HvCEBiP, a gene homologous to rice chitin receptor CEBiP, contributes to basal resistance of barley to Magnaporthe oryzae

Background Rice CEBiP recognizes chitin oligosaccharides on the fungal cell surface or released into the plant apoplast, leading to the expression of plant disease resistance against fungal infection. However, it has not yet been reported whether CEBiP is actually required for restricting the growth of fungal pathogens. Here we evaluated the involvement of a putative chitin receptor gene in the basal resistance of barley to the ssd1 mutant of Magnaporthe oryzae, which induces multiple host defense responses. Results The mossd1 mutant showed attenuated pathogenicity on barley and appressorial penetration was restricted by the formation of callose papillae at attempted entry sites. When conidial suspensions of mossd1 mutant were spotted onto the leaves of HvCEBiP-silenced plants, small brown necrotic flecks or blast lesions were produced but these lesions did not expand beyond the inoculation site. Wild-type M. oryzae also produced slightly more severe symptoms on the leaves of HvCEBiP-silenced plants. Cytological observation revealed that these lesions resulted from appressorium-mediated penetration into plant epidermal cells. Conclusions These results suggest that HvCEBiP is involved in basal resistance against appressorium-mediated infection and that basal resistance might be triggered by the recognition of chitin oligosaccharides derived from M. oryzae.


Background
To resist attack by microbial pathogens, plants have evolved to recognize them, triggering the expression of diverse defense reactions. The currently accepted model is that plants recognize conserved pathogen-associated molecular patterns (PAMPs) through corresponding pattern recognition receptors (PRRs) which in turn trigger plant immune responses [1][2][3]. The involvement of PRRs in disease resistance against bacterial pathogens is welldocumented. For example, the N-terminal amino acid sequence of bacterial flagellin (designated as flg22) can be recognized through the corresponding receptor FLS2 in Arabidopsis thaliana [4,5]. In addition, the N-terminal sequence of bacterial translational elongation factor Tu (designated as elf18) can be recognized through the corresponding receptor EFR [6,7].
In contrast to bacterial PAMP receptors, much less is known about the role of fungal PAMP receptors in plants. It is conceivable that oligosaccharides derived from chitin or glucan may function as PAMPs because they are major structural components of fungal cell walls and can induce the expression of several defenserelated genes when they are applied to plants [8,9]. The rice plasma membrane glycoprotein CEBiP (Chitin Elicitor Binding Protein) was shown to be an important component for chitin-derived signaling and is thought to be a receptor for fungal PAMPs [10]. CEBiP was identified as a chitin-binding protein from suspension cultured rice cells and contains two LysM (lysin) domains which mediate binding to oligosaccharides. Physiological experiments suggest that CEBiP is required for the production of reactive oxygen species by rice plants in response to treatment with chitin elicitor [10].
It is assumed that CEBiP recognizes chitin oligosaccharides present on the fungal cell surface or released into the plant apoplast, leading to the expression of plant disease resistance against fungal infection. However, it has not yet been reported whether CEBiP is actually required for restricting the growth of fungal pathogens in rice.
Magnaporthe oryzae is an ascomycete fungus that causes the devastating blast disease in rice [11]. In the previous report, we have generated ssd1 mutants in M. oryzae and the cucumber anthracnose fungus Colletotrichum orbiculare, in which infection of their respective host plants was restricted by cellular defense responses [12]. Subsequently, by inoculating the C. orbiculare ssd1 mutant onto Nicotiana benthamiana plants in which defense-related genes were silenced, we evaluated the involvement of those genes in basal defense. These experiments revealed that plants in which genes encoding specific MAPKK (MEK2) and MAPKs (SIPK/WIPK) had been silenced were susceptible to the ssd1 mutant, as well as the wild-type strain [13]. Furthermore, we revealed that these MAPKs were activated by fungal cell surface components during infection and that the level of MAPK activation induced by the ssd1 mutant was higher than by the wild-type strain, suggesting that MAPK signaling is required for enhanced basal defense and restriction of fungal infection. In addition, use of the ssd1 mutant together with gene-silenced plants allowed us to critically evaluate the involvement of specific defense-related genes in basal resistance by assessing whether the ssd1 mutant could produce disease lesions on the silenced plants.
In plants, RNA interference (RNAi) is a powerful tool for the evaluation of gene function [14]. For RNAi, it is necessary to generate transgenic plants that express a partial fragment of the target gene, but considerable time is required to obtain seeds from T 1 transformants. In contrast, virus-induced gene silencing (VIGS) is a simple, rapid method to transiently generate knockdown plants that avoids the need for stable transformation [15]. Although procedures for VIGS are not yet established for rice, there are reports that VIGS is applicable to barley through the use of barley stripe mosaic virus (BSMV) [16,17]. Barley is a susceptible host plant for M. oryzae, so that interactions between M. oryzae and barley provide a model for the molecular analysis of compatible interactions between monocot plants and fungal pathogens [18].
In this study, we have exploited the barley-Magnaporthe pathosystem to evaluate the involvement in basal resistance of genes encoding a putative PAMP receptor, namely HvCEBiP, which is homologous to the rice chitin receptor CEBiP. For this, we used the M. oryzae ssd1 mutant and BSMV-mediated gene silencing. We present evidence that HvCEBiP contributes to basal defense against appressorium-mediated infection by M. oryzae in barley.

Magnaporthe oryzae SSD1 is required for infection of barley
In previous work we showed that the SSD1 gene of M. oryzae is essential for the successful infection of susceptible rice plants, and that the failure of mossd1 mutants to infect was associated with the accumulation of reactive oxygen species (ROS) by host cells [12]. First, we examined whether the SSD1 gene is also essential for the infection of barley (Hordeum vulgare). When conidial suspensions of the wild-type strain Hoku-1 were inoculated onto leaves, necrotic lesions similar to those of rice blast disease could be observed at 4 days post inoculation (dpi). In contrast, leaves inoculated with the mossd1 mutants K1 and K4 did not show visible disease symptoms ( Figure 1A). When conidial suspensions were spotted onto intact leaf blades of barley, mutant K1 did not produce any disease symptoms, although the wildtype Hoku-1 forms typical blast lesions at inoculation sites at 4 dpi ( Figure 1B). To test whether the K1 mutant retained invasive growth ability, conidial suspensions were spotted onto wound sites on the surface of barley leaves. The mutant produced brown necrotic flecks at wound sites but disease symptoms did not spread further, in contrast to the wild-type Hoku-1 which could form typical blast lesions after infection through wounds ( Figure 1B). Overall, the pathogenicity of the M. oryzae ssd1 mutants was severely attenuated on barley, producing an infection phenotype similar to that seen previously on rice [12].
Microscopic analysis showed that the mossd1 mutant formed appressoria on the plant surface indistinguishable from those of the wild-type strain Hoku-1 ( Figure 2A). However, while Hoku-1 produced intracellular infection hyphae inside host epidermal cells, mutant K1 had formed no infection hyphae at 48 hpi ( Figure 2A). To observe the responses of H. vulgare cells to attempted infection by the mutant, inoculated leaves were stained with 3,3'-diaminobenzidine (DAB) to detect H 2 O 2 accumulation. However, no significant accumulation of H 2 O 2 was detectable in host cells after inoculation with Hoku-1 or K1 at 48 hpi (data not shown). Next, we attempted to detect the formation of autofluorescent papillae under appressoria using epi-fluoresence microscopy [18]. At sites of attempted penetration by the mossd1 mutant, autofluorescent papilla-like structures could be observed beneath approximately 80-90% of mutant appressoria ( Figure 2B), and intracellular infection hyphae were only rarely observed inside host cells ( Figure 2C). On the other hand, the frequency of papilla formation under appressoria of Hoku-1 was only 20% and infection hyphae developed from 60% of appressoria ( Figure 2C). These results suggest that the localized deposition of cell wall material (papillae) at attempted fungal entry sites forms part of the basal defense response of barley epidermal cells to appressorial penetration by M. oryzae.

Virus-induced gene silencing of HvCEBiP using barley stripe mosaic virus
Chitin is major structural component of fungal cell walls and is therefore likely to function as a PAMP [10]. We therefore searched for a gene homologous to the CEBiP chitin receptor of rice using a barley EST database (TIGR plant transcript assemblies; http://blast.jcvi.org/ euk-blast/plantta_blast.cgi) and found an assembled sequence TA30910_4513 which contains the putative full-length coding sequence. The predicted amino acid sequence showed 66% identity to rice CEBiP. Furthermore, this sequence contained a signal peptide at the N-terminus, and two LysM motifs and a transmembrane region in the C-terminal region, which are all present in  rice CEBiP ( Figure 3A). Therefore, we consider this gene is very likely to be orthologous to rice CEBiP, and accordingly designated the gene HvCEBiP. When we examined the expression of HvCEBiP during the course of infection of barley by M. oryzae ( Figure 3B), transcripts were detectable at all time points (3, 6, 12, 24, 48 hpi), indicating that HvCEBiP is likely to be constitutively expressed in barley. In addition, we also examined the expression of selected defense-related genes during infection. Genes homologous to phenylalanine ammonia lyase, respiratory burst oxidase homologue A and pathogenesis-related proteins 1, 2, and 5 were searched from the barley EST database, and designated as HvPAL, HvRBOHA, HvPR-1, HvPR-2a and HvPR-5, respectively. As shown in Figure 3C, transcripts of HvPAL, HvRBOHA and HvPR-5 could be detected at all time points, suggesting they are constitutively expressed. However, it should be noted that both PAL and PR5 generally belong to multi-gene families and we cannot exclude that gene members other than those evaluated in this experiment may be inducible by fungal infection.
HvPR-1 and HvPR-2a expression could not be detected at 0 hpi (no inoculation) but was detected from 6 hpi, suggesting the expression of HvPR-1 and HvPR-2a was induced by inoculation with M. oryzae. However, there were no major differences in plant defense gene expression induced by the wild type and mossd1 mutant K1. Next, to evaluate the involvement of HvCEBiP in basal resistance of barley, we attempted to perform virusinduced gene silencing (VIGS) using the barley stripe mosaic virus (BSMV) [17]. Before silencing HvCEBiP, we first confirmed the efficiency of BSMV-mediated gene silencing in barley by silencing a gene encoding phytoene desaturase (PDS). After BSMV:PDS genomic RNA was inoculated into the first developed leaves of barley plants, a photobleaching phenotype typical of PDS deficiency was visible on the third developed leaves of all inoculated plants, indicating that BSMV-mediated gene silencing of PDS was effective in barley (see Additional file 1: Figure S1). For silencing of HvCEBiP, we first amplified a 298 bp partial fragment of HvCEBiP from barley leaf cDNA and introduced it into plasmid pSL038-1 which carries the γ genome of BSMV. The resulting construct, in which a fragment of the target gene is introduced in the antisense orientation, was designated as pγ:HvCEBiPas ( Figure 4A). The sequence used for silencing HvCEBiP did not contain either of the two LysM motifs ( Figure 3A). In the EST data base background, we selected unique sequences to HvCEBiP, although without access to the complete barley genome, we could not exclude that there might be other  potential CEBiP homologs that are silenced. Next, we attempted to evaluate the silencing effect of HvCEBiP by RT-PCR. After inoculation of BSMV:HvCEBiP onto first-developed barley leaves, total RNA was extracted from the third-developed leaves and used for reverse transcription. Typical viral disease symptoms were observed in the third leaves of plants treated with BSMV (control) or BSMV:HvCEBiP genomic RNA ( Figure 4B). In these leaves, the expression of both BSMVCP, encoding the BSMV coat protein, and HvEF1a, encoding barley translational elongation factor, was detectable ( Figure 4C). On the other hand, the third leaves of plants treated with BSMV:HvCEBiP showed reduced transcription levels of HvCEBiP compared to control plants treated with BSMV ( Figure 4C). These results indicate that the transcript level of HvCEBiP was down-regulated by BSMV:HvCEBiP-mediated gene silencing in barley.

HvCEBiP contributes to restricting infection by mossd1 mutants
To examine whether HvCEBiP is involved in the basal resistance of barley to Magnaporthe, we inoculated the mossd1 mutant K4 onto the third-developed leaves of barley plants after inoculation of BSMV:HvCEBiP onto the first-developed leaves. To quantify the severity of disease symptoms produced by the mossd1 mutant, we classified disease symptoms as follows; Type I, no visible symptoms; Type II, brown necrotic flecks; Type III, blast lesions without brown necrotic flecks ( Figure 5A). On the leaves of BSMV-treated plants, most symptoms produced by mossd1 mutant K4 were classified as Type I ( Figure 5B), whereas on leaves of BSMV:HvCE-BiP-treated plants Type II symptoms were produced at approximately half of the sites inoculated with K4 ( Figure 5B). This tendency was confirmed in three independent experiments. When the wild-type strain Hoku-1 was inoculated onto leaves of BSMV:HvCEBiP-treated plants, the frequency of Type III symptoms was slightly but consistently higher compared to the control plant, although these effects were not statistically significant ( Figure 5B). When conidial suspensions were inoculated onto wound sites on the leaves of BSMV:HvCEBiP-treated plants, there was no significant difference in disease symptoms produced by Hoku-1 and K4 (data not shown), suggesting that the silencing of HvCEBiP does not affect invasive growth ability through wounds. Taken together, these results suggest that HvCEBiP is involved in basal defense responses of susceptible barley plants to appressorial penetration by M. oryzae.
To determine whether the mossd1 mutant was able to develop infection hyphae and colonize barley tissues, we observed leaf inoculation sites in BSMV:HvCEBiP-treated plants at 96 hpi. At sites showing brown necrotic flecks (Type II symptom), appressoria were present on the leaf surface, and infection hyphae developed from appressoria inside the initially infected epidermal cell, which appeared to undergo a cell death reaction ( Figure 5C). However, when we observed inoculation sites at 7 dpi, fungal hyphae had not colonized the neighboring host cells and hyphae were entirely confined to the first infected cell (data not shown). These observations suggest that mossd1 mutant appressoria could penetrate into HvCEBiP-silenced plants but subsequent growth of the infection hyphae became restricted by host defense responses. However, at the few inoculation sites showing severe lesions (Type III), infection hyphae were seen to develop from appressoria without visible host cell death ( Figure 5C). Taken together, these results suggest that HvCEBiP contributes to host defense responses expressed after invasion of epidermal cells by M. oryzae infection hyphae.
To evaluate whether HvCEBiP is also involved in nonhost resistance, we inoculated conidia of the nonadapted maize anthracnose pathogen C. graminicola onto the third leaves of BSMV:HvCEBiP-treated plants. Although C. graminicola formed appressoria on the leaves of both BSMV-and BSMV:HvCEBiP-treated plants, intracellular infection hyphae were not observed, and no disease symptoms were produced ( Figure 6). This suggests that HvCEBiP does not play a critical role in resistance to non-adapted pathogens such as C. graminicola.
Next, we evaluated the possible role in basal defense of selected barley genes required for penetration  resistance and R-gene mediated resistance to the powdery mildew fungus, Blumeria graminis f. sp. hordei. For this, we used barley mutant lines deficient in Ror1 and Ror2 (required for mlo-specified resistance) [19,20], Rar1 (required for Mla12 resistance) [21] and Rom1 (restoration of Mla12-specified resistance) [22]. After inoculating conidial suspension of mossd1 mutant K4 onto leaves of these barley mutants, no significant differences in symptom severity were observed compared to the respective wild-type barley cultivars (Figure 7). It therefore appears that none of these genes are involved in restricting infection by the mossd1 mutant.

Expression profiling of defense-related genes in HvCEBiPsilenced plants
To identify plant defense-related genes that may be regulated by HvCEBiP-mediated signaling, we evaluated the expression patterns of selected barley defense genes in the leaves of BSMV:HvCEBiP-treated plants (Figure 8

Barley expresses two layers of basal defense in response to infection by Magnaporthe oryzae
In our previous study, we generated an ssd1 mutant of M. oryzae, in which the infection of rice plants was restricted by a defense response involving death of the initially infected epidermal cell [12]. This cell death reaction expressed by rice in response to compatible isolates of M. oryzae has been termed 'whole-plant specific resistance' (WPSR), and is independent of R-gene mediated resistance in rice [23,24]. In the present study, infection assays revealed that the mossd1 mutant also showed attenuated pathogenicity on barley. However, the host defense responses expressed in barley to appressorial penetration by the mossd1 mutant took the form of papilla deposition at attempted fungal entry sites rather than host cell death. The phenomenon of papilla formation during M. oryzae infection of barley has also been reported by other authors [18]. In rice, papilla-like wall appositions were also observed beneath appressoria of M. oryzae, although these appeared small and thin with electron microscopy [25]. Therefore, the formation of papillae appears to be a general form of basal defense against attempted appressorial penetration by M. oryzae in barley. However, the efficiency of Figure 7 Pathogenicity test of the wild-type Hoku-1 and mossd1 mutant K1 on a range of barley mutants affected in various defense-related genes. Droplets of conidial suspension were applied onto leaves of genetic mutants of mlo5, Ror1, Ror2, Rar1 and Rom1. Ingrid is the wild-type cultivar for mlo5, ror1 and ror2 mutants. Sultan5 is the wild-type cultivar for rar1 and rom1. papillae in restricting appressorial penetration seems to be weak because the wild-type strain could successfully penetrate into plant cells with high frequency, as shown in Figure 2C. Apart from papilla formation, a localized cell death reaction was also observed in the initially penetrated host cells in which infection hyphae had developed. This cell death reaction was observed in the leaves of BSMV:HvCEBiP-treated barley plants after infection by both the ssd1 mutant and the wild-type strain of M. oryzae. The cell death reaction was associated with inhibition of fungal growth because infection hyphae had not developed beyond the first infected epidermal, even after 7 days. The barley cell death reaction resembles WPSR in rice [23] and conceivably it represents a basal defense response triggered after successful penetration by M. oryzae appressoria. It therefore appears that barley deploys two distinct layers of basal defenses against appressorium-mediated infection by M. oryzae, namely papilla formation and localized cell death. Two similar layers of plant defense were also shown to operate during non-host resistance of Arabidopsis to powdery mildew fungi [26].
HvCEBiP is involved in basal resistance to appressorial penetration by M. oryzae In our recent work, we used the C. orbiculare ssd1 mutant to show that a specific MAPK pathway in N. benthamiana plays a critical role in host basal defense but genes required for R-gene mediated resistance (RAR1, SGT1 and HSP90) do not [13]. Here, we used the M. oryzae ssd1 mutant to examine the role in basal defense of genes required for penetration resistance and R-gene mediated resistance. Ror1 and Ror2 were identified as genes required for mlo-specific resistance against the barley powdery mildew fungus Blumeria graminis f. sp. hordei and Ror2 shows functional homology to syntaxin AtSYP121 in Arabidopsis [27]. Rar1 was originally shown to be required for race-specific resistance triggered by resistance gene Mla12 against B. graminis f. sp. hordei expressing the avirulence gene AvrMla12 [28,29]. Rom1 was identified as a restoration of Mla12-specified resistance (rom1) mutant that restores disease resistance to B. graminis f. sp. hordei carrying the avirulence gene AvrMla12 [22]. However, infectivity of the mossd1 mutant was not significantly enhanced on any of these barley mutants compared to wild-type plants, suggesting that genes required for R-gene mediated resistance do not play a role in basal defense against M. oryzae, consistent with findings from the C. orbiculare-N. benthamiana interaction [13].
In contrast to mutations in these barley genes, the knock-down of HvCEBiP did enhance infection by the mossd1 mutant. Thus, on BSMV:HvCEBiP-treated plants mutant K4 produced more severe (Type II) symptoms, i.e. brown necrotic flecks, compared to BMSV-treated control plants ( Figure 5B). The silencing of HvCEBiP also increased the frequency of successful appressorial penetration by the mossd1 mutant. However, the formation of infection hyphae inside penetrated epidermal cells appeared to trigger localized host cell death, resulting in brown necrotic symptoms. These results suggest that HvCEBiP is involved in basal defense against appressorial penetration by M oryzae. In contrast to the mossd1 mutant, infectivity of the wild-type strain was not significantly enhanced on HvCEBiP-silenced plants but there was a slight increase in symptom severity. This suggests that although HvCEBiP contributes to basal defense in barley, the level of its contribution may be low, so that with the highly pathogenic wild-type strain differences in symptoms between non-silenced and HvCEBiP-silenced plants were hard to distinguish. One plausible explanation of these findings is that basal defense against appressorial penetration involves multiple PAMP receptors and signaling pathways, of which signaling via HvCEBiP is only one. A working model for the contribution of HvCEBiP to the dual-layered basal defense responses of barley to M. oryzae is presented in Figure 9.
In addition to the increased frequency of brown necrotic fleck symptoms induced by the mossd1 mutant on BSMV:HvCEBiP-treated plants, a few inoculation sites also showed formation of severe blast lesions (Type III symptom) as shown in Figure 5A. Lesion formation was not associated with localized cell death reactions and infection hyphae developed extensively, colonizing many host cells. This suggests that in some cases the mossd1 mutant was able to infect HvCEBiP-silenced plants without triggering cell death-associated defense responses. This raises the possibility that HvCEBiP might be involved in mediating the localized cell death response of barley epidermal cells to invasion by M. oryzae infection hyphae. Thus, HvCEBiP might contribute not only to papilla-based defenses but also to the hypersensitive cell death response to cell invasion. HvCEBiP does not appear to play a central role in nonhost resistance because the non-adapted pathogen C. graminicola produced no symptoms on silenced plants. In contrast, the LysM domain receptor kinase CERK1 was reported to contribute weakly to the resistance of Arabidopsis thaliana against the incompatible pathogen Alternaria brassicicola [30].
Is HvCEBiP a specific receptor for components of the mossd1 mutant?
In the interaction between cucumber anthracnose pathogen C. orbiculare and N. benthamiana, we reported previously that the altered fungal cell wall composition conferred by ssd1 gene disruption triggers plant basal resistance through the activation of a specific plant MAPK cascade [13]. We hypothesized that activation of the MAPK pathway might result from recognition of fungal PAMP(s) by corresponding plant receptor protein(s). In this study, we attempted to determine whether HvCEBiP is a specific receptor for PAMPs expressed uniquely by the mossd1 mutant, in which case pathogenicity of the wild-type strain should not be affected by the silencing of HvCEBiP. However, the wild-type strain Hoku-1 showed a slight increase in pathogenicity on HvCEBiP-silenced plants, suggesting that HvCEBiP is a receptor for component(s) shared by both the wild-type M. oryzae and mossd1 mutant.
Rice CEBiP is a receptor-like protein containing two LysM domains, which was originally identified in enzymes that degrade the bacterial cell wall component peptidoglycan [31]. Recent biochemical analysis showed that the LysM domain can also mediate binding to chitin oligosaccharides [32]. The genome of Arabidopsis contains five LysM domain-containing receptor-like kinases [33], among which CERK1 (At3g21630) was identified as a receptor-like protein required for chitin signaling in Arabidopsis [30]. Although the function of the other LysM domain-containing receptor-like kinases is unknown, it is tempting to speculate that plants possess multiple receptor proteins for the perception of particular classes of pathogen-derived oligosaccharides. It is likely that other PAMP receptors, in addition to HvCEBiP, are conserved in barley and contribute to basal resistance to M. oryzae.

Conclusions
Rice CEBiP recognizes chitin oligosaccharides derived from fungal cells leading to the expression of plant disease resistance against fungal infection. We evaluated the involvement of putative chitin receptor gene HvCE-BiP in barley basal resistance using the mossd1 mutant of Magnaporthe oryzae, which enhances host basal defense responses. The mossd1 mutant showed attenuated pathogenicity on barley and appressorial penetration was restricted by the formation of papillae at attempted entry sites. On HvCEBiP-silenced plants, the mutant produced small brown necrotic flecks or blast lesions accompanied by appressorium-mediated penetration into plant epidermal cells. Wild-type M. oryzae also produced slightly more severe symptoms on the leaves of HvCEBiP-silenced plants. These results indicated that HvCEBiP is involved in basal resistance against appressorium-mediated infection and that basal resistance could be triggered by the recognition of chitin oligosaccharides derived from M. oryzae.

Plant growth conditions and fungal strains
Hordeum vulgare wild-type cultivars Fiber-snow, Ingrid and Sultan5, and genetic mutants mlo5, mlo5ror1, mlo5ror2, rar1 and rom1 were grown in a controlled environment chamber (16 h photoperiod, 24°C). Magnaporthe oryzae Hoku-1 was used as the wild-type strain in this study. The mossd1 mutants K1 and K4 were generated as reported previously [12]. These fungal cultures were maintained at 24°C on oatmeal agar medium (6.0 g  powder oatmeal, 1.25 g agar per 100 ml distilled water) under continuous light. Colletotrichum graminicola isolate MAFF236902 was described previously [13].

Pathogen inoculation and cytological assays
To induce conidiation, two week-old cultures of M. oryzae were washed with sterile water to remove aerial hyphae and then incubated for a further 3 days. For inoculation, conidial suspension was sprayed (5 ml; 1 × 10 6 conidia/ml) or spotted (10 μl; 1 × 10 5 conidia/ml) onto the third leaves of H. vulgare and incubated in a humid plastic box at 24°C. For evaluation of invasive growth ability, the surface of barley leaves was scratched with a sterile plastic pipette tip and droplets of conidial suspensions were placed directly onto the wound sites. Cytological observations and the detection of papillae were performed as follows. Inoculated leaves were cut to 1 cm × 1 cm size and decolorized with a 3:1 mixture of ethanol:chloroform and mounted under a coverslip in lactophenol solution. Autofluorescent papillae formed beneath appressoria were visualized by epifluoresence. The accumulation of H 2 O 2 in host cells was detected by staining with 3,3'-diaminobenzidine [13].

Vector construction
A 298 bp partial fragment of HvCEBiP was amplified by primer pairs HvCBP1-S1 (5'-CCAAAGACCCTCAA-GAAGGA-3') and HvCBP1-AS1 (5'-AGCCGTTGGAA-TAACCACTG-3') from cDNA of H. vulgare and subcloned into the pGEM-T easy vector (Promega). The resulting construct was digested by NotI and a fragment containing the amplified sequence of HvCEBiP was introduced into the NotI site of pSL038-1 in the antisense orientation. This construct was designated as pγ: HvCEBiPas.

Virus-induced gene silencing
BSMV genomic RNAs were transcribed in vitro as previously described with some modifications [17]. The reaction was performed at 37°C for 60 min in 50 μl of reaction buffer containing 1 μg of linearized plasmids, 1 μl of T7 RNA polymerase (Takara), 10 μl of 50 mM DTT, 6 μl of 10 mM NTPs (rATP, rCTP, rUTP), 0.4 μl of 10 mM rGTP and 5 μl of 5 mM m 7 G(ppp)G RNA cap structure analog (New England Biolabs). After the reaction, 1.62 μl of 10 mM rGTP and 1 μl of T7 RNA polymerase were added to the reaction mixture, and further incubated at 37°C for 60 min. Transcribed α, β, γ genomic RNAs were mixed in a 1:1:1 ratio with 20 μl FES and inoculated onto the first-developed leaves of H. vulgare plants with gentle rubbing. The third-developed leaves were used for evaluating fungal infections.

Additional material
Additional file 1: Figure S1.  Table S1. Primers used for RT-PCR.