Fungal interactions of Fusarium verticillioides and F. graminearum in maize ears and the impact on fungal development and toxin accumulation were investigated in a 2-year field study at two locations in France. Maize ears were inoculated either with a spore mixture of F. graminearum and F. verticillioides or using a sequential inoculation procedure consisting of a first inoculation with F. graminearum followed by a second with F. verticillioides 1 week later. Toxin and fungal biomass were assessed on mature kernels, using HPLC and quantitative PCR. Correlation between the levels of DNA and toxin was high concerning F. graminearum DNA and deoxynivalenol (R² = 0·73) and moderate for F. verticillioides DNA and fumonisin (R²= 0·44). Fusarium graminearum DNA either decreased in mixed inoculations or was not influenced by subsequent inoculations with F. verticillioides, compared to single inoculations. In contrast, F. verticillioides DNA either significantly increased or was not affected in mixed and sequential inoculations. In two of the replicates, it can be assumed that natural contamination by F. verticillioides was favoured by previous contamination with F. graminearum. Overall, the results suggest that F. verticillioides has competitive advantages over the F. graminearum strains. Additionally, the data provide, for the first time, key evidence that previous contamination by F. graminearum in maize ears can facilitate subsequent infections by F. verticillioides.
Contamination of maize by Fusarium species leads to several diseases including seedling blight and root, stalk and ear rots. Among Fusarium spp., F. graminearum and F. verticillioides are the predominant species responsible for two distinct maize ear diseases, namely gibberella and fusarium ear rot, respectively (Munkvold, 2003). These fungi pose a serious threat to food safety because of their ability to produce mycotoxins. Fusarium verticillioides produces fumonisins, among which fumonisin B1 (FB1), B2 (FB2) and B3 (FB3) are the most abundant (Nelson et al., 1993). Fusarium graminearum can produce a wide range of mycotoxins, including deoxynivalenol (DON) and its acetylated forms: 15-acetyl-deoxynivalenol (15-ADON) and 3-acetyl-deoxynivalenol (3-ADON). The levels of these toxins in agricultural products have been recently subjected to European regulations, limiting the DON and FB1 + FB2 content to 1·75 and 4 mg kg−1, respectively, in unprocessed maize for human consumption (Commission Regulation (EC) No. 1126/2007). Because there is no available treatment that can efficiently reduce mycotoxin content during food processing, management strategies mainly focus on controlling these diseases in the field.
The best way for controlling gibberella and fusarium ear rot is probably to promote host-genetic resistance. However, one of the difficulties in developing resistant maize varieties is the possible co-occurrence of several Fusarium species within the same maize ear. Interspecific interactions between fungi may occur, affecting the fungal community structure and the subsequent amount of disease and toxin produced. The effects of these interactions on disease development are poorly understood and never accounted for by models aimed at predicting the mycotoxin risk. Moreover, disease is usually estimated through visual observation of symptoms, which does not allow distinction between different members of the Fusarium complex (Xu et al., 2007a).
Microbial ecologists have shown that different types of interactions can occur among species, depending on biotic and abiotic factors. These interactions play a major role in organizing the structure of fungal communities (Carroll & Wicklow, 1992). They range from antagonistic to mutualistic and can be negative, positive or neutral for each species involved. When several fungi occupy the same niche, different mechanisms can give advantage to one fungus at the expense of the others. For example, faster growth (competition for resource) and/or production of toxic secondary metabolites (allelopathy) can impede the development of other fungi (Carroll & Wicklow, 1992). In contrast, the growth of one species can make the environment ‘more hospitable’ for other species (Stachowicz, 2001). Ecological interactions can therefore significantly affect the conditions of fungal development and it is essential to take them into account to give a more accurate assessment of the risk of mycotoxin contamination in maize.
Evidence for interspecific interactions between Fusarium species or between Fusarium and other genera has already been investigated in vitro and in field conditions. It is generally recognized that negative interactions between species, particularly competition, are predominant (Xu et al., 2007a; Pan & May, 2009). Furthermore, the outcomes of fungal interactions are largely dependent on environmental conditions including temperature and water activity (aw) (Xu et al., 2007a; Pan & May, 2009). In mixed inoculations, F. verticillioides was found to dominate over Aspergillus flavus and Penicillium spp. both under field and in vitro conditions (Wicklow et al., 1988; Zummo & Scott, 1992; Marín et al., 1998; Zorzete et al., 2008). A first field study suggested that F. verticillioides reduced the growth of F. graminearum (Reid et al., 1999) in maize ears whereas in vitro studies yielded opposite results (Marín et al., 1998; Velluti et al., 2000, 2001). Divergence on the species dominance observed between F. graminearum and F. verticillioides might be strain- and/or environment-specific, making the comparisons difficult between in vitro and in vivo results. While the previous studies were based on mixed inoculations, it appears that the impact of sequential inoculations has rarely been investigated, despite the fact that time of sequential establishment is cited as a key factor affecting the infection and colonization processes (Wicklow et al., 1988). Interestingly, various field studies have described a severe fumonisin contamination in maize ears inoculated with F. graminearum alone (Schaafsma et al., 1993; Reid et al., 1999; Caron & Naïbo, 2007). These observations prompted these authors to hypothesize that F. graminearum may facilitate F. verticillioides infection of maize ears. However, to the authors’ knowledge, no published study has investigated the consequences of a previous inoculation with F. graminearum on F. verticillioides development.
The general goal of this study was to clarify the interactions between F. graminearum and F. verticillioides in maize ears and the impact of these interactions on disease development and toxin production with two main objectives: (i) to determine the type of interactions that occur between F. graminearum and F. verticillioides in maize ears, depending on the sequential time of establishment, and (ii) to evaluate the effects of these interactions on toxin contamination. Based on previously reported data, it was hypothesized that ears initially inoculated with F. graminearum would favour infection by F. verticillioides. To test this hypothesis, the levels of fungal DNA and toxin production in mature maize kernels were compared after single, mixed and sequential inoculations, using two strains of F. graminearum and one strain of F. verticillioides.
Materials and methods
The two F. graminearum strains (INRA 155 and INRA 159, referred to here as Fg1 and Fg2, respectively) and the F. verticillioides strain (INRA 63, referred to here as Fv) used in this study were previously isolated from naturally infected cereal grains and are maintained in the MycSA collection (INRA, Villenave d’Ornon, France). Based on previous field experiments, Fg2 was expected to be less aggressive than Fg1, in terms of symptom expression.
Spore suspensions were prepared by culturing the fungi in a fermenter (New Brunswick Scientific, BIOFLO III) in modified Bilay’s liquid medium supplemented with carboxymethyl cellulose (CMC) (Reid & Zhu, 2005). After cultivation for 7 days at 25°C, the medium was filtered through two layers of cheesecloth to harvest the spores, which were then enumerated with a haemocytometer. Alternatively, in the 2008 experiments, because of a weak spore production by Fg1 in the fermenter, the spores were obtained from sterilized barley grains inoculated with agar slants and incubated for 20 days at ambient temperature. The spore suspensions were then stored at 4°C for a maximum of 1 month or at −20°C in a 1:10 mixture of glycerol and milk. Prior to inoculation, the spore suspensions were diluted with distilled water to achieve the desired concentration.
Field experiments and inoculation procedures
The PR38H20 maize variety was sown in 2008 and 2009 in two locations, in the south west of France, 220 km apart (Montesquieu-Lauragais and Montardon) and at two sowing dates. The location × year experiments will be referred to as Laur08 and Laur09, for Montesquieu-Lauragais 2008 and 2009, and Mont08 and Mont09 for Montardon 2008 and 2009. The maize variety was selected for its susceptibility to both fungi, as assessed under natural contamination over 3 years and in more than 200 locations in France (ARVALIS Infos, 2007). The first sowing was carried out at the beginning of May and the second one 15–30 days later. For each location, a split-plot design was used with sowing date as the main plot and 13 subplot units corresponding to the 13 inoculation treatments, with three replicates. The subplot size was of four rows per approximately 50 plants. In one row of each subplot, the primary ears of 15 plants presenting the same development stage were inoculated.
The 13 inoculation treatments are summarized in Table 1. Two inoculation procedures were used depending on the time of inoculation. At 4 days after silking, 0·5 mL of spore suspension was injected through the silk channel, using a self-refilling syringe with an obtuse needle. At 11 days after silking, because silk senescence and ear development impeded silk-channel inoculation, the spore suspensions were inoculated at the top of the ear, using a chirurgical needle after lightly wounding the husk. Preliminary results performed in the laboratory indicated that the F. graminearum strains did not develop in ears inoculated with an equal number of spores of both Fusarium species (data not shown). Therefore, to allow the installation of the F. graminearum strains, the mixed inoculum used at 4 days after silking contained F. graminearum and F. verticillioides spores using a 9:1 ratio, corresponding to a concentration of 3·6 × 106 spores of F. graminearum mL−1 and 4 × 105 spores of F. verticillioides mL−1. To test whether a previous colonization by F. graminearum favoured the infection by F. verticillioides, the sequential inoculation consisted of a first inoculation with F. graminearum (at 4 × 106 spores mL−1) 4 days after silking and a second one with F. verticillioides (at 4 × 106 spores mL−1 and 4 × 105 spores mL−1), 1 week later. Single inoculations with F. graminearum were performed using 0·5 mL of a spore suspension at a concentration of 4 × 106 spores mL−1, 4 days after silking. Single inoculations with F. verticillioides were performed at 4 and 11 days after silking, using 0·5 mL of a spore suspension at 4 × 106 spores mL−1 and 4 × 105 spores mL−1. In addition, reference plants were inoculated with water, 4 days after silking.
Table 1. Inoculation treatments used in this study
Silking + 4 daysb
Silking + 11 daysb
aFg: Fusarium graminearum (strains Fg1 and Fg2), Fv: F. verticillioides (strain Fv).
bAll inoculations carried out using 0·5 mL of spore suspension at given concentration.
Single Fg inoculations
Fg1 4 × 106 spores mL−1
Fg2 4 × 106 spores mL−1
Single Fv inoculations
Fv 4 × 106 spores mL−1
Fv 4 × 105 spores mL−1
Fv 4 × 106 spores mL−1
Fv 4 × 105 spores mL−1
Mix Fg1/Fv (90/10) 4 × 106 spores mL−1
Mix Fg2/Fv (90/10) 4 × 106 spores mL−1
Fg1 4 × 106 spores mL−1
Fv 4 × 106 spores mL−1
Fg1 4 × 106 spores mL−1
Fv 4 × 105 spores mL−1
Fg2 4 × 106 spores mL−1
Fv 4 × 106 spores mL−1
Fg2 4 × 106 spores mL−1
Fv 4 × 105 spores mL−1
At the final mature stage (approximately 25% kernel moisture content), 10 inoculated ears were randomly hand-picked in each subplot and the kernels from the first third of each ear were hand-removed and pooled. Ears showing European corn borer damage were discarded. Kernels were then hot-dried for 48 h at 45°C to reach 10–12% moisture content and stored at 4°C. A 50 g subsample was ground to fine powder with an ultra- centrifugal mill (Zm200, Retsch) and stored at 4°C, before mycotoxin and fungal DNA quantification.
Meteorological data (mean monthly temperatures and total monthly precipitations) were obtained from two meteorological stations, one in each location.
The disease severity on each ear was rated as a percentage of the ear surface showing Fusarium disease symptoms, classically rot or white to pinkish-coloured mycelium. Kernels exhibiting white ‘starburst’ symptoms were also considered as diseased. Symptoms caused by F. graminearum and F. verticillioides could not be distinguished.
DON was extracted by agitating 5 g of ground maize kernels with 20 mL of acetonitrile/water (84:16) for 1 h. After centrifugation, 5 mL of the filtrate were purified using Trichothecene P columns (R-Biopharm) before evaporation to dryness at 70°C and dissolution in methanol/water (50:50, v/v). DON concentration was determined using HPLC-MS2 analyses. These analyses were performed using a QTrap 2000 LC/MS/MS system (Applied Biosystems) equipped with a 1100 Series HPLC system (Agilent), a Zorbax eclipse XDB C18 column (2·1 × 150 mm, 5 μm, Agilent) and a TurboIonSpray ESI source. Solvent A consisted of methanol (100%) and solvent B consisted of methanol/water (10:90, v/v). The flow rate was kept at 0·25 mL min−1. Gradient elution was performed with the following conditions: 8 min held at 10% A, 2 min linear gradient from 10% to 70% A, 2 min linear gradient from 70% to 100% A, 7 min held at 100% A, 1 min linear gradient from 100% to 10% A, 8 min held at 10% A. The injection volume was 10 μL. Detection was monitored at 230 nm. The electrospray interface was used in the negative ion mode at 400°C with the following settings: curtain gas, 20 p.s.i.; nebulizer gas, 30 p.s.i.; auxiliary gas, 70 p.s.i.; ion spray voltage, −4200 V; declustering potential, −30 V; entrance potential, −10 V; collision energy, −30 eV; collision-activated dissociation gas, medium. High concentrated samples were diluted when necessary. Quantification was performed using external calibration with DON, 15-ADON and 3-ADON standard solutions, ranging from 10 to 1000 ng mL−1.
Fumonisins were extracted as originally described by Shephard et al. (1990) with some modifications. Briefly, 10 g of ground maize kernels were agitated with 20 mL of methanol/water (75:25, v/v) for 15 min. After centrifugation, filtrates were adjusted to pH 6·5 and fumonisins were purified using Bond Elut Strong Anion Exchange (SAX) cartridges (Varian). Fumonisins were eluted with 10 mL of methanol/acetic acid (99:1, v/v) and evaporated to dryness under a nitrogen stream. Dried samples were dissolved in 200 μL of methanol. Fumonisin concentration was determined with a high-performance liquid chromatograph Agilent 1100 series (Agilent) equipped with an Equisorb ODS2 column and a fluorescence detector (excitation λ of 335 nm and emission λ of 440 nm). Ten microlitres of the sample were derivatized with 90 μL of o-phthalaldehyde reagent and 20 μL of this solution were injected into the HPLC system within 1 min after derivatization. Quantification was performed using external calibration with FB1, FB2 and FB3 standard solutions, ranging from 1 to 100 μg mL−1.
The sum of DON, 15-ADON and 3-ADON and the sum of FB1, FB2 and FB3 were calculated and the amount of toxin was expressed in μg of toxin per g of ground maize kernels.
DNA extraction and quantification of fungal DNA
Total DNA was extracted from 100 mg of each sample using the DNeasy® Plant Mini Kit according to the manufacturer’s instructions (QIAGEN). Grinding was performed with the TissueLyser System (QIAGEN-Retsch), using one stainless steel bead in an Eppendorf tube containing 400 μL of AP1 Buffer (QIAGEN), for 90 s at 30 Hz. After total DNA quantification using a NanoDrop® spectrophotometer (NanoDrop Technology®), each DNA sample was diluted to a final concentration of 10 ng μL−1. Quantification of fungal DNA was performed by quantitative PCR (Q-PCR), using a real-time PCR instrument (ABI PRISM 7300; Applied Biosystems).
For each analysed sample, DNA was amplified with primers designed to track fumonisin-producing fungi or F. graminearum DNA. One primer pair (forward sequence: 5′-GGATTGGCTTGATCTTCACAG-3′; reverse sequence: 5′- GAAGATGGCATTGATTGCCTC-3′) was designed from accession number AF155773 to amplify a 352 bp fragment (Tm: 65°C) from the polyketide synthase gene (FUM1) of fumonisin-producing fungi (Proctor et al., 1999). The primer pair used to track F. graminearum DNA (named Fg16N) was previously designed by Nicholson et al. (1998) (forward sequence: 5′-ACAGATGACAAGATTCAGGCACA-3′; reverse sequence: 5′-TTCTTTGACATCTGTTCAAC CCA-3′). The corresponding melting temperature was at 60°C and the amplified fragment length was 280 bp.
The reaction mix contained 500 nm of each primer and Power SYBR® Green PCR Master Mix (Applied Biosystems). For each assay, 20 μL of reaction mix were mixed with 50 ng of total DNA in 5 μL of water. For the amplification of a DNA fragment of the FUM1 gene, the amplification conditions consisted of a first denaturation step for 10 min at 95°C followed by 45 cycles of 15 s denaturation at 95°C, 60 s annealing at 65°C, and 30 s extension at 72°C. For the amplification of F. graminearum DNA, experiments were performed under the following conditions: a first denaturation step for 10 min at 95°C, then 45 cycles of 15 s denaturation at 95°C, 60 s annealing at 60°C and 30 s extension at 72°C. Assays for each pair of primers were performed in triplicate.
In order to check for contamination with F. proliferatum and F. culmorum DNA, 20 samples were amplified with specific primers designed to track those fungi. Quantification of these fungal species was performed by Q-PCR using a LightCycler® real-time detector (Roche Applied Science).
One primer pair (forward sequence: 5′-TCGTCATCCCTGATAG-3′; reverse sequence: 5′-GAAGATGGCATTGATTGCCTC-3′) was designed from accession number AY577458 to amplify a 216 bp fragment (Tm: 60°C) from F. proliferatum. The primer pair used to track F. culmorum DNA was previously designed by Naef & Defago (2006) (forward sequence: 5′-ATGGTGAACTCGTCGTGGC-3′; reverse sequence: 5′-CCCTTCTTACGCCAATCTCG-3′). The corresponding melting temperature was at 60°C and the amplified fragment length was 570 bp.
The reaction mix contained 500 nm of each primer and GoTaq® qPCR Master Mix (Promega). For each assay, 9 μL of reaction mix were mixed with 20 ng of total DNA in 1 μL of water. For the amplification of F. proliferatum DNA, the amplification conditions consisted of a first denaturation step for 15 min at 95°C followed by 45 cycles of 15 s denaturation at 95°C, 20 s annealing at 60°C, and 20 s extension at 72°C. For the amplification of F. culmorum DNA, experiments were performed under the following conditions: a first denaturation step for 15 min at 95°C, then 45 cycles of 15 s denaturation at 95°C, 20 s annealing at 60°C and 30 s extension at 72°C. Assays for each pair of primers were performed in triplicate.
Quantifications were determined using standard curves of F. verticillioides, F. graminearum, F. proliferatum and F. culmorum DNA extracted from pure cultures. Standard curves for the amplification of F. graminearum and F. verticillioides DNA were generated by using serial dilutions ranging from 10 to 10−4 ng μL−1. For the amplification of F. proliferatum and F. culmorum DNA, standard curve serial dilutions ranged from 20 to 0·002 ng μL−1 and 50 to 0·005 ng μL−1, respectively. PCR efficiency always ranged from 95% to 100% while R² of standard curves ranged from 0·98 to 0·99 (Fig. S1). Minimum quantification threshold for F. verticillioides and F. graminearum DNA equalled the last DNA concentration of standard curves (10−4 ng μL−1) with corresponding Ct value of approximately 33·3 and 38·6, respectively. Lack of nonspecific PCR amplification and dimer formation was checked by including a melting curve analysis in each run (Figs S2 & S3). The melting curve was generated using the following profile: 15 s at 95°C, 15 s at melting temperatures, and 15 s at 95°C. In addition, each run included no-template controls (NTCs), positive controls (DNA extracted from maize kernels inoculated with the fungus under laboratory conditions) and negative controls (DNA extracted from maize kernels free of Fusarium spp.). The NTCs and the negative controls mostly did not amplify or, for a few cases, amplified at a Ct > 43·1 with a different melting curve profile compared to that of positive controls.
Quantification of each fungal species was expressed as a fungal DNA (ng) to total DNA (ng) ratio.
Statistical analysis was performed using Splus® software (TIBCO Sofware Inc.). Because a preliminary analysis indicated significant interactions between ‘year’ and other factors, the 2008 and 2009 experiments were analysed separately. For each year, the effects of location, sowing date, inoculation treatments and replicates were tested on mycotoxin accumulation and fungal biomass in an analysis of variance (anova) (Table 2). Mycotoxin and fungal DNA were log-transformed as log10 (fumonisin + 0·1), log10 (DON + 1), log10 (F. graminearum DNA + 0·00001) and log10 (F. verticillioides DNA + 0·0001) to ensure the normal distributions of residues and homogeneity of variance. Differences between inoculation treatments were determined with multiple comparisons tests with the Sidak method. The level of significance was set at α = 0·05.
Table 2. Analysis of variance showing the effect of the main factors and their two-way interactions for each year of experiment on Fusarium verticillioides DNA (Fv), fumonisin (FB), F. graminearum DNA (Fg) and DON contents in mature maize kernels
Source of variationa
F-value for the 2008 experiment
F-value for the 2009 experiment
*P <0·05; **P <0·001.
aL: Location; SD: Sowing Date; IT: Inoculation Treatment; R(L) : Replicate as a nested factor inside each location.
bd.f.: degrees of freedom.
L × SD
L × IT
Natural contamination with noninoculated Fusarium spp.
The primers used to amplify F. verticillioides DNA can also amplify F. proliferatum, another important fumonisin-producing fungus. In addition, F. culmorum is potentially a DON-producing fungus that can contaminate maize ears. In order to evaluate the possibility that DON and fumonisin contaminations may be linked to these fungi, 20 of the samples showing high levels of fumonisin and/or DON were analysed with specific primers for F. proliferatum and F. culmorum. Results showed that the percentage of F. proliferatum DNA represented <0·5% of the fumonisin-producing strains DNA for 16 samples and between 1·4% and 7·6% for four samples. Fusarium culmorum was not detected. This indicates that contamination with F. proliferatum did not occur often and was negligible when it occurred, with regard to F. verticillioides and F. graminearum. Therefore, under the experimental conditions here, F. verticillioides can be considered as the main species quantified with the primers designed to track the FUM1 gene from fumonisin-producing strains.
Global analysis of data
Among the 13 inoculation treatments studied (summarized in Table 1), inoculation with F. verticillioides at two different spore concentrations (4 × 106 and 4 × 105 spores mL−1) and at two different dates (4 and 11 days after silking) always resulted in similar contamination in terms of both F. verticillioides DNA and fumonisin levels. Therefore, data concerning these treatments were pooled. Although chosen for their differences in aggressiveness, no significant effect relative to the F. graminearum strains was observed in 2008. Consequently, data concerning Fg1 and Fg2 were pooled in the 2008 experiments. In contrast, in 2009, as expected from previous field experiments, Fg1 was significantly more aggressive than Fg2, with a mean ear severity in single inoculations of 40 and 4·2%, respectively.
The anova, reported in Table 2, shows, for each year, the effect of single factors: locations, sowing dates, inoculation treatments, replicate (as a nested factor inside each experiment) and their two-way interactions on fungal DNA and toxin content in the mature maize kernels using the following model:
where Y represents the quantitative variables (fungal DNA or toxin content); explicative factors are: L: Locations, SD: sowing dates, IT: inoculation treatments and R(L): replicate as a nested factor in each location.
The anova indicated significant differences among inoculation treatments for all four quantitative variables. The two-way interaction: inoculation treatment × location was significant for both the levels of fungal DNA and/or associated toxins. Therefore, differences between inoculation treatments in terms of fungal DNA and mycotoxin content were determined separately for each location. In addition, for both years of experiment, a significant interaction was found between inoculation treatment and sowing date for the levels of F. verticillioides DNA and/or fumonisins, which led to a separate analysis of the two sowing dates.
Levels of fungal DNA and toxin obtained in the 2008 experiments are reported in Figures 1 and 2 for F. verticillioides and F. graminearum, respectively, while those obtained in the 2009 experiments are shown in Figures 3 and 4.
Fungal interactions in mixed and sequential inoculations
In mixed inoculations, levels of F. verticillioides DNA were either higher (Figs 1d & 3b,d for MFg1) or not significantly different than in single Fv inoculations. Conversely, the amount of F. graminearum DNA tended to decrease in mixed inoculations with regard to single Fg inoculations, with significant differences in Mont08 (Fig. 2a) and Mont09, for MFg2 (Fig. 4b).
In sequential inoculations, levels of F. graminearum DNA were never affected by the inoculation with F. verticillioides performed 1 week later, whatever the strain used (Figs 2 & 4). However, this previous inoculation with F. graminearum resulted in an increased contamination by F. verticillioides DNA in several replicates. In Mont08, the levels of F. verticillioides DNA were significantly higher in sequential inoculations compared to single Fv inoculations (Fig. 1c,d). Moreover, it should be noted that, for the second sowing date in Mont08, levels of F. verticillioides DNA were significantly higher in the plants inoculated with F. graminearum alone than in the plants inoculated with F. verticillioides alone (Fig. 1d). This result suggests that an external contamination occurred, that was favoured by the presence of F. graminearum. In 2009, the levels of F. verticillioides DNA were also significantly higher in sequential inoculations when Fg1 was previously inoculated (Fig. 3a,c,d), except in one plot for which increased levels were observed but without statistical significance (Fig. 3b).
In summary, F. graminearum either decreased (Figs 2a & 4b) in mixed inoculations or was found not influenced by subsequent inoculations with F. verticillioides. In contrast, F. verticillioides increased in mixed (Figs 1d & 3b,d) or sequential (Figs 1c,d & 3a,c,d) inoculations, or was not found significantly affected. In two cases, it can be suspected that natural contamination by F. verticillioides was favoured by the presence of F. graminearum in the ears (Figs 1d & 3c).
Impact on toxin production
Significant positive correlation (R² = 0·73, P <0·00001) was found between F. graminearum DNA and DON production (Fig. 5). Differences in DON content between treatments were consistent with those observed for F. graminearum DNA: there was less DON in the mixed treatments relative to the single Fg inoculations (Fig. 2) and more DON in the single inoculations with Fg1, relative to Fg2 in the 2009 experiments (Fig. 4). The correlation between F. verticillioides DNA and fumonisin production was lower (R² = 0·44, P <0·00001) than that observed for F. graminearum and DON but was nevertheless significant (Fig. 6). In 2008, differences between treatments for fumonisin content were consistent with those found for F. verticillioides DNA. In 2009, levels of fumonisin were very low in all treatments (Fig. 3). Although F. verticillioides DNA was found significantly higher in the mixed and sequential inoculation treatments with Fg1, surprisingly, this did not result in significantly higher amounts of fumonisin.
In Montardon, August 2008, September and October 2008 and 2009 were 2–3 times rainier compared to the other location, with cumulated rainfall of 438·6 mm in Montardon vs. 73·1 mm in Montesquieu-Lauragais (Table 3). The mean monthly temperatures were relatively similar across both locations. In 2008, the weather was colder than in 2009, with differences between months ranging from 0·5 to 2°C (Table 3).
Table 3. Meteorological data for each location × year combination
Mean monthly temperatures (°C)
Total monthly precipitation (mm)
In these experiments, although F. verticillioides was inoculated at a lower spore concentration than F. graminearum in the mixed inoculum, mixed inoculations never led to lower levels of F. verticillioides DNA, compared to single Fv inoculations. In contrast, levels of F. graminearum DNA were significantly reduced in several replicates of the mixed inoculations. It seems therefore that F. verticillioides had a competitive advantage over F. graminearum, when inoculated simultaneously. Reid et al. (1999) also found that F. verticillioides out-competed F. graminearum in 50:50 mixed inoculations. This competitive advantage may be related to a better growth rate and a better spore germination rate over a wider range of temperatures and water activities. Indeed, it has been shown that F. verticillioides is able to grow over a broader range of temperatures on silk tissues, with growth rates doubling those of F. graminearum at 24 and 30°C (Reid et al., 1999). Moreover, as reported by the latter author, the minimum aw requirement for F. graminearum spore germination at 25°C was 0·94–0·95 vs. 0·88 for F. verticillioides (Sung & Cook, 1981; Marín et al., 1996). In addition, it cannot be excluded that F. verticillioides negatively influences F. graminearum germination and development.
In sequential inoculations, the levels of F. verticillioides DNA were regularly found significantly increased, without affecting the levels of F. graminearum DNA. These results support the initial hypothesis that F. verticillioides benefits from a previous colonization by F. graminearum, without compromising the further development of F. graminearum. It also has to be underlined that the levels of F. verticillioides DNA in the sequential inoculations were sometimes found not significantly different than in single Fv inoculations, but never lower. This interaction could therefore be regarded as facilitation, which occurs if ‘host infection by one species leads to the host becoming more vulnerable to infection by another species’ (Pan & May, 2009). Significant increases in F. verticillioides DNA levels were observed in sequential inoculations in Mont08, in which a natural contamination of F. verticillioides occurred in the plots inoculated with F. graminearum alone, as already observed in previous field studies (Schaafsma et al., 1993; Reid et al., 1999; Caron & Naïbo, 2007). In addition, in 2009, levels of F. verticillioides DNA were increased in ears previously inoculated with Fg1 or when mixed with Fg1. In those ears, a very high level of disease symptoms was observed and a high content in F. graminearum DNA was detected. Based on these results, it can be hypothesized that, as for wounds caused by lepidopteran larvae, ear damage caused by F. graminearum colonization may act as a breach for F. verticillioides infection. From a breeding viewpoint, these results suggest that breeding for resistance to F. graminearum may indirectly select for resistance to F. verticillioides, by reducing a possible infection pathway. In previous breeding studies investigating resistance to maize ear rots after silk-channel inoculation with F. graminearum and F. verticillioides, it has been concluded that breeding for F. graminearum resistance may be useful in breeding for resistance to F. verticillioides (Reid et al., 2009; Löffler et al., 2010). The data presented here therefore provide another piece of evidence supporting this breeding recommendation.
The data show similar trends between toxin accumulation and fungal colonization (Figs 5 & 6), indicating that the toxin productivity (i.e. toxin production per unit of fungus, as defined by Xu et al., 2007a) did not vary much between treatments. The main exception was the sequential and mixed inoculations with Fg1 in 2009 (Fig. 3) in which the levels of F. verticillioides DNA were increased compared to single Fv inoculations but not those of fumonisin, suggesting a decrease in fumonisin productivity. The ears in those treatments were highly rotten by F. graminearum colonization and may therefore not have been conducive for fumonisin production. In another study (Xu et al., 2007b), an increase in trichothecene productivity was reported in mixed inoculations with combinations of Fusarium species in wheat heads, compared to single inoculations. In that case, the presence of competing fungi in the mixed inoculum enhanced the production of mycotoxin while the fungal biomass was reduced. The net outcome of fungal interactions between Fusarium spp. may therefore be species-specific and/or host-specific.
It is known that DON produced by F. graminearum may help the fungus to colonise maize ears (Harris et al., 1999). Although there is no compelling evidence that fumonisins participate in the disease development in maize ears, it has been suggested that fumonisin may somehow increase the fitness of F. verticillioides, maybe by increasing its competing ability (Reid et al., 1999; Munkvold, 2003; Bacon et al., 2008). It is suggested by an in vitro study that fumonisin has antifungal properties against F. graminearum (Keyser et al., 1999). However, the high concentration levels that were used (minimal inhibitory concentration ranging from 5 to 10 mm) do not reflect the fumonisin levels that were obtained in the field study here. In this experiment, it is unlikely that DON or fumonisin acted as inhibitors against F. verticillioides or F. graminearum. First, although high levels of DON were produced in ears inoculated with Fg1 in 2009, this did not impede better colonization of maize ears with F. verticillioides. Secondly, the higher levels of F. verticillioides DNA obtained in sequential or mixed treatments with Fg1 did not result in higher levels of fumonisin, suggesting that the enhanced colonization ability of F. verticillioides in such treatments did not rely on its ability to produce fumonisin.
The occurrence of Fusarium species is known to be largely influenced by the climatic conditions (temperature and relative humidity). In this study, the levels of fungal DNA and toxin in single inoculations varied between years and/or locations. In addition, large differences in disease severity in the Fg1 treatment were observed between years. This uncontrolled variability may be partly explained by different climatic conditions among years and locations. First, the weather in Montardon was rainier than in Montesquieu-Lauragais (Table 3). Levels of F. graminearum DNA and DON in single inoculations were generally higher in Montardon, in agreement with the fact that higher prevalence of gibberella ear rot requires wetness during summer and early autumn (Vigier et al., 1997, 2001). In contrast, fusarium ear rot is said to be favoured by dry and/or hot conditions, especially just before or at pollination (Shelby et al., 1994; Miller et al., 1995; Vigier et al., 1997; Pascale et al., 2002). In the current experiments, although pollination time (August) was drier in Montesquieu-Lauragais than in Montardon, no significant differences in F. verticillioides DNA and fumonisin contamination were found between the two locations in single Fv inoculations.
In conclusion, the results show a complex combination of competitive (F. graminearum was out-competed in mixed inoculations) and facilitative (infection by F. verticillioides was facilitated by a previous infection with F. graminearum) interactions that shape the F. graminearum–F. verticillioides community in maize. The outcome of these interactions depends on the temporal sequence of infection establishment and on the aggressiveness of F. graminearum (F1 versus F2 in 2009), with a probable effect of climatic conditions. Further studies are required to investigate the interactions with other fungal species recovered from the mycoflora of maize ears and the impact on toxin contamination. It was also found that injuries caused by F. graminearum colonization may provide an infection pathway for F. verticillioides. On wounded maize ears, F. verticillioides seems to act as an opportunistic saprophyte but this fungus is also known to have an endophytic development in maize. In that case, the fungal development is symptomless but may become pathogenic under certain conditions (Bacon et al., 2008). Based on these results, it can be hypothesized that this change in the biological comportment of F. verticillioides may be induced by the presence of F. graminearum or damage caused by this fungus.
The authors are grateful to ARVALIS-Institut du végétal and the ANRT (Association Nationale de la Recherche et de la Technologie) for their financial support as part of a PhD grant. This work was also partly supported by ANR (Agence Nationale de la Recherche) through the project ‘MoniMaize’ GPLA-07-043C. The authors would like to thank Alain Boué-Laplace, Florian Gays, Gilles Marque, Sonia Elbelt and Geania Gambarra for their technical assistance in the field experiments and Cécile Larchevêque for her technical assistance in the Q-PCR analysis.