A Phytophthora infestans RXLR effector targets plant PP1c isoforms that promote late blight disease

Abstract

Plant pathogens deliver effectors to alter host processes. Knowledge of how effectors target and manipulate host proteins is critical to understand crop disease. Here, we show that in planta expression of the RXLR effector Pi04314 enhances leaf colonization by Phytophthora infestans via activity in the host nucleus and attenuates induction of jasmonic and salicylic acid-responsive genes. Pi04314 interacts with three host protein phosphatase 1 catalytic (PP1c) isoforms, causing their re-localization from the nucleolus to the nucleoplasm. Re-localization of PP1c-1 also occurs during infection and is dependent on an R/KVxF motif in the effector. Silencing the PP1c isoforms or overexpression of a phosphatase-dead PP1c-1 mutant attenuates infection, demonstrating that host PP1c activity is required for disease. Moreover, expression of PP1c-1mut abolishes enhanced leaf colonization mediated by in planta Pi04314 expression. We argue that PP1c isoforms are susceptibility factors forming holoenzymes with Pi04314 to promote late blight disease.

Figure 1. The nuclear location of effector fusion GFP–Pi04314 is important for its ability to promote infection.

(a) Images are projections of confocal Z series of GFP–Pi04314 and the forms of the fusion with added targeting signals; _myrGFP–Pi04314 and _NLSGFP–Pi04314 to decrease or increase the level of the effector fusion in the nucleus, respectively. Inset images are single optical sections through the nuclei of the cells. Scale bar, 20 μm. (b) GFP–Pi04314 and _NLSGFP–Pi04314 are able to promote P. infestans growth following Agrobacterium-mediated expression compared with a GFP control. Lesion diameter was not significantly different between _myrGFP–Pi04314-expressing plants and the control. Error bars are standard error and the graph represents the combined data from three biological replicates (n=56 per construct). Letters on the graph denote statistically significant differences (ANOVA, P≤0.005).

Figure 2. Effector Pi04314 interacts with Solanum tuberosum PP1c isoforms in yeast-2-hybrid and immunoprecipitation assays.

(a) Yeast co-expressing the PP1c isoforms with Pi04314 grew on histidine (HIS) medium and yielded b-galactosidase (B-gal) activity, while those co-expressed with the control SFI3 or empty vector (EV) did not. (b) Immunoprecipitation (IP) of protein extracts from agroinfiltrated leaves using GFP–Trap confirmed that cMyc-tagged PP1c isoforms specifically associated with GFP–Pi04314 and not with the GFP–SFI3 control. Expression of constructs in the leaves is indicated by +. Protein size markers are indicated in kDa, and protein loading is indicated by Ponceau stain.

Figure 3. Co-expression of Pi04314 with the PP1c results in reduced nucleolar association of the latter.

(a) Image sets are single optical sections of typical nuclei co-expressing GFP–Pi04314 or the GFP–SFI3 control with each isoform of PP1c fused to mRFP. Scale bar, 5 μm. White arrows indicate mRFP fluorescence intensity plots shown in graphs to the right of each image. (b) Graphs of the average ratios of nucleolar to nucleoplasmic mRFP fluorescence from the PP1c isoforms alone or co-expressed with each effector, demonstrating that significant loss of mRFP–PP1c from the nucleolus occurred only in the presence of GFP–Pi04314. The averages were obtained from images of more than 50 nuclei for each condition. Error bars are s.e. and the graph represents the combined data from six biological replicates. Letters on the graphs denote statistically significant differences (ANOVA, P<0.001).

Figure 4. Nucleolar accumulation of PP1c-1 is reduced in nucleoli of the cell interacting with P. infestans haustoria.

(a) Graph showing that mean ratio of nucleolar to nucleoplasmic fluorescence of GFP–PP1c-1 is reduced in cells containing P. infestans haustoria (combined from three biological replicates with a total of 35 nuclei measured from cells expressing GFP–PP1c-1 fluorescence and showing red fluorescent haustoria from P. infestans strain 88069 tdt), compared with 22 GFP–PP1c-1-expressing nuclei in uninfected cells. Letters denote statistical significance (ANOVA, P=0.001). (b) Images are representative of the different patterns observed in GFP–PP1c-1 re-localization from the nucleolus. Nucleolar GFP–PP1c-1 fluorescence in haustoriated cells was either considerably attenuated (upper panel) or reduced (middle panel), compared with uninfected cells (lower panel). Corresponding fluorescence intensity plots (from arrowed lines indicated in each image) are shown in graphs to the right of the image. Red fluorescent haustoria are indicated with *; scale bar, 10 μm.

Figure 5. Mutation of the PP1c-binding motif in Pi04314 results in a loss of PP1c association.

(a) Yeast-2-hybrid assay following co-expression of PP1c-1 and wild-type (WT) Pi04314, which grew on −histidine (−HIS) medium and had β-galactosidase (B-gal) activity, while co-expression of PP1c-1 with Pi04314mut did not. (b) Immunoprecipitation of WT GFP–Pi04314 and GFP–Pi04314mut protein extracts from agroinfiltrated leaves using GFP–Trap confirmed that cMyc–PP1c-1 co-immunoprecipitated only with WT GFP–Pi04314. Expression of constructs in the leaves is indicated by +. Protein size markers are indicated in kDa, protein loading is indicated by Ponceau stain, and antibodies used are as indicated (αcMyc and αGFP). (c) Single optical section through co-expressing nuclei show that the mutated effector fusion GFP–Pi04314mut co-expressed with mRFP–PP1c-1 did not cause reduction of mRFP fluorescence in the nucleolus, whereas the WT GFP–Pi04314 did. Scale bar, 10 μm. White arrows indicate mRFP fluorescence intensity plots shown in graphs to the right of each image. (d) Graph shows the average ratio of nucleolar to nucleoplasmic mRFP fluorescence from the mRFP–PP1c-1 expressed alone, with the WT effector fusion GFP–Pi04314 and with the mutated effector GFP–Pi04314mut. The averages were obtained from a minimum of 30 nuclei for each sample. Error bars are s.e. and the graph represents the combined data from three biological replicates. Letters on the graph denote statistically significant differences (ANOVA, P<0.001). (e) The GFP fusion to the mutated Pi04314 is no longer able to significantly promote P. infestans infection as measured by lesion diameter. Error bars are s.e. and the graph represents the combined data from three biological replicates (n=108 per construct). Letters on the graph denote statistically significant differences (ANOVA, P≤0.022).

Figure 6. Virus-induced gene silencing of the NbPP1c isoforms reduces P. infestans leaf colonization.

(a) Graph showing the reduction in percentages of inoculations resulting in P. infestans lesions in plants expressing TRV–PP1c3′ or TRV–PP1c5′, compared with a TRV–GFP control. Error bars are s.e. and the graph represents the combined data from six biological replicates (n=50 per construct). Letters on the graph denote statistically significant differences (ANOVA, P≤0.009). (b) Graph showing the reduction in the average numbers of sporangia per ml recovered from infected leaves of plants expressing TRV–PP1c3′ or TRV–PP1c5′, compared with the TRV–GFP control plants. Error bars are s.e. and the graph represents the combined data from four biological replicates (n=32 per construct). Letters on the graph denote statistically significant differences (ANOVA, P<0.001). (c) Example leaves showing P. infestans lesion development on control and TRV–PP1c3′ or TRV–PP1c5′ plants.

Figure 7. Pi04314 does not inhibit PP1c-1 phosphatase activity.

(a) Graph of phosphatase activity measured directly on GFP–Trap_M beads following immunoprecipitation of GFP co-expressed with cMyc–Pi04314; or GFP–PP1c-1 co-expressed with either cMyc vector or cMyc–Pi04314. Error bars are s.e. and the graph represents combined data from three biological replicates. Letters on the graph denote statistically significant differences (ANOVA, P<0.001). (b) Immunoprecipitation of protein extracts from agroinfiltrated leaves using GFP–Trap confirmed that cMyc–Pi04314 is co-immunoprecipitated with GFP–PP1c-1. Expression of constructs in the leaves is indicated by +. Protein size markers are indicated in kDa, and protein loading is indicated by Ponceau stain. Antibodies used are indicated (αcMyc and αGFP). (c) Graph of phosphatase activity measured directly on GFP–Trap_M beads following immunoprecipitation of GFP, GFP–Pi04314 or GFP–Pi04314mut co-expressed with cMyc–PP1c-1. Error bars are s.e. and the graph represents the combined data from three biological replicates. Letters on the graph denote statistically significant differences (ANOVA, P<0.001). (d) Immunoprecipitation of protein extracts from agroinfiltrated leaves using GFP–Trap confirmed the cMyc–PP1c-1 co-immunoprecipitated with the GFP–Pi04314 effector, but not with GFP–Pi04314mut. Expression of constructs in the leaves is indicated by +. Protein size markers are indicated in kDa, and protein loading is indicated by Ponceau stain. Antibodies used are as indicated (αcMyc and αGFP).

Figure 8. A phosphatase-dead mutant of PP1c-1 reduced P. infestans leaf colonization, and the enhanced colonization promoted by expression of the effector.

(a) Immunoprecipitation of protein extracts from agroinfiltrated leaves using GFP–Trap. Expression of constructs in the leaves is indicated by +. Protein size markers are indicated in kDa, and protein loading is indicated by Ponceau stain. Antibodies used are as indicated (αcMyc and αGFP). (b) Graph of the phosphatase activity of the wild-type GFP–PP1c-1 and mutated GFP–PP1c–1mut showing that the mutant showed similar background activity to the GFP empty vector control. Error bars are s.e. and the graph represents the combined data from three biological replicates. Letters on the graph denote statistically significant differences (ANOVA, P<0.001). (c) Graph of the average ratio of nucleolar to nucleoplasmic GFP fluorescence from the GFP–Pi04314 fusion protein expressed with mRFP–PP1c–1mut or, as a control, free mRFP. Error bars are s.e. and the graph represents the combined data from three biological replicates (n=40 per construct). Letters on the graph denote statistically significant differences (ANOVA, P<0.001). (d) Single optical section through co-expressing nuclei show that the effector fusion GFP–Pi04314 was reduced in the nucleolus when co-expressed with mRFP–PP1c–1mut, but not with free mRFP. Scale bar, 10 μm. White arrows indicate mRFP fluorescence intensity plots shown in graphs to the right of each image. (e) Co-expression of the cMyc–PP1c-1 wild-type did not significantly change the enhancement of P. infestans growth caused by expression of the GFP–Pi04314 effector compared with the control. However, co-expression of the mutant cMyc–PP1c–1mut with GFP–Pi04314 caused a significant reduction in effector-induced enhancement of colonization. Error bars are s.e. and the graph represents the combined data from three biological replicates (n=102 per construct). Letters on the graph denote statistically significant differences (ANOVA, P<0.001). (f) Expression of wild-type cMyc–PP1c-1 did not significantly alter P. infestans infection compared with the control empty (cMyc) vector (EV), but expression of the phosphatase-dead mutant cMyc–PP1c–1mut significantly reduced lesion growth. Error bars are s.e. and the graph represents the combined data from three biological replicates (n=73 per construct). Letters on the graph denote statistically significant differences (ANOVA, P<0.001).

Figure 9. Transgenic potato expressing Pi04314 enhances P. infestans leaf colonization and attenuates induction of JA- and SA-responsive genes.

(a) Representative leaf images, exposed to UV, showing increased lesion sizes on Pi04314-expressing transgenic lines E-6 and OE-8, compared with the untransformed control, cv E3. (b) Mean lesion diameter is significantly increased in OE-6 and OE-8 lines compared with the control, cv E3. Letters denote statistical significance (ANOVA, P<0.001) from three biological replicates, each containing inoculation of three leaves from each of the six plants. (c) Relative expression of flg22 marker genes StWRKY8 and StACRE31 30 min after treatment with flg22 in E3, OE-6 and OE-8 lines, compared with untreated lines (which was given a value of 1). (d) Relative expression of JA-responsive genes StJAZ1-like (StJAZ1L) and StMYC2L 1 h after treatment with meJA in E3, OE-6 and OE-8 lines, compared with untreated lines (which was given a value of 1). (e) Relative expression of SA-responsive genes StWRKY40-like (StWRKY40L) and StWRKY16L 1 h after treatment with SA in E3, OE-6 and OE-8 lines, compared with untreated lines (which was given a value of 1). Results in B-E are the mean of three independent biological replicates. Error bars show s.e. and * denotes significantly reduced induction (ANOVA, P<0.001) of responsive genes in transgenic lines compared with the E3 control.