In this study, an isolate of Magnaporthe oryzae expressing the green fluorescent protein gene (gfp) was used to monitor early events in the interaction of M. oryzae with resistant rice cultivars harbouring a blast resistance (R) gene. In the resistant cultivars Saber and TeQing (Pib gene), M. oryzae spores germinated normally on the leaf surface but produced morphologically abnormal germ tubes. Germling growth and development were markedly and adversely affected in leaves of these resistant cultivars. Penetration of host cells was never seen, supporting the idea that disruption of germling development on the leaf surface might be one of the resistance mechanisms associated with Pib function. Thus, this particular R gene appeared to function in the absence of host penetration by the fungal pathogen. Confocal laser scanning microscopy of M. oryzae-infected susceptible rice cultivars showed the dimorphic growth pattern that is typically observed during the biotrophic and necrotrophic stages of leaf colonization in susceptible cultivars. The suitability of the gfp-expressing M. oryzae isolate for further research on R-gene function and identification of resistant genotypes in rice germplasm collections is discussed.
Magnaporthe oryzae is the causal agent of rice blast, the most economically important fungal disease of cultivated rice (Oryza sativa) worldwide (Ou, 1985). The blast fungus infects both temperate and tropical rice grown under different ecosystems (i.e. upland, lowland, irrigated and rainfed) and causes more damage in areas where high-input agricultural systems are adopted. Based on multilocus genealogy and mating experiments, M. oryzae was defined as a new species, separate from Magnaporthe grisea (Couch & Kohn, 2002). Thus, the Magnaporthe complex comprises two distinct clades, one that includes isolates infecting Digitaria (crabgrass), referred to as M. grisea isolates, and a second one that includes isolates pathogenic on rice, millet and other grasses, referred to as M. oryzae.
Magnaporthe oryzae has been described as a hemibiotrophic pathogen which first establishes a biotrophic interaction with the rice plant and later switches to a destructive necrotrophic lifestyle. Foliar infection is initiated by attachment of a three-celled spore of M. oryzae to the rice leaf cuticle. A polarized germ tube emerges from the spore and grows on the leaf surface, before differentiating into the dome-shaped appressorium. Once formed, the appressorium matures and generates enormous cellular turgor that is sufficient to rupture the plant cuticle (Wilson & Talbot, 2009). The appressorium forms a specialized hypha, the penetration peg, which in the lumen of epidermal cells expands to become a filamentous primary hypha. The infection proceeds with the differentiation of primary hyphae into bulbous invasive hyphae that fill the invaded cells (Kankanala et al., 2007; Wilson & Talbot, 2009). Several days after infection, blast lesions appear, in which the fungus sporulates profusely, thus allowing the disease to spread rapidly to adjacent rice plants. Very recently, the fungus M. oryzae was identified as the most damaging plant pathogenic fungus (Dean et al., 2012). At present, blast disease control worldwide relies on the combination of chemical fungicides and integrated cultural practices.
In recent years, plant breeding has achieved significant progress towards the enhancement of blast resistance with the identification of a broad array of blast resistance (R) genes. To date, more than 85 blast R genes have been identified (Ballini et al., 2008) and some of them have also been molecularly characterized: Pib (Wang et al., 2009), Pita (Bryan et al., 2000), Pi9 (Qu et al., 2006), Pi2/Pizt (Zhou et al., 2006), Pid2 (Chen et al., 2006), Pi36 (Liu et al., 2007), Pi37 (Lin et al., 2007), Pik-m (Ashikawa et al., 2008), Pi5 (Lee et al., 2009b), Pit (Hayashi & Yoshida, 2009), Pid3 (Shang et al., 2009) and Pi21 (Fukuoka et al., 2009). Several of these blast resistance genes have demonstrated their ability to confer resistance to various blast pathotypes, and are being effectively used in breeding programmes to increase blast resistance in rice. However, none of the blast resistance genes so far identified appears to confer resistance against all isolates of M. oryzae. Because of the genetic variability of M. oryzae and race changes in blast populations, resistant cultivars with single-gene resistance have been shown to lose effectiveness after a few years (Lee et al., 2009a). Durable resistance in breeding programmes requires gene pyramiding strategies for the simultaneous expression of more than one R gene in the same cultivar. To this end, molecular markers tightly linked to major blast resistance genes are being used for marker-assisted selection (MAS) in rice (Fjellstrom et al., 2004). Recently, molecular markers for a panel of 13 blast resistance genes from 25 rice donor parental genotypes were reported (Tacconi et al., 2010), this information being useful to address pyramiding strategies in rice breeding. However, R-gene-mediated mechanisms and processes involved in blast disease resistance require further investigation.
The usefulness of fungal isolates expressing the gfp (green fluorescent protein (GFP)) gene to monitor fungal infection of host plants is well demonstrated. Various fungi have been transformed with the gfp gene, including Ustilago maydis, M. oryzae, Fusarium oxysporum and Sclerotinia sclerotiorum, among others (Sesma & Osbourn, 2004; Campos-Soriano & San Segundo, 2009; de Silva et al., 2009; Zvirin et al., 2010). This approach offers the possibility of monitoring in planta the invasion process and growth of the fungus in living tissues, using fluorescence microscopy without any further manipulation.
In this work, a gfp-expressing isolate of M. oryzae (PR9) and confocal laser scanning microscopy (CLSM) were used to visualize the early stages of fungal growth and invasion of susceptible rice cultivars. Moreover, CLSM allowed the in vivo imaging of M. oryzae growth in leaves of rice genotypes carrying blast resistance genes. Spore germination and the fate of germlings were examined in more detail in resistant rice genotypes.
Materials and methods
Plant and fungal materials
A series of resistant and susceptible rice genotypes was chosen for investigation. They were: Vialone Nano and Maratelli, two Italian cultivars characterized as highly susceptible to blast; Senia, a susceptible Spanish cultivar; Katy, Saber and Kanto 51, three American cultivars resistant to blast; and TeQing, a highly resistant Chinese cultivar. The resistant genotypes were characterized by the presence of resistance genes: Pik in Kanto 51, Pita2 in Katy, and Pib in Saber and TeQing (Tacconi et al., 2010). Rice genotypes TeQing, Saber and Katy were obtained from Dr Harold Bockelman (USDA, Agricultural Research Service, National Small Grain Research Facility, National Small Grains Collection, USA), whereas Kanto 51 was provided by Dr Kazutoshi Okuno (National Institute of Agrobiological Sciences, Japan). Pure seed stocks came from the rice germplasm seed bank of the Consiglio per la Ricerca e la Sperimentazione in Agricoltura (CRA-Rice Research Unit, Vercelli, Italy).
The gfp-expressing M. oryzae isolate PR9 described by Campos-Soriano & San Segundo (2009) was used in this study. This isolate expresses the sgfp gene under the control of the P. tritici-repentis ToxA gene promoter (Sesma & Osbourn, 2004). The fungus was grown on oat medium (Difco) for 2 weeks. Spores were collected by adding sterile water to the surface of the mycelium. After filtration through sterile Miracloth (Calbiochem), spores were adjusted to the appropriate concentration with sterile water using a Bürker counting chamber.
Infection experiments with the gfp-expressing M. oryzae isolate (gfp-PR9) were carried out using the detached leaf assay as previously described (Coca et al., 2004). Briefly, the second leaves of rice plantlets at the three-leaf stage were placed into agar plates (1% w/v in water) containing 2 mg kinetin L−1. Whatman filter paper discs saturated with a M. oryzae spore suspension at the appropriate concentration were placed onto the adaxial leaf surface. The inoculated leaves were maintained in the dark in a chamber under high humidity conditions for 48 h, after which the filter paper discs were removed. Leaves were maintained at 28°C and 90% relative humidity under a 16-/8-h light/dark photoperiod for the required period of time. Microscopic examinations of the infected tissue were carried out with three different spore concentrations (104, 105 and 106 spores mL−1). Disease symptoms were evaluated 3 and 6 days post-inoculation (dpi). Four independent experiments were carried out for each rice cultivar. In each experiment, at least three leaves were inoculated with each spore concentration (five inoculation points per leaf). Lesion areas were measured by image analysis software assess v. 2.0 for plant disease quantification (Lamari, 2008).
Rice leaves were subjected to CLSM analysis at 3, 6, 30, 54, 60 and 72 h after inoculation with M. oryzae spores using an Olympus Fluoview FV1000. For visualization of GFP fluorescence, the excitation wavelength was 488 nm and the emission window was set at 500–550 nm. For visualization of chlorophyll autofluorescence, the same excitation wavelength was used and the emission window was set at 600–700 nm. Two independent experiments, with 10–20 leaves (five inoculation points per leaf) in each, were carried out.
Macroscopic evaluation of disease severity
During a plant–pathogen interaction, a complex set of reactions must occur in the two partners, host and pathogen, that ultimately influences the outcome of their interaction. The particular aim of this study was to characterize early events occurring during the rice–M. oryzae interaction. To this end, several rice cultivars were selected for which phenotypes of resistance or susceptibility to blast infection had been previously demonstrated under field conditions (Katsantonis et al., 2007; Tacconi et al., 2010; Faivre-Rampant et al., 2011). Among the resistant cultivars chosen were TeQing, Saber, Katy and Kanto 51, which each carry a Pi gene in their genome, namely Pib (Saber and TeQing), Pik (Kanto 51) or Pita2 (Katy). Both TeQing and Saber (Pib gene) showed a high level of resistance to various M. oryzae isolates at different plant developmental stages under natural field conditions (Tacconi et al., 2010). Kanto 51 and Katy (Pik and Pita2 genes, respectively) exhibited resistance to blast infection at the tillering and stem elongation stages, but a low level of infection at the heading and milk stages of plant growth (Tacconi et al., 2010). The susceptible cultivars assayed in the present study showed moderate (Senia) to high (Maratelli, Vialone Nano) sensitivity to blast.
For blast disease assays, rice leaves were locally inoculated with increasing doses of spores from the gfp-expressing M. oryzae isolate to give rise to a macroscopically visible lesion. As previously reported (Campos-Soriano & San Segundo, 2009), gfp expression did not affect pathogenicity nor the ability of the M. oryzae fungus to sporulate on the infected leaf. The infection process was followed visually and microscopically up to 6 dpi.
The resistant cv. TeQing did not develop disease symptoms on its leaves, not even when the highest inoculum dose was used (106 spores mL−1; Fig. 1a). Saber and Katy exhibited small necrotic spots, if any, appearing as tiny, pinpoint dots at the inoculated regions that did not further expand on the leaf surface (106 spores mL−1). However, cvs Saber and Katy never developed typical blast lesions on their leaves. Representative results are shown in Fig. 1b,c. Under the same experimental conditions, leaves from Kanto 51 exhibited lesions which were most common at the distal portions of the inoculated leaves (Fig. 1d). By contrast, Kanto 51 exhibited resistance to blast in the field (Tacconi et al., 2010). Previous studies have shown that the detached leaf assay enhances susceptibility to blast infection compared with inoculation of whole plants (Berruyer et al., 2006). This fact might explain the different responses observed in Kanto 51 depending on the method used for inoculation, namely the detached leaf assay (present work) versus infection at the whole-plant level (Tacconi et al., 2010). When infected with the gfp-M. oryzae isolate, the susceptible genotypes Senia, Vialone Nano and Maratelli showed severe symptoms of infection. With time, these lesions developed into typical blast lesions (Fig. 1e–g).
Leaf lesions were also observed by fluorescent microscopy, which allowed the in planta detection of the fungal mycelium. As depicted in Fig. 1e–g (lower panels), the inoculated regions showed bright GFP fluorescence. A large number of fluorescent hyphae grew and sporulated on the leaf surface in cvs Vialone Nano and Maratelli (Fig. 1f,g; lower panels).
The observed susceptibility to M. oryzae infection of the various rice genotypes was further assessed by image analysis. The percentage of leaf area affected by blast lesions at 3 and 6 dpi was determined. In agreement with the visual inspection of infected leaves, no evident blast lesions developed in infected leaves of cvs Saber and Katy (Fig. 2a,b). The small percentage quantified by image analysis at 6 dpi in Katy leaves (and to a lesser extent in Saber leaves) account for the small necrotic spots, but not for blast lesions. Presumably, the appearance of these necrotic spots in Saber and Katy would prevent tissue colonization which, in turn, would contribute to the resistance phenotype that is observed in these cultivars. Although infected leaves of Kanto 51 showed visible lesions at 3 dpi at the highest inoculum dose (106 spores mL−1), these lesions did not develop with time (3 and 6 dpi; Fig. 2c). In contrast to Kanto 51, a direct correlation between lesion size and inoculum concentration was observed in infected leaves of the susceptible cvs Senia, Vialone Nano and Maratelli, the areas of these lesions increasing with time (Fig. 2d–f).
CLSM analysis of susceptible rice cultivars
Confocal microscopy was used to investigate in detail the early stages of the infection process in the susceptible rice genotype Vialone Nano. Magnaporthe oryzae spores were easily visualized on the leaf surface and started to germinate by 3–6 h after inoculation (Fig. 3a). Typically, each spore produced a polarized germ tube, which emerged from the apical cell of the conidium (Fig. 3b). Penetration structures were identified by localized hyphal swellings corresponding to appressoria (Fig. 3b,c). After penetration into the host tissue, thin filamentous primary hyphae grew in the cell lumen and subsequently differentiated into bulbous and branched secondary hyphae. Confocal images illustrating the dimorphism of primary and internal hyphae are shown in Fig. 3c–f. A close inspection of CLSM images of the infected leaves also revealed specific locations at which hyphae constricted (Fig. 3e,f; arrowheads). These specialized hyphae, also referred to as invasive hyphae pegs, allow the cell-to-cell movement of the fungus. Bulbous infection hyphae were also detected inside the guard cells (Fig. 3g). By 3 dpi, contiguous mesophyll cells were invaded by hyphae forming green fluorescent lines along the leaf (Fig. 3h,i). By this time, patches of mycelial network had also formed on the surface of the leaf (Fig. 3j).
Confocal analysis of resistant rice cultivars
Leaves of cvs Saber and TeQing (Pib gene) were inoculated with spores of the gfp-expressing M. oryzae isolate and examined by CLSM during the early stages of fungal growth. Results are presented in Figure 4. Magnaporthe oryzae spores germinated freely, but the majority of the germlings produced highly abnormal germ tubes, with constrictions and bulges distributed along them. The spore and germ tube cells of germlings growing on Saber and TeQing were highly vacuolated (Fig. 4a–c) compared to germlings growing on the susceptible cv. Vialone Nano. Frequently, more than one germ tube formed from the same spore (Fig. 4d–f; arrowheads). Although appressoria formed on the leaf surface, there was no visible evidence of penetration events, either in Saber or TeQing. Infectious hyphae were not detected inside host cells, not even 60 h after inoculation. These observations support that growth and development of germlings is severely impaired in leaves of Saber and TeQing which might, in turn, prevent penetration into the host tissue.
Blast disease, caused by M. oryzae, is one of the most serious diseases of rice worldwide. This is apparently the first study describing blast infection in susceptible and resistant rice cultivars using a gfp-expressing isolate of M. oryzae. The infection process was followed at the macroscopic level up to 6 dpi (necrotrophic stage). The gfp-expressing isolate was also used to examine, by means of CLSM, the occurrence of reactions in the pathogen upon inoculation of resistant rice cultivars. In all cases, detached leaves were locally inoculated with increasing concentrations of M. oryzae spores. The detached leaf assay has proven to be a convenient method for the analysis of the early in planta growth stages of the rice blast fungus, yielding highly uniform and reproducible infection in the rice leaf (Berruyer et al., 2006). Nevertheless, detached leaves show enhanced susceptibility to the rice blast fungus compared with inoculation of whole plants (Berruyer et al., 2006). This would explain the difference between the results obtained in this study, where nectrotic lesions were observed on detached leaves of Kanto 51 (Pik gene), and those previously obtained with artificial inoculation of whole Kanto 51 seedlings, which showed resistance (Tacconi et al., 2010). It is also true that blast lesions in Kanto 51 leaves only developed at the highest inoculum dose (106 spores mL−1) in the present work, and that this inoculum dose was much higher than that used to spray-inoculate whole seedlings (104 spores mL−1; Tacconi et al., 2010). Although M. oryzae-infected Kanto 51 leaves developed necrotic lesions, sporulation in these lesions was relatively low. Contrary to this, a dense hyphal mycelium covered blast lesions in Vialone Nano at 3 dpi, and by 6 dpi the fungus was sporulating abundantly on these lesions.
Confocal laser scanning microscopy also revealed fungal penetration into the epidermal host cells and colonization of mesophyll cells in Vialone Nano. The primary hyphae filled the first-invaded cells, and moved into the neighbouring cells, presumably via plasmodesmata, after producing highly constricted hyphae that crossed the cell wall (Kankanala et al., 2007). The primary hyphae differentiated into bulbous hyphae that were thicker than the primary ones, thus confirming that the fungus had entered the necrotrophic phase. This morphology and behaviour has been already described in other compatible rice–M. oryzae interactions (Kankanala et al., 2007; Campos-Soriano & San Segundo, 2009).
Whereas the process of infection by M. oryzae in susceptible rice cultivars is documented in the literature, few studies have addressed the impact of R gene action on the early stages of fungal growth in resistant rice cultivars. Most blast disease resistance genes encode nucleotide-binding site–leucine-rich repeat (NBS–LRR) proteins with a predicted intracellular location. The predicted intracellular location of the cognate resistance gene products implies that the fungal virulence factors that interact with these resistance gene products enter into the cytoplasm of plant cells. For instance, the Pita gene was shown to mediate gene-for-gene resistance against M. oryzae isolates that express avirulent alleles of AVR-Pita (Bryan et al., 2000). It has been proposed that AVR-Pita is delivered into the cytoplasm of the rice cell where it binds to the Pita protein to initiate a Pita-mediated defence response (Jia et al., 2000). However, despite the large number of resistance genes that have been identified and mapped in the rice genome, only a few blast resistance genes have been functionally characterized (e.g. Pib, Pita, Pik and Pi54) and the prospect for durable control of rice blast based on the development of resistant genotypes carrying R genes still represents a substantial challenge.
The rice cultivars examined in this study were TeQing and Saber (Pib gene), Katy (Pita2 gene) and Kanto 51 (Pik gene). The Pib gene (distal end of the long arm of chromosome 2), was shown to confer a high level of resistance to most Japanese, some Chinese and all European isolates tested (Roumen et al., 1997; Wang et al., 2009). Rice breeding programmes incorporated Pib from the Chinese cultivar TeQing into Saber (McClung et al., 2004). As for the Pita gene (chromosome 12), it confers resistance to most of the European and two African M. oryzae races (Roumen et al., 1997; Jia, 2009). The Pita resistance gene from the Vietnamese landrace Tetep was incorporated into cv. Katy (Moldenhauer et al., 1990). This R gene is also considered an important source of resistance to blast, and has been effectively deployed to prevent blast in the southern USA for over a decade (Moldenhauer et al., 1990; Jia et al., 2004). Finally, the Pik locus (long arm of chromosome 11) includes at least five blast R genes (Pik, Pik-m, Pik-p, Pik-h and Pik-s) (Ashikawa et al., 2008; Wang et al., 2009). With the exception of Kanto 51, the resistant cultivars in the present study (TeQing, Saber and Katy) did not develop blast lesions.
To gain further insights into the basis of blast resistance, a detailed study of the initial stages of the interaction of M. oryzae with the resistant cvs Saber and TeQing was carried out by CLSM, focusing on spore germination and germling development. This revealed that germling growth was markedly and adversely affected in the resistant cultivars. Even though spores germinated on the leaf surface of Saber and TeQing, they produced morphologically abnormal germlings. Penetration could not be detected in any of these resistant genotypes. Presumably, the abnormal appearance of germ tubes, then becoming vacuolated, and the apparent lack of penetration would reflect R-gene activity. If so, these observations might be indicative that, at least in part, the resistance observed in Saber and TeQing results from interference in germling growth and/or the failure of the germ tube to differentiate into a functional appressorium. The increased vacuolation that accompanies morphological defects in germ tube cells might reflect a misregulation in processes mediating turgor generation in the appressorium, which then cannot be translated into physical force for penetration into the host tissue. Additionally, the possibility that recognition of the pathogen by the host cell leads to the production of chemicals or factor(s) being released to the leaf surface and disrupting germling development should be considered. Alternatively, host cell wall modification and creation of structural barriers in the rice cells may well cause the host to resist attack by the fungus. Whatever the explanation is, it is intriguing to observe that these cytological reactions of M. oryzae cells in resistant cultivars occur in the absence of host cell penetration. Disruption of germling development on the leaf surface might be part of the resistance mechanisms induced by Pib functioning. In other studies carried out on typical gene-for-gene interactions, major blast resistance genes have been shown to prevent disease by blocking fungal growth after the pathogen penetrates into the host tissue and begins to grow within rice cells (i.e. the interaction of the Pita and avirulence gene products in the cytoplasm of the plant cell; Jia et al., 2000).
On the other hand, elegant studies by Veneault-Fourrey et al. (2006) provided evidence that the formation of a functional appressorium requires, sequentially, the completion of mitosis, nuclear migration, and death of the fungal spore. Autophagic cell death of the fungal spore was found to be a prerequisite for infection. Thus, it is possible that impairment or failure in orchestrating these processes might contribute to resistance in cvs Saber and TeQing. Clearly, further studies are needed to shed light on the mechanisms responsible for the observed abnormalities in M. oryzae germlings growing on the leaf surface of these resistant rice cultivars.
Collectively, the results presented here illustrate the usefulness of the combination of fluorescently labelled fungal isolates and CLSM for studies on rice–M. oryzae interactions. The use of gfp-expressing M. oryzae will be valuable for identifying resistant genotypes in rice germplasm in future breeding programmes and for evaluating the effectiveness of R-gene-mediated resistance. Understanding processes and resistance mechanisms operating in the rice plant during its interaction with the rice blast fungus is of paramount importance for the development of effective and durable strategies to manage rice blast disease.
LCS was a recipient of a predoctoral fellowship from the Generalitat de Catalunya. We are grateful to A. Godó for her collaboration in parts of this work. This work was funded by grant BIO2009-08719 from MINECO and the Proyecto Intramural 200420E613 from CSIC to BSS, the Consolider-Ingenio CSD2007-00036 to CRAG, the VALORYZA project (DM 301/7303/06 Ministero delle Politiche Agricole, Rome, Italy) to EL, and the EU co-funded project EURIGEN (049 AGRI GEN RES). We also thank the Xarxa de Referencia en Biotechnologia and SGR (Support to Research Groups from the Agència de Gestió d'Ajuts Universitaris i de Recerca) from the Generalitat de Catalunya for substantial support. EL acknowledges a CRA grant as visiting scientist at CRAG, Barcelona.