Defining Fungicide Resistance Mechanisms in the Corynespora cassiicola Population from Mississippi Soybean

Target spot is a fungal disease affecting soybean (Glycine max [L.] Merr.) caused by Corynespora cassiicola (Berk. & M.A. Curtis) C.T. Wei. Target spot is primarily a disease of leaflets but can also affect petioles, stems, pods, seed, hypocotyls, and much less frequently the roots (Godoy 2015). The disease was first documented in cowpea (Vigna unguiculata [L.] Walp) and soybean plants in China in the early 1900s (Wei 1950). In 1944, target spot was observed in the United States causing defoliation of cowpea in Louisiana, North Carolina, and South Carolina as well as affecting soybean in Florida (Olive et al. 1945). Since then, target spot has been reported from soybean producing countries worldwide (Almeida et al. 1976; Godoy 2015; Ploper and Ramallo 1988; Seaman et al. 1965). The symptoms associated with target spot are generally reddish-brown, round to irregularly shaped lesions on leaflets that vary from pin-point specks to mature spots, and 10 to 15 mm or more in diameter (Godoy 2015). Mature lesions oftentimes develop concentric light and dark necrotic rings, giving them an appearance that leads to the common name target spot and are most often surrounded by a yellow halo. Severely infected leaflets prematurely senesce (Godoy 2015).

Recently, target spot has resurged as an economically important disease across the soybean production systems in Argentina, Brazil, and the southern United States due to monoculture farming, adoption of conservation tillage practices, the use of high-yield producing susceptible cultivars, lack of crop rotation, and subtle changes in rainfall patterns (Allen 2017; Edwards Molina et al. 2022; Koenning et al. 2006; Ploper et al. 2011; Xavier et al. 2013). Historically, yield losses between 18 and 32% were recorded for susceptible soybean cultivars in the Mississippi Delta (Hartwig 1959). However, more recently, yield losses in commercial soybean fields were estimated between 20 and 40% in South Carolina (Koenning et al. 2006). In 2019, the estimated yield losses reached 12,029 metric tons in 29 soybean producing states in the United States (Bradley et al. 2021). In Brazil, yield losses caused by target spot were reported to range from 11 to 42% in fungicide trials during the 2012 to 2016 growing seasons (Edwards Molina et al. 2019). The widespread resurgence of target spot is likely to pose a threat to soybean production in the southern United States. Therefore, management of target spot in the Mississippi soybean production system has recently gained attention.

Presently, target spot management relies on foliar fungicides since resistant cultivars are not commercially available (T. Allen, personal communication). Registered fungicides for target spot management belong to multiple chemical classes with different modes of action (MOA) including demethylation inhibitors (DMI), methyl benzimidazole carbamates (MBC), quinone outside inhibitors (QoI), and succinate dehydrogenase inhibitors (SDHI). QoI fungicides with broad-spectrum activity have been widely used by soybean farmers as either stand-alone products or a component of premix products generally with a DMI fungicide to manage foliar diseases and potentially provide yield benefits even in the absence of visible disease symptoms widely considered to be a “plant health” benefit (Allen 2013; Bradley and Sweets 2008; Dorrance et al. 2010; Wise and Mueller 2011). From 2005 to 2015, the usage of QoI fungicides including azoxystrobin, fluoxastrobin, picoxystrobin, pyraclostrobin, and trifloxystrobin increased by 3.3-fold to approximately 800 metric tons in 28 soybean producing states (Bandara et al. 2020). All QoI fungicides exhibit a common and highly specific biochemical MOA. Specifically, QoI fungicides inhibit mitochondrial respiration by binding to the quinone outside (Qo) site of the cytochrome (cyt) bc₁ complex III and block electron transfer between cyt b and cyt c₁, which significantly reduces ATP production in fungi (Bartlett et al. 2002; Fernández-Ortuño et al. 2010). Consequently, because of the specificity, the Fungicide Resistance Action Committee (FRAC) considers the QoI fungicide class to have a high risk for the development of fungicide resistance within targeted organisms (FRAC 2022).

In Mississippi, one crop production practice that has become more commonplace since the mid-2000s is an automatic foliar fungicide application timed between the beginning pod (R3) and full pod (R4) growth stages (Standish et al. 2015). Stand-alone QoI fungicides such as azoxystrobin (as Quadris, Syngenta Crop Protection) and pyraclostrobin (as Headline, BASF Corporation) were the products of choice at the R3/R4 application timing for general foliar disease management and perceived yield benefits (Allen 2013). However, QoI fungicide resistance has become an increasing threat to profitable soybean production likely as a result of automatic fungicide applications and the continued use of fungicides with a single-site MOA (Sierotzki and Stammler 2019; Standish et al. 2015). Since the introduction of azoxystrobin in 1996, field resistance to QoI fungicides has been reported in over 50 fungal species from 35 genera worldwide (FRAC 2020). For example, QoI resistance has been identified in soybean pathogens from the United States including Cercospora kikuchii (Tak. Matsumoto & Tomoy.) M.W. Gardner (Price et al. 2015) and C. cf. flagellaris Ellis and Martin (Albu et al. 2016) which cause Cercospora leaf blight, C. sojina Hara (Standish et al. 2015; Zhang et al. 2012) which causes frogeye leaf spot, C. cassiicola (Rondon and Lawrence 2019; Smith et al. 2021; Wang et al. 2022), and Septoria glycines Hemmi which causes Septoria brown spot (Neves et al. 2018, 2022). As a result, the continuous monitoring of QoI fungicide sensitivity is needed to help determine the distribution and frequency of resistant pathogen populations before failure to control soybean diseases occurs in the field.

Currently, three amino acid substitutions have been reported in the cytb gene that confer variable levels of resistance to QoI fungicides in plant pathogens (Sierotzki and Stammler 2019; Sierotzki et al. 2007). The most important and common alteration is a single nucleotide base change from guanine to cytosine at codon position 143, which leads to the substitution of glycine with alanine known as G143A (Ishii et al. 2007). The G143A substitution confers the greatest level of resistance and frequently results in a complete loss of sensitivity to QoI fungicides (Fernández-Ortuño et al. 2008; Gisi et al. 2002). The G143A substitution has been reported in several soybean fungal pathogens such as C. kikuchii (Sautua et al. 2020) and C. cf. flagellaris (Albu et al. 2016), C. sojina (Standish et al. 2015; Zhang et al. 2012), and S. glycines (Neves et al. 2018). More specifically, the G143A substitution has previously been reported and identified in C. cassiicola from Alabama, Arkansas, Mississippi, and Tennessee with the initial published report in 2019 from Alabama but the first identification occurring in Arkansas and Mississippi in 2016 (Brochard et al. 2017; Rondon and Lawrence 2019; Smith et al. 2021; Wang et al. 2022). The additional amino acid substitutions that can play a role in conferring resistance to fungicides include an amino acid substitution from phenylalanine to leucine at position 129, known as F129L, which is generally associated with moderate resistance. Fungal pathogens possessing the F129L substitution can be managed with either more frequent applications or increased rates of QoI fungicides (Fernández-Ortuño et al. 2008; Gisi et al. 2002). One last amino acid substitution from glycine to arginine at position 137, known as G137R, has been linked to a low level of QoI resistance (Sierotzki et al. 2007). Multiple amino acid substitutions relevant to QoI resistance can occur in the same species. For example, G143A, F129L, and G137R were all detected in Pyrenophora tritici-repentis (Died.) Drechsler in wheat (Sierotzki et al. 2007). However, the G143A substitution was considered having the greatest impact on field-level disease management based on its association with high resistance factors and the failure to control G143A-carrying isolates with QoI-based fungicides applied at a full labeled rate. At present, neither the F129L nor G137R substitution has been reported from the fungi responsible for any of the soybean diseases previously reported.

Corynespora cassiicola is considered by the FRAC to be a high-risk pathogen for fungicide resistance development (FRAC 2019). C. cassiicola isolated from soybean in Alabama, Arkansas, Mississippi, and Tennessee has recently been identified as containing the G143A substitution conferring resistance to the QoI fungicides (Rondon and Lawrence 2019; Smith et al. 2021; Wang et al. 2022), which is consistent with the previous findings in Brazil (Basso et al. 2015; Teramoto et al. 2017). However, limited research has been done to characterize QoI fungicide resistance in C. cassiicola throughout the Mississippi soybean production system over time and to consider the occurrence and severity of fungicide resistance given the widespread use of fungicides. Therefore, the objectives of this research were to (i) collect isolates of C. cassiicola from Mississippi soybean fields, (ii) determine the frequency and pattern of QoI resistance in the C. cassiicola population, and (iii) explore the nature of the molecular mechanisms responsible for fungicide resistance.

Materials and Methods

Target spot-infected leaf sampling and C. cassiicola isolation

Soybean leaflets exhibiting target spot symptoms, characterized by red-brown lesions surrounded by a faint yellow halo in the lower-to-middle plant canopy, were collected from commercial fields throughout Mississippi between 2019 and 2021 regardless of field history related to soybean production, target spot severity, or previous fungicide application practices. Soybean fields at reproductive growth stages between beginning pod (R3) and beginning maturity (R7) were scouted to collect representative samples. Additional isolates, previously collected during 2016 in a similar fashion as outlined above, supplemented the more current collections. Five Mississippi geographic regions, differing in number of counties as well as overall soybean production, were previously defined across the state to group counties geographically. The defined regions, number of counties within each region, and 3-year average in harvested soybean hectares in each region (2019 to 2021), based on the USDA – National Agricultural Statistics Service (USDA-NASS; www.nass.usda.gov), including the Capital (16 counties; 21,184 ha), Coast (six counties; 799 ha), Delta (15 counties; 610,503 ha), Hills (16 counties; 85,499 ha), and Pines (29 counties; 48,936), were adopted to relay information on soybean production and more clearly capture the potential distribution of fungicide resistance within each region (Standish et al. 2015). In general, regions were created based on geographical grouping within the state of Mississippi. General information related to each sampling location was collected including the county of collection, GPS coordinates, and sampling date. An attempt was made to sample several unique field locations from within each county. However, in some counties, a limited number of field locations made this difficult, especially in the Coast region. A composite sample from each distinct field consisted of approximately 10 to 15 individual leaflets exhibiting target spot symptoms randomly selected from the lower-to-middle plant canopy. In situations where a county contained a low number of soybean fields, samples were collected from separate fields as delimited by road or turn-row. Sampled leaflets were placed in a plastic freezer storage bag, transported to the laboratory on ice in a cooler, and preserved at 4°C until pathogen isolation, generally within 7 days.

The section of the leaf tissue at the margin of a lesion was excised (approximately 2 × 2 mm), surface-disinfested in a 0.05% sodium hypochlorite solution for 5 min, rinsed three times with sterile distilled water (SDW), and subsequently blotted dry with a sterile paper towel. Two sections from each representative leaflet sample were separately placed onto 2% water agar media (20 g agar per liter of water) amended with streptomycin (100 μg/ml) in Petri dishes (60 × 15 mm) (Fisher Scientific Inc., Pittsburgh, PA). Multiple isolates were recovered from each location to help in obtaining a broader picture of fungicide efficacy. Culture plates were incubated in a growth room under 24 h of continuous fluorescent light at 25°C for 3 days. Sporulating C. cassiicola isolates were subcultured to a Petri dish containing quarter-strength potato dextrose agar (qPDA; 9.75 g of PDA powder and 11.25 g of agar per liter of water; Fisher Scientific Inc., Pittsburgh, PA) and incubated in the growth room with an environment as described above to enhance production of conidia (Sumabat et al. 2018). Species confirmation of all the isolates was initially based on conidia morphology using light microscopy (×400; Ellis and Holliday 1971) and subsequently confirmed by ITS-based sequencing (ITS1/ITS4; White et al. 1990). Five-day-old C. cassiicola subcultures were used to prepare a conidial suspension; 3 ml of SDW was pipetted onto a qPDA plate containing each isolate, and the mycelial surface was scraped to dislodge conidia using a sterile cell spreader. The conidial suspension of each isolate was filtered through four layers of cheese cloth and deposited into a 15-ml Falcon tube (Fisher Scientific Inc., Pittsburgh, PA). After vortexing at 3,400 rpm for 5 s, 500 μl of conidial suspension representing each isolate was dispensed onto a water agar plate (100 × 15 mm) and spread across the surface using a sterile cell spreader. Excess suspension was poured out of the plate prior to incubation. Culture plates were incubated at 25°C and examined after 4 h for evidence of conidial germination under a microscope placed in the biosafety cabinet. A germinating conidium of each C. cassiicola isolate was aseptically transferred onto a PDA (39 g PDA per liter of water) plate and incubated in the growth room with an environment as described above for 5 days.

For isolate preservation, each monoconidial C. cassiicola isolate was subcultured to two qPDA plates (60 × 15 mm) and stored at 4°C. In addition, six 6-mm-diameter agar plugs from the leading edge of a PDA plate containing mycelia and conidia were removed and placed in a sterilized 1.5-ml microcentrifuge tube. The agar plugs were covered with a 15% glycerol solution and placed in a −80°C freezer for long-term storage. The preservation of each isolate was repeated in at least two centrifuge tubes.

Genomic DNA extraction from all C. cassiicola isolates was conducted using a fungi/yeast genomic DNA isolation kit (Norgen Biotek Corp., Ontario, Canada). The manufacturer’s instructions were followed with some modifications. Briefly, 2 ml of 0.9% (w/v) sodium chloride collection solution was poured onto each 6-day-old culture plate, and the mycelial surface was subsequently scraped using a sterile cell lifter (Corning Inc., Corning, NY). One milliliter of conidia and mycelia solution was transferred to a 1.5-ml microcentrifuge tube (Fisher Scientific Inc., Pittsburgh, PA) and pelleted by centrifugation at 14,000 rpm for 1 min. Five hundred microliters of lysis buffer was added to each tube, and the mixture was transferred to a bead tube and vortexed for 5 min at maximum speed followed by incubation in a water bath at 65°C for 10 min. After centrifugation for 2 min at 14,000 rpm, the supernatant was transferred to a fresh microcentrifuge tube with an equal volume of 96% ethanol and vortexed. The lysate was placed in the spin column and centrifuged at 10,000 rpm for 1 min followed by two washes with 500 μl of wash buffer. The DNA was subsequently eluted in 50 μl of elution buffer by spinning at 10,000 rpm for 2 min, and the extracted DNA was quantified using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific Inc., Waltham, MA). DNA samples were stored at −20°C prior to additional further processing.

Polymerase chain reaction (PCR) was conducted to amplify the DNA of C. cassiicola using forward primer ITS1 (5′-TCCGTAGGTGAACCTGCGG-3′) and reverse primer ITS4 (5′-TCCTCCGCTTATTGATATGC-3′; White et al. 1990). Fifty-microliter reactions consisting of 28.0 μl nuclease free water, 10.0 μl 5× Green GoTaq Flexi buffer (Promega, Madison, WI), 4.0 μl MgCl₂ (25 mM, Promega, Madison, WI), 0.8 μl PCR nucleotide mix (10 mM, Promega, Madison, WI), 3.0 μl of each primer (5 μM, Eurofins Genomics LLC, Louisville, KY), 0.2 μl GoTaq DNA polymerase (5 u/μl, Promega, Madison, WI), and 1.0 μl genomic DNA (10 to 100 ng/μl) were prepared. PCR was performed in a T100 Thermal Cycler (Bio-Rad Laboratories Inc., Hercules, CA) with the following cycling conditions: an initial denaturation at 94°C for 2 min; 35 cycles of 94°C for 30 s, 55°C for 30 s, and 72°C for 45 s, followed by 72°C for 5 min. Amplification of fragments of the expected sizes was confirmed by electrophoresis in a 1.2% (wt/vol) agarose gel run in 1× Tris-Borate-EDTA (TBE) buffer, staining with ethidium bromide, and visualization of PCR products under UV light. PCR product purification was done using ExoSAP-IT PCR product cleanup reagent (Thermo Fisher Scientific Inc., Waltham, MA). DNA fragments of all C. cassiicola isolates were submitted to Eurofins Genomics LLC (Louisville, KY) for Sanger sequencing. The ITS1 and ITS4 primers were used for sequencing. The amplicons of 20 total isolates were sequenced in both directions with the remainder of the isolates sequenced using only ITS1 due to overall cost concerns. The obtained sequences were compared with those currently available in GenBank using BLASTn (http://www.ncbi.nlm.nih.gov/BLAST).

Characterization of C. cassiicola by PCR-restriction fragment length polymorphism

Genomic DNA, previously extracted as described above, of all C. cassiicola isolates was used. A specific primer set of cct-cytb-F (5′-TATTATGCGGGATGTAAATAATGG-3′)/cct-cytb-R (5′-TAATGAGAAGAATCTATTTAATGTAGCA-3′; MacKenzie et al. 2020) was used to amplify an approximately 346-bp long region of the cytb gene that includes all three codons (127, 139, and 143) reported to be involved in conferring resistance to the QoI fungicides. The PCR master mix included 28.8 μl nuclease free water, 10.0 μl 5× Green GoTaq Flexi buffer, 3.0 μl MgCl₂ (25 mM), 1.0 μl PCR nucleotide mix (10 mM), 3.0 μl each primer (5 μM), 0.25 μl GoTaq DNA polymerase, and 1.0 μl genomic DNA in a final reaction volume of 50 μl. PCR was performed in a T100 Thermal Cycler with the following cycling conditions: an initial denaturation period of 95°C for 2 min; 35 cycles of denaturation at 95°C for 30 s, annealing at 54°C for 30 s, and elongation at 72°C for 40 s; and a final elongation period of 72°C for 5 min. An 8-μl aliquot of each PCR reaction was electrophoresed in a 1.5% (wt/vol) ethidium-bromide agarose gel in 1× TBE buffer to visually confirm the presence of the amplicon using UV light illumination.

A restriction fragment length polymorphism (RFLP) method was used to investigate the presence or absence of a nucleotide substitution that confers the resistance to the QoI fungicide class. Each RFLP reaction consisted of 10 μl of the cytb PCR product combined with 17 μl of nuclease free water, 2 μl of 10× FastDigest Buffer (Thermo Fisher Scientific Inc., Waltham, MA), and 1 μl of FastDigest restriction enzyme AluI (Thermo Fisher Scientific Inc.). The mixture was incubated at 37°C for 15 min per the manufacturer’s instructions. The process was repeated for every C. cassiicola isolate. The resultant RFLP products were separated via electrophoresis on a 2% (wt/vol) agarose gel and visualized using UV light illumination as previously described.

To confirm the accuracy of the PCR-RFLP method, PCR products from a total of 79 randomly selected isolates were purified using ExoSAP-IT and submitted to Eurofins Genomics LLC for Sanger sequencing. Nucleotide sequences were edited and aligned using DNASTAR Lasergene (DNASTAR Inc., Madison, WI). The QoI-sensitive (C6-2) and QoI-resistant (ST-20S-1) C. cassiicola sequences previously published were used as the reference sequences to determine the presence or absence of additional nucleotide substitutions at codons 129 (F129L), 137 (G137R), or 143 (G143A) of the cytb gene (Ishii et al. 2007).

In vitro azoxystrobin mycelial growth inhibition assay

An azoxystrobin stock solution at the concentration of 100 mg/ml was prepared by dissolving 0.1508 g of technical grade azoxystrobin (99.5% purity; Chem Service, West Chester, PA) in 1.5 ml acetone, and a serial dilution of 50, 25, 10, 1, and 0.25 mg/ml was made from the stock solution. The final concentration in media was attained by adding 1 ml of each fungicide concentration into 1 liter of PDA after cooling to 55°C. A stock solution of salicylhydroxamic acid (SHAM) at 100 mg/ml was prepared by dissolving 1.01 g of SHAM (99% purity; Millipore Sigma, St. Louis, MO) in a 10-ml methanol solution. Aliquots of 1 ml of the SHAM stock solution were incorporated into all treatments for a final concentration of 100 μg/ml (Teramoto et al. 2017). SHAM was added to each concentration to inhibit the alternative oxidase pathway some fungi have been reported to use for in vitro QoI assays (Cruz Jimenez et al. 2018). A non-azoxystrobin-amended control contained 1 ml/liter of both acetone and SHAM at 100 mg/ml. Amended PDA was poured into 9-cm Petri plates in 15-ml aliquots and stored at 4°C in the dark prior to use.

A subset of 25 C. cassiicola isolates representing a random selection of isolates collected from 2016, 2019, and 2020 and consisting of 14 isolates containing the G143A amino acid substitution and 11 wild-type isolates representing four of the five regions in Mississippi was used for in vitro growth assays. Isolates and their respective genotype were selected based on the PCR-RFLP results. Each of the 25 isolates was removed from −80°C long-term storage by placing agar plugs from centrifuge tubes onto a fresh plate of full-strength PDA amended with streptomycin (100 μg/ml of streptomycin sulfate). The plates containing the agar plug were incubated in the growth room with an environment as described above for 7 days. Once mycelia were observed, a mycelial transfer was made onto PDA amended with streptomycin. The PDA plates containing mycelial transfers were incubated in the growth room with an environment as described above for 7 days. A 6-mm plug from the margin of a 7-day-old actively growing colony of each C. cassiicola isolate was transferred mycelia side down to azoxystrobin-amended and nonamended PDA. Each azoxystrobin concentration by C. cassiicola isolate was replicated three times and incubated in the growth room at 25°C in the dark for 8 days. Four radial colony measurements were recorded for each Petri plate following incubation, and 3 mm was subtracted for each radial measurement to exclude the size of the initial plug. Experiments were conducted twice.

Data analysis

To make comparisons between Mississippi regions (five regions) as well as sampling years (2016 and 2019 through 2021), the isolate genotypes based on the PCR-RFLP analysis were assigned binary codes whereby G143A-containing isolates were assigned a “1” and wild-type isolates were assigned a “0”. Since data between regions were not uniformly normally distributed, data from each region were analyzed with the nonparametric Kruskal-Wallis test using the NPAR1WAY procedure in SAS (version 9.4; SAS Institute, Cary, NC). Means separation was conducted on least square means using pairwise t tests (α = 0.05) following the assignment of ranks in PROC GLM. Similarly, since data between sampling year were nonnormally distributed, a Wilcoxon Signed-Rank test was performed in PROC UNIVARIATE to determine whether there were differences between collection year based on the results of the PCR-RFLP analysis. Means separation procedures were conducted on the least square means (α = 0.05) following the assignment of ranks in PROC GLM.

Proportions of the number of isolates that contained the G143A substitution were calculated based on the results of the RFLP-PCR using the basic formula ([G143A-containing isolates/total isolates] × 100%). In addition, to make comparisons between collection years, proportions of the number of isolates identified as either containing the G143A substitution or wild-type were created in a similar method. The bulk of the data presentation in results, tables, and figures are based on proportions.

Relative growth was calculated as a percentage of radial growth compared with the control (0 μg/ml azoxystrobin). The EC₅₀ value, the effective concentration to inhibit mycelial colony growth by 50% of each C. cassiicola isolate, was determined using a linear regression between the relative growth and log concentrations of azoxystrobin. The EC₅₀ value for each isolate was calculated using the following equation: EC₅₀ = e ( [50 – b₀]/b₁), where e = 2.71828, b₀ = intercept, and b₁ = slope (Wong and Wilcox 2002; Young et al. 2010). Data from two assays were combined for statistical analysis representing six replications per isolate-fungicide concentration. The EC₅₀ values were subjected to analysis of variance using the PROC GLM procedure in SAS (Version 9.4, SAS Institute Inc., Cary, NC). Mean EC₅₀ values of all 25 isolates were compared using Fisher’s protected least significant difference with a significance level of α = 0.05.

Results

Target spot-infected leaf sampling and C. cassiicola isolation

Over the 4-year sampling effort, 228 distinct geographic soybean field locations yielded a total of 819 C. cassiicola isolates. Sanger sequencing of the ITS region of all 819 isolates confirmed they were C. cassiicola as they shared more than 99% mutual nucleotide identity with the corresponding sequences available in GenBank. In particular, identities ranged between 99% (accession MT919847) and 100% (accession MT228954). Furthermore, all sequences generated in the current study were determined to be identical. Therefore, ITS sequences of isolates 511 and 517, representatives of the G143A-containing and wild-type genotypes, respectively, were deposited in GenBank (accession numbers OQ071215 and OQ071216).

In all, isolates were collected from 75 counties, representing approximately 98.7% of the soybean production area in the state. The Delta accounted for the greatest percentage of isolates (39.3%), resulting in a greater percentage of isolates than the Coast (1.7%), Hills (15.1%), and Pines (19.5%) regions combined (Table 1). The Capital region accounted for 24.3% of the isolates collected.

Table 1. The region and number of Corynespora cassiicola isolates collected and confirmed to contain the G143A substitution from Mississippi soybean fields during 2016 and between 2019 and 2021

Year	Number of G143A-containing isolates/total number of isolates (number of counties)x,y
	Capital (16 counties)z	Coast (6 counties)	Delta (15 counties)	Hills (16 counties)	Pines (29 counties)	Total
2016	10/13 (3)	0	45/51 (6)	1/6 (3)	1/10 (3)	57/80 (15)
2019	15/21 (2)	0	48/61 (7)	5/5 (1)	9/10 (2)	77/97 (12)
2020	78/79 (9)	0	192/206 (15)	65/98 (15)	75/89 (17)	410/472 (56)
2021	86/86 (11)	13/14 (4)	4/4 (1)	15/15 (6)	41/51 (9)	159/170 (31)
Total	189/199 (16)	13/14 (4)	289/322 (15)	86/124 (16)	126/160 (24)	703/819 (75)

^x Regions were previously defined by Standish et al. (2015) and were generally defined by the number of counties and soybean production (in ha) within each region.

^y Isolates were determined to either contain the G143A substitution or not contain the substitution at position 143 using a PCR-RFLP analysis and confirmed by the Sanger sequencing of 81 isolates.

^z Numbers contained within parentheses refer to the number of unique counties sampled in each region by year. Note that county values within parentheses are not additive over years and are simply presented to represent the number of counties sampled within each region in a given year.

View as image HTML

Characterization of C. cassiicola by PCR-restriction fragment length polymorphism

The primers of cct-cytb-F and cct-cytb-R used in the current study amplified the cytb gene through PCR resulted in a single band from genomic DNA of 346 bp (Fig. 1A). The digestion of PCR products with the restriction enzyme AluI resulted in two distinct patterns (Fig. 1B). In isolates containing a GCT triplet coding for alanine at position 143 (G143A-containing genotype), digestion of the cytb amplicons and separation by electrophoresis produced four DNA fragments of 121, 87, 84, and 54 bp, respectively (Pattern A). The PCR products from isolates that remained undigested at position 143 presented a different pattern with only three fragments of 205, 87, and 54 bp (Pattern B). Due to small difference in size (87 versus 84 bp) the two middle bands in Pattern A were comigrating under conditions applied in the current study and could not be distinguished in the gel (Fig. 1B). The analyses of PCR-RFLP patterns from all 819 isolates indicated that the amplicons of the cytb gene of 703 C. cassiicola isolates were digested by AluI, resulting in a Pattern A due to the presence of the GCT codon confirmed by the sequence analysis. The remainder of the isolates (116) were undigested by AluI at codon 143 and exhibited Pattern B. The results revealed the presence of molecular indicators for QoI resistance in the majority (85.8%) of the C. cassiicola isolates collected from Mississippi soybean fields between 2016 and 2021.

Download as PowerPoint

The percentage of isolates containing the G143A substitution differed between the five geographical regions with the Capital having the greatest proportion of C. cassiicola isolates containing the G143A substitution (95.0%). Isolates containing the G143A substitution accounted for 69.4, 78.8, 89.8, and 92.9% of the isolates collected from the Hills, Pines, Delta, and Coast, respectively (Fig. 2 and Table 1). Of the 819 isolates evaluated, 85.8% contained the G143A amino acid substitution. G143A-containing isolates were present in at least one field in 74 of the 75 Mississippi counties sampled; however, the isolates recovered from Newton County (in eastern Mississippi) were only of the wild-type genotype. Over time, the percentage of isolates containing the G143A substitution increased from 71.3% in 2016 to 93.5% in 2021 (Table 1).

Download as PowerPoint

Based on the results of the isolates by location sampled, the greatest number of locations was dominated by C. cassiicola isolates containing the G143A substitution (Fig. 3). Isolates containing the G143A substitution were the dominant isolate population and greater than the number of wild-type isolates recovered regardless of year sampled, with the greatest proportion of G143A-containing isolates observed in 2021 (Fig. 3A). When considering all locations sampled, 7.0% of the locations were observed to contain only wild-type isolates. In addition, the number of locations containing what could be deemed a mixed population, containing both wild-type and G143A-containing isolates, decreased from 33.3% in 2016 to 11.1% in 2021 (Fig. 3A). In addition, over time, there was a significant increase in the percentage of locations that contained isolates with the G143A substitution, with significant increases between 2016 and 2020 and 2021 (Fig. 3A). In general, the greatest number of locations contained isolates exhibiting the G143A substitution, with 77.2% of the locations surveyed, followed by 15.8% of the locations determined to contain a mixed population (Fig. 3B). In addition, the mixed populations differed across regions, with the greatest mixed population identified in the Hills, followed by the Pines and the Coast (Fig. 3B). Moreover, the proportion of isolates identified as containing the G143A substitution were significantly greater from the Capital than the Hills and Pines and significantly greater in the Coast and Delta than the Hills (Fig. 3B).

Download as PowerPoint

Sequence analysis of PCR products from a subsample of 81 isolates, representing 10% of the total number of isolates collected, revealed that the G143A substitution was the only amino acid substitution observed in C. cassiicola from Mississippi. Of the 81 C. cassiicola isolates, 36 wild-type isolates exhibited the triplet GGT at codon 143 of the cytb gene, which showed the absence of the AluI digestion site. Conversely, 45 isolates that contained the G143A substitution exhibited the GCT combination at codon 143 of the cytb gene that was digested during the RFLP. All 81 isolates had conserved codons 129 and 137 which demonstrated that no additional amino acid substitutions were associated with QoI resistance. Taken together, nucleotide sequence comparisons of these isolates confirmed that the PCR-RFLP analysis was a robust and accurate method to detect the G143A substitution and differentiate between the two genotypes as either G143A-containing or wild-type. No differences in nucleotide sequences other than GGT → GCT that are responsible for the G143A substitution were observed between resistant and sensitive isolates of C. cassiicola (Fig. 3). Finally, nucleotide sequences of two isolates from the current study, representative of the two RFLP profiles, were deposited in GenBank (accession numbers OQ102328 and OQ102329).

In vitro azoxystrobin mycelial growth inhibition assays

EC₅₀ values varied significantly among isolates (P < 0.0001; Table 2). Moreover, QoI-sensitive isolates had significantly lower EC₅₀ values than QoI-resistant isolates. Specifically, EC₅₀ values of the 11 QoI-sensitive isolates occurred from 9.93 to 46.14 μg/ml with a mean of 25.88 μg/ml. Isolate MS-Cc-522 was least sensitive, and isolate 58 was the most sensitive among QoI-sensitive isolates. The remaining 14 isolates were resistant to azoxystrobin, with 50% of the isolates exhibiting EC₅₀ values > 100.00 μg/ml and the remaining 50% of the isolates exhibiting EC₅₀ values from 52.89 to 97.47 with a mean of 73.66 μg/ml. Notably, C. cassiicola isolates MS-Cc-320, MS-Cc-321, and MS-Cc-322 originated from the same field in Coahoma County but reacted differently to azoxystrobin, with MS-Cc321 being 4.6 and >4.7 times more sensitive than isolate MS-Cc-320 and MS-Cc-322, respectively. Similarly, C. cassiicola isolates MS-Cc-566 and MS-Cc-567 from the same field in Benton County also reacted differently to azoxystrobin, with isolate MS-Cc-566 being 3.6 times more sensitive than isolate MS-Cc-567.

Table 2. The effective concentration (EC₅₀) of azoxystrobin at which 50% of colony growth was inhibited for 25 Corynespora cassiicola isolates collected from Mississippi soybean

Isolateu	Collection location and yearv			EC50 (μg/ml)w	Isolate reaction
	County	Region	Year		Resistantx	Sensitivex
MS-Cc-16	Sharkey	Delta	2016	>100.00y	×
MS-Cc-29	Leflore	Delta	2016	69.30 cz	×
MS-Cc-58	Lee	Hills	2016	9.93 k		×
MS-Cc-64	DeSoto	Hills	2016	14.10 jk		×
MS-Cc-86	Coahoma	Delta	2019	52.89 e	×
MS-Cc-89	Coahoma	Delta	2019	62.07 d	×
MS-Cc-91	Sunflower	Delta	2019	88.45 b	×
MS-Cc-121	Hinds	Capital	2019	>100.00	×
MS-Cc-132	Tallahatchie	Delta	2019	16.58 ij		×
MS-Cc-168	Washington	Delta	2019	17.22 ij		×
MS-Cc-194	Claiborne	Capital	2020	>100.00	×
MS-Cc-216	Madison	Capital	2020	>100.00	×
MS-Cc-242	Lincoln	Capital	2020	72.67 c	×
MS-Cc-317	Coahoma	Delta	2020	17.20 ij		×
MS-Cc-320	Coahoma	Delta	2020	97.47 a	×
MS-Cc-321	Coahoma	Delta	2020	21.27 i		×
MS-Cc-322	Coahoma	Delta	2020	>100.00	×
MS-Cc-366	Panola	Delta	2020	>100.00	×
MS-Cc-473	Carroll	Delta	2020	33.31 h		×
MS-Cc-522	Yalobusha	Hills	2020	46.14 f		×
MS-Cc-552	Union	Hills	2020	43.05 fg		×
MS-Cc-566	Benton	Hills	2020	27.59 hi		×
MS-Cc-567	Benton	Hills	2020	>100.00	×
MS-Cc-610	Attala	Pines	2020	72.80 c	×
MS-Cc-674	Winston	Pines	2020	38.28 gh		×
LSD				6.58
F value				333.44
P value				<0.0001

^u Mississippi county, region, and year of collection where soybean leaflets exhibiting symptoms consistent with target spot were collected and C. cassiicola isolates originated.

^v C. cassiicola isolates were collected during the 2016, 2019, and 2020 growing season in Mississippi and represent a small subset of the isolates collected. Isolates were selected at random based on whether or not they contained the G143A substitution or not following PCR-RFLP analysis.

^w Colony measurements were made following an 8-day incubation period at 25°C in the dark.

^x Resistant isolates exhibited the G143A amino acid substitution which has been reported to confer resistance to the QoI fungicides. Isolates determined to be sensitive, referred to as wild-type, did not contain the G143A substitution based on PCR-RFLP analysis.

^y Values >100 μg/ml express EC₅₀ values that exceeded the tested azoxystrobin concentration.

^z Mean EC₅₀ values (n = 6) within each column followed by the same letter are not significantly different based on Fisher’s protected least significant difference test (LSD, α = 0.05).

View as image HTML

Discussion

Currently, QoI fungicides including azoxystrobin are registered for use on many crops including soybean because of their broad-spectrum activity against disease-causing fungi (Bartlett et al. 2002). For example, azoxystrobin is marketed as either a solo product such as Quadris (or Aframe, Syngenta Crop Protection) and several generic products or included in premixed products such as Quadris Top, as well as several additional products, to manage frogeye leaf spot, target spot, and many additional potentially yield-limiting soybean diseases. In addition to disease management, QoI fungicides are oftentimes marketed for their potential nonfungicidal physiological effects (Dorrance et al. 2010; Wise and Mueller 2011). However, widespread occurrence of QoI-resistant isolates of Cercospora sojina, the causal agent of frogeye leaf spot of soybean has been reported in the United States since initially detected in 2010 (Zhang et al. 2012, 2018). While there is likely a direct correlation between the increased use of QoI fungicides for disease management or plant health benefits since 2010 and QoI-resistance in Cercospora sojina, care should be taken to reduce the increased development of fungicide resistance within additional organisms. Coincidentally, over the same period, a resurgence of target spot has occurred in the southern United States. Moreover, the increased reports of QoI resistance within C. cassiicola across the southern United States have been made from states that documented Cercospora sojina QoI-resistant populations. Our current study is the first large-scale characterization of C. cassiicola isolates from soybean conducted in Mississippi and the southern United States.

The sensitivity of C. cassiicola isolates to QoI fungicides, based on the results of the PCR-RFLP analysis, significantly varied among years and Mississippi regions. A general reduction in the wild-type population sensitivity to QoI fungicides and a shift to isolates containing the G143A amino acid substitution was observed throughout Mississippi soybean fields. In all, 89.8% of the C. cassiicola isolates collected in the Delta region were observed to contain the G143A substitution which is considered to confer resistance to the QoI fungicides, whereas only 69.4% of the C. cassiicola isolates collected in the Hills region contained the same amino acid substitution. The overall percentage of isolates identified as either G143A-containing or wild-type genotypes differed greatly between the Delta and Hills probably due to varying soybean management strategies as well as greater number of hectares in soybean production and more widespread fungicide use in the Delta than the Hills (Standish et al. 2015). In general, soybean fields in the Delta are typically early planted (April/May) with high-yield producing MG IV cultivars, predominantly in continuous soybean production and for the most part are furrow irrigated (Allen 2018; Heatherly 1999). Moreover, fields in the Delta typically receive an automatic foliar fungicide application for yield enhancement, rather than for disease management, a regular practice dating back to the mid-2000s. Historically, this practice utilized a single application of a QoI fungicide such as azoxystrobin or pyraclostrobin between the R3 and R4 growth stages at a reduced rate of 292 ml/ha instead of a full rate of 438 ml/ha (Standish et al. 2015). Conversely, soybean fields in the Hills tend to be planted later than those in the Delta (May/June) and are predominantly nonirrigated, which oftentimes leads to reduced yield potential. Furthermore, only 12% of Mississippi’s soybean hectares in 2020 were planted in the Hills as compared with 78% in the Delta (NASS 2021). Fungicide applications are also much less frequent in the Hills due to longer crop rotations, reduced yield potential caused by dryland production practices, and the perception that fungicide applications may not be as profitable in the absence of disease. The reduced proportion of C. cassiicola isolates containing the G143A substitution occurring in the Hills could be the result of a reduced reliance on QoI fungicides for general yield benefits or disease management or simply as a result of a reduction in the total number of fungicide applications over time. Regardless of region, C. cassiicola isolates collected from most soybean fields across Mississippi were identified as containing the G143A amino acid substitution which confers resistance to QoI fungicides. Interestingly, both wild-type and G143-containing C. cassiicola isolates were detected from the same sampling locations in a few counties including Benton, Coahoma, Covington, and Jones, representing the Hills, Delta, and Pines regions, respectively. The associated pattern demonstrates the shift to isolates containing the amino acid substitution at position 143 has occurred over time with different frequencies in response to varying levels of selection pressure. The shift that occurred over time was observed in the current study where the percentage of locations with C. cassiicola containing the G143A substitution were observed to significantly increase between 2016 and 2019 and 2020. Moreover, in vitro research stemming from this study of distribution determined C. cassiicola isolates from Mississippi containing the G143A substitution appeared to have fitness benefits by producing a greater number of conidia and having enhanced growth characteristics (Wang et al. 2023). In addition, the inherent fungal population structure may complicate future fungicide management strategies based on the EC₅₀ values observed following exposure of C. cassiicola to members of the DMI and MBC classes (Wang 2022). For example, G143A-containing C. cassiicola isolates required on average of approximately 41 and 115% more DMI and MBC-containing fungicide, respectively, compared with wild-type isolates to achieve equal control. How potential fitness benefits as well as how the elevated EC₅₀ values may impact future management strategies should continue to be explored especially given that most soybean farmers make fungicide applications with products that contain multiple MOAs and no longer rely on stand-alone QoI products.

Our results revealed that isolates of C. cassiicola containing the G143A substitution occur throughout the Mississippi soybean production system. The magnitude of isolates containing the G143A substitution is disturbing since fungicides are such an important management tool for target spot in Mississippi as well as additional southern states. Moreover, 85.8% of the C. cassiicola isolates collected from soybean fields in Mississippi carry only the G143A amino acid substitution, which is reported to confer complete resistance to QoI fungicides. Our results aligned with previous studies on several related fungi within the Pleosporales order such as Alternaria alternata (Fr.) Keissl in pistachio (Pistacia vera L.), Didymella bryoniae (Fuckel) Rehm in watermelon (Citrullus lanatus [Thunb.] Matsum. & Nakai), and Venturia inaequalis (Cooke) G. Winter in apple (Malus domestica L. Borkh), in which a direct correlation was identified between the G143A substitution in the cytb and QoI resistance (Finger et al. 2014; Lesniak et al. 2011; Ma and Michailides 2004). The F129L and G137R amino acid substitutions have previously been reported to occur in a limited number of pathogens (Pasche et al. 2005; Sierotzki et al. 2007). However, neither the F129L nor G137R substitution was identified in a subsample of the C. cassiicola isolates collected in the current study, which was supported by previous studies on C. cassiicola as well as Cercospora sojina (Duan et al. 2019; Rondon and Lawrence 2019; Smith et al. 2021; Standish et al. 2015; Wang et al. 2022; Zhang et al. 2018).

The PCR-RFLP method has been successfully used for reliable detection of the G143A substitution in several important pathogens (Fernández-Ortuño et al. 2012; Forcelini and Peres 2018; Standish et al. 2015). The specific molecular method was based on an additional AluI digestion site created by the nucleotide substitution at codon 143 (GGT → GCT). The current study demonstrated not only the effective utility of a PCR-RFLP method to detect the G143A substitution responsible for QoI resistance in C. cassiicola isolates collected in Mississippi but also its application for monitoring C. cassiicola field populations in the future. The accuracy of the PCR-RFLP method was subsequently corroborated by Sanger sequencing of a total of 81 randomly selected PCR products from both RFLP profiles (36 wild-type and 45 containing the G143A substitution). Therefore, such a method can provide a robust and reliable tool for detecting isolates that contain the amino acid substitutions that confer QoI resistance. Moreover, the presence of additional amino acid substitutions, specifically the F129L or G137R, were not revealed by nucleotide sequences in any of the 81 analyzed isolates. Taken together, this will enable the monitoring of C. cassiicola populations from commercial production soybean fields for the presence of the amino acid substitutions responsible for fungicide resistance in the future, in addition to helping define the efficacy of current fungicide management strategies for target spot.

The differences observed in the EC₅₀ values of C. cassiicola isolates also supported that resistance to QoI fungicides was present in soybean fields across Mississippi. The EC₅₀ values > 100 μg/ml were observed for azoxystrobin on C. cassiicola isolates, which has previously been reported from tomato (Solanum lycopersicum L.) in Florida (MacKenzie et al. 2020). The phenotypic responses of C. cassiicola isolates observed in vitro were supported by the nucleotide sequences detected in the cytb gene. Translated amino acid sequences of the 14 isolates with EC₅₀ values from 52.9 to >100 μg/ml revealed that these azoxystrobin-resistant isolates were characterized by the G143A substitution, while the 11 isolates with EC₅₀ values from 9.9 to 46.1 μg/ml maintained the wild-type amino acid, glycine. Therefore, C. cassiicola isolates exhibiting EC₅₀ values < 50 μg/ml following exposure to azoxystrobin could be phenotypically categorized as QoI sensitive, and those with EC₅₀ values between 50 and 100 μg/ml and > 100 μg/ml could be phenotypically categorized as moderately resistant and highly resistant to the QoIs, respectively. The same level of sensitivity was previously reported in studies to evaluate the in vitro sensitivity of C. cassiicola isolated from soybean in Brazil to DMI and MBC fungicides based on mycelial growth inhibition (Avozani et al. 2014; Teramoto et al. 2017).

Our results suggest that a shift in the C. cassiicola population toward resistance to the QoI fungicides may have occurred throughout Mississippi given the widespread nature of isolates containing the G143A substitution. Notably, the C. cassiicola isolates from Mississippi only carried the G143A substitution that can confer complete field resistance and the complete failure of QoI products when relied upon for target spot management (Fernández-Ortuño et al. 2008). In light of the widespread occurrence of C. cassiicola isolates containing the G143A substitution, management of target spot may become a more complex endeavor. Therefore, alternative disease management practices such as crop rotation, planting cultivars with documented resistance, and tillage should be explored and integrated into the current target spot management strategies. Moreover, efforts to monitor fungicide sensitivity in C. cassiicola populations to the labeled fungicides should be continued to improve our understanding of where resistance is occurring and therefore minimize the selection pressure imposed by fungicide applications in the future.

Acknowledgments

The authors thank J. Standish for aid in the nonparametric statistical analyses.