Introduction

Fusarium head blight (FHB) infection of small grain cereals may lead to significant economic losses due to reduction in grain yield, and more importantly, impairment of grain quality due to contamination of the grain with mycotoxins, many of which are detrimental to human and animal health (Berthiller et al. 2013; Escriva et al. 2015). Among the predominant Fusarium species in the Nordic region are F. langsethiae, F. graminearum, F. avenaceum, F. poae, and F. culmorum (Parikka et al. 2012).

F. langsethiae is among the least aggressive of the FHB-causing pathogens, leaving few symptoms of disease on the host (Torp and Adler 2004). In addition to a high preference for oats (Divon et al. 2012; Imathiu et al. 2010; Opoku et al. 2018), it has a narrower niche that F. graminearum. It solely infects the grain (Divon et al. 2012), has limited ability to penetrate plant cell wall (Imathiu et al. 2009), and does not cause seedling blight in wheat and oat seedlings (Imathiu et al. 2010). Nonetheless, F. langsethiae is considered the main causal agent for HT-2/ T-2 mycotoxin contamination in oats in the Nordic region and is the main problem for this crop in the UK (Edwards 2009; Edwards et al. 2012).

The understanding of F. langsethiae epidemiology is increasing but still limited compared to that of F. graminearum. Some reports point to inherent differences between the two. Growth optimum for F. langsethiae in vitro is similar to F. graminearum (20–25 °C; (Medina and Magan 2010; Torp and Nirenberg 2004), however, infection experiments from durum wheat and oat indicate that optimum temperature for F. langsethiae during infection is lower, in the range of 15–20 °C (Divon et al. 2012; Nazari et al. 2014). It has been suggested that dry late summers increase HT-2/ T-2 accumulation, but humid weather pre-and/ or post flowering seems to be a prerequisite for infection (Hjelkrem et al. 2018; Opoku et al. 2013; Xu et al. 2014). In a field trial of several oat varieties it was shown that F. langsethiae has a higher demand for relative humidity than F. poae (Martin et al. 2018). In vitro studies also indicate that F. langsethiae has a water activity profile similar to that of both F. avenaceum and F. graminearum, and may even be more sensitive to reduced water activity (Medina and Magan 2010). Many of the issues concerning cultivation practices of oat with regard to F. langsethiae infection have been addressed (Edwards 2009; Hofgaard et al. 2016b), however, whereas F. graminearum infection is optimal during anthesis, there is conflicting evidence regarding the optimal time of infection for F. langsethiae (Opoku et al. 2013; Hjelkrem et al. 2018; Xu et al. 2014). Importantly, control measures such as fungicide treatments successful against F. graminearum tend to fail against F. langsethiae, emphasizing a need for better understanding of the disease (Hofgaard, pers. comm.).

There are so far no reports describing the F. langsethiae infection processes at the microscopic level. Such studies are informative for understanding the colonization mechanisms, and could help explain F. langsethiae behavior in comparison to other FHB fungi. F. culmorum, F. avenaceum and F. graminearum were studied in detail microscopically. Their infection processes share resemblance in that they mainly enter the grain through natural openings. Similarly to many other plant pathogenic fungi the infection is accompanied by secretion of cell wall degrading enzymes and development of penetration structures (Boenisch and Schafer 2011; Kang and Buchenauer 2000; Kang et al. 2005; Wanjiru et al. 2002). F. graminearum develops composite infection structures during infection, and DON production is necessary for spreading within the rachis tissue in wheat (Boenisch and Schafer 2011; Jansen et al. 2005).

The role of anthers in development of FHB disease has long been a focus of investigation. Early studies indicated that anthers possess fungal growth stimulants resulting in prolific growth of F. graminearum on anther tissue, and these compounds were characterized as betaine and choline (Strange and Smith 1978). More recent studies confirm a high preference for anthers and pollen of F. graminearum and F. culmorum (Kang and Buchenauer 2000; Miller et al. 2004; Tekle et al. 2012), and anther extrusion in wheat is associated with increased resistance to FHB and DON contamination (Buerstmayr and Buerstmayr 2015; Lu et al. 2013). Given the differences in oat and wheat anatomy and physiology it is relevant to ask whether the same requirements for anthers apply to F. langsethiae.

Understanding of the colonization process of F. langsethiae at the microscopic level is one as yet unexplored feature of the disease. In the present work, we have studied the F. langsethiae infection process on oats and wheat in detail with the aim to better understand what makes this fungus a “weak” pathogen and what factors determine the preference for oats. The main focus was to capture the early events of the F. langsethiae infection, especially by resolving two key questions: whether it is able to develop penetration structures, and whether infection is aided by pollen in a similar manner as for F. graminearum. Infected oat grains and fungal proliferation were traced at 3, 6, 10, and 14 days post infection (dpi) using scanning electron microscopy (SEM). Finally, the F. langsethiae interaction with oat was compared to F. langsethiae infection in wheat under the same conditions.

Material and methods

Fungal strains and culture conditions

F. langsethiae strain 9821-16-1 (IBT9951) (Schmidt et al. 2004; Thrane et al. 2004; Torp and Adler 2004) was used in all infections. The fungal strain was revived from sand culture made as follows: glass vials were filled ¾ with sand, sterilized by autoclaving, and 1 ml spore suspension was added to the vial with sterile sand. After growth for 2 weeks vials were capped and stored at 4 °C. Fungal cultures were kept for short term storage on SNA plates at 4 °C until microspore production. Inoculum for plant infection was prepared as described in (Divon et al. 2012).

Plant material, infection and sampling

Oat cultivar Gere and wheat cultivar Zebra were grown in a glasshouse using 1.5 L plant pots (15 cm in diameter; LOG A/S, Norway) and perlite-added commercially available “P-jord” composed of peat mixed with sand, limestone flour and a multimix of micronutrients (LOG A/S, Norway). Plants were grown in white light (HPI) with 16 h photoperiod and watered upon demand. Fertilized water (ionic strength 1.1 mMho mSiemens−1) was used after approximately 1 month of growth and for the rest of the experiment. Temperatures at sowing were 25 °C day and 18 °C night, and relative humidity (RH) 50%. Within the first month temperatures were gradually lowered to 18 °C day and 14 °C night, RH 75%. One week before infection plants were moved to S3 chambers maintaining the same conditions.

Spray inoculation of panicles was done as previously described (Divon et al. 2012) using a spore suspension of 106 conidia ml−1 and 0.1% Tween20, and bagging for 6 days, maintaining the same conditions for light, temperature and humidity. Time of infection was shortly after anthesis, at Zadoks scale Z69–77 for oat and Z71–80 for wheat. These growth stages were chosen because we have previously had best success with F. langsethiae infection on oat during stages later than anthesis (Z60–65; Divon et al. 2012).

One full replicate experiment of each oat and wheat consisted of two treatments: i) mock control (sterile distilled water (SDW) + 0.1% Tween20), and ii) F. langsethiae wild type (wt), each with a total of five pots, each pot containing five plants (Fig. 1). On each plant only the first panicle was directly infected and used for sampling. For each sampled time point (0, 3, 6, 10, and 14 days post inoculation, dpi) one pot (five panicles/ heads) was taken, and from this approximately 40 spikelets were randomly selected for fixation (Fig. 1). The whole infection experiment was carried out twice for both wheat and oat, with 3 weeks delay between replicates, using freshly made spore suspensions and plant at approximately the same growth stage.

Fig. 1
figure 1

Experimental setup and sampling regime. Each treatment (oat/ mock; oat/ F.l wt, wheat/ mock, wheat/ F.l) was performed as shown in the figure using 5 pots each with five plants. One pot was sampled per time point. Each pot contained five plants, and only the first head/ panicle from each plant was sampled. The top and base of the head were removed by 2 cm, and approximately eight spikelets were sampled randomly from the middle part and distributed in vials (1–3 spikelets per vial). Hence, a total of approximately 40 spikelets were sampled per treatment/ timepoint, and all five plants within a treatment/ timepoint were sampled equally. The whole experiment was repeated once (with a delay of 3 weeks) with fresh spore suspension made from fresh SNA plates as described in (Divon et al. 2012)

Full size image

Successful infection was confirmed by re-isolation of fungus from a small random sample of infected and mock grains at 14 dpi, using surface sterilization with 70% ethanol, and incubation on moist filter paper in room temperature for 7 days. In the first replicate infection in oat the percent of infected grains was 70.4% (19/27). None of the 19 tested oat grains from mock-inoculated control were infected. Infection and fungal identification was done morphologically as there were little chance of co-contaminations under the strictly controlled conditions. The “powdery” appearance of sporulating fungus on infected grains was easily recognized as F. langsethiae.

Post-infection treatments and microscopy

At random, approximately 40 grains were sampled for each treatment and time-point as shown in Fig. 1. Floral organs were separated and grains were divided in two or three by cross sectioning to allow for rapid fixation of the tissue. Fixative before staining with trypan blue or aniline blue staining was 96% ethanol. Fixative for scanning electron microscopy (SEM) was 4% paraformaldehyde (PFA) in 1xPBS buffer (pH 7). Samples were left in fix at room temperature over night and then for 2 days in 4 °C. Fixative was then replaced with 1x phosphate buffer saline (PBS; pH 7) and samples were kept at 4 °C.

For staining and examination with light microscope fixed samples were stained with either trypan blue solution (0.1% w/v trypan blue, 10% v/v acetic acid) to visualize dead cells and fungal hyphae, or aniline blue solution (0.05% aniline blue water soluble, 40% glycerol, 20% lactic acid) staining callose and fungal structures. Samples were submerged in staining solution for 20 min, rinsed twice with SDW, and examined using an Olympus SZH10 stereo microscope with additional light source.

For SEM fixed samples were dehydrated through a graded ethanol series, critical-point dried, and mounted on stubs. Upon sputter-coating with gold/palladium the samples were viewed in a Zeiss EVO - 50 - EP scanning electron microscope (Carl Zeiss SMT Ltd., Cambridge, UK) operating at 15–25 kV.

Results

Basipetal spreading of infection via natural openings on oat

We followed the spreading of fungus throughout the oat spikelet from infection until 14 dpi. Spreading was traced by visual inspection of grains showing browning and with trypan blue or aniline blue staining. Scanning electron microscopy (SEM) was used to confirm observations and study fungal structures in more detail. Degree of infection varied considerably from grain to grain, which is usual for Fusarium and expected with a 70% infection rate (at 14 dpi). The same variation was true for the replicated experiment, producing grains from no, to heavy, infection, however the structures and patterns described were similar. Staining of hyphae was easiest spotted on heavily infected grains. In most cases initial infection started at the apex, or apically in the overlap between palea and lemma, and proceeded basipetally as shown in Fig. 2. Also laterally, along the palea wings overlapping with lemma, hyphae gained access to the floret cavity. This was mainly evident on the distal half of the grain. At 3 and 6 dpi fungal growth on abaxial sides of palea and lemma was hard to observe, and we noticed only occasionally blue staining of aerial hyphae at 6 dpi and on the distal half. Staining of stomatal cells, pollen grains and pistils was also observed in control plants, thus staining was not in itself a reliable sign of infection unless accompanied by hyphae. Occasionally we observed growth of hyphae through stomata on the glumes, but this was not a typical trend. More often extending hyphae would grow past stomatal openings.

Fig. 2
figure 2

Spreading of F. langsethiae on oat grain. a Overview of spreading pattern. Infection typically started at the apex (arrowhead) or apically in the overlapping wings between palea and lemma (ventral; short arrows). Proliferation of the fungus then proceeded basipetally (long arrows). Apical, ventral and lateral parts of the grain were colonized before the dorsal side. b Cross section showing the overlapping wings between palea and lemma (arrowheads). le; lemma, pa; palea, se; starchy endosperm

Full size image

At 6 dpi tissues in the ventral crease and the epidermal tissue of lemma were showing slight browning even before hyphae were detected. As of 6 dpi, infected caryopses showed in general a faint blue stain of aniline blue apically, spreading basipetally along the sides and in the ventral crease (data not shown). Control oat grains at 6 dpi showed no staining in palea and caryopsis, no hyphal structures, and no browning in any tissue.

Plastic bags were removed from the oat panicles and wheat heads at 6 dpi. Further proliferation of fungus from this point on was found exclusively on internal surfaces such as adaxial side of palea and lemma. Hyphae were now growing more vigorously and were easily spotted on heavily infected grains, forming networks that grew throughout the adaxial surfaces of palea and lemma and extended basipetally. Hyphae present on abaxial surfaces started at this point to decay (data not shown). At 10 dpi hyphae were observed on the caryopsis surface. These were usually thicker (2–3.25 μm in diameter versus 1.35–1.9 μm on abaxial surfaces) and associated closely with the pericarp surface. Frequent runner hyphae extended over the caryopsis surface and then started to branch. On heavily infected grains a thick mat of hyphae developed at the apical tip, extending basipetally between the brushhairs. At 14 dpi the fungus continued to spread along the same axes; mainly basipetally over adaxial surfaces of palea and lemma and the caryopsis. Only heavily infected caryopses were colonized on the basal end. Where we noticed colonization at the basal end we found no indication of F. langsethiae spreading into the rachis (data not shown).

Pollen speeds up growth of F. langsethiae

Proliferation of fungal hyphae on abaxial surfaces was primarily associated with the presence of pollen (Fig. 3). At 3 dpi development of hyphae was only observed in the vicinity of pollen. Microconidia without contact with pollen had barely germinated at this point both in oat and wheat, whereas fungal hyphae extended more than 300 μm when growing out from pollen grains (Fig. 3b). Oat spikelets maintained the anthers at the apex, intertwined with stigma and quenched between palea and lemma. This part of the spikelet was usually overgrown with hyphae (Fig. 3a). Individual pollen grains found elsewhere on abaxial surfaces were similarly overgrown with hyphae. The fungus had already at 3 dpi completely colonized the pollen grain and radiated hyphae in all directions (Fig. 3b–d). The vigorous growth was in contrast to microconidia which were not associated with pollen, and which barely germinated by 3 dpi. Hyphae radiating from pollen grains on abaxial surfaces seemed to deposit secreted substances on the interface between the hyphae and the plant surface, and they were covered with a mucilage-like layer surrounding the hyphae (Fig. 3e, f). During infection of wheat hyphal growth on the outer surfaces was very scarce. Anthers were extruded, and fungal growth was almost exclusively in conjunction with remaining pollen grains. At 6 dpi we occasionally observed extending hyphae without contact with pollen on abaxial surfaces in both wheat and oat, but their growth was considerably attenuated, extending just short distances (50–100 μm) and almost without branching.

Fig. 3
figure 3

Fungal growth on abaxial (outer) surfaces on oat at 3 dpi. a Anther and stigma tissue overgrown with fungal hyphae (arrowheads). b Fungus colonizing pollen grain on abaxial side of glume at 3 dpi. Hyphal diameter ranged from 1.35 to 1.9 μm, the thickest being observed closest to pollen grains. Asterisks indicate stomata. c-f Close-up of b showing complete colonization of pollen grain. c Vigorous growth of fungus in association with pollen grain. Hyphae associate with each other and grow in all directions. Note that microconidia in the vicinity of, but without access to, the pollen grain barely germinate at this point (arrowhead). d Fungal hyphae folding back on themselves and intertwining. e Secreted material (arrowheads) on hypha and plant surface at a hyphal branch point. f Mucilage-like layer surrounding hypha (arrowhead). Hyphae were rarely found to enter through stomata (asterisk). an; anther, st; stigma, br; brush hair

Full size image

Penetration of internal surfaces on oat

At 10 dpi colonization of all internal (adaxial) surfaces was detected, although the amount of hyphal growth varied considerably and some grains remained un-infected. On adaxial surfaces of palea and lemma runner hyphae were developing and forming focal points (Fig. 4a). On the pericarp tissue the hyphal network was denser, with thicker hyphae and more branching (Fig. 4b). Trypan blue staining at 10 dpi indicated that penetration took place in the focal points (Fig. 4a). At 14 dpi we could distinguish several infection structures and penetration points on the caryopsis surface using SEM (Fig. 4c–f). Runner hyphae were frequently observed extending over the caryopsis surface by several 100 μm on grains with low infection. Along the runner hyphae short infection hyphae were branching off and directly penetrating the pericarp (Fig. 4d). Runner hyphae could also end in swollen foot structures (Fig. 4c). On heavily infected caryopses, branching was frequent, and branches ended shortly in formation of dense compound structures resembling lobate appressoria and infection cushions, only 20–50 μm apart (Fig. 4e, f). Number of infection cushions varied with the degree of infection. Penetration holes were visible along the edges of infection cushions, and where fungal structures were torn off during tissue preparation (Fig. 4f). Hyphae were observed throughout several layers of the pericarp (data not shown), however, the separate layers were not investigated in more detail. Penetration of the aleurone layer or embryo, and fungal growth in the endosperm, was not detected, but was not investigated in detail.

Fig. 4
figure 4

Colonization and penetration structures of F. langsethiae on oat. a Trypan blue staining showing hyphal networks and focal points on the adaxial side of lemma at 10 dpi (8x magnification). b Branching hyphae forming on the caryopsis surface around the base of brush hairs at 14 dpi. Hyphae were often irregularly shaped, with a diameter ranging between 2 and 3.25 μm. c Lobate foot structures forming at the end of a runner hypha on the pericarp surface at 14 dpi. d Infection hypha branching off runner hypha and penetrating the pericarp at 14 dpi. e Infection cushion on pericarp surface at 14 dpi. Composite structures were typically 10–25 μm in diameter. f Hyphal network on pericarp surface at 14 dpi showing infection cushions as hyphal focal points. Inset: close-up showing penetration holes where infection structures have been torn off. Penetration holes were of similar diameter as the infection hyphae (1.15–1.46 μm). bh; brush hair, fs; foot structure, ih; infection hypha, ic; infection cushion

Full size image

F. langsethiae infection on wheat leads to degradation of hyphae and profuse sporulation

Using the oat infection as reference we investigated infected wheat grains at 3, 6, 10 and 14 dpi (SEM was only used on 6 and 14 dpi). Infected wheat kernels typically stained much weaker than oat using trypan blue and aniline blue staining, and the infection was not as evident and easy to follow. Blue staining arising from infected tissue was only seen at the apical tip of the grain no earlier than 6 dpi (data not shown). On higher magnification occasional hyphae could be seen developing along the wings of palea and lemma almost exclusively around the floret mouth, in conjunction with individual pollen grains. Overall the hyphal growth observed on outer surfaces at 6 dpi was very scarce compared to oat. At 14 dpi hyphal networks were clearly visible on the caryopsis and adaxial surfaces of palea and lemma, but were not as dense as in oat. Similarly to oat, we did occasionally observe runner hyphae and indications of penetration points on the adaxial surface of lemma. Trypan blue staining at the penetration point indicated plant cells undergoing necrosis, and plant tissue close to the penetrations showed browning (data not shown). We observed structures resembling compound penetration structures on the adaxial side of lemma, but these were simpler than the typical infection cushions observed in oat (data not shown).

In general the hyphal networks we observed on the adaxial side of lemma and caryopsis at 14 dpi were loosely associated with the plant surface as compared to oat, and were accompanied by massive sporulation (Fig. 5). From SEM photos the loosely associated fungal structures (i.e. standing out from the plant surface) appeared brighter due to the increased reflection of electrons. In particular, the hyphal focal points seemed detached from the plant surface. Interestingly, further characterization using SEM revealed that these focal points resembled sporodochium-like structures rather than compound appressoria (Fig. 5c). The nodes consisted of conidiophore-like structures where we could distinguish phialides, mainly on the adaxial surface of lemma, but also on the caryopsis surface (Fig. 5d, e). Spores were also seen budding off the phialides (Fig. 5c).

Fig. 5
figure 5

F. langsethiae infection on wheat showing loose association of hyphae and sporulation on surfaces at 14 dpi. a, b Comparison of typical dorsal sides of infected caryopses in oat (a) and wheat (b). Denser hyphal network is developing on oat. In wheat fungal growth is scarce and aggregates resembling sporodochia can be seen (asterisk). c Close-up of sporodochium on the wheat caryopsis surface. Phialide-like struktures and budding spores (asterisk) are indicated. d Aggregates of spore producing structures on adaxial side of lemma in wheat. The hyphal network is loosely attached and growing mostly in the open space between lemma and the caryopsis. e Close-up of hyphal branching structures on the wheat caryopsis surface resembling lobate appressoria with close association to the plant surface, and possibly sporodochium primordia (asterisk) branching away from the surface. ph; phialide, sp.; microspore, tr; trichome, la; lobate appressorium

Full size image

We found that where hyphae were associated closely with the wheat pericarp surface, they often underwent degradation (Fig. 6). Due to the scanty hyphal growth and uneven surface of the wheat caryopsis, this was only visible using SEM at higher magnifications and with careful examination. Hyphae where observed with holes in the membrane and disintegrating until they were barely visible (Fig. 6a, b), and foot structures seemed to shrink and loose cellular content (Fig. 6c, d). Hence, the scanty hyphal growth on the caryopsis surface can in part, be explained by the simultaneous degradation of hyphae.

Fig. 6
figure 6

Degradation of hyphae on wheat surfaces. SEM photographs showing degradation of hyphae tightly associated with the wheat surface. a, b Desintegrating hyphae (asterisks) showing holes in the membrane and loss of volume. c, d Hyphal foot structures shrinking and loosing cellular content. fs; foot structure

Full size image

Discussion

The main goal for this study was to elucidate the microscopic structures of the F. langsethiae infection of oat to understand the slow disease picture and relate it specifically to the role of pollen and ability to penetrate plant cell wall. Therefore, infection was conducted in climate chambers under optimal infective conditions with enough humidity and excess inoculum. We achieved a high degree of infection and disease symptoms were clearly visible on infected grains, whereas some grains were found to be not infected. The same infection conditions have been used previously and this confirms that the infection is reproducible and reliable (Divon et al. 2012). Well aware of that this is an artificial state, a high degree of infection was necessary in order to produce enough material for microscopy analyses. Under field conditions F. langsethiae infection is rarely accompanied by symptoms (Martin et al. 2018).

Our results show that many aspects of F. langsethiae infection are similar to other Fusarium species such as Fusarium graminearum. Fungal hyphae enter the kernel through natural openings, mainly at the apex and apically in the overlap between palea and lemma. Hyphal growth on outer surfaces is scanty, whereas hyphal networks develop on all internal surfaces, and infection spreading basipetally. The same patterns have been described for wheat infection with F. culmorum and F. avenaceum, and F. graminearum infection on barley and oat (Kang and Buchenauer 2000; Kang et al. 2005; Skadsen and Hohn 2004; Tekle et al. 2012). Studies from F. graminearum, F. avenaceum, and F. culmorum indicate secretion of cell wall degrading enzymes during infection (Kang and Buchenauer 2000; Kang et al. 2005; Wanjiru et al. 2002). Secretion of various substances is common for phytopathogenic fungi. From SEM photos we could clearly see hyphae covered in a mucilage substance both on external and internal surfaces. The mucilage substance may play a role in hyphal adherence to the plant surface as well as to other hyphae (Deacon 2006; Huang et al. 2008; Matsuura 1986), and may serve as protection against desiccation. Additionally, judging from SEM photos, we detected secreted substances on the adaxial surface of lemma in immediate vicinity of hyphae. This needs to be further characterized in F. langsethiae. Colonization of wheat and barley by F. graminearum has been studied using GFP-tagged strains (Skadsen and Hohn 2004; Miller et al. 2004). We have successfully introduced GFP to F. langsethiae for the same purpose (personal commun., Divon), however the strong autofluorescence in oat falls within the emission spectrum of GFP, making it an unfit marker for tracing oat infection.

The most striking similarity we found to other head blight fungi was the formation of penetration structures on internal surfaces. This is unexpected since F. langsethiae is known as a weak pathogen and has only limited ability to cause wounding on detached leaves (Imathiu et al. 2009). Infection structures were observed once the fungus gained access to internal surfaces. Early in colonization of the interior surfaces we observed frequent runner hyphae along which short infection hyphae that were branching off and penetrating the epidermis. Runner hyphae were occasionally observed to end in foot structures or lobate appressoria. In more heavily colonized tissue dense hyphal networks formed in which the nodes consisted of composite infection structures resembling infection cushions. Numerous infection pegs were visible along the edges and where infection structures were torn off during handling. This was best visualized with SEM on the caryopsis surface, but light microscopy indicated similar structures forming on adaxial sides of palea and lemma. The infection structures observed in F. langsethiae have stunning resemblance to those of the much more aggressive relative, F. graminearum (Boenisch and Schafer 2011). Similar structures have also been described for F. culmorum and F. avenaceum, as well as other phytopathogenic fungi (Armentrout and Downer 1987; Huang et al. 2008; Kang and Buchenauer 2000; Kang et al. 2005; Matsuura 1986). In our study we did not observe any penetration of the outer surfaces of palea or lemma.

The main difference in the F. langsethiae infection on oat relative to other FHB fungi was the time frame for the infectious events. We detected only scanty hyphal growth on the outer surfaces at 3 dpi and access to internal surfaces at 6 dpi. In contrast, F. graminearum, F. culmorum and F. avenaceum will germinate and develop hyphae within 1 dpi, and form hyphal networks on internal surfaces by 2 dpi (Kang and Buchenauer 2000; Kang et al. 2005; Skadsen and Hohn 2004; Tekle et al. 2012). Inoculum, humidity and mode of infection (spray, point or dipping) are factors that influence the infection rate and may jeopardize a direct comparison of experiments. F. langsethiae is characterized by slow growth (Torp and Nirenberg 2004), hence a slower disease development is expected. It is also our experience that germination of F. langsethiae microspores in PDB is 10–15 h delayed compared to F. graminearum (Divon, personal observation).

Humidity and the presence of pollen grains seem to be prerequisites for a successful F. langsethiae infection. Several studies have concluded that the need for humidity is equally or even more important in F. langsethiae compared to other Fusarium species (Hjelkrem et al. 2018; Martin et al. 2018; Medina and Magan 2010; Opoku et al. 2013). Humidity in the microclimate is in our study achieved by the extended period of bagging upon spray inoculation of panicles (6 dpi vs 2 days in F. culmorum and F. graminearum) (Divon et al. 2012; Kang and Buchenauer 2000). In our study the presence of anthers/ pollen grains seems to be a key factor in F. langsethiae infection. We observed that access to pollen greatly accelerates the growth rate of F. langsethiae, making the fungus able to enter internal surfaces of the kernel faster. Similarly, it was demonstrated that anthers offer a more conducive environment for infection of F. graminearum on oat (Tekle et al. 2012). Oat varieties with more anther extrusion had fewer infected florets (12–20%), compared to “normal varieties” with 40–50% infected florets. In a study by Martin et al. (2018) only a hull-less oat variety exhibited partial resistance to HT-2/T-2 accumulation. The combination of a high degree of anther retention in oat, together with the hanging anatomy of the oat spikelet, maintaining humidity for an extended period of time at the apex, may be factors that contribute to a favorable infection environment and to the observed preference of F. langsethiae for oat rather than wheat in the field (Edwards et al. 2012; Hofgaard et al. 2016a; Opoku et al. 2018).

According to our observations, we conclude that F. langsethiae gains access to the inside of the wheat grain through mainly the same routs as in oat and similar to what is described for other FHB species. The progress of infection in wheat is slower and weaker, although we detected hyphae on all internal surfaces by 14 dpi under our optimal infecting conditions. The weaker and slower infection can be a result of fewer germinating spores reaching internal surfaces. This coincides with anthers being extruded providing little pollen to grow on. In our experiments humidity is not expected to be a limiting factor, however, the absence of anthers due to anther extrusion may explain the attenuated infection. Hyphal growth on wheat was seen exclusively in association with pollen grains and only a relatively small amount of individual pollen grains, and no anthers, were available on the wheat spikelets.

The degradation of fungal hyphae that we noted on internal surfaces in wheat may further explain the weaker disease picture in wheat compared to oat. Degradation of tightly associated hyphae may indicate that wheat exudates or other components in the wheat tissue inhibit, slow down, or modulate the fungal growth on internal surfaces. In a study comparing F. langsethiae infection in wheat, barley and oat under identical conditions Opoku et al. (2018) concluded that genetic differences between the cereal species are likely to play a part in the strong preference of F. langsethiae for oat. It has been suggested that a stronger resistance to F. langsethiae infection in wheat may be explained by a greater ability to metabolize the HT-2/ T-2 toxins (Lattanzio et al. 2012; Opoku et al. 2018); however, other active and passive defense mechanisms are likely to be present. In a detached leaf assay, Imathiu et al. (2009) showed that F. langsethiae is unable to cause lesions on wheat leaves without prior wounding, as opposed to oat leaves where no prior wounding was needed. Wheat and other members of the Triticeae, are also known to produce exudates containing allelochemicals that serve in protection against various pathogens (Niemeyer et al. 1992; Wu et al. 2002). Wheat root exudates contain allelochemicals with antifungal properties against Gaeumannomyces graminis var. tritici (wheat take-all disease), Cephalosporium gramineum, Gaeumannomyces graminis var. graminis, and F. culmorum (Martyniuk et al. 2006; Schalchli et al. 2012). This area is largely unexplored with regard to F. langsethiae infection.

Few contact points with the host tissue and a loose association of hyphae with the wheat surface are likely factors that will cause nutrient stress on the fungus. Nutrient stress is a common inducer of sporulation in fungi and may be the trigger for the formation of sporodochia and the massive sporulation we noticed in wheat. Glucose availability plays a role in conidiation in F. culmorum (Larmour and Marchant 1977), however, this warrants further investigations for F. langsethiae.

Conclusion

In this study we investigate the infection route and penetration mechanisms of F. langsethiae in oat and conclude that they are in general similar to that of F. graminearum using pollen as a conducive medium to speed up growth and having the ability to form multiple penetration structures and penetrate plant cell wall. A combination of plant anatomy (hanging spikelets retaining humidity), anther retention, preference for colder temperatures during infection, and absence of plant resistance mechanism are likely factors that collectively may explain the infection niche that F. langsethaie has on oat. Infection of F. langsethiae on wheat was delayed compared to oat, and fungal growth on internal surfaces seems attenuated and stressed. However, given optimal conditions infection in wheat is also likely to happen. Infection on wheat was accompanied by profuse conidiation, which, in this study, was not noted in oat. Resolving more of the biochemical and molecular mechanisms associated with oat and wheat infection would be useful in future design of preventive strategies against F. langsethiae infection in small grain cereals.