The present study aimed to determine the potential for Bd to act as a host to Fg and Fc and ascertain whether this interaction might serve as a model of that between Fusarium species and wheat. The results clearly demonstrated the compatibility of interaction between Bd and the two Fusarium species of greatest relevance to FHB, the major Fusarium-associated disease of wheat. Moreover, the development of disease symptoms closely resembled those reported in wheat.
With respect to FHB, after a short asymptomatic period, Bd spikes spray inoculated with Fg conidia displayed small brown spots, which first appeared at the middle or base of the lemma, highly reminiscent of the initial symptoms in wheat . Lesions spread to infect adjacent florets, often provoking the bleaching of the upper part of the spikelet in a manner similar to that observed in wheat [32, 10] and infection extended down the rachis to adjacent spikelets and even colonised peduncles as seen during infection of wheat. Overall, Fusarium infection of Bd spikes results in the development of symptoms that strikingly resemble those described in wheat heads infected with Fg and Fc .
Microscope analysis of floral tissues highlighted the potential role played by specific epidermal cell types during the early stages of infection. Fusarium hyphae were repeatedly observed entwined about voluminous macro-hairs that displayed a characteristic amber-brown discolouration. Globose fungal structures were repeatedly observed at the base of these hairs, suggesting that these cell types are favoured targets for penetration. Two components of resistance to FHB are widely recognised; resistance to initial infection (type I) and resistance to spread within the head (type II) . The palea and lemma tissues of barley have been shown to express different levels of type I resistance with the former being more susceptible than the latter . Similar differential type I susceptibility of the palea and lemma tissues of Bd was observed in the present study along with differences in type I susceptibility of the two tested inbred lines. Type II resistance is assessed by point inoculation of individual florets in wheat heads . Following point inoculation of Bd florets, both Fg and Fc successfully colonised Bd spikelet tissues and spread through the rachis into neighbouring spikelets and down the peduncle, closely resembling the pattern of colonization in heads of susceptible wheat cultivars . The bleaching of spikelets above the inoculation site in wheat heads is another characteristic symptom of FHB . Bleaching has been correlated with the production of DON by the fungus within infected wheat heads and is also induced following injection of DON into wheat heads . Our observation of bleaching of infected spikes of Bd thus resembles the situation in FHB of wheat more closely than does barley, which has an inherent type II resistance restricting Fusarium symptoms to the area of initial infection .
DON has been shown to function as a virulence factor in wheat, inhibiting the development of cell wall fortification within the rachis during FHB development  and aiding stem colonisation during development of crown rot . In contrast DON appears to play no discernable role in disease development in heads of barley [34, 7] or floral tissues of Arabidopsis . Amendment of the conidial inoculum with DON significantly enhanced both disease symptoms and conidial production by Fg and Fc on wounded detached leaves of Bd. DON amendment similarly influenced symptom development and conidial production in detached wheat leaves following inoculation with Fg and Fc (Additional file 4). This strongly suggests that DON functions in Bd as it does in wheat, where it is understood to act as a virulence factor [34, 35].
The detection of high concentrations of DON in Bd21 and Bd3-1 flowers following inoculation with Fg indicates that these tissues support DON production in Fusarium species. The levels of DON in Bd spray-inoculated spikes were similar to those reported previously following inoculation of wheat under controlled conditions [37, 38]. The high levels of DON observed in floral tissues of Bd differs markedly from the situation with Arabidopsis where the reported levels are generally extremely low [23, 21]. Trichothecene production has been shown not to be uniformly induced during infection of wheat but, rather, is tissue specific with induction in developing kernels and the rachis node . It is probable that the necessary components to induce trichothecene production are present in Bd and wheat whereas they are absent in Arabidopsis, making Bd an attractive model for wheat. The current experiment could not provide information on kernel resistance as whole floral tissues were sampled because the high infection pressure resulted in extremely shrivelled seeds. However, reducing infection pressure and dissection of floral parts could provide insight onto resistance to kernel infection in future experiments.
Following spray inoculation of whole Bd plants, symptoms developed on virtually all above-ground plant parts (stems, leaf sheaths and leaves). Unexpectedly, intact leaves from spray inoculated plants also developed necrotic and chlorotic symptoms as did inoculated unwounded detached leaf sections. The presence of Fusarium within Bd tissues was confirmed by CLSM observation of GFP-expressing fungus. This is, to our knowledge, the first report to date of a successful infection on intact foliar tissue by a Fusarium species. Detached leaf assays have been used previously to identify components of resistance related to FHB but these experiments, although using unwounded inoculation, were carried out using Microdochium majus, a non-toxin producing FHB species . We have determined that Fg and Fc can infect floral and foliar tissues of Bd allowing the mycotoxin-producing species to be used in comparative assays on these tissues. The susceptibility of intact Bd leaves therefore provides the first opportunity to establish the relationship between foliar and floral components of resistance to Fusarium species and identify those foliar components of relevance to FHB resistance. The unique susceptibility of Bd to foliar penetration by Fusarium spp. also provides the potential to undertake high throughput genetic screening of Bd mutant collections to identify lines altered in susceptibility to penetration. Having observed disease symptoms on all tested Bd tissues, histological examination was undertaken to determine how Fg and Fc gain entry into this host. Direct stomatal penetration of wheat head tissues by Fg and Fc has been previously reported [41–43]. Despite observing multiple instances of direct contacts between Fg and Fc germination hyphae and stomatal apertures, we did not obtain evidence for entry into Bd via stomata. Overall, our results suggest that, although penetration may occur through stomatal apertures, it is not likely to be the main mode of entry. In numerous instances, hyphal contact with stomata resulted in guard cells becoming very dark brown, indicating the possible deposition of phenolic compounds. Interestingly, phenolic compounds have been previously shown to play a role in FHB disease resistance in wheat  and a similar situation may occur in the guard cells of Bd. Light microscopy images of the first visible symptoms developing on leaves revealed a characteristic amber-brown discolouration (distinct from the colour of contacted guard cells), of the macro-hair base and directly adjacent cells that was correlated with the presence of the fungus and attempted penetration of the host. Although this amber-brown colour is also indicative of phenolic compounds, the results from coleoptile infection studies showed that its accumulation at the site of attempted fungal penetration is not effective in preventing infection. Similar appositions have been observed during infection of wheat by Fg and were more pronounced in resistant than in susceptible cultivars . During infection of Bd coleoptiles Fg appeared to produce infection pegs and gain entry via growth between cells. Again, this is similar to infection observed on wheat . SEM analysis of intact Bd leaf surface indicated that penetration of hair cells may be the preferred route of entry for the pathogen. We observed penetration of the cuticle, growth and branching at the base of the macro-hair. Macro-hairs are located above the vascular bundles, and targeting their base for initial penetration provides the pathogen almost direct access to the vascular bundles enabling rapid spread to adjacent tissues . This is an interesting finding in relation to previous studies made on detached wheat glumes where Fg was observed to penetrate and invade host tissue through short hair cells (termed prickle hairs ). Association between Fg hyphae and prickle hairs (also referred to as papilla cells) on wheat was also noted by Pritch and colleagues , although they did not undertake detailed investigation of the interaction. The comparison of microscope images of infected floral and foliar Bd tissues revealed striking similarities. Fusarium hyphae were observed to specifically target hairs in both tissues, where globose hyphal structures developed about BMH. Accumulation of phenolic compounds of unknown composition occurred in both floral and foliar tissues as a host response to penetration attempts. These similarities support the idea that investigating the mechanisms of Fusarium infection on foliar tissues may have direct relevance to the mechanisms of resistance of the floral tissues to FHB.
Root tissues were also successfully infected following inoculation by contact with mycelial plugs. The infection pressure generated by conidia, however, failed to induce infection and it remains to be determined whether infection can proceed directly from conidia or whether infection requires hyphae. Infection was indicated by discolouration and confirmed by observation of inter- and intracellular fungal hyphae in the cortex at an early stage of infection. Even at late stages of infection fungal hyphae were excluded from the stele, a situation similar to that recently reported in wheat . Together with observation of symptoms developing on the stem base, these results suggest that Bd can also be used for modelling crown rot and root rot.
Differential responses among Bd accessions to biotic and abiotic stresses have been observed by others indicating that naturally occurring allelic variation in Bd accessions may provide insights into mechanisms underlying responses to agronomically important traits [48, 49]. Inoculation of Fg conidia on detached Bd florets revealed quantitative differences in fungal development between Bd21 and Bd3-1 lines. Interestingly, the two lines also differed in susceptibility in the detached leaf assay with the most notable difference between them being the extensive chlorosis that developed in Bd3-1. Interestingly, DON application to wounded Bd21 and Bd3-1 leaves also resulted in a difference in response with respect to the development of chlorosis indicating that the differential response of the two lines to Fg is, at least in part, a result of differential susceptibility to DON. The availability of the population derived from a cross between Bd21 and Bd3-1 (http://www.modelcrop.org), will permit genetic mapping of the differential susceptibility of these lines to DON and foliar infection. Additionally, investigating the wide range of di-, tetra- and hexaploid Bd accessions would be expected to reveal different levels and mechanisms of resistance to Fusarium.
Bd was previously reported as a model for rice in order to study resistance to Magnaporthe grisea . The current study provides the first detailed report of Bd as a potential model for a wheat disease caused by a necrotrophic fungus.