Fusarium head blight (FHB) is recognized as a major threat for oat production [1
]. The disease is caused by different species belonging to the genus Fusarium
, and leads to the accumulation of mycotoxins in grains that are toxic for humans and animals [4
]. In oats from Scandinavia, the most dominant FHB-causing species is F. graminearum
, producing deoxynivalenol (DON) [6
]. However, in other European areas, a considerable increase in F. poae
(FP) and F. langsethiae
(FL) infections has been observed in oats [9
]. These pathogens produce highly noxious toxins, putting food and feed safety in jeopardy. FP is the main producer of nivalenol (NIV), a vomitoxin that is more than 10 times more toxic than DON [14
]. FL has been associated with the recent increases of oat grain contaminations with hazardous T-2 and HT-2 toxins [11
]. In fact, T-2/HT-2 contaminations caused an outbreak of alimentary toxic aleukia that was responsible for the death of a thousand people in Russia in the 1940s [16
]. Guidance levels have been defined to limit the amount of T-2/HT-2 entering the food chain (European Commission 2013/165/EC [17
]); however, limits for NIV have not yet been defined. FP and FL are emerging pathogens, and in contrast to F. graminearum
, very little is known about the life cycle, the epidemiology, and environmental requirements for infection and toxin production. Without such information, it remains difficult to forecast FL and FP infection, and thus avoid toxin contaminations [13
Oat has long been regarded as more resistant to FHB than wheat or barley. This may be attributed to the lack of visual symptoms on infected panicles, and the presence of long pedicels between spikelets that usually prevent the spread of fungal mycelia throughout the panicle [6
]. Yet, oat grains can accumulate considerable mycotoxin amounts. The use of resistant oat varieties would be the most sustainable way to overcome such contaminations [19
]. However, the selection of FHB resistant oat varieties has been strongly delayed due to a lack of knowledge about FP and FL infection processes and their impact on oat tissues. Nevertheless, it has been demonstrated that oats, much like other small grain cereals, are susceptible to primary infection during anthesis [21
]. Fungal hyphae enter via the tips of the glumes, quickly colonize the anthers, and then reach the developing kernels, which are frequently aborted [23
]. Besides toxin contamination, the consequences of Fusarium
infections on oats are not yet clearly identified. Some epidemiological studies relate a significant reduction of grain weight and germination ability [24
], whereas others report minimal consequences on crop yields and grain properties [6
]. Moreover, the vast majority of studies aiming to describe FHB symptoms only focus on plants infected with F. gramineraum
or F. culmorum
]. As of yet, it is not known how oat grains react to the infection with FP and FL.
As in other small grain cereals, oat resistance against FHB has a complex, quantitative character that is based on the cumulative effect of several genes [26
]. Up to now, all of the resistance studies in oats aimed to limit the toxin contamination in grains, which is defined as type V resistance in wheat [28
]. It has been observed that naked grains accumulate lower amounts of DON than hulled grains [3
]. Recently, several minor quantitative trait loci (QTLs) associated with resistance to DON accumulation have been identified in oats [26
]. It is not clear if resistance against DON contamination also provides resistance against NIV and T-2/HT-2 toxins. Besides resistance type V, other resistance types have not yet been described in oat. The resistance type II, which is described in wheat as the resistance against disease propagation in head tissue [30
], is inherent to the panicle shape [6
]. Due to the limited knowledge of oat grain responses to FHB, so far it has not been possible to examine the resistance against kernel infection (type III) and the tolerance of grain in terms of yield (type IV).
The aim of this study is to investigate the responses of oat grains to infection by FP and FL. Seven oat genotypes were artificially inoculated at three different field sites. Each experiment was composed of plots inoculated with either FP or FL, as well as non-inoculated plots. The success of the infections was determined by the observation of grains infected by the pathogens and measurements of fungal DNA in three varieties of grains. The contents of NIV and T-2/HT-2 toxins were measured in all of the grains from the inoculated plots. The impacts of FP and FL on grain properties were assessed by comparison of the grain weight, β-glucan contents, and protein contents between inoculated and non-inoculated grains.
The presence of different types of mycotoxins in oat grains is a menace for the whole value chain. The mycotoxins found in the current study are produced by several Fusarium
species. In particular, F. poae
(FP) and F. langsethiae
(FL) caused contamination of grains with highly toxic NIV and T-2/HT-2 [13
], respectively. In some geographic areas, F. graminearum
and its mycotoxins deoxynivalenol (DON) and zearalenone are also found in oat grains [6
]. This study aims to better understand the response of oat grains in the presence of FP, FL, NIV, and T-2/HT-2 toxins. Using artificial inoculations, we intended to trigger the response of oat grains and to challenge the resistance of seven oat genotypes. Despite artificial inoculations, no symptoms were detected on the panicles in all of the experiments. This finding is in agreement with other studies on oats [2
]. Severe symptoms such as blighted spikelets, the discoloration of glumes, floral abortions, and browning of rachis attributed to Fusarium
infections are rare, and mainly observed in Northern Europe [6
]. Divon et al. [23
] observed these symptoms on oat panicle strongly infected with FL under controlled conditions. The presence of such FHB symptoms on panicles is probably due to long exposure to high humidity [24
]. In the present study, without visible signs of FHB outbreak, the success of infections was assessed by recovering the inoculum on the grains using a seed health test (SHT) [32
] and quantitative PCR for the amount of fungal DNA. Infections with FP were successful at all three field sites, whereas FL infections could only be tracked in Reckenholz. The average temperature of about 20 °C was rather similar at the three field sites, but the meteorological conditions in Reckenholz, with a high relative humidity and a long rainy period during and after anthesis, might have favored the infection by FL. Indeed, in vitro studies showed that FL requires longer periods of high humidity, compared with FP and other Fusarium
Once the grains are infected, both FP and FL produce mycotoxins. The infection of FP caused an accumulation of NIV in similar contents across all of the field sites. In more contrasting environments across the United Kingdom and Canada, Edwards et al. [10
] and Tekauz et al. [18
], respectively, noticed the strong influence of environmental conditions on toxin accumulation in grains. In the present study, it was not possible to investigate the impact of environmental conditions on T-2/HT-2 accumulation. T-2/HT-2 were detected in considerable amounts in all of the FL-infected grains from Reckenholz, with a maximum value of 180 μg/kg. Yet, the results of the SHT and the quantification of FL DNA revealed a low degree of kernel colonization by FL, whereas FP caused more severe grain infections. This result suggests a high toxigenicity of FL and indicates that FL infections on oat crops, although less dominant than FP, nevertheless constitute a real threat to the safety of oat products. In plants, T2/HT2 are transformed in a glycosylated form [40
] that may not be detectable by the ELISA kit (manufacturer, personal communication).
Besides the accumulation of toxins, FP and FL altered the β-glucan content in grains. Depending on the genotypes and environmental conditions, the β-glucan contents increased or decreased subsequent to FP and FL infections. In particular, the β-glucan content in grains of the naked variety Samuel substantially increased following infections by both FP and FL (+48% and +64%, respectively), whereas the changes in β-glucan content caused by FHB were limited for the other genotypes (approximately +20%). Hence, with the current data set, no clear tendency in β-glucan content variation could be identified. In a very similar study on barley grains, a 10% reduction of β-glucan content was observed in six barley varieties after inoculation with F. graminearum
]. On one hand, the reduction of β-glucan content might be attributed to the β-glucanase activity observed from Fusarium
]. On the other hand, since β-glucan are mainly concentrated in the outer layers of oat grains [43
], such variations in β-glucan content might be attributed to changes in grain morphology caused by the infection. With respect to the TKW of grains, FP and FL infections showed only a weak impact in the present study. Over all three environments, only a weak increase of TKW of grains from Triton was measured when infected with FP. This indicates that both FP and FL had very little to no impact on grain filling, suggesting a minimal impact of FHB on oat yield. However, these observations are likely conditioned by the prevailing environmental conditions in Switzerland. In Brazil, with temperatures above 20 °C and relative humidity frequently exceeding 90%, Martinelli et al. [44
] observed a significant reduction of oat yield, up to 25%, due to FHB caused by F. graminearum
. Furthermore, Mielniczuk et al. [25
] detected TKW reductions of 32% in oats after field inoculations with F. crookwellense
in Poland. We also assume that a reduction of grain filling and yield is specific to the particular Fusarium
species causing FHB. In the present study, we did not observe a significant impact of FP and FL on the protein content in oat grains. Finally, our grain analyses showed that even if FP and FL have an impact on some of the grain properties, these alterations are not reliable and useful symptoms of FHB in breeding for resistant varieties.
For the understanding of resistance against FHB in oats, we compared the types of resistance with those for wheat, where type I and type II comprise the resistance against the primary infection and the spread of the infection throughout the spike, respectively [30
]. In the present experiments with oats, we could not distinguish these types of resistance. Types III to VI describe the resistance of the kernel, where type III is the resistance against colonization, type IV is the tolerance against FHB in terms of yield, and type V is the resistance against the accumulation of trichothecenes [28
]. The three cultivars Canyon, Husky, and Triton displayed a similar degree of fungal incidence, and hence, a type III resistance was not detected. Nevertheless, Gagkaeva et al. [36
] noticed significant differences in F. sporotrichioides
DNA in oat grains among 105 genotypes, indicating the presence of type III resistance elements that remains to be investigated. In the present study, there was no significant weight loss or changes in TKW between the varieties and therefore, no detectable type IV resistance. Among the panel of oat cultivars, none of them was able to completely avoid contaminations with NIV or T-2/HT-2 (type V resistance). Nevertheless, we observed significant differences in toxin accumulation between the genotypes. This result is consistent with Yan et al. [3
], who previously suggested that, to date, oat FHB resistance can only be reflected by type V resistance. However, the authors detected only small differences in mycotoxin accumulation between the genotypes, and none displayed a high type V resistance. The general susceptibility of oat to toxin contamination was also reported in several previous studies on both modern oat varieties and oat landraces [18
]. In the current study, we observed that resistance of oat genotypes to toxin contamination was strongly impacted by environmental conditions, with strong GxE interactions explaining 46% of NIV accumulation in grains infected with FP. This finding suggests a weak stability of resistance in oat genotypes, which could impede the breeding for FHB resistance. Moreover, when comparing T-2/HT-2 and NIV accumulation in FL and FP-infected grains from Reckenholz, we noticed that the variety Triton did not accumulate NIV, whereas considerable amounts of T-2/HT-2 toxins (90 μg/kg) were detected. In fact, Edwards et al. [10
] found no correlation between HT-2 and DON content in oat grains. Hence, it is likely that type V resistance depends on the nature of the toxin, and the QTLs associated with low DON accumulation [26
] will not provide enhanced resistance to other mycotoxins. In our study, only the variety Samuel was resistant to both T-2/HT-2 and NIV across all field sites. This variety produces naked grains, suggesting a role of hulls in mycotoxin accumulation [29
]. Interestingly, in grains from Samuel, we also measured high increases of β-glucan content. We suggest that this reaction of the grain also contributes to the low accumulation of toxins.
Overall, the results presented here show that oats at the panicle level are constitutively resistant to FHB. However, the pathogen colonizes the glumes and grains depositing mycotoxins in the developing grain. Only a weak resistance against mycotoxin accumulation was observed, which was probably due to the absence of glumes (naked grains) in the variety Samuel. Improving resistance against mycotoxin accumulation may be achieved by the use of resistant wild oats (i.e., Avena sterilis
) and other germplasms as new source of resistance for breeding [6
4. Materials and Methods
4.1. Plant Material
Seven oat varieties were used for this study. The varieties differ in their morphological properties, favored environmental conditions for cultivation, and resistance to DON accumulation (Stefan Beuch, personal communication). Plant height, panicle characteristics, and grain characteristics are presented in Table 4
]. Descriptions of these characteristics were carried out in the field at Changins in 2015. Ten plants of each variety from non-inoculated plots were randomly chosen, and the averages of plant height and number of seeds in a panicle were calculated.
4.2. Fungal Material
Artificial inoculations were carried out with three strains of FL and three strains of FP, all isolated from infected oat grains (Table 5
Strains were retrieved from long-term storage in a 1:1 mix of water with glycerol at −80 °C. For the mass production of each isolate, strains were cultured on potato dextrose agar (PDA, BD Difco, Le Pont de Claix, France) for one week at 18 °C with 12 h/12 h UV light/darkness. Subsequently, two discs (5-mm diameter) from the outer margin of a well-grown colony were transferred to 200 mL of liquid V8-medium in a 1L capacity Erlenmeyer flask. A V8 medium consisted of a 1:5 mix of V8 juice (Campbell Soup Company, Camden, NJ, USA) and distilled water with 2 g sodium carbonate per liter. Cultures were incubated on a shaker at 200 rpm for 7 days at 24 °C in the dark. Cultures were then filtered through sterile cheesecloth to remove all mycelia. Finally, the culture medium was removed by centrifugation at 4500 rpm for 10 min, and the pellet was re-suspended in sterile distilled water. These preparations were either used immediately or stored at −20 °C.
4.3. Field Experiments and Artificial Inoculations
Field experiments were conducted in 2015 at three locations in Switzerland: Changins (Canton Vaud, southwest [46°24′36′′/6°14′06′′]), Reckenholz (Canton Zürich, northeast [47°16′30′′/8°26′45′′]), Cadenazzo (Canton Ticino, south of the Alps [46°09′00′′/8°57′00′′]). In Changins and Reckenholz, oat grains were sown in March 2015; in Cadenazzo, oat grains were sown in November 2014. At all of the field sites, planting and management was conducted with the same experimental design and protocol. The varieties were sown with a Seedmatic seeding machine (Hege Maschinen, Eging am See, Germany) in 1 m2 microplots with five rows and 15 cm space between rows. Artificial inoculations took place when 50% of the plants within each microplot were at mid-anthesis (BBCH65). Inoculum was prepared immediately before the inoculation by mixing an equal proportion of the liquid culture of each of the three strains of FP or FL, and adding 0.0125% of Tween®20 (Sigma-Aldrich Chemie GmbH, Darmstadt, Germany). A concentration of conidia and volume of prepared suspension were adjusted to 7.5 × 106 conidia per microplot. Suspensions were applied with a hand-held sprayer (Spray-Matic 1.25P, Birchmeier, Stetten, Switzerland) at dawn. If required and according to weather conditions during inoculations, the plots were irrigated to maintain humidity on the panicles for at least 24 h, with a maximum of 600 L of water per hectare. A high pressure/low volume overhead spray irrigation system was available in Changins, while in Cadenazzo, an equivalent of 300 L/ha of water was sprayed manually with a backpack sprayer. Additional irrigation was not required at the Reckenholz site. The additional irrigation enhanced the infection pressure. The humidity regime for the plants increased by 30 mL/m2, which was an insignificant change in the humidity regime for the plants. Three artificial inoculations were carried out at three-day intervals.
The meteorological conditions between anthesis and harvest were measured with the meteorological stations of Meteoswiss [33
] based at each experimental site. The meteorological stations were at a distance of 500 m (Changins and Reckenholz) and 150 m (Cadenazzo).
4.4. Harvest and Sample Collection
All of the oat plots were harvested at full maturity (BBCH 89) during the first week of August 2015 at the three field sites. A combine harvester (HEGE 140, Eging am See, Germany) was used, with a reduced airflow to recover a maximum of kernels. Grains were dried to a maximum of 14% moisture, and cleaned using a vertical airflow (Baumann Saatzuchtbedarf, Waldenburg, Germany) to remove dust and other debris. At this stage, clean grains with husks were obtained. Two hundred gram grain samples were extracted after three times homogenization with a rifle divider (Schieritz & Hauenstein AG, Laufen, Switzerland). Fifty grams were extracted from these subsamples and milled separately, with a sample mill using a 0.25 mm screen (1093 Cyclotec Sample Mill, FOSS, Höganäs, Sweden) to obtain whole meal flour samples. All of the grain and flour samples were stored at −20 °C until further analyses.
4.5. Analyses of Grain Infection
The presence of FP and FL on oat grains confirmed the success of the artificial inoculations. The infection of grains was assessed by the incidence of grains infected by FP and FL (%), and by the quantification of fungal DNA in grains.
4.5.1. Proportion of Grains Infected by F. poae and F. langsethiae
The proportion of grains colonized with Fusarium
was determined using the seed health tests method described in Vogelgsang et al. [32
] and used in Schöneberg et al. [13
] and Martin et al. [41
]. Essentially, the seed health test consists in the culture of surface sterilized grains on the Fusarium
selective media containing pentachloronitrobenzene [13
]. For three to 10 days, the cultures were observed, and growing Fusaria were scored and taxonomically identified. Samples from the varieties Canyon, Husky, and Triton from both inoculated and non-inoculated plots were analyzed to determine the infection of grains not only from FP and FL, but also from other Fusarium
species. The different Fusarium
species were identified according to the Fusarium
laboratory manual by Leslie and Summerell [49
4.5.2. Quantification of Fungal DNA
The severity of FP and FL infection was measured by quantitative PCR (qPCR) of fungal DNA in the grains. Protocols for DNA extraction, total DNA quantification, and quantitative PCR reactions are described in Schöneberg et al. [13
]. In every performed assay, all standards as well as the negative control (double distilled water) were run as triplicates.
Briefly, for FP qPCR, the primer pairs ACL1-F160 and ACL1-R330 and TaqMan-Probe (ACL1_poae1_probe) (Qiagen AG, Hombrechtikon, Switzerland) were used. For FL qPCR, the protocol specifications and thermocycling parameters were employed, as described in Edwards et al. [50
], and adapted to the available reaction mixes and laboratory devices as detailed in Schöneberg et al. [13
]. The amplification mix consisted of the primer pairs FlangF3 and LanspoR1 [51
] and IQ SYBR®
Green Supermix (Bio-Rad, Cressier, Switzerland).
4.6. Mycotoxin Analysis
NIV contents were examined in all of the samples from plots inoculated with FP, and three non-inoculated plots. The identification and quantification of NIV was done using LC-MS/MS. Sample preparation and LC-MS/MS measurements followed the protocols detailed in Schöneberg et al. [13
Contents in T-2/HT-2 toxins in grains were measured in all of the samples from plots inoculated with FL, and three non-inoculated plots. Contents were determined in flour using an enzyme immunoassay kit for the quantitative screening of toxins (Ridascreen® T-2/HT-2 Toxin, R-biopharm AG, Darmstadt, Germany), according to the manufacturer’s instructions. The protocol specific for the extraction of T-2/HT-2 in oat flour was followed. The detection limit was 30 μg/kg, according to the manufacturers’ manual.
All of the mycotoxin concentrations in grains were expressed as μg/kg of dry weight.
4.7. Analyses of Grain Properties
Grain weights were measured for all of the grain samples using the thousand kernel weight (TKW) indicator. TKW were measured with the MARVIN optical grain counter (Digital Seed Analyser, GTA Sensorik GmbH, Neubrandenburg, Germany) and a balance (Mettler PM2000, Mettler-Toledo, Greifensee, Switzerland).
The content of β-glucan was determined in all of the samples (inoculated and non-inoculated plots) with the mixed-linkage β-glucan kit (Megazyme International Ireland Ltd., Wicklow, Ireland). The assay procedure of mixed-linkage β-glucan in oat flour—streamlined method—(ICC Standard Method No. 166 [52
]) was followed and adapted to the laboratory facilities.
The protein contents (%) were examined in all of the samples of hulled oat grains using near-infrared reflectance spectroscopy (NIRS) with a NIRFlex N-500 (Büchi Labortechnik AG, Flawil, Switzerland). The protein calibration of NIRFlex was regularly adjusted, and basis analyses were made with the Kjeldahl method, according to ICC standard method No.105/2 [53
]. The coefficient of confidentially of the calibration was R2
= 0.61 (C. Brabant and C. Oberson, personal communication).
4.8. Experimental Design and Statistical Analyses
Field design, inoculation methodology, harvest preparations and sample preparations were similar across the three field sites. Each field site consisted of three replicates with FP inoculations, three with FL inoculations, and three without artificial inoculations, planted in a split-plot design. Proportion of FP and FL-infected grains and measurements of fungal DNA were carried out on grains from inoculated plots of the three field sites for three selected oat genotypes (Canyon, Husky, and Triton). Mycotoxin measurements were performed on grains from inoculated plots in the three field sites for all seven oat genotypes. TKW, β-glucan content, and protein content were measured in all of the grain samples.
Statistical analyses were carried out using the statistical software R [54
]. The effect of FP and FL infection on grains were separately analyzed. Analyses of variance ANOVA were performed on the proportion of grains infected by FP or FL, fungal DNA, and mycotoxin contents using the environment (field site) as the main factor, and the genotypes as the sub-factor. Data were normally distributed. The impact of FP or FL infections on TKW, β-glucan content, and protein content were analyzed by ANOVA using the environment as the main factor, the presence of artificial inoculation as the sub-plot, and the genotypes as the sub-sub-plots. After ANOVA, a multiple comparison was performed on significant factors using Tukey’s test with β-glucan (package “agricolae”, De Mendiburu [55
]. Pearson correlations were used to verify correlations between toxin contents, fungal DNA concentration, and the proportion of Fusarium