Natural occurrence of fusarium head blight, mycotoxins and mycotoxin-producing isolates of Fusarium in commercial fields of wheat in Hubei

Authors


E-mail: yucailiao@mail.hzau.edu.cn

Abstract

Combined analyses of the natural occurrence of fusarium head blight (FHB), mycotoxins and mycotoxin-producing isolates of Fusarium spp. in fields of wheat revealed FHB epidemics in 12 of 14 regions in Hubei in 2009. Mycotoxin contamination ranged from 0·59 to 15·28 μg g−1 in grains. Of the causal agents associated with symptoms of FHB, 84% were Fusarium asiaticum and 9·5% were Fusarium graminearum, while the remaining 6·5% were other Fusarium species. Genetic chemotyping demonstrated that F. asiaticum comprised deoxynivalenol (DON), 3-acetyldeoxynivalenol (3-AcDON), 15-acetyldeoxynivalenol (15-AcDON) and nivalenol (NIV) producers, whereas F. graminearum only included DON and 15-AcDON producers. Compared with the chemotype patterns in 1999, there appeared to be a modest shift towards 3-AcDON chemotypes in field populations during the following decade. However, isolates genetically chemotyped as 3-AcDON were present in all regions, whereas the chemical 3-AcDON was only detected in three of the 14 regions where 3-AcDON accounted for 15–20% of the DON and acetylated forms. NIV mycotoxins were detected in seven regions, six of which also yielded NIV chemotypes. The number of genetic 3-AcDON producers was positively correlated with amounts of total mycotoxins (DON, NIV and acetylated forms) or DON in wheat grains. Chemical analyses of wheat grains and rice cultures inoculated with different isolates from the fields confirmed their genetic chemotypes and revealed a preferential biosynthesis of 3-AcDON and 4-AcNIV in rice. These findings suggest the importance of chemotyping coupled with species identification for improved prediction of mycotoxin contamination in wheat.

Introduction

Fusarium head blight (FHB) or scab, caused by Fusarium spp., is an economically devastating disease of wheat and other small grain cereal crops worldwide, and is particularly favoured by conditions of high humidity and warm temperatures (Parry et al., 1995; Chen et al., 2000; Lu et al., 2001). FHB epidemics in wheat occur frequently in central China, especially along the middle and lower reaches of the Yangtze River, including the province of Hubei, where there are often abundant rains paired with high temperatures during the wheat flowering period (Chen et al., 2000). In 1985, a FHB epidemic outbreak in wheat affected up to 373 million hectares in Henan province alone, causing a yield loss of 900 000 t. Recently, global climate change has aggravated the spread and severity of FHB to an even wider region, extending northward in China (Liang et al., 2007). Since the mid-1990s, FHB has re-emerged as a serious problem for agriculture in North America and Europe (Parry et al., 1995; Windels, 2000).

The causal agents of FHB produce various trichothecene mycotoxins that are harmful to humans, domestic animals and plants (D’Mello et al., 1999). Trichothecene type B mycotoxins such as deoxynivalenol (DON), nivalenol (NIV) and their acetylated derivatives are considered the main toxic compounds that accumulate in contaminated wheat grains entering food/feed chains (Chen et al., 2000). Mycotoxicosis caused by the consumption of FHB-affected wheat flour has been reported in China (Chen et al., 2003) and continues to pose a serious threat to human health.

Molecular characterization of trichothecene mycotoxin biosynthesis pathways in Fusarium has revealed mycotoxin gene clusters and their independent evolution from the rest of the fungal genome (O’Donnell et al., 2000; Ward et al., 2002). These findings led to the development of methods for the genetic chemotyping of mycotoxin-producing Fusarium species to identify their toxigenic potential based on the sequence polymorphisms of the Tri genes that are involved in the biosynthesis of trichothecene mycotoxins (Ward et al., 2002; Li et al., 2005). Genetic chemotyping has been extensively used to analyse population structures and chemotype distributions of Fusarium isolates from Argentina (Pinto et al., 2008), Brazil (Scoz et al., 2009), China (Zhang et al., 2007), Europe (Jennings et al., 2004), Iran (Haratian et al., 2008), Korea (Lee et al., 2009), Japan (Suga et al., 2008) and North America (Guo et al., 2008), revealing its wide applicability and reliability.

Fusarium graminearum is the predominant species causing FHB in wheat in China (Chen et al., 2000; Gale et al., 2002; Qu et al., 2008) and has recently been shown to comprise 14 phylogenetic species (O’Donnell et al., 2008; Wang et al., 2011). Earlier studies of Fusarium isolates from scabby wheat in China in the 1980s showed that the trichothecene mycotoxins DON, 3-acetyldeoxynivalenol (3-AcDON) and 15-acetyldeoxynivalenol (15-AcDON), but not NIV, were produced by Chinese F. graminearum isolates (Wang & Milier, 1994). Characterization of isolates collected from wheat in 1999 and subsequently showed that the predominant fungal species derived from FHB-infected wheat in China were F. asiaticum and F. graminearum (Gale et al., 2002; Zhang et al., 2007; Qu et al., 2008). These two species co-occurred in the middle and lower reaches of the Yangtze River. Recent genetic chemotypings together with chemical analyses showed that both DON and NIV producers were present in the pathogen population of wheat FHB (Zhang et al., 2007; Wang et al., 2008). However, there has been no systematic study on the natural occurrence of FHB in commercial fields of wheat in epidemic regions in China or the types of mycotoxins produced, their concentrations, and the specific Fusarium species involved. In addition, during the last few decades, new wheat cultivars tolerant to FHB have been grown in the regions with frequent epidemics, in conjunction with altered crop rotation systems and constant application of fungicides in every season (Yuan & Zhou, 2005). These practices may have had a profound impact on FHB development, mycotoxin production and species distribution. Thus, current knowledge on the occurrence of FHB along with its associated pathogen population structure and mycotoxin chemotypes will enable a better understanding of the nature of recent FHB epidemics and mycotoxin contamination, which are essential for developing effective control strategies.

The objectives in this study were to: (i) investigate the current natural occurrence of FHB in field wheat in the epidemic regions in China; (ii) chemically determine trichothecene mycotoxins of wheat grains; (iii) study phylogenetic species and genetic chemotypes of mycotoxin producers causing FHB and mycotoxin contamination; and (iv) reveal correlations among mycotoxins and mycotoxin producers in the epidemic regions of China.

Materials and methods

FHB surveys and disease indices in field wheat

In 2009, FHB surveys were carried out when commercial fields of wheat were at growth stages (GS) 81–82 (Zadoks et al., 1973), the growth stage at which FHB symptoms on wheat spikes are very clear and easy to score. Fourteen regions (locations) in Hubei province with a history of FHB epidemics were surveyed (Fig. 1, Table 2). Geographic data such as latitude and longitude were taken on the sites with a Global Positioning System (GPS) viewer (UniStrong Science & Technology Co. Ltd). In each region, five fields (400 m apart) were surveyed and 500 random wheat spikes per field were visually inspected for FHB symptoms. For each spike, total and infected spikelets were scored and used for analyses of FHB incidence and severity. A total of 2500 spikes (500 spikes for each of five fields) were scored in one region and used for the percentage analysis of FHB spikes (FHB incidence) in each region. For estimation of FHB severity for each field, a FHB disease index based on the 500 spikes was calculated using a five-point scale of disease severity for each spike: 0 = no visible disease; 1 = ≤25% infected spikelets; 2 = 26–50% infected spikelets; 3 = 51–75% infected spikelets; 4 = ≥76% infected spikelets. To calculate a FHB index for each region, the spike number for each scale point was recorded and designated as N with a footnote corresponding to the scale (i.e. N0, N1, N2, N3 and N4). Data from five fields were considered as five replicates and used to calculate standard errors. These data were then used to calculate the FHB disease index for one region (Lu et al., 2001):

image
Figure 1.

 Geographical locations of wheat spikes collected in Hubei. Triangles and numbers (1–14) indicate the sampling regions and dots represent main cities.

Chemical determination of mycotoxins and a pathogenicity assay were performed with 14 representative F. graminearum isolates obtained from wheat fields in 2009 (Table S1).

Gas chromatography–mass spectrometry analysis

For analysis of trichothecene mycotoxins by gas chromatography–mass spectrometry (GC/MS), wheat grains were bulked by hand from 500 spikes from five fields (100 spikes per field) of each region in 2009 or from 40 wheat spikes artificially inoculated in 2011 for each of 14 isolates (see below), and 10 g of each bulked grain sample were milled in a commercial blender for 5 min. Rice cultures were also inoculated with the same set of 14 isolates used for the artificial inoculation, and 10 g of each rice culture were milled. The milled samples were extracted for trichothecene mycotoxins with acetonitrile/water (84:16, v/v) as previously described (Wang et al., 2008).

The extracts were analysed by a GC/MS system consisting of Agilent 7890A Network gas chromatography, an Agilent-5975C quadrupole mass spectrometer and an Agilent G4513A auto injector (Agilent Technologies Inc.) as previously described (Wang et al., 2008), with the following modifications. The analysis was performed with a programmed temperature of 120°C for 1 min, then raised by 20°C min−1 to 280°C, with the final temperature being held for 8 min. The helium flow rate was held constant at 1 mL min−1. DON, 3-AcDON, 15-AcDON, NIV, 4-acetylnivalenol (4-AcNIV), T2-toxin (T2), hydroxy-T2 toxin (HT2) and diacetoxyscirpenol (DAS) purchased from Sigma-Aldrich were used to construct standard curves for each chemical during the GC/MS analysis. The following ions were used for trichothecene detection: DON, m/z 235 and 422; 3-AcDON, m/z 117 and 392; 15-AcDON, m/z 193 and 392; NIV, m/z 289 and 379; 4-AcNIV, m/z 450 and 480; T2, m/z 350 and 436; HT2, m/z 347 and 466; and DAS, m/z 350 and 378. The first ion in each set was used for quantitative analysis. Wheat grain mycotoxins were assigned to chemotype groups based on their major metabolite profiles compared with commercial mycotoxins purchased from Sigma-Aldrich as standards. The standards were also used to measure mycotoxin concentrations in grain samples.

Fusarium isolation and culture

Fusarium isolates were obtained from wheat grains collected from scabby spikes in fields in 2009. Thirty wheat grains per region (Talas et al., 2011), which were selected from the bulked grains as prepared above for the GC/MS analysis, were surface-sterilized with HgCl2 (0·1%) and, after extensive washing, placed on potato dextrose agar (PDA) medium at 28°C in the dark for 3 days. Mycelia growing from the grains were transferred onto fresh PDA and mycelium plugs from the newly formed fungal colonies were cultured onto further fresh PDA until uniform mycelial cultures were obtained. Macrospores were produced for each culture in CMC medium (7·5 g carboxymethyl cellulose, 0·5 g NH4NO3, 0·5 g KH2PO4, 0·25 g MgSO4·7H2O and 0·5 g yeast extract per litre) medium (Wu et al., 2005), and a single spore of each culture was derived after serial dilutions of a spore suspension derived from the CMC medium. Single-spore cultures were then grown on PDA for subsequent use.

DNA extraction

Fusarium isolates from single-spore cultures were grown on sterile glass-membrane paper overlaying PDA at 25°C for 5 days. Mycelia (100 mg) of each isolate were harvested and ground to fine powder using a TissueLyser II system (QIAGEN). Total fungal genomic DNA was extracted using the CTAB method as previously described (Nicholson et al., 1998).

Species identification

All purified isolates were identified morphologically by their conidiospores and hyphae (Nelson et al., 1983) and molecularly with genomic DNA extracted from each. Species-specific primers for molecular identification were synthesized as previously described (Nicholson et al., 1998; O’Donnell et al., 2000; Glynn et al., 2005; Jurado et al., 2005, 2006; Kulik, 2008) and are listed in Table 1.

Table 1. Primers and their sequences used for fungal species identification
SpeciesPrimerSequenceDNA fragment (bp)Reference
Fusarium asiaticum Fg16-F5′-CTCCGGATATGTTGCGTCAA-3′400–500 Nicholson et al., 1998
Fg16-R5′-GGTAGGTATCCGACATGGCAA-3′
Fusarium graminearum Fg16-F5′-CTCCGGATATGTTGCGTCAA-3′400–500 Nicholson et al., 1998
Fg16-R5′-GGTAGGTATCCGACATGGCAA-3′
Fusarium poae FPS-F5′-CGCACGTATAGATGGACAAG-3′400 Jurado et al., 2005
FPO-R5′-CAGCGCACCCCTCAGAGC-3′
Fusarium tricinctum FT-F5′-CGTGTCCCTCTGTACAGCTTTGA-3′215 Kulik, 2008
FT-R5′-GTGGTTACCTCCCGATACTCTA-3′
Fusarium proliferatum Fp3-F5′-CGGCCACCAGAGGATGTG-3′230 Jurado et al., 2006
Fp4-R5′-CAACACGAATCGCTTCCTGAC-3′
Microdochium nivale var. nivaleEFNiv-F5′-GTTCCCCTGTCTGACTGTTGT-3′491 Glynn et al., 2005
EFMic-R5′-GTCTCGATGGAGTCGATGG-3′
M. nivale var. majusEFMaj-F5′-CCCCTTCTCCCTATCGC-3′487 Glynn et al., 2005
EFMic-R5′-GTCTCGATGGAGTCGATGG-3′

The polymerase chain reaction (PCR) amplifications were performed in a volume of a 25 μL solution consisting of 1× PCR buffer, 1·5 mmol L−1 MgCl2, 0·25 mmol L−1 each deoxynucleoside triphosphate, 1 U Taq DNA polymerase and 80 ng genomic DNA template, with 0·1 μmol L−1 forward and reverse primers (Li et al., 2005). A negative control omitting the DNA template was used in every set of reactions. The thermal cycler (MyCycler, Bio-Rad) conditions used were 94°C for 4 min, followed by 25 cycles of denaturation at 94°C for 30 s, annealing at 58°C for 30 s (54°C was used for the primers EFNiv-F/R) and extension at 72°C for 30 s, and then a final extension at 72°C for 5 min. Amplified PCR products (2 μL) were separated by electrophoresis on 1·5% (w/v) agarose gels. Gels were stained with ethidium bromide and photographed under UV light using the Bio-Imaging system (Bio-Rad).

Genetic chemotyping

A pair of generic primers derived from the Tri13 gene sequences were used for genetic chemotyping of all F. graminearum sensu lato isolates as previously described (Wang et al., 2008). These primers, Tri13P1 (5′-CTCSACCGCATCGAAGASTCTC-3′) and Tri13P2 (5′-GAASGTCGCARGACCTTGTTTC-3′), generate an 859-bp fragment from NIV-producing isolates, a 644-bp fragment from 3-AcDON producers, and a 583-bp fragment from 15-AcDON producers. 15-AcDON chemotypes of F. graminearum (with an 11-bp repeat within the Tri7 gene) and N-15-AcDON producers from F. asiaticum (without the 11-bp repeat) were identified as described by Zhang et al. (2007). PCR amplifications were performed as described above.

Mycotoxins in rice culture

Fourteen representative F. graminearum isolates from wheat fields in 2009 (Table 5; Table S1) with different chemotypes were selected to inoculate rice as described by Tanaka et al. (1985). The rice cultures were incubated at 28°C for 21 days and extracted for trichothecene mycotoxins as described above for GC/MS analysis (Goswami & Kistler, 2005).

FHB indices and yield assessment in artificially inoculated wheat

Fourteen Fusarium isolates selected for inoculation of rice cultures were used for artificial inoculation of an FHB-susceptible wheat cultivar, Annong 8455, which was sown in autumn 2010 in an experimental field in Wuhan. At 50% anthesis in spring 2011, 10-μL droplets of conidial suspension (5 × 105 spore mL−1) were injected into single florets of the middle spikelets of awn-cut spikes, and the heads were covered with polyethylene bags for 72 h to ensure constant high humidity. Forty spikelets on different heads were injected for each isolate (Wu et al., 2005). FHB disease severity was scored using the above-described five-point scale 21 days post-inoculation (dpi) and used to calculate the FHB index as described above (Lu et al., 2001).

Relative yield loss caused by infection was assessed for each isolate. Wheat grains from 40 spikes at stages GS 85–87 were harvested and weighed (N0). Grains from the same numbers of non-inoculated healthy Annong 8455 wheat spikes served as controls (N1). The relative yield losses for each isolate with the measured N0 and N1 were calculated with the following formula: yield loss (%) = (N1 − N0)/N1 × 100%. Grains bulked from the 40 inoculated spikes were used for mycotoxin analysis by GC/MS as described above.

Pathogenicity assays of non-F. graminearum isolates

Conidiospores of seven representative non-F. graminearum isolates from field wheat in 2009 (three isolates of F. tricinctum and four of F. poae; Table 4, Table S1) were cultured and used for single-floret injection of FHB-resistant wheat cv. Sumai 3 (40 spikes per isolate) in Wuhan in 2011 as described above. Two isolates (5035 and B51066) of F. asiaticum were used as highly aggressive controls (Zhang et al., 2007). FHB disease severity was scored 21 dpi and used for calculation of percentage of infected spikelets.

Statistical analysis

Data analysis was performed by sas v. 8 (SAS Institute) and Pearson’s correlation coefficient test on a regional basis.

Results

FHB occurrence in field wheat

Diseased wheat spikes and indices were surveyed in 2009 in 14 regions in Hubei, China with a history of frequent FHB epidemics (Fig. 1, Table 2). FHB incidence (percentage of diseased wheat spikes) ranged from 36.2% to 98·7%, with 12 regions having an incidence higher than 50% (Table 2), the threshold of a severe FHB epidemic (Lu et al., 2001). In Qianjiang and Xiantao (regions 13 and 14 in Table 2), almost all spikes displayed clear FHB symptoms.

Table 2. Percentage of fusarium head blight (FHB)-infected spikes, disease indexes, and mycotoxin contents of grains in field wheat in 14 regions in 2009
Region no.OriginLocationFHBMycotoxins (μg g−1)
Incidence (%)IndexDON3-AcDONNIVTotal
  1. –, Not detected.

1Xiaonan30°55·140′N, 113°52·102′E69·31 ± 1·5429·25 ± 2·213·633·63
2Yunmeng130°57·220′N, 113°49·478′E80·87 ± 2·2643·44 ± 2·073·340·13·44
3Yunmeng230°57·220′N, 113°49·478′E58·14 ± 2·7726·14 ± 1·361·251·25
4Anlu31°14·589′N, 113°41·270′E36·19 ± 2·5813·17 ± 1·250·520·140·66
5Guangshui31°25·349′N, 113°36·360′E83·92 ± 3·4433·87 ± 1·879·91·72·2713·87
6Zengdu31°37·262′N, 113°31·305′E50·18 ± 2·8325·29 ± 1·1811·12·951·2315·28
7Zaoyang32°07·788′N, 112°42·881′E35·95 ± 2·1211·28 ± 1·511·091·09
8Xiangcheng31°55·495′N, 112°08·966′E69·63 ± 4·1930·46 ± 2·000·590·59
9Yicheng31°40·809′N, 112°15·403′E90·02 ± 1·4733·29 ± 1·864·010·134·14
10Dongbao31°16·742′N, 112°14·683′E87·81 ± 4·2641·65 ± 2·542·360·212·57
11Duodao30°57·347′N, 112°13·941′E71·70 ± 3·8327·20 ± 1·671·371·37
12Shayang30°46·341′N, 112°29·831′E73·05 ± 1·5535·01 ± 1·971·721·72
13Qianjiang30°27·548′N, 112°49·760′E93·26 ± 2·9651·40 ± 7·858·448·44
14Xiantao30°22·844′N, 113°15·725′E98·68 ± 0·2549·01 ± 2·248·291·540·129·95

FHB indices, scored on a five-point scale that reflected the extent of fungal spread on individual diseased spikes, ranged from 11·3 to 51·4 and were significantly correlated with incidence (R = 0·934, P < 0·001) (Table 6), and thus both methods can be used to evaluate FHB severity in field wheat.

Mycotoxin chemotypes and concentrations in wheat grains

Mycotoxin contamination in wheat grains collected from different regions was determined by GC/MS analyses, which revealed mycotoxin chemotypes and concentrations. All 14 samples contained high levels of DON, and 12 of the 14 contained concentrations higher than the limit for consumption (>1 μg g−1). 3-AcDON mycotoxins were detected in the three samples (from regions 5, 6 and 14) that had the highest total mycotoxins. The relative level of 3-AcDON was low in the samples, accounting for 15–20% of the total DON and acetylated DON mycotoxins. No 15-AcDON was detected in the grains.

Seven samples contained varying amounts of NIV, accounting for 1·2–21% of total type B trichothecene mycotoxins. DON was still predominantly present in all the seven samples, three of which (nos 5, 6 and 14) contained a higher or comparable level of 3-AcDON (>1 μg g−1) compared to that of NIV. Neither T2 nor HT2 mycotoxins were detected in the grains. These results suggest that different regions apparently have different mycotoxin-producing fungal species in their respective wheat fields. Therefore, the mycotoxin-producing fungi were isolated from the wheat grains for further analysis.

Mycotoxin-producing Fusarium isolates in wheat grains

In total, 168 isolates were obtained from wheat grains from wheat fields in 2009 and then identified morphologically and molecularly. Based on the results, 141 isolates were F. asiaticum, 16 were F. graminearum and only 11 were other Fusarium species (Table 3; Table S1).

Table 3. Species identification and genetic chemotyping of fungal isolates from field wheat grains in 14 regions of Hubei in 2009
Region no.Origin Fusarium asiaticum Fusarium graminearum Other species (n)
3-AcDONN-15-AcDONaNIV3-AcDON15-AcDONNIV
  1. a15-AcDON chemotypes identified with the Tri13 primers (Wang et al., 2008) but which contained no 11-bp repeat in the Tri7 gene sequence were referred to as N-15-AcDON (Zhang et al., 2007).

1Xiaonan9   5  
2Yunmeng19 1    
3Yunmeng25      
4Anlu12     F. poae (4); F. proliferatum (3)
5Guangshui1523    
6Zengdu9 1    
7Zaoyang6   4  
8Xiangcheng3   2  
9Yicheng11 3    
10Dongbao9   5  F. tricinctum (2); Microdochium nivale var. majus (1)
11Duodao66     
12Shayang9      
13Qianjiang831    F. tricinctum (1)
14Xiantao17 2    
Total1681171311016011

Non-F. graminearum or -F. asiaticum species consisted of F. poae (four isolates), F. proliferatum (three isolates), F. tricinctum (three isolates) and Microdochium nivale var. majus (one isolate). Thus, six different species associated with FHB in field wheat were identified in these regions and showed species-specific DNA patterns on agarose gels (data not shown).

A pathogenicity assay of seven representative non-F. graminearum or -F. asiaticum isolates revealed that all were able to infect wheat, with greatly varying rates of spread (Table 4). One isolate of F. tricinctum (W51504) was more aggressive than the two F. asiaticum isolates used as highly aggressive controls (Zhang et al., 2007). The percentages of infected spikelets (Table 4), not FHB index, were used here to assess pathogenicity because all the scores of infected spikelets in wheat cv. Sumai 3 were <25% and were all on scale point 1 according to the five-point scale of FHB index. These results suggest that these non-F. graminearum or -F. asiaticum species were pathogens of wheat and among the members in the population of FHB pathogens in commercial fields of wheat in the regions studied.

Table 4. Pathogenicity of non-Fusarium graminearum or -F. asiaticum species isolated from field wheat in 2009 on wheat cv. Sumai 3 after floret inoculation with single isolates in Wuhan in 2011
SpeciesIsolate codeaInfected spikelets (%)
  1. aCodes and origins of isolates are presented in Table S1.

  2. bThese isolates are used as controls in pathogenicity assays (Zhang et al., 2007).

F. poae W568051·41 ± 0·09
W568061·55 ± 0·07
W568091·48 ± 0·02
W568102·02 ± 0·25
F. tricinctum W5150422·60 ± 2·44
W595170·80 ± 0·26
W595181·99 ± 0·22
F. asiaticum 5035b10·11 ± 0·89
B51066b21·81 ± 1·56

Genetic chemotyping of mycotoxin producers from wheat grains

To obtain more information on the population structure of mycotoxin-producing isolates and their association with mycotoxins detected in the grains and regions, a thorough genetic chemotyping of all F. graminearum and F. asiaticum isolates was carried out. Results revealed that F. asiaticum consisted of isolates producing different mycotoxins, including various combinations of DON, 3-AcDON, 15-AcDON and NIV (Table 3). 3-AcDON producers were the predominant chemotypes, accounting for 83%, whereas 15-AcDON- and NIV-producing isolates accounted for only 9% and 8%, respectively (Table 3). All F. graminearum isolates identified produced only DON and 15-AcDON. Within F. graminearum and F. asiaticum, 75% of isolates (117) were 3-AcDON producers, 18% (29) were 15-AcDON producers and 7% (11) produced NIV.

Chemotypes were unequally distributed in the surveyed regions. For F. asiaticum, 3-AcDON producers were present in all regions, while 15-AcDON and NIV producers were identified in four and six regions, respectively (Table 3). In F. graminearum, 15-AcDON producers were isolated from only four regions.

Association of mycotoxins and mycotoxin producers in wheat grains

Further associative analyses of chemically determined mycotoxins (Table 2) and genetic chemotyping of Fusarium isolates (Table 3) showed a clear correlation between the mycotoxin chemotypes of wheat grains determined by GC/MS and the genetic chemotyping of Fusarium species isolated from the grains. DON was detected as the predominant mycotoxin in all the samples (Table 2), while DON-producing Fusarium species were indeed present in all 14 regions and were dominant in 13 regions (Table 3). The number of potential 3-AcDON producers identified in 14 regions was positively correlated with the amounts of DON and 3-AcDON (R = 0·698, P < 0·01) or total mycotoxins (R = 0·675, P < 0·01) measured in the grains (Table 6). These results suggest that the Fusarium isolates obtained from the different regions represented the chemotype population structures of the isolates producing the respective mycotoxins in the grains.

NIV was detected in seven regions, six of which yielded NIV-producing Fusarium isolates. This included F. asiaticum identified in five regions (2, 5, 6, 9 and 14) and F. poae detected in one region (no. 4). There was one region that showed the presence of NIV, but where no NIV producer was isolated (no. 10, Table 3). In addition, one NIV-producing F. asiaticum isolate was obtained in Qianjiang (no. 14, Table 3), but no NIV was detected in the wheat grains sampled from this region (no. 14, Table 2).

Chemical analysis of mycotoxins in rice culture

To confirm the genetic chemotyping of F. graminearum isolates from field samples of wheat identified above, 14 isolates (five each of the 3-AcDON and 15-AcDON chemotypes, and four of the NIV chemotype) from different regions were selected to inoculate rice cultures and FHB-susceptible wheat cv. Annong 8455. As shown in Table 5, all the isolates were indeed capable of producing two or more mycotoxins in both rice and wheat, in greatly varying amounts. In rice cultures, four of five 3-AcDON producers produced more 3-AcDON mycotoxins (1·3- to 3·1-fold) than DON, and one produced comparable levels of 3-AcDON and DON. All 3-AcDON chemotypes produced a low level of 15-AcDON. As for 15-AcDON producers, DON was predominantly produced, in 6.4 to 56·1-fold amounts relative to 15-AcDON (Table 5). All 15-AcDON chemotypes also produced low levels of 3-AcDON. Four NIV chemotypes produced NIV and a high quantity of 4-AcNIV that was not detected in the samples from field wheat in 2009.

Table 5. Mycotoxins in rice culture, and mycotoxins, fusarium head blight (FHB) index and yield loss in wheat cv. Annong 84 55 after floret inoculation with single Fusarium isolates in Wuhan in 2011
Isolate codeaMycotoxin (μg g−1) in rice cultureChemotypebMycotoxin (μg g−1) in wheat grainFHB indexYield loss (%)
DON3-AcDON15-AcDONNIV4-AcNIVTotalDON3-AcDON15-AcDONNIV4-AcNIVTotal
  1. –, Not detected.

  2. aCodes and origins of isolates are presented in Table S1.

  3. bChemotype of each isolate was identified by the PCR assay in this study and is summarized in Table S1.

  4. cF. asiaticum.

  5. dF. graminearum.

W51512c0·955·938·895·6NIV0·364·134·4969·93 ± 6·4970·48 ± 6·14
W52517c1·6828·045·0874·6NIV0·3827·2727·6568·57 ± 4·6169·30 ± 3·05
W52611d1352·612·464·11429·115-AcDON24·251·6325·8882·34 ± 4·1782·50 ± 3·19
W52701c1129·23455·7166·34751·23-AcDON79·435·4984·9296·44 ± 5·5188·45 ± 2·04
W54601c1010·7970·7162·32143·73-AcDON14·030·8914·9286·25 ± 5·7275·98 ± 4·13
W54606c2·4381·3721·71105·4NIV0·211·381·5931·61 ± 3·1446·46 ± 3·22
W54616c1003·822·592·31118·615-AcDON20·701·0421·7476·61 ± 3·3683·39 ± 2·64
W55501c1278·91636·677·72993·23-AcDON19·471·2820·7579·82 ± 3·5682·86 ± 4·72
W56604c1·2618·2321·3940·7NIV0·233·243·4755·19 ± 5·5452·88 ± 8·22
W56701c1030·82286·5184·73502·03-AcDON21·471·6523·1284·53 ± 3·0767·79 ± 4·40
W56708d1818·4162·6186·82167·815-AcDON11·530·8412·3760·12 ± 4·1458·07 ± 5·86
W56801d1032·17·118·41057·615-AcDON40·122·9443·0677·69 ± 5·6873·28 ± 4·91
W59501c1869·94319·0127·16316·03-AcDON26·831·5328·3661·07 ± 6·6465·01 ± 7·14
W59610d1626·2118·5255·72000·415-AcDON17·881·0218·973·88 ± 6·0977·73 ± 7·86

FHB index, mycotoxins and yield loss in artificially inoculated wheat

To assay pathogenicity and to further verify the genetic chemotyping of F. graminearum isolates from field samples of wheat, the same set of 14 isolates as used for rice culture described above were used to artificially inoculate FHB-susceptible wheat cv. Annong 8455 in Wuhan in 2011. As shown in Table 5, DON was also the predominant mycotoxin produced by either 3-AcDON or 15-AcDON chemotypes, accounting for 93·8% of the total DON and acetylated DONs. All four NIV-producing isolates also produced a small proportion of DON (3·2%) but did not produce 4-AcNIV at all. The amounts of DON and its acetylated forms in wheat were significantly correlated with those produced in rice culture (R = 0·630, < 0·05), whereas the total mycotoxins from wheat and rice were just below a significant level (R = 0·512, P = 0·061) (Table 6) because of the presence of 4-AcNIV only in rice. Different isolates displayed a vast variation of rates of spread on wheat spikes with floret inoculation, potentially reflecting genetic variation in their pathogenicity on wheat in these regions. A significant correlation was seen between mycotoxins (DON, 3-AcDON, 15-AcDON and total mycotoxin amounts) and FHB index (R = 0·642, < 0·05; R = 0·640, < 0·05) (Table 6).

Table 6. Correlation coefficients (P values) between fusarium head blight (FHB) and mycotoxin parameters from natural field wheat in 2009 and wheat artificially inoculated with Fusarium isolates in Wuhan in 2011
 FHB incidence in 2009aDON + 3-AcDON + 15-AcDONbDON + NIV + acetylated formsbFHB index in 2011c
  1. –, Not determined.

  2. aData from natural field wheat in 2009.

  3. bChemical determination of mycotoxins of samples from natural field wheat in 2009 or artificial inoculation with 14 Fusarium isolates in Wuhan in 2011.

  4. cData from artificial inoculation in Wuhan in 2011.

  5. dData from rice cultures were analysed for correlations with data from artificial inoculation with the same set of 14 isolates in Wuhan in 2011.

  6. *< 0·05; **< 0·01; ***< 0·001; ****< 0·0001.

FHB index in 2009a0·934***0·446 (P = 0·110)0·340 (P = 0·234)
3-AcDON producer in 2009a0·698**0·675**
FHB index in 2011c0·642*0·640*
Yield loss in 2011c0·612*0·611*0·843****
DON + NIV + acetylated forms in riced0·630*0·512 (P = 0·061)

Heavy yield losses were seen for floret-inoculated wheat cv. Annong 8455, ranging from 46.46% to 88·45% (Table 5). The lowest FHB disease index was 31·6, with a 46·5% reduction in yield and a mycotoxin concentration of 1·59 μg g−1. Yield losses were significantly correlated with DON and acetylated DON (R = 0·612, < 0·05) or total mycotoxins (R = 0·611, < 0·05), and FHB index (R = 0·843, < 0·0001) (Table 6).

Discussion

Combined analyses of natural FHB occurrence, mycotoxins, and mycotoxin-producing isolates in field wheat in this study illustrate the profile of FHB development, and provide correlation of mycotoxin contamination with the chemotyping patterns of the causal fungal agents in Hubei, a main agriculture area in China with frequent FHB epidemics. Further chemical analyses confirmed mycotoxin production potentials of Fusarium isolates that were revealed by genetic chemotyping.

Correlation between total mycotoxin amounts (for DON and 3-AcDON) and numbers of 3-AcDON producers suggests that these chemotype producers are mainly responsible for the mycotoxin contamination in wheat grains. The same number of wheat grains from each sample was initially used for isolation of mycotoxin-producing fungi. However, more isolates (Table 3) were obtained from grains that accumulated more mycotoxins as determined by chemical analyses (Table 2). For instance, two samples from Anlu (no. 4, Table 2) and Xiangcheng (no. 8, Table 2) contained less than the limit of mycotoxins for consumption (0·66 and 0·59 μg g−1, respectively). From these two samples, only one and three 3-AcDON-producing isolates, respectively, were obtained (Table 3). The remaining samples contained more than 1 μg g−1 mycotoxins, and from each of these samples at least five isolates were obtained (Tables 2 and 3). These results suggest that five 3-AcDON-producing isolates per 30 seeds may be considered as a cut-off threshold for prediction of heavy DON contamination in grains in the studied regions. Thus, total DON appeared to be a predictive value of 3-AcDON producers, or vice versa. In addition, because of the high susceptibility of current wheat germplasm to initial infection and the presence of a rich source of FHB pathogens in the environment, it is essential to breed wheat varieties highly resistant to initial infection to effectively reduce FHB incidence, mycotoxin accumulation and yield loss in wheat.

Patterns of acetylated DON mycotoxins chemically determined in wheat grains differed from genetic chemotyping of mycotoxin producers. 3-AcDON producers, which were present in F. asiaticum but not in F. graminearum, were the predominant chemotypes genetically identified in all regions. However, 3-AcDON mycotoxins were detected at a relatively low level in only three regions, all of which yielded high levels of total DON (about 10 μg g−1 or more, Table 2). This appears to suggest that acetylation of DON at the 3C position in nature takes place only when a large quantity of DON is synthesized in the wheat grains. 15-AcDON producers were identified in eight regions (four for each of two species, F. asiaticum and F. graminearum), accounting for 18% (29 isolates, Table 3) of total DON producers. However, no 15-AcDON was chemically detected in the grains, even when total mycotoxins were high and 15-AcDON producers were identified (no. 5, Tables 2 and 3). For artificial inoculation of FHB-susceptible cv. Annong 8455 with single Fusarium isolates (Table 5), a similar proportion of 3-AcDON (5·4–7·1% of total mycotoxins; isolates W52701, W54601, W55501, W56701 and W59501) and 15-AcDON (5·4–6·8% of total mycotoxins; isolates W54616, W56708, W56801 and W59610) was chemically detected in the grains, confirming the comparable production potential of these mycotoxins conferred by the two chemotypes when a single isolate was inoculated and no fungal competitor was present at the same infection site. These results imply that wheat grains in nature may favour the biosynthesis of 3-AcDON over that of 15-AcDON, or that producers of the latter may be less competitive on wheat grains. The mechanism and regulation for this preferential synthesis of one acetylated-DON chemotype in wheat in nature requires more investigation.

Genetic chemotyping of NIV showed a pattern similar to that revealed by chemically determined NIV in wheat grains. NIV producers (15 isolates, Table 3) were isolated from six regions, in all of which NIV was found. However, NIV-producing species appeared to differ among the regions. Fusarium asiaticum was present in five regions (2, 5, 6, 9 and 14, Table 3) while F. poae, which produces NIV (Stenglein, 2009), was identified in Anlu (region no. 4, four isolates). In one region (no. 10, Tables 2 and 3), NIV was detected but no NIV producers were isolated, probably because only a limited number of isolates was obtained and identified. One NIV producer was isolated from a region where no NIV was detected (no. 13). This was probably as a result of low productivity of NIV in the grains. NIV producers accounted for only 9% (11 F. graminearum sensu lato and four F. poae) of the Fusarium population (168 isolates), giving rise to a low accumulation of mycotoxins in the grains. No NIV producers were identified in F. graminearum sensu stricto, and a minority of NIV-producing isolates seemed to be F. asiaticum-specific mycotoxin producers within F. graminearum sensu lato in the regions, a chemotype pattern similar to that observed with isolates collected in 1999 (Zhang et al., 2007). However, the percentage of NIV-producing F. asiaticum isolates accounted for 23% of total F. graminearum sensu lato in 1999 (19 of 84 isolates in the regions described by Zhang et al., 2007) and only 7% in 2009. 15-AcDON producers also fell from 31% in 1999 (26 of 84 isolates) to 18% in 2009 (Zhang et al., 2007). Conversely, 3-AcDON producers increased from 46% in 1999 (39 of 84 isolates) to 75% in 2009, an apparent shift favouring the 3-AcDON chemotype in the F. graminearum population on wheat during the decade, as reported recently in North America (Guo et al., 2008).

One F. tricinctum isolate, W51504, isolated from Qianjiang, was found to be highly virulent on wheat. Previous studies showed that F. tricinctum was a minority species among FHB pathogens and had a low pathogenicity on wheat, so pathogenicity assays were focused on isolates of the predominant F. graminearum-clade (Wu et al., 2005). The presence of highly virulent non-F. graminearum-clade species suggests that more attention should be paid to non-predominant species within FHB pathogen populations in wheat, especially for those species with high virulence.

Isolates of the same genetic chemotype produced different amounts or types of mycotoxins depending on the conditions under which they grew. In general, greater amounts of mycotoxins were produced in rice than in wheat (Table 5). Moreover, rice apparently favoured the biosynthesis of 3-AcDON and 4-AcNIV. For instance, 3-AcDON accounted for 45·3–72·7% of total DON mycotoxins in rice for five 3-AcDON producers (W52701, W54601, W55501, W56701 and W59610), whereas only 5·4–7·1% 3-AcDON was accumulated in wheat, i.e. a 10-fold difference between rice and wheat. In addition, all 15-AcDON producers produced 3-AcDON at a level comparable to that of 15-AcDON in rice. A large quantity of 4-AcNIV, accounting for 5·1–65·3% of total NIV, was produced in rice by four NIV producers that all belong to F. asiaticum, but this acetylated NIV was not detected in wheat. In contrast, wheat seemed to favour the conversion of 3-AcDON into DON, with a rather low level of 3-AcDON in both artificially inoculated wheat and field wheat (Tables 2 and 5). As for 15-AcDON, a relatively low level was detected in both wheat and rice for all isolates belonging to F. asiaticum and F. graminearum. It is likely that wheat and rice may favour the conversion of 15-AcDON into DON. The lack of detection of 15-AcDON in field wheat in 2009 (Table 2) may have been the result of the small proportion of 15-AcDON producers (20%, 29/146, Table 3) within the DON-producer population. Nevertheless, these results imply that Fusarium mycotoxin producers were capable of preferential biosynthesis or conversion of mycotoxins under favourable conditions/environments. The mechanisms for this differential biosynthesis of mycotoxins by genetically stable chemotypes of FHB pathogens await further investigation.

Wheat spikes of cv. Annong 8455 artificially inoculated in 2011 with 14 different isolates were collected at GS 85–87 for analysis of yield loss that was close to natural losses caused by FHB. DON is considered the pathogenicity factor that enhances fungal spread on wheat spikes, thus damaging grains and reducing yields. A significant correlation between DON amounts and yield losses suggests that DON content in wheat grains may predict overall yield losses. More importantly, a very significant correlation was observed between yield loss and FHB index (P < 0·0001, Table 6). FHB index reflects the fungal development and disease spread on wheat spikes. FHB pathogens have been proposed to carry additional unknown virulence factors other than DON mycotoxin, based on the expression pattern of a cytochrome P450 gene induced by a Tri5 gene-deleted isolate (Li et al., 2010). If this holds true, DON together with other unknown virulence factors may synergistically impair plant tissues during fungal infection, giving rise to FHB disease the severity of which is reflected by FHB index and final yield loss. Therefore, FHB index appears to more accurately predict yield loss than DON content.

The current results suggest the reliability of both evaluation methods, FHB incidence and index, for FHB disease in commercial fields of wheat. Quick inspection of infected spike numbers in the field during GS 81–82 of wheat is usually carried out in FHB breeding programmes and this does not permit the assessment of variation in fungal spread on wheat spikes (Lu et al., 2001). Disease indices, however, are able to show the differences of individual diseased wheat spikes, reflecting fungal development and disease spread. With regard to simplicity and breeding practice, FHB incidence is easily surveyed and can be used for FHB resistance assays, at least for regions similar to Hubei where wheat varieties resistant to the spreading of FHB pathogens have been planted.

This study shows that F. asiaticum species, especially 3-AcDON producers, are mainly responsible for the current FHB epidemics in commercial fields of wheat and for DON contamination of grains in Hubei, China. The numbers of genetic 3-AcDON producers in grains predict total mycotoxin contents. These findings may provide insight into understanding the population structures of FHB mycotoxin producers and their interactions with their hosts and agroenvironments.

Acknowledgements

This work was financially supported by the National Basic Research Program of China (2009CB118806), the National Natural Science Foundation of China (30571160, 30771337), the Ministry of Agriculture of China (2008ZX08002-001, 2009ZX08002-001B), the Ministry of Education of China (20090146120013), the Chinese–Belgian joint project of BELSPO (BL/02/C58) and MOST of China (S2012GR0016).

Ancillary