The genetic diversity and pathogenicity of isolates of Fusarium graminearum and F. asiaticum isolated from wheat heads in China were examined and compared with those of isolates of F. graminearum, F. asiaticum and F. meridionale from Europe, USA and Nepal. Genetic diversity was assessed by SSCP (single strand conformation polymorphism) and AFLP (amplified fragment length polymorphism) analysis and by molecular chemotyping. SSCP analysis of the Fg16F/Fg16R PCR amplicon differentiated F. graminearum, F. asiaticum and F. meridionale and revealed three haplotypes among sequence-characterized amplified region (SCAR) type 1 F. graminearum isolates. AFLP analysis showed a high level of genetic diversity and clustered the majority of Chinese isolates in one group along with other isolates of Asian origin. The second cluster contained F. graminearum isolates from China, Europe and the USA. Of the Chinese isolates, 79% were F. asiaticum and 81% of these were of the 3-AcDON chemotype, with only 9·5% of either chemotype 15-AcDON or NIV. All the Chinese and USA isolates of F. graminearum were 15-AcDON, whereas among the isolates from Europe, 21% were NIV and 8% were 3-AcDON chemotype. No evidence was found for possible differences in aggressiveness between F. graminearum and F. asiaticum. Highly aggressive isolates were present in each region and no evidence was found for any association between aggressiveness and geographical origin or chemotype among the isolates examined. No difference was observed in pathogenicity towards wheat seedlings between Chinese isolates and those from Europe, the USA or Nepal.
In addition to causing yield loss, FHB is of concern because several of the species involved produce a variety of potent mycotoxins that accumulate in grains during infection. The principal mycotoxins associated with FHB are the 8-ketotrichothecenes deoxynivalenol (DON) and nivalenol (NIV) (Placinta et al., 1999). Based upon the production of different trichothecenes, isolates of F. graminearum sensu lato have been split into chemotypes (Ichinoe et al., 1983; Miller et al., 1991). NIV types produce predominantly NIV and 4-acetylnivalenol (4-AcNIV) and DON-producing isolates have been further subdivided into 3-AcDON types and 15-AcDON types, depending upon the predominantly acetylated version produced alongside DON. Two genes, Tri7 and Tri13 have been found to be non-functional in all DON-producing isolates (Brown et al., 2002; Lee et al., 2002). The Tri7 gene and flanking sequence are deleted in 3-AcDON chemotypes of F. graminearum sensu lato and F. culmorum (Chandler et al., 2003; Kimura et al., 2003). PCR assays have been developed to enable chemotype determination of F. graminearum sensu lato, F. culmorum and F. crookwellense (syn. F. cerealis) (Chandler et al., 2003).
Numerous studies revealed significant genotypic variability within F. graminearum sensu lato (Carter et al., 2000, 2002) and seven biogeographically structured lineages within the F. graminearum clade were proposed by O'Donnell et al. (2000). More recently, 11 phylogenetically distinct species were proposed within the F. graminearum clade (O'Donnell et al., 2004; Starkey et al., 2007). Fusarium graminearum sensu stricto (formerly lineage 7) is the predominant species associated with FHB in North America, and is increasing in prevalence in Europe (Waalwijk et al., 2003). It was previously shown that F. asiaticum (formerly lineage 6) is much more prevalent than F. graminearum on wheat in China (Qu et al., 2008). The reasons for the predominance of F. asiaticum in China are not known and may be epidemiological or other factors. One possibility is that Chinese isolates of the two species differ in their relative aggressiveness towards wheat. Similarly, it is unclear whether Chinese isolates of the two species differ from those in other geographic regions. No comparative study of aggressiveness or other characteristics of Chinese isolates has been reported to date. Furthermore, since the recognition that F. graminearum (sensu lato) represents a species complex, very few studies have been undertaken to examine genetic diversity and pathogenicity among isolates of Fusarium species associated with FHB from different countries (Miedaner et al., 2001; Goswami & Kistler, 2005; Toth et al., 2005). Such analyses are important to determine whether differences in aggressiveness related to species, chemotype or geographic origin are present that may pose a threat to particular wheat-producing regions.
Although sequence-characterized amplified region (SCAR) analysis can resolve F. graminearum, the Fg16F/R amplicon size of F. meridionale is identical (497 bp) to that of F. asiaticum (Carter et al., 2002). The sequence of the F. meridionale and F. asiaticum amplicons differs, making them potentially amenable to discrimination on the basis of single-strand conformational polymorphism (SSCP). The objectives of this study were (i) to differentiate species and haplotypes of Fusarium isolates from China and representative isolates from other countries by SSCP and AFLP analysis and (ii) to compare the aggressiveness of isolates of F. asiaticum and F. graminearum from China and to determine whether they differ markedly from those of isolates from other regions.
Materials and methods
Chinese Fusarium isolates were collected from wheat spikes showing FHB symptoms in 1999 in regions with a known history of FHB epidemics within 12 provinces in China. Isolates were obtained by single-spore isolation and identified as described previously (Qu et al., 2008). Forty-three isolates were arbitrarily selected from 12 provinces representing all the regions in those provinces (Table 1). In addition, Fusarium isolates from Nepal (7), Japan (1), USA (7) and Europe (24) were obtained from the fungal collection of the John Innes Centre, UK (Carter et al., 2002). The European isolates were from France (6), Germany (8), Italy (3), Sweden (3) and the UK (4) (Table 1).
Table 1. Origin, host, SCAR/SSCP type, chemotype and disease index of Fusarium graminearum, F. asiaticum and F. meridionale isolates from China, Europe, Japan, Nepal and the USA
All the Fusarium isolates were grown in Petri dishes on sterile cellulose-membrane paper overlaying potato dextrose agar at 23°C for 1 week. Mycelium was harvested and ground to a fine powder in liquid nitrogen. Total genomic DNA was extracted as described by Nicholson et al. (1997).
SCAR and SSCP analysis
SCAR analysis was used to identify the Fusarium isolates. It was previously demonstrated that SCAR typing, based on the size of the PCR product obtained with primer set Fg16F/Fg16R, resolved isolates of F. graminearum (type 1) and F. asiaticum (type 5) from China (Qu et al., 2008). This method also resolved F. meridionale (formerly termed lineage 2) and F. asiaticum (formerly termed lineage 6) from Nepal (Carter et al., 2002; Chandler et al., 2003). Primer set Fg16F/Fg16R was used for SCAR analysis as described previously (Nicholson et al., 1998). Amplifications were carried out using 50 ng of DNA template with thermal cycling consisting of 30 cycles of denaturation (94°C, 1·5 min), annealing (60°C, 1 min) and extension (72°C, 2 min). PCR products were separated on 2% agarose gels. For SSCP analysis, PCR amplification was performed using the same primer set. SSCP electrophoresis of the amplified target DNA was performed using Sequa Gel® MD (National Diagnostics UK Ltd) with TTE buffer (glycerol tolerant buffer, National Diagnostics UK Ltd). Samples (5 µL of PCR product) were denatured for 5 min at 95°C and placed on ice for 10 min prior to loading. Electrophoresis was carried out at 8 W for 16 h. Following electrophoresis, products were visualized by silver staining (Promega Inc.) according to the manufacturer's instructions. Gel images were made with Kodak C/RA duplicating film and developed using a Fuji RGII Full X-ray film processor.
AFLP analysis was performed according to Vos et al. (1995). The primers were provided by Keygen. Amplification with non-selective primers E00/M00 was carried out with 29 cycles of 94°C for 30 s, 56°C for 60 s and 72°C for 60 s. The selective bases at the 3′-end of the EcoRI primers were AA, CC or TG, while those on the MseI primers were CAT, CTT, GAT or TCG. Four primer pairs were used for selective amplification: AA-CAT (E11/M50), AA-GAT (E11/M66), CC-TCG (E16/M85) and TG-CTT (E25/M62) as described previously (Qu et al., 2008). Selective amplification with the different primer combinations was carried out using 12 cycles of 94°C for 30 s, 65°C for 30 s, 0·7°C touchdown and 72°C for 60 s; followed by 22 cycles of 94°C for 30 s; 56°C for 30 s and 72°C for 60 s. AFLP products were denatured in formamide prior to electrophoresis through 5% polyacrylamide gels followed by silver staining as described for SSCP analysis. The use of a combination of two plus three selective nucleotides, rather than two plus two, improves the robustness of AFLP analyses while reducing the complexity of the profiles.
Polymorphic AFLP fragments were manually scored as binary data with presence as ‘1’ and absence as ‘0’. Cluster analysis based on dice coefficients was performed on the similarity matrix employing the ‘unweighted pair group method using arithmetic means’ (upgma) algorithm (Sneath & Sokal, 1973) provided in the software program ntsyspc, version 2·10e (Exeter Software Co.).
Products from PCR amplifications were purified using the Concert‘ Rapid PCR purification system (Gibco-BRL) according to the manufacturer's instructions, and sequenced from both strands using the Big Dye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems). Sequences were analysed using the staden package (Staden, 1996) version 2000. Alignment to previously obtained sequences was carried out using gcg (Wisconsin package 10·2, Genetics Computer Group).
Isolates were chemotyped on the basis of polymorphisms in Tri7 and Tri13, two genes involved in trichothecene biosynthesis. PCR assays were carried out using the primer sets developed and according to the method described by Chandler et al. (2003). Isolates were chemotyped as either NIV, 3-acetyldeoxynivalenol (3-AcDON) or 15-acetyldeoxynivalenol (15-AcDON) producers according to the classification proposed by Miller et al. (1991). The latter two chemotypes also produce DON in addition to the acetylated product.
A total of 72 isolates, 36 from China, 24 from Europe, 6 from USA and 6 from Nepal (Table 1) were assayed for aggressiveness towards seedlings of Avalon, a commercially successful winter wheat variety in the UK used as a parent in the Defra-supported WIGN reference genetic map (http://www.wign.org.uk). Seedlings were grown in John Innes no. 2 compost (pH 8·0) in 7-cm pots (five seeds per pot) at 20°C. As coleoptiles emerged they were sleeved with a 2-cm collar of PVC tubing (3-mm internal diameter). When the second leaf was fully emerged, macerate (200 µL) of each isolate was injected into the space between the seedling and the collar, according to the method of Simpson et al. (2000). Each experiment consisted of six replicates and was repeated once using a total of 60 seedlings with each isolate. Inoculated seedlings were kept at high humidity to prevent desiccation of the inoculum. Disease symptoms were assessed 17 days after inoculation. Pathogenicity was estimated by multiplying the lesion length by the penetration score as determined using the method of Simpson et al. (2000) to produce a disease index (DI). Statistical analyses used a generalized linear model to assess experimental variability, confirm normal distribution and indicate significant differences among isolates. All statistical analyses were performed using genstat release 9 (copyright 2006, Lawes Agricultural Trust). DI data were log10-transformed prior to analysis because of the non-independence of mean and variance and all factors were assigned as fixed in the model.
All isolates of F. graminearum from China, USA and Europe were SCAR type 1 (0·41 kb), except for isolate G5 from Germany and the three isolates from Sweden, which were type 6 (0·38 kb). Isolate N4 from Nepal was also SCAR type 1 (Table 1). All 34 Chinese isolates of F. asiaticum were type 5, as was isolate Z5047 from Japan (Table 1). In contrast, only isolate WL1 of F. asiaticum from Nepal was type 5, while the other two Nepalese F. asiaticum isolates, N1 and N3, were SCAR type 3 and 4, respectively (Table 1). Isolate N4 was SCAR type 1. The differentiation of type 2 and type 5 SCAR products (sensu Carter et al., 2002) was dependent upon gel-running conditions and not considered to be sufficiently robust for unequivocal identification. For this reason, SCAR products were analysed subsequently by SSCP assay.
The products of the Fg16F/Fg16R SCAR PCR assay for each isolate were subjected to SSCP analysis to determine whether further haplotypes could be resolved using this system. Results indicated that SCAR type 1 consisted of three haplotypes: 1A, 1B and 1C, while the remaining SCAR types displayed the same SSCP patterns with no polymorphism (Table 1). All F. graminearum isolates from China and Europe, and five of the seven isolates from the USA were haplotype 1A (Table 1). The remaining two American F. graminearum isolates (A2 and A5) were haplotype 1B (Table 1). DNA sequencing of the Fg16 product confirmed that haplotype 1B differed from that of 1A at nucleotide 348 (C-T) and 412 (T-A) (data not shown). The Nepalese F. graminearum isolate N4 was clearly resolved by SSCP from the other two F. graminearum haplotypes 1A and 1B and was grouped as haplotype 1C. All the Chinese isolates of F. asiaticum had a similar SSCP haplotype (5), in common with the isolate from Japan (Z5047) and one from Nepal (WL1) (Table 1). The two other Nepalese isolates of F. asiaticum (N1 and N3) had distinct SSCP haplotypes (3 and 4, respectively) (Table 1). Furthermore, SSCP analysis clearly resolved the products of SCAR type 5 from those of type 2 (characteristic of F. meridionale), which appeared very similar on agarose gels. Therefore, SSCP analysis with primer set Fg16F/Fg16R was able to differentiate F. graminearum, F. asiaticum and F. meridionale isolates and revealed polymorphism within F. graminearum and F. asiaticum species (Table 1).
A total of 106 polymorphic AFLP fragments were observed among the 82 isolates using the four AFLP primer combinations. Fragments ranged in size from 150 to 700 bp. Analysis of the AFLP dataset by upgma resolved two main clusters as shown in Fig. 1. Cluster A contained only isolates that originated from China, Nepal and Japan, including all except two isolates (7107 and 13063) of F. asiaticum. This cluster contained isolates of F. asiaticum with different Fg16 SSCP types, viz. 3 (N1), 4 (N3) and 5 (WL1, Z-5047 and Chinese isolates). Also within this cluster were two isolates of F. meridionale (N5 and N6). Cluster B consisted exclusively of isolates of F. graminearum and included isolates of Chinese, European and USA origin. There was no evidence for any subgroups relating to geographic origin within this cluster. Four isolates did not cluster within either of the main groups. These included two isolates of F. asiaticum (7107 and 13063) of Chinese origin, a F. meridionale isolate (N2) and isolate N4, both from Nepal.
Of the 31 chemotyped isolates of F. asiaticum from China, 25 were chemotype 3-AcDON. Only four isolates (4020, 5035, 6009 and 7069) were 15-AcDON and three (2012, 12003 and 13033) were chemotype NIV. In contrast, the four isolates of F. asiaticum from Nepal and Japan were all chemotype NIV. All Chinese (9) and American (7) and most European (17) isolates of F. graminearum were chemotype 15-AcDON. Two isolates of F. graminearum from Europe (G5 and I1) were chemotype 3-AcDON and two (F6 and F8) were NIV.
Pathogenicity towards wheat seedlings
Scores for penetration and spread were significantly correlated (r = 0·95), indicating that disease index (DI) provides a robust measure of aggressiveness. Isolates differed significantly (P < 0·001) in aggressiveness, with DI scores ranging from 36 to 326 (Table 1). The overall disease level across the 72 isolates was significantly less in the second experiment (mean DI = 235) than in the first (mean DI = 131) (Table 2), but results from the two tests were significantly correlated (r = 0·72) (P < 0·001). No significant differences in aggressiveness were observed among Chinese isolates in relation to species (F. asiaticum or F. graminearum) (P = 0·056) or chemotype (NIV, 3-AcDON or 15-AcDON) (P = 0·241). Similarly, when additional isolates from elsewhere were included, there was no difference in aggressiveness between F. graminearum and F. asiaticum (P = 0·449) (Table 2). In addition, an overall comparison of all isolates used in this work failed to reveal significant links between aggressiveness and the four geographic regions (P = 0·342) or chemotypes (P = 0·053) (Table 2).
Table 2. Analysis of variance of origin, species and chemotype of Fusarium isolates for disease index (DI) determined using generalized linear modelling
Source of variation
Degrees of freedom
Isolate × Trial
Origin: (China, USA, Europe and Nepal)
Origin × Trial
Species: (F. graminearum, F. asiaticum)
Species × Trial
Chemotype: (NIV, 3AcDON, 15 AcDON)
Chemotype × Trial
This study assessed genetic diversity among F. graminearum (sensu lato) isolated from wheat in China in order to assign isolates to either F. graminearum or F. asiaticum, and examined links between aggressiveness, species designation and chemotype. In addition, Chinese isolates were compared with those from other geographic regions to determine any grouping based on molecular markers or aggressiveness.
SCAR analysis using Fg16F/Fg16R was previously used to distinguish F. graminearum and F. asiaticum isolates from China (Qu et al., 2008) However, in the present study, differentiation of F. asiaticum from F. meridionale on the basis of SCAR fragment size was not reliable. The use of SSCP analysis enabled unequivocal resolution of F. meriodionale from F. asiaticum. Of the four haplotypes (1A, 1B, 1C and 6) observed among isolates of F. graminearum, only 1B was present in the USA and 6 was present in Europe. Haplotype 1C was unique for isolate N4, which was proposed as the new ‘lineage 9’ (Chandler et al., 2003), with its Tri101 sequence differing from the other species recognized to date (O'Donnell et al., 2004). Locating additional isolates resembling N4 and sequencing additional genes will be required to determine if isolate N4 represents a new species. SSCP provides a rapid method for differentiation of F. graminearum, F. asiaticum and F. meridionale. The Fg16F/Fg16R target is located at the XM381603 locus of chromosome 1 in the F. graminearum genome (http://www-genome.wi.mit.edu/annotation/fungi/fusarium/index.html) coding for a hypothetical protein (FG01427·1) whose function is not yet known. SSCP assays of a further 440 isolates from China indicated that 90 were F. graminearum (haplotype 1A) and 350 were F. asiaticum (haplotype 5) with no polymorphism within the species (data not shown). This high level of uniformity contrasts markedly with isolates from elsewhere. For example, three haplotypes (3, 4 and 5) were observed among F. asiaticum isolates from Nepal (Carter et al., 2002) while two haplotypes were observed among the limited number of isolates from both Europe (1A and 6) and the USA (1A and 1B) examined in the present study.
AFLP analysis revealed a high level of genetic diversity among the isolates examined. Individual isolates displayed distinct AFLP profiles (Fig. 1). The most similar isolates were N1 and N3, which differed by a single fragment. Two large groups were revealed by cluster analysis, with cluster A containing SSCP types 3, 4 and 5, characteristic of F. asiaticum, and type 2, characteristic of F. meridionale, while AFLP cluster B consisted of only SSCP types 1A, 1B and 6, characteristic of F. graminearum. Four isolates did not cluster within these groups. These were two Chinese F. asiaticum isolates (7107 and 13063), a Nepalese F. meridionale isolate (N2) and the Nepalese isolate N4. This high level of diversity concurs with results from large-scale studies of F. graminearum within the USA (Zeller et al., 2004; Schmale et al., 2006) and F. asiaticum in China (Gale et al., 2002), which revealed the presence of single large, interbreeding populations in each case. In the present study, isolates of F. graminearum from China clustered with those from Europe and the USA and no evidence for clustering associated with geographic origin was observed. Fusarium asiaticum has not been found in Europe or the USA, but the F. asiaticum isolates from China and Nepal clustered together without any clear differentiation along geographic regions (Fig. 1). This finding is similar to findings from Europe, where no correlation was found between origin or chemotype and random amplified polymorphic DNA profiles (Toth et al., 2005). Further studies are required to establish the degree of genetic isolation, or otherwise, among isolates of F. graminearum and F. asiaticum from different geographic regions. Diversity among Asian populations of F. graminearum sensu lato was reported to be greater than for those from Europe (Gagkaeva & Yli-Mattila, 2004). This may reflect the presence of only F. graminearum sensu sticto in Europe (and the USA), whereas both F. asiaticum and F. graminearum are prevalent on wheat in China, with the former being predominant (Qu et al., 2008). It is unclear why F. asiaticum is absent from Europe and the USA, while F. graminearum is present in all continents. If the two species evolved in sympatry in China, as suggested by O'Donnell et al. (2000), then perhaps some selective advantage has permitted the spread of F. graminearum, while F. asiaticum remains restricted in its distribution. Fusarium asiaticum predominates in the warmer regions of China while F. graminearum was isolated form cooler regions; perhaps ecological factors have had a significant effect on the distribution of these species (Ji et al., 2007; Qu et al., 2008).
In this study, the majority of F. graminearum isolates from China and elsewhere were of the 15-AcDON chemotype (Table 1). In contrast, the majority of F. asiaticum isolates were 3-AcDON, with a few 15-AcDON and NIV. These results are largely in agreement with those of Ji et al. (2007), who observed three chemotypes among populations of F. graminearum and F. asiaticum, except that some F. graminearum isolates were of the 3-AcDON chemotype. Again, further studies with many more isolates of the two species from diverse origins will be essential to confirm these trends. Wang & Miller (1994) observed that 3-AcDON-producing isolates (presumably F. asiaticum) came from warmer regions of China, while 15-AcDON-producing isolates (predominantly F. graminearum) came from colder regions, supporting the view that temperature may influence the distribution of the two species. If seed movement is important in the distribution of these species, as suggested for Canada (Fernando et al., 2006), it is likely that the distribution pattern will change with the increasing trade and exchanges between continents. Joint international efforts may be required for proper and timely monitoring of global population structures of F. graminearum sensu lato. Such action would be particularly important, as would be evidence of differences in pathogenicity associated with geographic regions, species or chemotypes.
Within the current collection, no evidence was found that Chinese isolates of F. asiaticum and F. graminearum differed in their aggressiveness towards wheat seedlings. In addition, the aggressiveness of Chinese isolates of F. graminearum was also found to be similar to that of F. graminearum isolates from other geographic regions. In the present work, although only three isolates of F. meridionale were examined, they were significantly less aggressive towards wheat (P = 0·003) than isolates of F. graminearum or F. asiaticum (results not shown). This concurs with the limited amount of previous work with this species. Isolates of F. meridionale were not found to be highly aggressive towards wheat heads (Goswami & Kistler, 2005). Isolates of F. graminearum were found to be slightly, but significantly, more aggressive towards wheat seedlings than Nepalese isolates of F. asiaticum or F. meridionale, with the effect being greater on maize seedlings (Carter et al., 2002). Wide ranges in aggressiveness were observed within F. graminearum sensu lato in other studies involving different assays. Significant differences in aggressiveness in causing FHB of wheat and rye were observed among Australian and German populations, respectively (Miedaner et al., 2000; Akinsanmi et al., 2006). Recent studies also found a wide variation in aggressiveness among isolates of F. graminearum sensu lato from China and elsewhere (Goswami & Kistler, 2005; Wu et al., 2005). Aggressiveness was observed to be an isolate-rather than species-specific characteristic. Similarly, although only a limited number of isolates were examined in the present study, no evidence was found to suggest that isolates from different geographic regions differed significantly in their aggressiveness. However, the present study examined aggressiveness towards wheat seedlings and it is conceivable that isolates differ in aggressiveness towards other tissues or hosts. Mesterhazy (1984) found that aggressiveness towards wheat seedlings and heads were correlated, but further study is required to confirm this for the seedling assay used in the present study.
Within the current collection of isolates, differences in aggressiveness of different chemotypes towards wheat fell just short of statistical significance (P = 0·053) and much of this effect was caused by the low aggressiveness of the F. meridionale isolates, which were all chemotype NIV. DON and NIV were shown to act as “virulence” factors on wheat (Proctor et al., 2002; Maier et al., 2006). Carter et al. (2002) observed no difference in aggressiveness towards wheat of DON and NIV chemotypes of F. asiaticum, but found that NIV chemotypes were significantly more aggressive than DON chemotypes towards maize. This observation was supported by results using trichothecene-deficient mutants which suggested that NIV, but not DON, contributed towards “virulence” on maize (Maier et al., 2006). However, studies with other trichothecene-deficient mutants provided evidence that DON may play a role in aggressiveness towards maize (Harris et al., 1999). Pathogenicity towards wheat heads was found to be correlated with the amount of mycotoxin produced, but not influenced by the chemotype of the isolate (Goswami & Kistler, 2005). It is probable that aggressiveness is determined by numerous factors other than chemotype, but that this may play a role in relative competitive ability on particular host species. As pointed out by Miedaner et al. (2000), the high level of genetic variation in aggressiveness and other characteristics suggests that these species possess a high level of genetic plasticity that may threaten resistant host varieties. Continued monitoring of populations is required to detect such events, which might pose a threat to the FHB-resistant varieties being produced in different countries relying on a limited number of resistance genes.
This work was supported by EU-INCO (ERBIC18 CT98 0312), the National Natural Science Foundation of China (30530510, 30571160, 30771337) and the Ministry of Science and Technology of China (2006BAK02A12).