Diversity of Fusarium species isolated from UK forage maize and the population structure of F. graminearum from maize and wheat

Maize sample collection

Sampling of maize plants was undertaken in 2011 and 2012 in the UK to assess the diversity of Fusarium species at harvest in five major production regions (North, North Central, South East, South West and West Central) at fifteen fields. The fifteen forage maize fields sampled were trail grounds for maize variety testing which were managed by Limagrain and the British Society of Plant Breeders (Table 1). Sixty-six plants were sampled at each field from the surrounding maize around each trail ground which allowed for the identification of Fusarium species present at greater than 5%. Data on the mean daily temperature and precipitation was acquired from the closest weather station to each sampling site from the British Atmospheric Data Centre (National Centre for Atmospheric Science, 2012).

The maize was sampled when dry matter (DM) content was between 28 and 32% as determined by five randomly sampled plants at each field. DM of 28–32% occurs when the maize plant has entered into the reproductive stages of plant development and the kernel begins to turn yellow on the outside while the inner fluid becomes milky white due to accumulating starch. The harvested maize plants were commercial hybrids naturally infected with Fusarium. The maize was precision sown in rows with 0.75–0.80 m spacing between rows with an expected density of 100,000 plants per hectare. Sampled plants were selected in a diagonal manner beginning at an arbitrary point at the edge of the field and working towards the center. Each plant was harvested at the base of the stalk, 2–3 cm above the ground, and retained in separate collection bags to avoid cross contamination of samples. Plant samples were stored at 4 °C until they were analyzed for Fusarium colonization.

Isolation and identification of Fusarium species in maize

Fusarium isolates from maize were recovered from kernels, and stalk tissues utilizing a standardized isolation technique. In a laminar flow bench, maize stalk sections were taken in 5 mm sections from either the base of the plant, below the primary ear or above the primary ear for the selected plant. Stalk samples were longitudinally split and tissues of the vascular bundle were removed aseptically from the innermost leaf of the stalk. Three pieces of stalk tissue from the same plant were transferred to 90-mm Petri plates with Spezieller Nährstoffarmer Agar (SNA). Under the same aseptic conditions as the stalk samples, maize kernels were manually removed from the ear and ten kernels were randomly selected for analysis. The maize kernels were surface sterilized in sodium hypochlorite (1% available chlorine) for 2 min and rinsed twice with sterile distilled H₂O and dried on sterilized paper. After drying completely the maize kernels were transferred to 90-mm Petri plates with SNA at 5 kernels per plate. All samples were incubated at 20 °C under 12 h near-UV and 12 h dark cycle for 3–5 days. Hyphal tips of potential Fusarium colonies were transferred aseptically to new 90-mm Petri plates of SNA and incubated under the same conditions for 7–10 days. A conidia and mycelium solution from each sample was preserved in 1.5 ml of 25% glycerol-water and stored at −80 °C. All potential Fusarium isolates from the maize samples were retained and identified to species.

Fusarium species were initially identified to species based on the morphological characteristics (Leslie & Summerell, 2006) of the macroconidia, microconidia and general mycelium presentation from a single spore isolate grown for 7–10 days on SNA with an Olympus BH-2 light microscope. When macroconidia, microconidia and mycelial characteristics from SNA were insufficient for identifying the species further examination of the samples were done on different agar media. Potato Dextrose Agar (PDA) was used to identify colony pigment characteristics of aerial mycelium on the agar (Leslie & Summerell, 2006). Carnation Leaf Agar (CLA) was used to identify macroconidia, chlamydospores and the presentation of aerial mycelium.

DNA was amplified in multiplex PCR by Goertz et al., (2010) to confirm the identity of ten Fusarium species. The ten Fusarium species identified in the multiplex PCR are Fusarium avenaceum, F. culmorum, F. cerealis, F. equiseti, F. graminearum, F. oxysporum, F. poae, F. proliferatum, F. subglutinans and F. verticillioides. Single spore isolates were grown on PDA plates for 3 days and genomic DNA was extracted from harvested aerial mycelia with a cetyltrimethylammonium bromide (CTAB) method (Leslie & Summerell, 2006). For each amplification reaction 2 µl of template DNA (40 ng/µl) was mixed with 18 µl PCR master mix containing 12 µM of each primer, 6.0 µl of 10× PCR buffer plus 20 mM MgCl2 (Roche, Diagnostics, Mannheim, Germany), 400 µM dNTPs, 0.16 µl of Fast start Taq DNA polymerase (Roche Diagnostics, Mannheim, Germany). The PCR consisted of one cycle of 5 min at 94 °C; 35 cycles at 30 s at 94 °C (denaturing), 30 s at 60 °C (annealing), 1 min at 72 °C (extension); and one cycle of 10 min at 72 °C. After the PCR, 5 µl of the amplification product was used to reveal the presence/absence under UV light on 1.5% agarose gel (Sigma Aldrich, Darmstadt, Germany) containing ethidium bromide (0.5 µg/ml) and scored relative to a 100-bp HyperLadder (Bioline, London, UK).

Trichothecene genotyping for F. graminearum and F. culmorum

Multiplex PCR assays were used to characterize the trichothecene genotypes NIV, 3AcDON and 15AcDON based on the polymorphisms in the Tri5 gene cluster of the trichothecene biosynthetic pathway. Targeting more than one tri gene was undertaken to reduce misidentification due to gene cluster recombination (Pasquali et al., 2011). The first multiplex assay (Quarta et al., 2006) used primer pairs based on the Tri3 and Tri7 genes. The primers for the Quarta et al. multiplex assay are Tri3F971, Tri3F1325, Tri3R1679, Tri7F340, TriR965 with three DNA fragments for 15AcDON (708 bp), 3AcDON (354 bp) and NIV (625 bp). The second multiplex assay (Ward et al., 2008) used primer pairs based on the Tri12 gene. The primers for the Ward et al. multiplex assay are 12CON, 12NF, 12-15F, 12-3F with three DNA fragments for 15AcDON (670 bp), 3AcDON (410 bp) and NIV (840 bp). Both multiplex PCR reactions for determining the trichothecene chemotype were done in 20 µl reaction volumes with 30 ng DNA as described for Fusarium species-specific assays (Goertz et al., 2010). The amplification protocol consisted of one cycle of 5 min at 95 °C; 35 cycles at 30 s at 95 °C (denaturing), 30 s at 55 °C (annealing), 1 min at 72 °C (extension); and one cycle of 10 min at 72 °C. After the PCR, the amplification product was revealed as described for Fusarium species-specific assays (Goertz et al., 2010).

Statistical analysis

The incidence of Fusarium species was determined from the predicted mean at each field from a restricted maximum likelihood (REML) analysis to assess the significance of regional locations, chemotypes, and differences between years using Genstat (GenStat, 2011). Multiple strains of the same species recovered from the same plant were assessed qualitatively as their independence could not be critically assessed.

Sampling and isolation of F. graminearum in wheat

For the evaluation of the population structure of F. graminearum in maize and wheat the isolates from wheat were recovered from wheat kernel samples which were harvested in 2012 from wheat producers in the UK. A total of thirty-two isolates were recovered from thirty-two fields with one isolate representing each field. The locations of the UK fields are identified in an EU survey of Fusarium from 2001 to 2013 (Pasquali et al., 2016). The wheat kernels were screened for F. graminearum by surface sterilizing the kernels in sodium hypochlorite (0.37% available chlorine) with 0.1% Tween20 for 5 min and then rinsing twice with sterile distilled water and dried on sterilized paper. 20 wheat kernels per sample were then transferred to 9 cm Petri plates with SNA. Wheat kernel samples were then incubated under the same conditions as maize samples with F. graminearum isolates identified as described above.

Population genetics analysis with variable number tandem repeat markers

Nine VNTR markers (Suga, Gale & Kistler, 2004) were used for the analyses of population genetic structure in the F. graminearum populations which were; HK630, HK913, HK917, HK957, HK967, HK977, HK1043, HK1059 and HK1073. There were 245 F. graminearum isolates from the maize collection and 32 isolates from wheat collection used for the population genetics analysis. VNTR amplifications were performed with 20 ng of DNA per reaction in 10 µl of master mix as described in the Fusarium species-specific assays. PCR was performed as described by Ward et al. (2008) which consisted of 2 min at 95°C; 25 cycles of 1 min at 95 °C, 1 min at 59 °C and 1 min at 72 °C with a final extension stage of 10 min at 72 °C. Reaction products were scored relative to GS500 ROX (Applied Biosystems) internal size standard using an ABI 3730xl Genetic analyzer with Genemapper 4.0 software. Irresolvable alleles and missing data were treated as missing data.

Genetic diversity and F_ST based genetic distances (Slatkin, 1995) were estimated with ARELQUIN 3.5 (Excoffier, Laval & Schneider, 2005). Pairwise F_ST distance estimates were determined based on the number of different alleles and the significance of pairwise F_ST estimates using 1,000 permutations. The analysis of the admixture was performed using Bayesian method with the software STRUCTURE 2.3.4 (Pritchard, Stephens & Donnelly, 2000). The estimation of the number of populations was calculated based on the second order rate of change of the likelihood (Evanno, Regnaut & Goudet, 2005).

Incidence study in maize

The two year maize sampling resulted in a collection of 1,761 Fusarium isolates which were identified to species from the sampling of 2,970 maize stalk pieces (967 Fusarium isolates) and 9,900 maize kernels (794 Fusarium isolates). There were only three ears and one stalk sample that had observable Fusarium mycelium out of all of the samples evaluated. This study identified fifteen Fusarium species in maize with the predominant species being: F. graminearum (32.9%), F. culmorum (34.1%), F. solani (19.8%), F. cerealis (3.5%), F. avenaceum (3.0%) and F. poae (2.8%). Other Fusarium species identified in the collection were: F. oxysporum (1.2%), F. verticillioides (1.1%), F. proliferatum (0.4%), F. tricinctum (0.3%), F. subglutinans (0.3%), F. langsethiae (0.2%), F. napiforme (0.2%), F. equiseti (0.2%) and one isolate of F. scripi (Table 2). The species-specific assay for F. avenaceum (Turner et al., 1998) may not differentiate between F. avenaceum and the closely related F. tricinctum (Kulik, 2008; Tan & Niessen, 2003). Differentiation of F. avenaceum and F. tricinctum was not always conclusive with macroconidia, therefore we confirmed the species based on differences in the formation of microconidia.

Across all fields sampled in the UK, the incidence of Fusarium colonized maize plants with at least one Fusarium isolate recovered in either stalk of kernels was 85% ± 2.7 SD for 2011 and was 86% ± 0.08 SD for 2012. F. solani regularly recovered as a saprophyte was found to have co-occurred with another Fusarium species for 85% ± 0.18 SD of the samples for both years. Of the 990 maize plants sampled, F. graminearum (44.0% ± 0.14 SD) or F. culmorum (31.0% ± 0.16 SD) were present in 63.9% ± 0.12 SD of the plants. The incidence of F. graminearum and F. culmorum between fields was insignificant within the same year. The incidence per field of F. graminearum was significantly (P = 0.002) greater in 2011 than in 2012 as was the incidence of both F. graminearum and/or F. culmorum (P < 0.001). The incidence per field by F. culmorum was not significantly (P = 0.139) different between 2011 and 2012. The mean maximum temperatures and mean precipitation for the sites in Yorkshire, Shropshire, Devon and Gloucestershire are shown in Fig. 1 for the growing seasons in 2011 and 2012.

Trichothecenes for F. graminearum and F. culmorum

Trichothecene types for F. graminearum and F. culmorum isolates in maize had consistent results for all isolates in both multiplex PCR assays. The F. culmorum and the F. graminearum isolates from maize chemotypes of DON (3AcDON-type and 15AcDON-type) and NIV-type are shown in Table 3. The F. graminearum isolates were predominantly the 15AcDON-type (84.1%) followed by NIV-type (15.0%) and three isolates of 3AcDON-type. The F. culmorum isolates were predominantly the NIV-type (75.1%) followed by 3AcDON-type (24.0%) and three isolates of 15AcDON-type. The F. graminearum isolates from wheat used in the population genetic study were all determined to be 15AcDON-type.

Population structure and diversity of F. graminearum populations

Three genetic clusters (K = 3) was determined to capture the major genetic structure in our data set based on the highest $\mathrm{\Delta }K$ (<10). The three clusters were determined from the evaluation of the genetic variation based on the VNTR loci utilizing a Bayesian clustering method for admixture in the 245 (217 DON-type and 28 NIV-type) F. graminearum isolates from maize and the 32 DON-type isolates from wheat. The clusters corresponded and were organized based on the host species and the TCT-type of the isolates. The three groups were NIV-type from maize (NIVm), 15AcDON-type from maize (DONm) and 15AcDON-type from wheat (DONwh). The admixture estimates for each isolate is shown in Fig. 2. The pairwise F_ST distance estimates indicated significant differences between the groups based on trichothecene chemotype; NIVm and DONm (F_ST =0.2207, P < 0.001) and NIVm and DONwh (F_ST =0.2038, P < 0.001). There were also differences between host species with the same trichothecene chemotype; DONm and DONwh (F_ST =0.033, P < 0.05).

Admixture probabilities greater than 95% were obtained for 39 of the 277 isolates analyzed. Bayesian posterior mean estimates for the proportion of the genome for the DONm isolates that is derived from NIVm ranged from 13 to 98%. DONwh isolates had a significant (P < 0.05) admixture with genetic material shared with DONm isolates. The DONm isolates with high admixture shared a significant (P < 0.05) amount of genetic material with NIVm isolates. A visual representation of isolate distribution based on VNTR data for the F. graminearum populations minus the isolates with admixture probabilities greater than 95% is shown in Fig. 3.