were then excised from the margins of the colonies and individually placed in the center of carrot agar plates (60 mm in diameter). The plates were incubated in growth chambers under bright fluorescent lights for 7 days at 25°C. After this period, the aerial mycelia were gently removed with a sterile toothpick and an aliquot of 1 ml of 2.5% Tween 20 solution was added to the surface. The plates were then returned to the growth chamber. The formation of perithecia was checked at 3 and 6 days after removal of the aerial mycelium and photographs of the plates were taken. For photography, the plates were opened and photographed in a laminar flow cabinet using an iPhone XR camera (12 MP) at a standard distance between the mobile device and the plates. Photographs were analyzed in ImageJ where the area occupied by perithecia in 1 cm² of the culture medium per plate was quantified and expressed as the percentage of perithecia (%) (Stewart et al. 2016). Two replicates (plates) were used per isolate. The experiment was performed once.

Ascospore production on carrot agar.

This experiment was conducted as previously described for perithecia production. The production of ascospores was evaluated from the sixth day of removal of the aerial mycelia of the plates. Three plugs (1 cm diameter) from each plate were cut out of the medium with a sterile cork borer. Each plug was sliced in half, and the half-plugs were placed on a glass microscope slide with three plugs per slide (Cavinder et al. 2012). Slides were placed in a humidity chamber overnight under lights, and the accumulated spores were washed off the slide with water and quantified using a Neubauer chamber. There were two replicated humid chambers with 31 slides each (one slide per isolate). Two replicates (slides) were used per isolate. The experiment was performed twice.

Fungicide sensitivity assay.

The sensitivity of 31 isolates of F. graminearum to tebuconazole and metconazole was assessed by measuring mycelial growth on PDA amended with increasing concentrations of both fungicides (Becher et al. 2010; Spolti et al. 2014b). A stock solution (100 μg of active ingredient [a.i.]/ml) was prepared by dilution of commercially formulated Folicur 3.6F (38.7% a.i.; Bayer CropScience) and Caramba 90 (90% a.i.; BASF Corporation). Concentrations tested for both fungicides were 0 (nonamended agar-PDA), 0.01, 0.1, 0.5, 1, 10 μg/ml. A mycelial agar plug (6 mm of diameter) from the edge of a culture of each isolate was placed in the center of a Petri dish (90 mm diameter). The plates were incubated under continuous darkness at 25°C. The colony diameter was measured on the fourth and fifth day in two perpendicular directions using a digital caliper, and the agar plug diameter (6 mm) was subtracted. The mycelial growth of each isolate was calculated as the difference between the diameter colonies measured on the fourth and third days. For each combination of isolate-dose-fungicide, two replicates (plates) were used. The experiment was performed twice. All the fitness trait experiments were conducted as a completely randomized design with two replicates, and data from assays conducted two times were combined for analysis.

Data analysis, software, and reproducibility.

All data processing and analyses, as well as graphical work, were performed in R version 3.6.0 (2019-04-26) (R Core Team 2019). Descriptive statistics summarized the frequency of trichothecene genotypes among hosts, years, and locations. Fisher’s exact test (for small sample size) was used to evaluate whether there was a significant association between the trichothecene genotype composition (3ADON, 15ADON, NIV, and NX-2) of the isolates obtained from naturally FHB-affected spikes among years, locations, and host of origin. The same test was also used to compare the frequency of the trichothecene genotypes from the saprophytic and pathogenic phase of the fungal life cycle. The Student’s t test was used to compare mean fitness traits of the 15ADON and 3ADON populations. Statistical significance was considered at the P < 0.05 level. The ‘drm’ function of the ‘dcr’ package of R (Ritz et al. 2015) was used to estimate the effective concentration leading to a 50% reduction of mycelial growth (EC₅₀). To compare the distribution of the EC₅₀ values for tebuconazole and metconazole fungicides among 15ADON and 3ADON genotypes, a nonparametric Kolmogorov–Smirnov test (P = 0.05) was used. A multivariate analysis of variance (MANOVA) was performed in order to compare the overall saprophytic fitness of the two trichothecene genotype groups. In addition, a principal components analysis (PCA) was performed using the mean values of the variables for each isolate, averaged over replicates and experiments. The contribution of each fitness trait to each principal component was also estimated. A correlogram was made using the overall means for each pair of variables using the package ‘corrplot’ (Wei and Simko 2017). The packages ‘FactoMineR’ (Lê et al. 2008) and ‘factoextra’ (Kassambara and Mundt 2017) were used for the PCA. All data gathered for this study as well as the fully annotated computational codes, prepared in R Markdown, were organized as a research compendium publicly available at https://github.com/PSUPlantEpidemiology/paper-FHB-Pennsylvania. A website was generated to facilitate visualization of the commented scripts (https://psuplantepidemiology.github.io/paper-FHB-Pennsylvania).

RESULTS

Species and trichothecene genotypes.

Of the 461 total isolates, 331 and 130 were obtained from 2018 and 2019, respectively. The number of isolates per field ranged from 1 to 140 (mean = 17.73 and median = 9). Most of the isolates (78%) were obtained from wheat, followed by spelt (11%), barley (10%), and rye (1%).

The vast majority of the isolates (n = 418) were molecularly identified as F. graminearum based on the presence of a ∼450-bp amplicon with the Fg16 primer. Thirty-nine isolates (8%) that did not produce the ∼450-bp amplicon were identified as Fusarium avenaceum (n = 26); Fusarium sporotrichioides (n = 10), Fusarium poae (n = 2), and Fusarium fujikuroi species complex (n = 1) based on EF-1α or RPB2 genes sequence data. The identity of four isolates (2%) could not be resolved either based on the sequences of the EF-1α gene or RPB2.

The F. graminearum isolates were mainly of 15ADON type (94.9%; 393/414), followed by the 3ADON (4.1%; 17/414), NIV (0.7%; 3/414), and NX-2 (0.3%; 1/414) types, respectively (Table 1). Fisher’s exact test showed that the frequency of toxin types did not depend on the year (2018 and 2019) (P = 0.055). In 2018, the overall frequency was 93.9% (292/311) for the 15ADON genotype, 5.1% (16/311) for the 3ADON genotype, and 1.0% (3/311) for the NIV genotype. In 2019, 98.0% (101/103) of the isolates were characterized as the 15ADON genotype, 1.0% (1/103) as the 3ADON genotype, and 1.0% (1/103) as the NX-2 genotype. When genotype frequencies were compared across locations within each year, differences were only observed in 2018 (Fisher’s exact test, P < 0.001). With the exception of samples obtained from Potter County in 2018, where the proportion of 15ADON and 3ADON genotypes were similar (46.7% and 40.0%, respectively), 15ADON was the predominant genotype in other locations in 2018 (Table 1). In 2019, there were no differences in the frequency of the genotypes across locations (P = 0.651). The NIV genotype was detected only in 2018 (0.7%; 3/414). No difference in the trichothecene genotype frequency was found among sampled hosts (P = 0.541), where the 15ADON represented the majority of the isolates collected from wheat (94.8%; 308/325), followed by 3ADON (4.0%; 13/325), NIV (0.9%; 3/325), and NX-2 (0.3%; 1/325). Only the 15ADON genotype was recovered from spelt, barley, and rye hosts.

Among the 77 isolates obtained from overwintering residues, 72 were collected from maize stubble, two from wheat residue, and three from Johnsongrass (Sorghum halepense). All 77 F. graminearum isolates possessed the 15ADON type (Supplementary Table S3). The frequency of the toxin types did not depend on the source (spikes or residues) (P = 0.309).

Asexual and sexual fertility.

The 3ADON isolates grew significantly faster in vitro and produced more macroconidia than 15ADON isolates (Fig. 1A and B; Table 2). Mean perithecia production did not differ between the two genotypes when compared at 3 (P = 0.378) and 6 days (P = 0.598), but there was a difference between the two evaluation days (P < 0.01) (Fig. 1D). Similarly, no statistical difference was found for mean ascospore production between the two trichothecene genotypes (P = 0.449) (Fig. 1C; Table 2).

Download as PowerPoint

TABLE 2. Summary statistics and comparison of measures of saprophytic traits of 31 Fusarium graminearum isolates obtained from naturally Fusarium head blight-affected spikes of small-grain cereals in 2018 possessing either a 3-acetyl-deoxynivalenol or a 15-acetyldeoxynivalenol trichothecene genotype

	Trichothecene genotypea
Variable	n	15-ADON	n	3-ADON	P valueb
Mycelial growthc	60	6.87 ± 4.23	64	9.00 ± 4.61	0.008
Spore productiond	60	26.58 ± 16.53	64	41.07 ± 18.89	0.000
Ascospore productione	60	38.42 ± 31.77	64	42.89 ± 33.76	0.449
Perithecia productionf
3 DAIg	30	5.11 ± 6.79	32	6.95 ± 9.40	0.378
6 DAI	30	11.49 ± 10.04	32	12.98 ± 12.00	0.598
EC50 tebuconazoleh	15	0.14 ± 0.14	16	0.11 ± 0.11	0.847
EC50 metconazolei	15	0.02 ± 0.03	16	0.02 ± 0.02	0.498

^a Isolates were identified to trichothecene genotype using PCR assays targeting TRI3 and TRI12 genes (Starkey et al. 2007). In total, 31 Fusarium graminearum isolates (n_15ADON = 15; n_3ADON = 16) were used in this study. Data shown are means ± standard deviation. 3-ADON, 3-acetyldeoxynivalenol; and 15-ADON, 15-acetyldeoxynivalenol.

^b P value for comparisons of trichothecene genotypes based on Student’s t test (mycelial growth, spore production, ascospore production, perithecia production) and Kruskal–Wallis test (EC₅₀ tebuconazole, EC₅₀ metconazole).

^c Mycelial growth (centimeters squared) was determined on potato dextrose agar at 25°C incubated for 4 days under continuous darkness.

^d Macroconidia production (×10⁴ macroconidia/ml) on synthetic nutrient agar media for 7 days under a cycle of 12 h of light and 12 h of darkness at 25°C.

^e Ascospores (×10⁴ ascospore/ml) produced from carrot agar and discharged and harvested from a glass microscope slide.

^f Perithecia production (percentage) in 1 cm² of CA after 3 and 6 days of incubation at 25°C under constant lights.

^g Days after incubation (DAI) of plates after removal of aerial mycelium.

^h Effective concentration of tebuconazole fungicides (micrograms per milliliter) that reduces 50% of the mycelial growth of each of the isolates.

ⁱ Effective concentration of metconazole fungicides (micrograms per milliliter) that reduces 50% of the mycelial growth of each of the isolates.

View as image HTML

Fungicide sensitivity.

EC₅₀ values ranged from 0.01 to 0.41 μg/ml (0.13 ± 0.12) for tebuconazole and from 0.00 to 0.13 μg/ml (0.02 ± 0.03) for metconazole. EC₅₀ estimates for tebuconazole were greater (P < 0.01) than metconazole (Fig. 2A). However, no statistical differences were observed between the two trichothecene genotypes for the EC₅₀ estimates for tebuconazole (P = 0.847) and metconazole (P = 0.498) (Table 2).

Download as PowerPoint

Multivariate analysis.

Based on MANOVA, there was no difference between the 3ADON and 15ADON trichothecene genotypes when examining the combined dependent variables (P = 0.239). The correlation analysis for each pairwise comparison (n = 21) for all seven variables used to characterize fitness of the two trichothecene groups (Table 2) indicated that there were seven significant (P < 0.01) correlations (Supplementary Fig. S1). Overall, mycelial growth rate was positively correlated with macroconidia production (P = 0.01). However, the mycelial growth rate and macroconidia production were negatively associated with the EC₅₀ estimates for tebuconazole and metconazole. As expected, perithecia production on day 3 was highly correlated with perithecia production on day 6, as well as the EC₅₀ estimates for tebuconazole and metconazole.

The PCA suggested that EC₅₀ estimates for tebuconazole, mycelial growth rate, and EC₅₀ estimates for metconazole contributed the most to PC1: 22.7, 21.4, and 18.5%, respectively, of the total, with 34.4% of all variation in the analysis explained by this first PC (Fig. 2B). The second PC explained 26.9% of the variation, with perithecia production on days 3 and 6 contributing to 33.0 and 31.9%, respectively. Thus, the first two PCs explained 61.3% of the variation among the isolates. The third PC explained 15.1% of the variation among isolates, with a contribution of 58.0% from the variable ascospore production.

DISCUSSION

This is the first comprehensive survey conducted in PA to determine the composition and spatial distribution of FHB pathogens obtained from small grain cereals, mainly wheat, and overwintering residues. Our results corroborate previous findings of the large dominance of F. graminearum of the 15ADON genotype in the Eastern United States (Cowger et al. 2020; Kuhnem et al. 2015; Schmale et al. 2011), which was consistent in samples from both symptomatic spikes and overwintering residues. These isolates may be part of a genetically divergent North American population of F. graminearum sensu stricto that causes FHB and typically produces the 15ADON mycotoxin (Gale et al. 2007; Kelly and Ward 2018).

The non-FGSC isolates were found in very low frequency (<10%), but they are known to produce different types of mycotoxins, which could be a concern if their frequency increases in the future. For example, F. avenaceum is the most important species causing FHB within the F. tricinctum species complex (FTSC) and has the ability to produce what is referred to as “emerging mycotoxins,” such as enniatins, moniliformin, and beauvericin (Beccari et al. 2018; Jestoi 2008; Logrieco et al. 1998, 2002; Yli-Mattila et al. 2002). Differently, F. sporotrichioides, which is known to belong to a broader Fusarium sambucinum species complex (FSAMSC), within which FGSC is included (O’Donnell et al. 2013), is able to produce type A trichothecene mycotoxins, including T-2 and HT-2, and these are considered to be more toxic to humans and animals than type B trichothecenes (Krska et al. 2001; Logrieco et al. 1990). Another species is F. poae, which also belongs to the FSAMSC and are known to produce mostly NIV mycotoxin (Stenglein 2009; Thrane et al. 2004; Vogelgsang et al. 2008). Finally, members of the F. fujikuroi species complex (FFSC) are commonly found infecting rice grains, and like the FTSC species, are also able to produce a wide range of mycotoxins including fumonisins, moniliformin, beauvericin, fusaproliferin, enniatins, and fusaric acid (Bacon et al. 1996; Logrieco et al. 1998; Marasas 1986; Moretti 1996; Nicholson et al. 2004; Rheeder et al. 2002).

With regards to their spatial distribution, only three of these other species (nonF. graminearum) were found in southern PA, where most of the wheat is grown. Similar results were observed by Bec et al. (2015), where 5 of 68 isolates were identified as members of the FTSC in Kentucky. In this situation, winter wheat crop is traditionally planted in a rotation following corn. More recently, Cowger et al. (2020) showed that FTSC species account for 11.3% of the Fusarium isolates obtained from symptomatic wheat spikes in the 2013 to 2014 growing seasons in North Carolina, United States. In that case, members of the FTSC, especially F. avenaceum, were frequent or even dominant in some fields. Clearly, data on non-FGSC pathogens causing FHB in wheat in the United States is lacking. The recent reports reinforce the need for continued monitoring of regional FHB pathogen populations in PA, especially to characterize possible trends in pathogen diversity that may result from changes in cropping systems and agricultural practices across the state.

The frequency of the trichothecene genotypes reported in our study is in agreement with previous reports in the Northeastern United States where 15ADON was the predominant genotype, followed by 3ADON and NIV (Cowger et al. 2020; Fulcher et al. 2019; Kuhnem et al. 2015; Schmale et al. 2011). Schmale et al. (2011) reported 15ADON as the dominant genotype across the Eastern United States (92%). Interestingly, the frequency of the 15ADON (91.9%) and 3ADON (8.1%) types of 62 isolates collected from four commercial winter wheat regions in PA, more than a decade ago in the Schmale et al. (2011) study, was similar to that in our study. Moreover, only one 3ADON type isolate was found in the Southern fields. These consistencies in the results may be an indication that FHB pathogen populations in PA are following a unique evolutionary trajectory. Correspondingly, both reports from PA contrast with western Canada and the Upper Midwest of the United States, where an increased frequency of a F. graminearum population with the 3ADON genotype has been observed in studies conducted over years (Kelly et al. 2015; Liang et al. 2014; Puri and Zhong 2010; Ward et al. 2008). The low recovery of NIV genotype isolates in our study is in agreement with previous findings for the Northeastern United States, where a north-south cline of increasing frequency of the NIV genotype was observed, with a large population of NIV producers observed in Louisiana (Fulcher et al. 2019; Gale et al. 2011; Schmale et al. 2011). A recent study conducted in North Carolina with 2,197 isolates also showed lower frequency of the NIV producing isolates (2.2%), with no more than 12% observed in a single field (Cowger et al. 2020). Our results also contrast the study conducted by Lofgren et al. (2018) in neighboring New York State, where 42.1% of a population of 133 isolates previously identified as 3ADON genotype had a TRI1 allele genotype consistent with NX-2 producing strains, while only one isolate was identified as NX-2 in our study.

It has been suggested that the distribution of trichothecene genotypes and FGSC species in certain geographic regions may be shaped by differences in host and/or ecological preferences and cropping system (Dill-Macky and Jones 2000; Lee et al. 2009; Yang et al. 2018). In South America, Fusarium meridionale and Fusarium cortaderiae species, both possessing the NIV genotype, were more frequently obtained from maize, and F. graminearum with the 15ADON or 3ADON genotypes from wheat (Del Ponte et al. 2015; Kuhnem et al. 2016; Sampietro et al. 2011). The indication of host preference was also reported by Boutigny et al. (2011) in South Africa, where Fusarium boothii was the exclusive species associated with maize, and F. graminearum accounted for more than 85% of the FGSC isolates obtained from wheat and barley. Conversely, our results showed similar frequencies of the 15ADON genotype obtained from maize stubble in 2012 and from symptomatic spikes in the 2018 and 2019 growing season. While in the NY study (Kuhnem et al. 2015), the 3ADON genotype was also recovered from maize stubble, only the 15ADON genotype was found in the stubble-borne isolates from PA.

In this study, the Fusarium isolates were collected from locations with unique climate conditions and cultural practices. For example, Lancaster, Lebanon, and York counties are in a warmer region of southern PA, where traditional corn-wheat-soybean rotations are very common. In this area, only one isolate of F. graminearum with the 3ADON genotype and one with the NX-2 genotype were recovered. In contrast, Potter County is located in a colder region and at higher altitudes in Northern PA where different crop species are included in the rotations compared with the other regions. We found the greatest frequency of the 3ADON genotype in Potter County. The ecological factors driving dynamics of FHB pathogen populations in PA are still not clear. However, it has been suggested that the geographic distribution and population dynamics among North American F. graminearum populations are influenced by a complex ecological and adaptive landscape (Kelly et al. 2015). In addition, inherent regional differences in the landscape also seem to have influenced the composition of the trichothecene genotypes in NY (Kuhnem et al. 2015).

A previous study suggested that 15ADON genotype is dominant in corn-producing regions (Kelly et al. 2015). The authors reported higher frequency of the 15ADON genotype (78%) in the eastern Canadian provinces of Québec and Ontario, which produces 26 and 63% of the corn crop in the country, respectively (USDA 2020), than in the Maritime provinces and western provinces. Likewise, the 15ADON genotype was predominant (94.7%) among 713 isolates collected from the corn belt region in the Upper Midwestern United States (Gale et al. 2007). In that case, the 3ADON genotype was only detected in a transitional agricultural land area located east of the Red River Valley that forms part of the border between North Dakota and Minnesota. Similar results were observed in the studies conducted in NY (Fulcher et al. 2019; Kuhnem et al. 2015), where the 15ADON genotype predominated on symptomatic wheat spikes in the major maize and wheat production areas of Central NY, while the 3ADON genotype was more frequently associated with natural sites compared with agricultural sites. A similar trend was observed in our study with predominance of the 15ADON genotype in Southern PA, where maize plays a major role in the crop rotation system with wheat. In addition, for many farmers who neighbor small grains growers in PA, continuous corn cropping is common, which may promote inoculum build-up with 15ADON toxin.

In general, the F. graminearum isolates of the two toxin types could not be differentiated by many of the saprophytic fitness traits evaluated in this study. However, mycelial growth rate was faster for isolates belonging to the 3ADON genotype. They also produced more macroconidia in vitro than 15ADON isolates. Ward et al. (2008) suggested that the greater fecundity and growth rate of the 3ADON population that supposedly displaced the 15ADON population in western Canada were indicative of an adaptive fitness advantage. However, Milgroom (2015) argues that in vitro assay estimates are often poor predictors of fitness in natural field environments. Contradictory results were reported by Spolti et al. (2014a), who found that the mycelial growth rate did not differ between the 3ADON and 15ADON populations from NY at any tested temperature ranging from 15 to 30°C. In our study, the two toxin types were similar with regard to sexual fertility, as suggested by measures of perithecia formation and ascospore production on carrot agar. Spolti et al. (2014a) also observed similar ascospore production on corn stalk between both genotypes, but this was not true for ascospore production on carrot agar, where 3ADON isolates were more fertile.

Although the differential sensitivity to fungicides could be one of the factors related to possible adaptive fitness advantages of 3ADON isolates compared with 15ADON isolates (Spolti et al. 2014a), we found no statistical differences in EC₅₀ estimates for tebuconazole and metconazole between the two toxin types. However, a greater variability in the EC₅₀ was evident for the sensitivity levels to tebuconazole. It is not clear whether the extended use of tebuconazole against F. graminearum populations is a factor leading to the more dispersed EC₅₀ values or if the populations are naturally variable (Becher et al. 2010; Spolti et al. 2012, 2014b). Regardless of the individual effect of each saprophytic trait and sensitivity of triazole-based fungicides to the 3ADON and 15ADON genotypes, results from the multivariate analysis showed that there was no clear distinction in fitness attributes between isolates of the 3ADON and 15ADON genotypes. These results agree with those by Spolti et al. (2014a) who compared 14 different attributes of saprophytic and pathogenic fitness and reported no differences between populations of F. graminearum with 15ADON and 3ADON genotypes from NY.

This study provides the first full report on the distribution and composition of the trichothecene genotypes of F. graminearum isolates obtained from two fungal habitats in PA, including naturally FHB-affected spikes of different small grain hosts (wheat, barley, spelt, rye), and overwintering residues, especially maize stubble. Results indicated that the regional population of F. graminearum is composed mainly of the 15ADON genotype in frequency that agrees with reports from more than a decade ago, suggesting no shift in the population causing FHB in PA. Whether the north-south cline of increasing frequency of 3ADON genotype in PA depends on cropping system and/or environmental factors merits further investigation. It would be useful to further investigate the presence of the recently described NX-2 chemotype among North American F. graminearum populations (Liang et al. 2014; Varga et al. 2015). Our findings expand knowledge of the trichothecene diversity in PA, highlight the need to further explore the genetic diversity and population structure analysis in future studies in order to assist in answering questions about the sustainability of resistant cultivars and fungicides used to manage FHB in PA.

ACKNOWLEDGMENTS

We thank the members of the Penn State Field and Forage Crops Extension Team for collecting small grains samples across Pennsylvania. E. M. Del Ponte is thankful to CNPq for the research fellow’s support. This is a cooperative project with the U.S. Wheat & Barley Scab Initiative. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors and do not necessarily reflect the view of the U.S. Department of Agriculture.