A single fungal MAP kinase controls plant cell-to-cell invasion by the rice blast fungus

See allHide authors and affiliations

Science  23 Mar 2018:
Vol. 359, Issue 6382, pp. 1399-1403
DOI: 10.1126/science.aaq0892

Keeping the channels open

When the rice blast fungus enters a rice cell, the plasma membrane stays intact, so the rice cell remains viable. The fungus then moves to adjacent cells via plasmodesmata, the plant's intercellular channels. Sakulkoo et al. used a chemical genetic approach to selectively inhibit a single MAP (mitogen-activated protein) kinase, Pmk1, in the blast fungus. Inhibition of Pmk1 trapped the fungus within a rice cell. Pmk1 regulated the expression of a suite of effector genes involved in suppression of host immunity, allowing the fungus to manipulate plasmodesmal conductance. At the same time, Pmk1 regulated the fungus's hyphal constriction, which allows movement into new host cells.

Science, this issue p. 1399


Blast disease destroys up to 30% of the rice crop annually and threatens global food security. The blast fungus Magnaporthe oryzae invades plant tissue with hyphae that proliferate and grow from cell to cell, often through pit fields, where plasmodesmata cluster. We showed that chemical genetic inhibition of a single fungal mitogen-activated protein (MAP) kinase, Pmk1, prevents M. oryzae from infecting adjacent plant cells, leaving the fungus trapped within a single plant cell. Pmk1 regulates expression of secreted fungal effector proteins implicated in suppression of host immune defenses, preventing reactive oxygen species generation and excessive callose deposition at plasmodesmata. Furthermore, Pmk1 controls the hyphal constriction required for fungal growth from one rice cell to the neighboring cell, enabling host tissue colonization and blast disease.

Blast diseases of cereals are caused by the filamentous fungus Magnaporthe oryzae (synonym of Pyricularia oryzae), destroying sufficient rice each year to feed 60 million people (1), and wheat blast disease now threatens wheat production in South America and, most recently, Asia (2). Plant infection requires an infection cell, called an appressorium, which uses a pressure-driven mechanism to breach the tough cuticle of the leaf (3, 4). Once inside plant tissue, the fungus elaborates pseudohypha-like invasive hyphae that rapidly colonize living host cells, secreting effector molecules to suppress host immunity and facilitate infection (5). M. oryzae effectors are delivered into host cytoplasm by means of a biotrophic interfacial complex (BIC), a plant-derived membrane-rich structure in which effectors accumulate during transit to the host (58). Hyphae then appear to locate pit fields, composed of plasmodesmata, which are traversed by constricted, narrow hyphae, enabling the spread of the fungus to adjacent host cells (9). The fungus rapidly colonizes host tissue, and disease lesions appear within 4 to 5 days of initial infection by spores.

In this study, we investigated how M. oryzae colonizes host tissue and, in particular, how it spreads from one plant cell to the next. We first performed ultrastructural analysis of rice sheath cells infected with the pathogenic strain Guy11. This analysis confirmed constriction of hyphae from an average diameter of 5.0 μm to 0.6 μm during traversal of rice cells (Fig. 1, A and B, and fig. S1). The rice plasma membrane in the second invaded cell remained intact as an electron-dense lining near the rice cell wall, continuous with the plant plasma membrane around hyphae (Fig. 1B) (5, 8, 9). By contrast, the rice plasma membrane in the first invaded cell lost integrity upon exit of the fungus to the next rice cell (Fig. 1, B and E) (7, 9, 10).

Fig. 1 Cell-to-cell invasion and plasmodesmal manipulation by M. oryzae.

(A) Transmission electron micrograph of an invasive hypha (IH) traversing a rice cell wall (RCW) at 42 hpi. Scale bar, 0.5 μm. (B) High-magnification view of the crossing site. The rice plasma membrane (RPM), fungal plasma membrane (FPM), and fungal cell wall (FCW), and extrainvasive hyphal membrane (EIHM) are indicated. Scale bar, 20 nm. (C) Callose deposition in an infected rice cell at 34 hpi, shown by aniline blue staining. Arrowheads indicate plasmodesmal callose deposition. Arrows indicate callose collars that form around hyphae after they enter adjacent cells (asterisks). Scale bar, 5 μm. (D) Hyphae traversing the cell wall at a pit field (arrow). Scale bar, 0.5 μm. (E) Difference in host cytoplasmic contents between the first and second invaded cells (for a larger image, see fig. S1G). VC, vacuole. Scale bar, 1 μm. (F) Localization of a septin Sep5-GFP collar at rice cell crossing points (arrow and high-magnification inset) at 40 hpi. Scale bars, 10 μm (main panel); 5 μm (inset). (G) Diffusion of single and double mCherry fluorescent proteins in uninfected leaf tissue (left and middle panels, respectively) and in leaves infected with M. oryzae (right panel) at 24 hours (9). Asterisks indicate bombarded cells. Scale bars, 10 μm. (H) Percentages of bombarded cells showing diffusion of mCherry proteins in blast-infected and uninfected rice tissues (*P = 0.05; n = 100 cells; error bars indicate SE).

One of the plant’s defenses against infection is to close intercellular plasmodesma channels by deposition of callose (11, 12). We visualized callose using aniline blue staining of rice cells (Fig. 1C and fig. S3). Callose papillae often form at appressorium penetration sites, but no callose occlusions were initially observed at plasmodesmata during infection of the first rice cell at 27 hours postinoculation (hpi) (fig. S3). Later callose deposition was observed at plasmodesmata by 30 hpi, consistent with the onset of cell death among initially invaded rice cells (figs. S1 and S2). Callose collars then formed around the base of invasive hyphae after they invaded adjacent cells, at 34 hpi (Fig. 1C and fig. S3). These observations suggest that M. oryzae is able to clear plasmodesmal occlusion materials before penetrating pit fields (Fig. 1D). We also observed a switch from polarized to isotropic fungal growth by invasive hyphae at rice cell junctions. The polarisome marker Spa2–green fluorescent protein (Spa2-GFP) localized to hyphal tips and then disappeared as hyphal tips swelled upon contact with the host cell wall (fig. S2) (5). Spa2-GFP then appeared again at hyphal tips in newly colonized cells (fig. S2). Cell wall crossing was accompanied by reorganization of fungal septins and F-actin into an hourglass shape at the point of maximum hyphal constriction (Fig. 1F, figs. S2 and S10, and movie S1) (3, 5, 13).

Plasmodesmata are dynamic structures through which proteins diffuse between plant cells (11). To test whether the blast fungus can manipulate plasmodesmata by increasing their size exclusion limit to facilitate effector diffusion to adjacent plant cells, we bombarded rice tissue infected with M. oryzae at 24 hpi with single- and double-sized mCherry expression vectors. In uninfected rice tissue, a 28.8-kDa single mCherry protein moved into neighboring rice cells but a 57.6-kDa double mCherry fusion protein generally did not, owing to its larger size (Fig. 1, G and H). By contrast, in blast-infected tissue, double mCherry protein diffused to adjacent rice cells. The gating limit of rice plasmodesmata is therefore relaxed during early M. oryzae infection. We then carried out time-lapse imaging of the fungal apoplastic effector Bas4 (biotrophy-associated secreted protein 4)–GFP (8), expressed under its native promoter, during plant infection (fig. S2 and movie S2). We observed that upon exit of the fungus to an adjacent cell, Bas4-GFP leaked into the host cytoplasm of initially colonized rice cells. However, fluorescence did not diffuse into the newly colonized cells. Plasmodesmata therefore remain open at early stages of infection (24 to 27 hpi) but are closed at later stages, consistent with the increase in plasmodesmal callose deposition at 30 to 34 hpi (fig. S3) (7, 14), suggesting that the blast fungus is able to overcome callose deposition at pit fields to invade neighboring cells. Plasmodesmata may lose their structural integrity after this time, as initially infected rice cells lose their viability.

To study regulatory mechanisms controlling invasive growth, we characterized Pmk1, a fungal mitogen-activated protein kinase (MAPK) essential for appressorium development and pathogenicity that is conserved in many plant-pathogenic fungi (15). Pmk1 null mutants of M. oryzae cannot infect rice plants even when mutants are inoculated onto wounded leaves (15). We decided to conditionally inactivate the Pmk1 MAPK using a chemical genetic approach. We generated an analog-sensitive (AS) allele of PMK1 (pmk1AS) by mutating the gatekeeper residue of the kinase adenosine triphosphate (ATP)–binding site into a small amino acid residue, glycine. The equivalent (Shokat) mutation previously reported in yeast fus3-as1 confers susceptibility to the ATP analog 1-naphthyl-PP1 (1NA-PP1) (16). Expression of the pmk1AS allele, under control of its native promoter, restored pathogenicity to a Δpmk1 mutant (fig. S4). Addition of 1NA-PP1 selectively inhibited the function of Pmk1AS mutants, preventing appressorium development (fig. S4), a result that was identical to the effects of PMK1 deletion or expression of a kinase-inactive allele (17). We also observed that inhibiting Pmk1 after appressorium formation blocked cuticle penetration by preventing assembly of the septin ring at the appressorium pore (fig. S5).

To test the role of Pmk1 in tissue invasion, we allowed a pmk1AS mutant to invade the first rice epidermal cell before adding 1NA-PP1 at 26 hpi. This treatment blocked invasion of adjacent epidermal cells, resulting in the first infected cells becoming filled with fungal hyphae (Fig. 2, A and B). Pmk1 inactivation did not affect the structure of BICs or the morphology of invasive hyphae (fig. S6A). In the presence of 1NA-PP1, hyphae of the pmk1AS mutant still formed terminal swellings at host wall contact points (Fig. 2C) (9) but could not breach adjacent cells. Inhibition of Pmk1 also blocked rice cell invasion in a second rice cultivar, Mokoto, and in barley (fig. S6, B and C). Pmk1 inhibition was accompanied by rapid derepression of reactive oxygen species (ROS) generation, a key plant defense response, and increased ROS-dependent callose deposition (Fig. 2, D and E, and fig. S6). To confirm the role of Pmk1 in cell wall crossing, we performed live-cell imaging of a functional Pmk1-GFP protein, which showed expression of the protein during appressorium-dependent cuticle penetration (Fig. 2F) (17) but also upon contact of invasive hyphae with rice cell walls just before invasion of the neighboring cell (Fig. 2G and movie S4). M. oryzae has two additional MAPKs: Osm1, which is dispensable for virulence (18), and Mps1, which regulates cell integrity and is necessary for fungal infection (19). We therefore generated an analog-sensitive mps1 (mps1AS) mutant and found that selective inactivation of Mps1 enhanced host defenses (20) but did not block cell-to-cell invasion (fig. S7).

Fig. 2 Pmk1 MAPK-dependent regulation of cell-to-cell spread by M. oryzae.

(A and B) Effect of Pmk1 inhibition on host colonization at 48 hpi. Infected rice tissues were treated with 5 μM 1NA-PP1 at 26 hpi. Asterisks indicate appressorium penetration sites. Error bars in (B) indicate SE. (C) Formation of swollen hyphae (arrows) by the pmk1AS mutant upon contact with the host cell wall in the presence of 1NA-PP1, imaged at 40 hpi. (D and E) Induction of ROS production, shown by 3,3′-diaminobenzidine (DAB) staining, after the addition of 1NA-PP1 at 26 hpi. Images were taken at 48 hpi. n = 300 infection sites; *P < 0.05, **P < 0.01, ***P < 0.001, unpaired Student’s t test. Error bars in (E) denote SE. Scale bars in (A), (C), and (D), 20 μm. (F) Transient accumulation of Pmk1-GFP in an appressorium during emergence of a penetration hypha. (G) Transient accumulation of Pmk1-GFP in hyphae at rice cell crossing points (asterisks).

To investigate how Pmk1 regulates invasive growth, we performed RNA sequencing (RNA-seq) analysis to compare M. oryzae gene expression levels during infection with the pmk1AS mutant in the presence and absence of 1NA-PP1. Using an adjusted P value of ≤0.05 for differential gene expression, we found 1457 fungal genes with altered expression during Pmk1 inhibition, accounting for 11.5% of total protein-encoding genes. Of these, 715 fungal genes were up-regulated and 742 were down-regulated (table S2). A subset of effector genes implicated in plant immunity suppression were positively regulated by Pmk1, including genes for Avr-Pita1 (6); Slp1, which suppresses chitin-triggered immunity (21); Avr-Pik; and several Bas effectors, including Bas2 and Bas3 effectors (8) that putatively function at cell wall crossing sites (Fig. 3A and fig. S8). We expressed Bas2-GFP and Bas3-GFP in the pmk1AS mutant and found that they were not expressed in the presence of 1NA-PP1 (Fig. 3, B and C). We also generated a pmk1AS strain expressing cytosolic GFP under control of the Bas3 promoter (Fig. 3, D to F). The addition of 1NA-PP1 inhibited GFP expression during appressorium-mediated infection (Fig. 3D) and during invasive growth (Fig 3, E and F). Localization of other fungal effectors was not affected by Pmk1 inactivation (fig. S8), and Pmk1 therefore affects expression of a subset of fungal effectors involved in suppression of host immunity.

Fig. 3 Pmk1-dependent regulation of effector gene expression during biotrophic growth.

(A) Bar chart showing relative expression levels [fragments per kilobase of transcript per million mapped reads (FPKM)] of known effector genes differentially regulated during Pmk1 inhibition in RNA-seq analysis. (B and C) Localization of Bas2-GFP and Bas3-GFP effectors in a pmk1AS mutant with and without 1NA-PP1. BICs are indicated by arrows. Scale bars in (B), (C), and (E), 20 μm. (D to F) Expression of cytosolic GFP driven by the BAS3 promoter (Bas3p:GFP) in a pmk1AS mutant. (D) Strong GFP expression at 24 hpi in appressoria undergoing infection and lack of GFP expression in appressoria treated with 1NA-PP1 at 8 hpi. Scale bar, 10 μm. (E) GFP expression in pmk1AS hyphae exposed to 1NA-PP1 at 26 hpi and imaged at 30 hpi. (F) Dot plot of Bas3p:GFP fluorescence in a pmk1AS mutant treated with 5 μM 1NA-PP1 at 26 hpi or left untreated. Images were taken at 30 hpi, and background was subtracted before measurement of fluorescence levels. ****P < 0.0001, unpaired Student’s t test; three biological replicates.

To test whether suppression of host immunity, particularly at plasmodesmata, could explain the role of Pmk1 in cell invasion by M. oryzae, we suppressed host immune reactions simultaneously with Pmk1 inhibition. We found that chemical suppression of plant ROS or disruption of salicylic acid regulation did not reverse the effects of Pmk1 inactivation (fig. S9). To suppress host immunity completely, we therefore killed plant tissue by ethanol treatment before rehydration and inoculation with the pmk1AS mutant. In the presence of 1NA-PP1, the mutant still remained trapped in the first dead plant cell (Fig. 4A). We therefore hypothesized that Pmk1-dependent hyphal constriction must be critical for cell wall crossing at pit fields. This relationship would correspond with the role of Pmk1 in septin-dependent appressorium repolarization (fig. S5) (3, 13). Consistent with this idea, RNA-seq analysis revealed several morphogenetic regulators down-regulated during Pmk1 inhibition. Genes for Chm1, a homolog of Cla4 p21-activated protein kinase that phosphorylates septins (22), and a putative F-actin cross-linking protein, alpha-actinin, for example, are among genes positively regulated by Pmk1 during plant infection (table S2). We therefore investigated septin organization during Pmk1 inhibition. Sep5-GFP still accumulated at cell wall contact points but as a disorganized mass instead of the septin collars that normally form at cell wall crossing sites (Fig. 4, figs. S10 and S11, and movies S3 and S5). Finally, we investigated the ability of septin mutants to invade plant tissue. Septin mutants do not penetrate the rice cuticle efficiently because of the roles of septins in appressorium repolarization and the development of penetration hyphae (3). However, a small proportion of penetration events are successful. In these rare instances, the Δsep6 mutant, in particular, showed a reduction in its ability to spread between rice cells, consistent with a role for septins in cell invasion (fig. S12).

Fig. 4 Pmk1 controls septin-dependent morphogenesis of narrow invasive hyphae traversing cell walls.

(A) Micrographs showing that the pmk1AS mutant, growing in ethanol-killed rice tissue treated with 1NA-PP1 at 14 hpi, remained trapped inside initially invaded rice cells in all 200 infection sites examined at 48 hpi. Asterisks indicate appressorium penetration sites. Scale bar, 20 μm. (B) Confocal micrographs showing localization of Sep5-GFP in a pmk1AS mutant at 42 hpi after the addition of 1NA-PP1 (5 μM) at 26 hpi or in the absence of 1NA-PP1. Arrows indicate accumulation of Sep5-GFP at rice cell contact points. Scale bars, 10 μm. Red arrows in high-magnification panels show line scans used to generate corresponding fluorescence intensity distribution graphs.

Taken together, our results demonstrate that the Pmk1 MAPK pathway controls plant tissue invasion by controlling the constriction of invasive hyphae to traverse pit fields in order to invade new rice cells while maintaining the cellular integrity of the host. To accomplish this feat, the MAPK also regulates expression of a battery of effectors to suppress plant immunity, thereby preventing plasmodesmal closure until the fungus has invaded neighboring cells. Plant tissue invasion by the blast fungus is therefore orchestrated, rapid, and necessary for the devastating consequences of the disease.

Supplementary Materials

Materials and Methods

Figs. S1 to S13

Tables S1 and S2

References (2340)

Movies S1 to S5

References and Notes

Acknowledgments: We thank S. Yoshida (RIKEN, Japan) for providing Nipponbare and NahG rice lines. Funding: This work was funded by a Halpin Scholarship for rice blast research to W.S., a European Research Council (ERC) Advanced Investigator Award to N.J.T. under the European Union’s Seventh Framework Programme (FP7/2007-2013) and ERC grant no. 294702 GENBLAST, and Agriculture and Food Research Initiative Competitive grant no. 2012-67013-19291 to B.V. from the U.S. Department of Agriculture National Institute of Food and Agriculture. This is contribution number 18-174-J from the Kansas Agricultural Experiment Station. Author contributions: W.S., B.V., and N.J.T. conceived of the project. W.S., M.O.-R., and E.O.G. performed experimental work. D.M.S. provided bioinformatic data analysis. G.R.L. carried out laser confocal imaging. C.H. and A.C. performed ultrastructural analysis. W.S., M.O.-R., B.V., and N.J.T. wrote the paper with the assistance and input of all coauthors. Competing interests: The authors declare no competing interests. Data and materials availability: All data are available in the article or in the supplementary materials. Differential gene expression analysis data can be accessed at Gene Expression Omnibus ( under accession number GSE106845.

Stay Connected to Science

Navigate This Article