Introduction

Fusarium Head Blight is one of the most devastating and intensively studied wheat diseases worldwide. Most recent work focuses on F. graminearum (sensu lato) with various trichothecene genotypes and genetic lineages dominating in different regions of the world (van der Lee et al. 2015).

At present in South Africa, wheat prices are low and production costs are high, with the import of lower quality wheat into the country further reducing the demand for locally grown wheat (Janeke 2014). Price, weather conditions and natural disasters are major contributing factors to the reduction in wheat acreage, but diseases, especially Fusarium Head Blight, which reduces both yield and grain quality, also can be a contributing factor. Increased incidence of Fusarium Head Blight provides an additional economic disincentive for planting wheat and could be one of the reasons why South Africa has become a net importer of wheat.

Fusarium Head Blight is most severe when moist weather occurs during the flowering period in the growing season (Bai and Shaner 1994; McMullen et al. 1997). South Africa is unusual in that much of the wheat grown in the country is irrigated. Careful control of irrigation is required during this time to limit the conditions favorable for the development of Fusarium Head Blight. This disease not only reduces crop yield and seed quality, but also is associated with contamination of the grain by mycotoxins such as nivalenol, deoxynivalenol and zearalenone (Desjardins 2006; Leslie and Summerell 2006; McMullen et al. 1997).

Fusarium Head Blight may be caused by several species of Fusarium, with lineage 7 of F. graminearum (also known as F. graminearum sensu stricto) regarded as the primary causal agent in most geographic regions where the disease occurs. The fungus is homothallic (Leslie and Summerell 2006) with sexual development and the production of ascospores as inoculum thought to be a critical component of the disease cycle (Trail 2009). Based on morphological characters, i.e., spore and cultural characteristics, and biological characters, i.e., cross-fertility, F. graminearum is a single genetically diverse, widely distributed species (Leslie and Bowden 2008). If phylogenetic species descriptions are used, however, then F. graminearum can be subdivided into at least 15 species (O’Donnell et al. 2004, 2008; Sarver et al. 2011; Starkey et al. 2007; Yli-Mattila et al. 2009).

The objectives of this study were to identify the morphological species causing Fusarium Head Blight in the irrigated wheat-growing areas of South Africa and to determine the trichothecene production genotype(s) of the recovered isolates. We hypothesized that: i) F. graminearum (sensu lato) dominates in wheat afflicted with Fusarium Head Blight, and ii) that one of the deoxynivalenol genotypes will dominate within the F. graminearum population, as seen in wheat fields in other temperate wheat-growing regions. This study is the first evaluation of Fusarium Head Blight populations from irrigated wheat in South Africa, and differs significantly from reports of Fusarium populations from rain-fed crops grown in other parts of the country (Boutigny et al. 2011a).

Materials and methods

Field samples

Eight-hundred-and-sixty Fusarium Head Blight-infected wheat heads from commercial no-till wheat fields were collected from seven locations in the main irrigated wheat-growing regions in South Africa during two consecutive growing seasons (Fig. 1). Locations were selected based on past and present histories of Fusarium Head Blight epidemics. One-hundred-and-eighty wheat heads were collected at each of three locations (Prieska, Barkly West and Orania) that commonly report Fusarium Head Blight, and 80 wheat heads per site were collected at the other four locations, which do not regularly report Fusarium Head Blight. The grain samples were threshed manually.

Fig. 1
figure 1

Sampling locations in South Africa’s main irrigated wheat-growing areas. The map delineates South Africa’s nine provinces and identifies some of the major cities in each province

Full size image

Fungal isolation and identification

One Fusarium isolate was selected from each infected wheat head, even if multiple isolations could have been made. Seeds with Fusarium Head Blight symptoms were surface sterilized with 1.25 % (v/v) aqueous sodium hypochlorite solution for 3 min, rinsed three times for 30 s with sterile distilled water and dried at room temperature (20–25 °C) in a laminar flow hood. Surface-sterilized seeds were placed on a semi-selective medium for Fusarium (van Wyk et al. 1986) and incubated at 25 °C under a 12 h light/dark cycle for 4–7 days. After 4–7 days fungal hyphae were transferred to potato dextrose agar (PDA) [3.9 % (w/v) PDA bacteriological grade (BioLab®)] plates and incubated in a sterile growth chamber with a 12 h light/dark cycle at 25 °C for an additional 4–7 days. These fungal isolates were cultured on Spezieller Nährstoffarmer agar (SNA) (Nirenberg 1976) plates until conidia formed. Cultures originating from SNA plates were purified by subculturing single spores for identification purposes following dilution plating onto fresh PDA, SNA and carnation leaf agar (CLA) (Fisher et al. 1982; Leslie and Summerell 2006). All cultures used for morphological identification or trichothecene genotyping originated from a culture that grew from a single subcultured macroconidium. Fusarium species were identified morphologically based on the characters described by Leslie and Summerell (2006).

DNA extraction from fungal isolates

All isolates morphologically identified as Fusarium species were subcultured onto PDA plates and grown for 7–10 days at 25 °C. Mycelia were scraped from the PDA plates and freeze dried for three days at -60 °C with a Viritis Advantage Freeze Mobile II (New York City, New York, USA). The freeze-dried fungal material was ground to a fine powder with a Qiagen TissueLyser (Haan, Germany). Total genomic DNA was isolated by using a modified CTAB (hexadecyltrimethylammonium bromide) method (Herselman 2003). The quantity and quality of the DNA samples were determined by measuring absorbance at A260 and A280. Samples were diluted with 1× TE buffer to a working concentration of 200 ng/μl for subsequent experiments.

Trichothecene genotype analyses

The trichothecene genotype of the Fusarium isolates was determined by PCR amplification of the Tri7 and Tri13 genes in the trichothecene biosynthetic gene cluster (Chandler et al. 2003). These genes encode proteins that convert deoxynivalenol to nivalenol (Tri13) and nivalenol to 4-acetyl-nivalenol (Tri7). The selected primer pairs can be used to identify deoxynivalenol (DON) and the nivalenol (NIV) genotypes (Tri7F + Tri7R and Tri13F + Tri13R) or just one genotype at a time (DON: Tri7F + Tri7DON and Tri13F + Tri13DONR; NIV: Tri7F + Tri7NIV and Tri13NIVF + Tri13R). Primer pair combination Tri13F + Tri13R targets the Tri13 gene and can be used to identify both genotypes in a single reaction by amplifying a DNA fragment of 1075 base pairs (bp) associated with the NIV allele and/or a 799-bp fragment associated with the DON allele. Amplified fragment sizes between 458 and 535 bp resulting from amplification with the Tri7F + Tri7R primer pair amplification are associated with the DON genotype, while an amplified fragment size of 436 bp is associated with the NIV genotype.

PCR amplification reactions were performed in a total volume of 20 μl, containing ~1 μg of fungal genomic DNA, as described by Chandler et al. (2003). The optimized cycling conditions for the primers used were: one cycle at 94 °C for 2 min, followed by 35 cycles of 94 °C for 30 s, 58 or 60 °C (depending on primer combination) for 30–45 s and 72 °C for 30 s followed by a final extension of 72 °C for 5 min and a 10 °C hold (Chandler et al. 2003).

Results

Morphological characterization

Based on morphological characterizations of 860 isolates, six Fusarium species, F. chlamydosporum (0.3 %), F. crookwellense (1 %), F. culmorum (2 %), F. equiseti (3 %), F. graminearum (93 %) and F. semitectum (0.7 %) were recovered. Three of these species, F. crookwellense, F. culmorum and F. graminearum, have previously been widely associated with Fusarium Head Blight. Fusarium graminearum was the dominant species associated with Fusarium Head Blight in South Africa and was present at high frequencies at all seven sampled locations (Table 1). The other two species known to cause Fusarium Head Blight, F. crookwellense and F. culmorum, were limited to Cradock, Greytown and Potchefstroom, in the less temperate wheat production regions (Fig. 1, Table 1).

Table 1 Frequency of deoxynivalenol (DON) and nivalenol (NIV) genotypes in various Fusarium species present in seven field populations of irrigated wheat in South Africa
Full size table

Fusarium equiseti often is regarded as a saprophyte, but has been recovered at low frequencies from Fusarium Head Blight-infected grains in some European countries and Australia (Akinsanmi et al. 2004; Bottalico and Perrone 2002; Ioos et al. 2004). The highest percentage (6 %) of F. equiseti was found in the Barkly West field population, although this species was recovered at four of the seven sampled locations. Barkly West was the only location at which F. chlamydosporum (1 %) was recovered. The composition of Fusarium field populations at Prieska, Orania and Douglas were similar with only two species detected, F. graminearum and F. equiseti. The Prieska field population had the highest percentage (99 %) of F. graminearum isolates.

Four Fusarium species, F. graminearum, F. culmorum, F. crookwellense and F. semitectum were detected at Greytown, Potchefstroom and Cradock (Table 1). The Greytown field population had the highest observed frequencies for F. crookwellense (9 %) and F. semitectum (3 %). The Greytown and Potchefstroom field populations had the lowest percentage (84 %) of F. graminearum isolates. The Potchefstroom field population contained the highest frequency of F. culmorum (11 %). The lowest frequencies of F. crookwellense and F. semitectum were observed in the Cradock field population, but this population contained the second-highest frequency of F. culmorum (8 %).

Trichothecene genotype analysis

The trichothecene production genotype (DON or NIV) was determined for all isolates of F. crookwellense, F. graminearum, and F. culmorum, by using PCR assays. Trichothecene genotypes were not determined for isolates of F. chlamydosporum or F. semitectum. Most (89 %) of the F. graminearum isolates were DON. Fusarium equiseti isolates were ~50:50 DON:NIV. The NIV genotype was slightly favored (56 %) amongst the F. culmorum isolates. All of the F. crookwellense isolates were NIV, which is consistent with previous reports that isolates of F. crookwellense produce only nivalenol (Desjardins 2006).

Trichothecene genotype frequencies varied by location (Table 1). The potential for production of deoxynivalenol was highest at Prieska, where 99 % of the strains had the DON genotype. The F. graminearum populations at all locations had a higher frequency of DON (63–99 %) than of NIV alleles. There were more DON isolates than NIV isolates of F. equiseti at Prieska, Douglas and Orania, and more NIV isolates than DON isolates at Barkley West. For F. culmorum, there were more DON isolates at Cradock and Potchefstroom and more NIV isolates at Greytown.

Discussion

Fusarium Head Blight in South Africa

The first report of Fusarium Head Blight in South Africa was in 1980 in the North West Province along the Vaal River (Scott et al. 1988). The farm, in the Potchefstroom district (North West Province) from which we sampled infected wheat heads, also is located on the Vaal River. In 2006 Fusarium Head Blight was detected in this area for the first time since the original 1980 report.

Fusarium Head Blight was epidemic in the early 1990s on farms along the Orange River in the Northern Cape where wheat is grown under center-pivot irrigation (Boshoff et al. 1999). Four of the seven locations (Prieska, Barkly West, Orania and Douglas) from which we collected are in this portion of the Northern Cape. Prieska, in 1991, was one of the first locations in the Northern Cape where Fusarium Head Blight was detected on wheat under center-pivot irrigation (Boshoff et al. 1999). Since 1991 epidemics of Fusarium Head Blight have continued in the irrigated wheat fields of the Northern Cape, with factors such as tillage practices and a wheat/maize crop rotation as major contributing factors.

Elsewhere, Fusarium Head Blight outbreaks in KwaZulu-Natal were first reported in the 1985/86 season and included Greytown, one of the locations where we sampled. The most severe outbreaks in KwaZulu-Natal occurred near Winterton, an area where irrigated wheat often is rotated with maize in a long-standing maize production area. Follow-up reports of Fusarium Head Blight outbreaks near Greytown resumed in 2004 and reached epidemic status in 2006. In the Eastern Cape province, wheat farms are irrigated with water from the Great Fish River. Farmers in the Cradock district of the Eastern Cape were first plagued with Fusarium Head Blight in 2006.

Our results are consistent with previous reports that F. graminearum (sensu lato) is responsible for most of the outbreaks of Fusarium Head Blight in South Africa (Boshoff et al. 1999; Boutigny et al. 2011a; Scott et al. 1988). However, we focused on irrigated wheat, an unusual method for cultivating wheat, which usually is grown as a rainfed crop. The present study includes the bulk of the irrigated wheat-growing regions in South Africa and evaluated more isolates from more locations than did previous reports. Three of the six identified species, F. crookwellense, F. culmorum and F. graminearum, have been associated previously with Fusarium Head Blight in South Africa. These three species plus F. equiseti all have been associated previously with Fusarium Head Blight in other parts of the world (Scott et al. 1988; Snijders and Snijders and Perkowski 1990; Sutton 1982). The present report of F. equiseti associated with Fusarium Head Blight is the first of this association for South Africa. The available data do not suffice to claim that this fungus causes Fusarium Head Blight in South Africa as Koch’s postulates have not been completed with the isolates that we recovered. Fusarium equiseti was recovered only in temperate regions in conjunction with F. graminearum.

Distribution of Fusarium species and toxin genotypes

The toxin genotypes are predictive of, but not guarantors of, mycotoxin production. Mycotoxin production depends on environmental conditions and the function of all of the enzymes in the biosynthetic pathway. The relative frequency of toxin genotypes varied somewhat by species and by location, with 85 % of the strains evaluated having a DON genotype and 15 % having a NIV genotype. From a food safety perspective, nivalenol is approximately ten fold more toxic to mammals than is deoxynivalenol (Desjardins 2006). If all of the strains produce toxins at a similar level, then the potential threat to food safety is probably larger from nivalenol than it is from deoxynivalenol, as strains with a NIV genotype compose more than 10 % of the total population.

The distribution of Fusarium species varied by location. Four locations – Prieska, Barkly West, Orania and Douglas – have a history of Fusarium Head Blight epidemics since the early 1990s. The wheat-growing areas of this region are all under center-pivot irrigation and most of the farms, including the ones where we sampled, follow a double-cropping rotation system that probably contributes to the high disease incidence observed. The no-till maize fields are hypothesized to provide the primary inoculum for Fusarium Head Blight on wheat grown under center-pivot irrigation, with warm day temperatures and somewhat cooler night temperatures encouraging sexual reproduction and the production of ascospores, which serve as the primary inoculum for wheat (Boshoff et al. 1999).

Fusarium graminearum was the dominant species in the Fusarium Head Blight population from the main irrigated wheat-growing areas in South Africa constituting 93 % of the isolates recovered. This finding is consistent with the results of a recent study of F. graminearum from multiple hosts in South Africa (Boutigny et al. 2011a). Those authors identified several morphologically indistinguishable genetic lineages of F. graminearum (sensu lato) in their studies – lineages 2 (F. meridonale), 3 (F. boothii), 5 (F. acacia-mearnsii), 7 (F. graminearum sensu stricto), 8 (F. cortaderiae) and 9 (F. brasilicum) (Leslie and Summerell 2006).

Two other crops, maize and barley, may alter the fungal communities responsible for Fusarium Head Blight of wheat. Isolates of F. graminearum lineages 3 and 7 have been reported from barley in South Africa (Boutigny et al. 2011a, b). With the exception of one strain from several hundred, all of the South African barley isolates produced deoxynivalenol rather than nivalenol.

The lineage to which most isolates of F. graminearum from maize in South Africa belonged varies based on the portion of the plant that was sampled. Isolates from lineage 7 were common on maize crowns and roots (Boutigny et al. 2011a; Lamprecht et al. 2011), but were not recovered at all from maize ears (Boutigny et al. 2011a). Isolates from lineages 2 and 3 were both recovered from maize roots (Boutigny et al. 2011a; Lamprecht et al. 2011), but only lineage 3 isolates were reported from maize ears (Boutigny et al. 2011a). The lack of isolates of lineage 3 in our large set of samples suggests that strains belonging to this lineage are unable to produce ascospore inoculum either in sufficient quantity or at the right time to cause major problems in irrigated wheat fields. Most isolates from maize produced deoxynivalenol, with isolates from lineages 2, 5 and 8 responsible for most of the nivalenol isolates identified in these previous collections. Given the dominance of lineage 3 isolates in the maize samples, more work is needed to confirm that maize field debris left in no-till fields is the inoculum source for Fusarium Head Blight of wheat (Schaafsma et al. 2005) in the locations we evaluated. The population composition of F. graminearum (sensu lato) varies by location (van der Lee et al. 2015) and it is possible that multiple genetic lineages/phylogenetic species could be coexisting in the irrigated wheat fields of South Africa.

Two other Fusarium species, i.e., F. crookwellense and F. culmorum, associated with Fusarium Head Blight, were restricted to less temperate regions of the country in our study. Fusarium equiseti and F. poae also have been recovered from barley in South Africa (Boutigny et al. 2011b). These results are consistent with previous hypotheses that F. graminearum dominates in temperate, humid areas of the world, whereas F. culmorum and F. crookwellense occupy the corresponding niche in cooler regions (Parry et al. 1995; Waalwijk et al. 2003

Fusarium graminearum dominated at all seven locations sampled in the present study. This geographically widespread domination suggests that a shift has occurred in the composition of the Fusarium populations that cause Fusarium Head Blight, especially in the less temperate irrigated regions. The frequency of F. graminearum also has increased relative to F. culmorum in several European countries (Brennan et al. 2005; Waalwijk et al. 2003) and has been attributed to increased maize production.

We also find that the frequency of F. graminearum has increased, but changes in cropping pattern, which appear relatively stable, seem unlikely to be the reason for the observed shift in South Africa. No-till farming systems are common in the surveyed areas and could increase Fusarium Head Blight incidence and severity, as the debris from a previously infected crop could serve as an inoculum source for subsequent crops. Rotation with maize has been suggested to increase disease incidence and severity in wheat, as maize is an excellent host for F. graminearum. The fungus can colonize the entire maize plant and also can cause economically important stalk and ear rots.

Spores of F. graminearum also may be carried long distances in the atmosphere (Schmale et al. 2006), but these spores alone probably do not suffice to induce a widespread epidemic, although they could be important on a local scale (Schmale et al. 2005). Countering this airborne spore dispersal requires regional changes in tillage and crop rotation practices to effect significant disease control. Even if atmospheric spore movement is insufficient to initiate a major epidemic, it could move enough spores for migration to mix local subpopulations and prevent them from becoming genetically isolated.

Cropping practices probably are the most important influence on Fusarium Head Blight in the cooler regions of the country, where temperature increases could change the spectrum of Fusarium species present, as has been hypothesized to occur in Europe (Waalwijk et al. 2003). The relatively low frequencies of F. culmorum and F. crookwellense identified in this study are consistent with the hypothesis that an increase in average temperature could be changing the species composition of the fungal community in South African wheat fields.

Fusarium Head Blight epidemics are accompanied by the accumulation of mycotoxins, which can have a detrimental effect on food and feed quality and safety. The trichothecene genotypes, DON and NIV, can be distinguished in these South African populations by using allele-specific PCR assays (Lee et al. 2001, 2002) for the Tri7 and Tri13 genes (Chandler et al. 2003). The DON genotype (85 %) dominated in populations of all Fusarium species evaluated in this study. The DON genotype was more common in the F. graminearum population (89 %) than it was in the Fusarium population as a whole. The high frequency of the DON genotype is consistent with selection for this genotype in these populations. Other studies of F. graminearum have found that deoxynivalenol is important in the host plant/pathogen interaction in cereal grains (Brown et al. 2001; Champeil et al. 2004; Desjardins et al. 1996), which could suffice to explain the observed differences in the frequencies of the DON and NIV genotypes. Twenty DON isolates from the Prieska population were tested in a preliminary study and all were highly aggressive on wheat (de Villiers 2009).

In summary, we found that F. graminearum is the dominant fungal species associated with Fusarium Head Blight in South African wheat fields under center-pivot irrigation. The DON genotype dominated in the individual populations and in the population as a whole, a result that is consistent with the hypothesis that isolates with the DON genotype are more aggressive pathogens on wheat than are those with the NIV genotype. The relatively constant composition of the various populations, at least in terms of species present and the frequency of the DON genotype, suggests that selection pressures are similar across all irrigated wheat regions in South Africa, and that there is sufficient movement of isolates between locations to keep the local communities heterogeneous in spite of significant geographic separation and climatic differences. The potential shift in pathogen community composition due to climate change will be important to monitor in the coming years.