The Magnaporthe oryzae effector Pwl2 alters HIPP43 localization to suppress host immunity

Abstract

The rice blast fungus Magnaporthe oryzae secretes a battery of effector proteins to facilitate host infection. Among these effectors, pathogenicity toward weeping lovegrass 2 (Pwl2) was originally identified as a host specificity determinant for the infection of weeping lovegrass (Eragrostis curvula) and is also recognized by the barley (Hordeum vulgare) Mla3 resistance protein. However, the biological activity of Pwl2 remains unknown. Here, we showed that the Pmk1 MAP kinase regulates PWL2 expression during the cell-to-cell movement of M. oryzae at plasmodesmata-containing pit fields. Consistent with this finding, we provided evidence that Pwl2 binds to the barley heavy metal-binding isoprenylated protein HIPP43, which results in HIPP43 displacement from plasmodesmata. Transgenic barley lines overexpressing PWL2 or HIPP43 exhibit attenuated immune responses and increased disease susceptibility. In contrast, a Pwl2SNDEYWY variant that does not interact with HIPP43 fails to alter the plasmodesmata localization of HIPP43. Targeted deletion of 3 PWL2 copies in M. oryzae resulted in a Δpwl2 mutant showing gain of virulence toward weeping lovegrass and barley Mla3 lines, but reduced blast disease severity on susceptible host plants. Taken together, our results provide evidence that Pwl2 is a virulence factor that suppresses host immunity by perturbing the plasmodesmatal deployment of HIPP43.

Figure 1.

PWL2 expression is regulated in a Pmk1-dependent manner during host infection. A and B) Micrographs and line scan graph showing Pwl2 secreted through the BIC in rice cells during early infection. Conidial suspension at 1 × 10⁵ mL⁻¹ of M. oryzae strain coexpressing 2 BIC-localized effectors; Bas1-GFP and Pwl2-mRFP were inoculated onto a susceptible cultivar Moukoto rice leaf sheath and images captured at 26 hpi. Fluorescence of the 2 effectors was observed as small punctate signals in the same BIC. C and D) Coinfection assay of rice leaf sheath with 2 different M. oryzae strains, one expressing Pwl2-mRFP and the other Pwl2-GFP at 30 hpi. Micrographs and line scan graph show there is an absence of mixed fluorescence confirming the BIC does not contain Pwl2 transferred from rice cells. BICs indicated by magenta arrowheads for mRFP and green arrowheads for GFP. E) Schematic illustration to describe the workflow used to test genes regulated in a Pmk1-dependent manner. Rice leaf sheaths were infected with a M. oryzae pmk1^AS mutant spores before mock or 1NA-PP1 treatment. Treated and mock treated, infected leaf sheaths were trimmed and used for RNA isolation followed by sequencing. Figure created with BioRender https://biorender.com/. F and G) Bar charts to show that a subset of effectors is regulated in a manner that requires Pmk1. Gene expression is shown as transcripts per million at 32 hpi, F) showing lowly expressed effectors and G) highly expressed. Error bars represent SEM, and individual points represent 3 independent biological replicates. H to M) Micrographs showing expression of Pwl2-GFP by M. oryzae pmk1^AS in leaf sheaths of a susceptible rice line CO39 using conidial suspension at 1 × 10⁵ mL⁻¹. Fluorescence of Pwl2 at different stages of infectious hyphal progression, starting with early stage of infection, Stage 1 (30 to 36 hpi) where a newly differentiated bulbous hyphae is formed as per (Sakulkoo et al. 2018). Analysis was also carried out at later stages of infection including Stage 2 (36 to 40 hpi), where a primary invaded cell is filled with differentiated bulbous hyphae), and Stage 3 (40 to 48 hpi) where colonization of primary invaded cell is complete and there is full invasion of secondary invaded cells. Arrowheads indicate fluorescence in the BIC. The primary and secondary invaded cells are indicated with 1 and 2, respectively. Scale bars represent 20 µm. To test Pmk1 inhibition, inoculated leaf sheaths were treated with 5 µm 1NA-PP1 at 26 hpi. N) Inhibition of Pwl2 by 1NA-PP1 is quantified as percentage of cells showing Pwl2 fluorescence in the BIC (100 invaded cells were counted per replicate). Error bars represent SEM, and individual points represent independent biological replicate. For F, G, and N), significance between groups of samples was performed using unpaired Student's t-test with Welch correction. ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05; NS, no significant difference; AP, appressorium.

Figure 2.

Pwl2 is highly conserved in isolates of M. oryzae and displays conserved structural features. A) Phylogenetic analysis and PWL gene family distribution in M. oryzae. A maximum parsimony tree (branch lengths are not drawn to scale) was generated using kSNP3 to include isolates from different host-limited forms of M. oryzae including isolates that infect O. sativa (rice), Eleusine spp. (finger millet), H. vulgare (barley), Setaria spp. (foxtail millet), Triticum aestivum (wheat), Lolium spp. (rye grass), Brachiaria spp. (armgrass millet), Panicum spp. (torpedo grass), Eragrostis spp. (weeping lovegrass), Stenotaphrum spp. (St. Augustine grass), Cynodon spp. (Bermuda grass), and Urochloa spp. (signal grass) (Ou 1980; Talbot 2003; Cruz and Valent 2017; Inoue et al. 2017), as well as Magnaporthe species that infect Digitaria sanguinalis (crabgrass) and Pennisetum spp. (pearl millet). We used Pwl1 (BAH22184.1), Pwl2 (QNS36448.1), Pwl3 (AAA80240.1), and Pwl4 (AAA80241.1) protein sequences to query the presence or absence of each gene using tblastn. The heatmap indicates the presence/absence of genes in the PWL family. PWL1 is predominantly present in group EC-1I (Asuke et al. 2020) of Eleusine-infecting isolates and some Oryza-infecting isolates, but largely absent from other host-specific lineages, except one Eragrostis-infecting isolate, EtK19-1, and one Cynodon-infecting isolate, Cd88215, but not in Digitaria and Pennisetum spp. lineages. PWL3 is present in most Oryza-infecting isolates, some Setaria-, Brachiaria-, Stenotaphrum-, some Lolium-, and some Eleusine-infecting isolates but largely absent in Eragrostis and Triticum lineages, as well as in Digitaria and Pennisetum lineages. PWL4 is present in most Eleusine-, Eragrostis-, most Lolium-, most Triticum-, Pennisetum-, Digitaria-infecting isolates, but in only in 3 Oryza-infecting isolates. PWL4 is however missing in Setaria-, Brachiaria- and Stenotaphrum-infecting isolates. PWL2 is found in most host-limited forms of M. oryzae and related Magnaporthe species but absent in Brachiaria, Setaria, Panicum, Cynodon, and Urochloa. B) Superimposition of different Pwl2 variants predicted using AlphaFold3 onto the resolved Pwl2 structure (magenta) indicating region of polymorphism. Different colors represent different variants as follows: Pwl2-2 (teal), Pwl2-3 (brick red), Pwl2-TH3 (dark olive green), Pwl2-IN15 (dark teal), Pwl2-KE210 (blue), Pwl2-CKF3584 (dark gray), Pwl2-EtKY19-1 (orange), and Pwl2-DS9461 (light gray). The superimposition shows overall structural conservation in MAX-fold, without the signal peptide and the C-terminus. The variants of Pwl2 are named with corresponding isolate name and grouped according to host species. Varying residues are colored according to the genome from which they were identified.

Figure 3.

PWL2 has undergone copy number expansion in M. oryzae field isolates. A) Schematic diagram showing the estimated chromosomal location of PWL2 on multiple loci (Chr3 and 6) of the reference genome 70-15 and on equivalent region of assembled contigs of laboratory strain Guy11 genome. PWL2 is flanked by POT2 and MGR583 repeated sequences suggesting a possible involvement in translocation of events into different loci in the genome. Arrows indicate location of POT2 and MGR58, while PWL2 is labeled in magenta. B) A k-mer analysis on sequenced raw reads was used to determine copy number variation of PWL family genes in different M. oryzae isolates. Plot shows high copy number of PWL2 in analyzed genomes (n = 286) compared with the other gene family members, selected effectors AVR-Piz-t, BAS1, BAS4, and SLP1, and selected control genes EX070, MPG1, and SEC5. Similarly, AVR-Pik shows multiple copies in different isolates. Copy number of PWL2 in selected isolates Guy11, KE002, BF48, 70-15 is indicated. Ina168 that lacks PWL2 was used as a NC.

Figure 4.

CRISPR-Cas9 Δpwl2 mutants demonstrate Pwl2 as both host range and a virulence factor. A) Schematic illustration of the CRISPR-Cas9–mediated gene-editing process for inserting the HPH gene cassette at the PWL2 locus. The guided sequence (sgPWL2) directs Cas9 to introduce a double-stranded break at the PWL2 locus, with the DNA repair template containing the hygromycin resistance gene cassette and flanking regions of the PWL2 gene. This leads to mutations in the form of indels or gene replacements. Figure created with BioRender https://biorender.com/. Positive mutants were identified by amplifying the hygromycin cassette from transformants with the 3 copies deleted. B) Comparison of disease symptoms in weeping lovegrass (E. curvula) infected with different M. oryzae isolates. Top panel shows typical disease lesions produced by E. curvula-infecting isolate G17 (−PWL2) and a rice-infecting isolate Ina168 (−PWL2). In the middle panel, CRISPR-Cas9 deletion mutants exhibit gain of virulence on weeping lovegrass compared with Guy11. In the bottom panel, complemented strains with both native and constitutive RP27 promoters showed a loss of virulence on weeping lovegrass. Observations were consistent in 3 independent biological replicates (n = 3 plants). Scale bar represents 5 cm. C) Δpwl2 mutants display reduced pathogenicity on rice cultivar CO39. Conidial suspensions from Guy11, Δpwl2 (T6), and complemented Δpwl2(T6) + PWL2p:PWL2 were used to inoculate 21-d-old seedlings of the blast-susceptible cultivar CO39, and disease symptoms were recorded after 5 dpi. The boxplots show the mean lesion density recorded per 5 cm D) and lesion area in mm²  E). Data points of different colors represent different biological replicates. F) The Δpwl2 mutant exhibits enhanced pathogenicity on barley cultivar Baronesse (+Mla3). Conidial suspensions from Guy11 and Δpwl2 T6 were used to inoculate 10-d-old seedlings of barley lines Golden Promise (+Mla8), Siri (+Mla8), Nigrate (−Mla3), and Baronesse (+Mla3), and disease symptoms were recorded after 5 dpi. Conidial suspensions at 1 × 10⁵ mL⁻¹ spores/mL were used for infection assays. G) The boxplot shows lesion size of barley seedlings infected with Guy11 and Δpwl2 (T6) (individual points represent lesion area in mm²). The lower border and upper border of the box show the lower quartile and upper quartile, respectively. The line in the box shows the median. Significance between groups of samples in D, E, and F) were performed using unpaired Student's t-test with Welch correction. ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05; NS, no significant difference. Error bars represent SEM in 3 independent biological replicates. Scale bar represents 5 mm.

Figure 5.

Pwl2 suppresses PAMP-induced ROS in transgenic barley lines. A) Schematic illustration to describe the workflow used to generate transgenic plants, test for PTI response, and identify putative Pwl2 interactors using discovery proteomics. ROS production in leaf disks collected from 4-wk-old 2 independent stable transgenic lines expressing Pwl2-YFP compared with cv. Golden Promise induced by 1 μm flg22 B and C) or 1 mg/mL chitin D and E) (n > 8). For ROS assay, line graph B and D) points represent mean per time point and error bars represent SEM. Dot plots C and E) show integrated ROS production over 60 min; error bars represent mean ± Sd. The lower horizontal line shows the minimum value, the upper horizontal line shows the maximum value, and the middle line shows the mean value. F) Leaf drop infection on 2 independent barley transgenic lines expressing Pwl2-YFP compared with infection on wild-type cv. Golden Promise 3 to 4 dpi. Scale bar represents 5 mm. G) The boxplot represents the disease lesion area in mm² on cv. Golden Promise compared with 2 independent Pwl2-YFP transgenic lines (data points of different colors represent different biological replicates). The lower border and upper border of the box show lower quartile and upper quartile range of the data, respectively. The line in the box shows the median. Unpaired Student's t-test with Welch correction was performed to determine significant differences: ****P < 0.0001, ***P < 0.001, **P < 0.01. Error bars represent SEM. These experiments were repeated 3 times to obtain consistent result.

Figure 6.

Pwl2 interacts with HIPP43 and its orthologs from other grass species. A) Putative Pwl2 interacting peptides were immunoprecipitated from protein extracts of 3-wk-old stable cv. Golden Promise transgenic lines expressing Pwl2-YFP or free cytoplasmic YFP using anti-GFP antibodies, and LC-MS/MS was performed to identify unique putatively interacting peptides. The scale 0 to 20 represents the Log2fold change in peptides when compared with the control. B) One-to-one yeast 2-hybrid between Pwl2 and selected top candidates from IP-MS analysis, PPIPs. HMA integrated in rice NLR Pikm-1 and Pikp-1 were used as specificity controls. Simultaneous cotransformation of pGADT7-Pwl2 (prey vector) together with pGBK-PPIPs (bait vector) or PGBKT7-53 and pGADT7-T (positive control) pGADTT7-T and pGBKT7-Lam (NC) into Y2H gold strain was carried out. Positive interaction resulted in the activation of 4 reporter genes and growth on high-stringency medium (−Ade, −Leu, –Trp, –His + X-α-gal and 3AT). Cotransformation also activates the expression of MEL1, which results in the secretion of α-galactosidase and the hydrolysis of X-α-gal in the medium, turning the yeast colonies blue. HIPP43 exclusively interacts with Pwl2 in SD/-L-W-H-A X-α-gal medium with 3AT added. These experiments were repeated several times over 3 yrs obtaining consistent result. C) Co-IP of Pwl2-YFP and Myc-HIPP43 or D) mCherry-HIPP43 in N. benthamiana leaves. C- or N-terminal GFP-tagged Pwl2 and C-terminal Myc-HIPP43 was cloned into the vector pGW514 and transformed into Agrobacterium strain GV3101 and coinfiltrated into N. benthamiana leaves and left to incubate for 48 h. Immunoprecipitates were obtained with anti-GFP affinity matrix beads and probed with anti-GFP-peroxidase, anti-mCherry-peroxidase, and anti-Myc-peroxidase, Horseradish Peroxidase (HRP)-conjugated antibodies. Total protein extracts were also probed with appropriate (HRP-conjugated) antibodies. Magenta asterisks indicate expected band sizes. E) Y2H analysis shows in upper panel, HvHIPP43 interacts with PWL gene family products Pwl1, Pwl3, Pwl4, and variant pwl2-3. F) Pwl2 interacts with HIPP43 orthologs from rice (OsHIPP43), wheat (TaesHIPP43), weeping love grass (EcHIPP43), foxtail millet (SitiHIPP43), and wild wheat (HIPP1-V). These experiments were repeated several times over 3 yrs obtaining consistent result.

Figure 7.

HvHIPP43 suppresses PAMP-induced ROS in transgenic barley and is stabilized by Pwl2. A to D) ROS production measured from leaf disks collected from 4-wk-old stable transgenic line expressing YFP-HIPP43 and cv. Golden promise (control) in the absence and presence of 1 μm flg22 or 1 mg/mL Chitin (n > 8). Line graph A and C) points represent mean per time point, and error bars represent SEM. Dot plots B and D) show integrated ROS production over 60 min, and error bars represent mean ± Sd. The lower horizontal line shows the minimum value, the upper horizontal line shows the maximum value, and the middle line shows the mean value. E) Leaf drop infection on barley transgenic lines expressing YFP-HIPP43 compared with wild-type Golden Promise. Conidial suspensions at 1 × 10⁵ mL⁻¹ spores/mL from Guy11 were used for inoculation. Disease symptoms were recorded after 4 dpi. Scale bar represents 5 mm. F) Box plots show lesion area in mm². All experiments were repeated 3 times giving consistent results. G and H) Micrographs and boxplot showing mCherry-HIPP43 localizing as small puncta on the plasma membrane when expressed in N. benthamiana. White arrowheads indicate regions of HIPP43 localization at the PD. Staining of callose using aniline blue overlaps with mCherry-HIPP43, confirming mCherry-HIPP43 localizes exclusively at the PD localization and cytoplasmic mobile bodies (MBs) localization is absent. I and J) Micrographs and boxplot showing the presence of YFP-Pwl2, alters mCherry-HIPP43 localization, which was observed to translocate to the cytoplasm, as mobile cytoplasmic bodies. White arrowheads indicate cytoplasmic puncta of HIPP43 and Pwl2 colocalization. K and L) Micrographs and boxplot showing mCherry-HIPP43 remains in the PD in the presence of cytoplasmically localized AVR-PikE. M and N) Micrographs and boxplot showing mCherry-HIPP43 remains in the PD in the presence of cytoplasmically and PD localized MEP3. O and P) Micrographs and boxplot showing mCherry-HIPP43 remains in the PD in the presence of Pwl2^SNDEYWY, a mutant that does not interact with HIPP43. For K, M, and O), white arrowheads indicate regions of HIPP43 localization at the PD. Dotted lines in the overlay panels correspond to distance and direction of line intensity plot. Scale bars represent 20 µm. To determine colocalization, 50 or more transformed cells were counted per replicate. For box plots, the lower horizontal line shows the minimum value, the upper horizontal line shows the maximum value, and the middle line shows the median value. Data points of different colors represent different biological replicates. Unpaired Student's t-test was performed to determine significant differences ****P < 0.0001, ***P < 0.001, **P < 0.01. Error bars represent SEM. Microscope imaging experiments were repeated several times over 2 yr with consistent results. Q) Western blot to show that Pwl2 does not degrade HIPP43. Total protein extracts from N. benthamiana leaves coinfiltrated with YFP-Pwl2, MEP3 or AVR-PikE, and mCherry-HIPP43 were immunoblotted and probed with anti-GFP-peroxidase (left) or anti-mCherry-peroxidase (right) (HRP-conjugated) antibodies. Magenta asterisks indicate expected band sizes.

Figure 8.

Pwl2^SNDEYWY does not complement Mla3 recognition and virulence on a blast-susceptible rice cultivar CO39. A and B) Schematic representation to show how complemented Δpwl2 + PWL2p:Pwl2^SNDEYWY-positive transformants were screened using PCR to amplify PWL2 coding sequence. C) Boxplots showing relative expression as log2 fold change of Pwl2^SNDEYWY in 2 selected transformants, C6 (left panel) and C14 (right panel) using qRT-PCR. The lower border and upper border of the box show lower quartile and upper quartile range of the data, respectively. The line in the box shows the median. Unpaired Student's t-test with Welch correction was performed to determine significant differences. Detached leaves of 10-d-old seedlings of barley were inoculated with Δpwl2 + PWL2p:Pwl2^SNDEYWY, and infected tissue was collected 40 hpi and used for RNA isolation (3 biological replicates); cDNA was synthesized and samples were used for qRT-PCR. D and E) Δpwl2 + PWL2p:Pwl2^SNDEYWY-complemented strains C6 and C14 produced compatible disease lesions on barley cultivar Baronesse (+Mla3) D) and Nigrate (−Mla3) E). Conidial suspensions at 1 × 10⁵ mL⁻¹ spores/mL from Guy11, Δpwl2, and complemented Δpwl2 + PWL2p:Pwl2^SNDEYWY were used to inoculate 10-d-old seedlings of barley, and disease symptoms were recorded after 5 dpi. F to H) Complemented Δpwl2 + PWL2p:Pwl2^SNDEYWY display reduced pathogenicity on rice cultivar CO39. Mean lesion density recorded per 5 cm G) and lesion area in mm²  H) of Guy11 compared with Δpwl2 and complemented Δpwl2 + PWL2p:Pwl2^SNDEYWY following leaf spray infection. Box and whisker plots with individual data points from leaves collected in 3 independent replicates. Error bars represent SEM. Data points of different colors represent different biological replicates. Conidial suspensions at 1 × 10⁵ mL⁻¹ spores/mL from Guy11, Δpwl2 T6, and complemented Δpwl2 + PWL2p:Pwl2^SNDEYWY were used to inoculate 21-d-old seedlings of the blast-susceptible cultivar CO39, and disease symptoms were recorded after 5 dpi. The lower border and upper border of the box show lower quartile and upper quartile range of the data, respectively. The line in the box shows the median. Unpaired Student's t-test with Welch correction was performed to determine significant differences. **P < 0.01, *P < 0.05; NS, no significant difference. Scale bars represent 5 mm.