HopW1 from Pseudomonas syringae disrupts the actin cytoskeleton to promote virulence in Arabidopsis

Abstract

A central mechanism of virulence of extracellular bacterial pathogens is the injection into host cells of effector proteins that modify host cellular functions. HopW1 is an effector injected by the type III secretion system that increases the growth of the plant pathogen Pseudomonas syringae on the Columbia accession of Arabidopsis. When delivered by P. syringae into plant cells, HopW1 causes a reduction in the filamentous actin (F-actin) network and the inhibition of endocytosis, a known actin-dependent process. When directly produced in plants, HopW1 forms complexes with actin, disrupts the actin cytoskeleton and inhibits endocytosis as well as the trafficking of certain proteins to vacuoles. The C-terminal region of HopW1 can reduce the length of actin filaments and therefore solubilize F-actin in vitro. Thus, HopW1 acts by disrupting the actin cytoskeleton and the cell biological processes that depend on actin, which in turn are needed for restricting P. syringae growth in Arabidopsis.

Figure 1. HopW1 forms complexes with actin in plants.

Actin was detected by immunoblotting after immunoprecipitation (IP) of HopW1-HA complexes. (A) HopW1-HA-actin complexes in N. benthamiana transiently transformed with HopW1-HA (W1) using Agrobacteria. V is vector control. (B) HopW1-actin complexes in dexamethasone (dex)-treated Arabidopsis stable transgenics that carry dex:HopW1-HA. Input was 2% of extract used for each IP. These experiments were each repeated twice with similar results.

Figure 2. HopW1 disrupts the actin cytoskeleton during infection and promotes PtoDC3000 growth on Arabidopsis.

(A) Actin cytoskeleton changes in cotyledons after infection with PtoDC3000/HopW1. Col/Lifeact-GFP seedlings grown on MS plates were infected with PtoDC3000/empty vector or PtoDC3000/HopW1 at OD₆₀₀ = 0.01 or treated with 100 µM LatB (actin cytoskeleton-disrupting control). Cotyledons were imaged by laser-scanning confocal microscopy at 6, 24 and 48 h after infection/treatment. Representative micrographs shown are Z-series maximum intensity projections. Bar = 18 µm (B) Percent occupancy (density) of actin filament signals shown in (A) was measured using ImageJ software as described in , in 18 samples per treatment/time-point, from three biological repeats from picture regions without stomata. (C) Growth of PtoDC3000/vector or PtoDC3000/HopW1 (OD₆₀₀ = 0.0001) three days post-inoculation of 3-week old soil grown Arabidopsis Col leaves was monitored in the presence or absence of LatB. These experiments were performed three times with similar results. In B and C, average of all results is shown, different letters indicate significant difference (ANOVA/Tukey's test, P<0.05, n = 18 in B and n = 9 in C) and bars represent SEM.

Figure 3. HopW1 disrupts the actin cytoskeleton when expressed in plant cells.

Localization and effect of HopW1 on actin cytoskeleton was monitored in transiently transformed cells using laser scanning confocal microscopy. Representative micrographs shown are Z-series maximum intensity projections. (A) Expression of HopW1-RFP and Lifeact-GFP F-actin marker in Nicotiana benthamiana leaves 36–40 h after co-transformation with Agrobacteria. Micrographs show localization of cytoplasmic mCherry (control, upper panel) and HopW1-RFP (lower panel) together with Lifeact-GFP. GFP/RFP fluorescence is shown in green/red, respectively. Dotted line shows cells expressing HopW1-RFP. Bar = 30 µm. (B) Arabidopsis Col protoplasts from a transgenic line that expresses Lifeact-GFP were transfected with HopW1-CFP (lower panel) or without DNA (control, upper panel). Micrographs show Lifeact-GFP and HopW1-CFP 15 h after transfection. GFP/CFP fluorescence is shown in green/magenta, respectively, and chloroplast (chl) autofluorescence in blue. These experiments were repeated twice with similar results. The actin cytoskeleton was not detectable in all cells in which HopW1-RFP/CFP was observed (at least 30 N. benthamiana cells and 40 Lifeact-GFP Arabidopsis protoplasts, respectively, with HopW1 signal were observed).

Figure 4. HopW1-C disrupts F-actin in vitro.

(A) HopW1-C (HopW1^407–774) causes a reduction in the size of actin filaments in vitro as assayed by sedimentation ultracentrifugation. Non-muscle F-actin (10 µM) was incubated with HopW1-C (W1-C, asterisks) for 30 min, partitioned into supernatant (S) and pellet (P) fractions by ultracentrifugation, separated by SDS-PAGE and stained with Coomassie blue. Greater than 90% of actin was found in the pellet in the absence of HopW1-C. In contrast, after incubation with 0.5 µM HopW1-C, >70% of actin was in the supernatant (note concomitant decrease of actin in the pellet). (B) F-actin disruption by HopW1-C is dose dependent. Preassembled F-actin (10 µM) was disrupted in the presence of different amounts of HopW1-C, but not with the controls: phosphate buffer (PB), BSA, and E. coli BL21 extract (EXT) after 30 min incubation. (C) Visualization of F-actin disruption. Mock (50 mM phosphate buffer) (i), BSA (0.5 µM) (ii), and different amounts of HopW1-C (iii to vii) were incubated with 5 µM F-actin for 1 h. Filaments were stained with TRITC-phalloidin and observed by epifluorescence microscopy. (D) Quantitation of the reduced F-actin lengths. Actin filaments (≥100) from (C) were measured in each treatment. The distribution of lengths of F-actin was different for 0.03 µM to 0.5 µM HopW1-C (W1-C) treatments compared with buffer control (mock), as determined by χ ² tests (P<0.0001). These experiments were repeated three or more times with similar results.

Figure 5. Inhibition of the vacuolar and ER trafficking by HopW1.

(A) Wild-type and dex:hopW1 Arabidopsis Col protoplasts were transfected with AALP:GFP or SPO:GFP and incubated with 0.2 µM dex for the indicated times and imaged using fluorescence microscopy. AALP:GFP was targeted to the central vacuole and SPO:GFP localized to the ER and vacuole in wild-type protoplasts, as previously documented ; upper rows in each panel show examples of vacuole and ER localization, respectively. However, in the presence of HopW1, many protoplasts transfected with the AALP:GFP and SPO:GFP showed notable punctate patterns. (B) Localization patterns were quantified from at least 100 images, such as those in (A); see also Figure S2A for comparison with the effect of LatB. Bars indicate SEM. χ ² tests indicated that the distributions were significantly different between the wild-type and dex:hopW1 at each time point (P<0.0001, n≥100 per genotype/fusion construct). This experiment was performed three times using at least two transgenic dex:HopW1 lines, with similar results.

Figure 6. Endocytosis inhibition by HopW1 in Arabidopsis.

(A) Representative microscopic images that show the effects of HopW1 and LatB on endocytic vesicle formation. Wild-type and dex:hopW1 Arabidopsis Col protoplasts were treated with 0.2 µM dexamethasone, stained with FM4-64 and visualized by fluorescence microscopy. 10 µM LatB was used to disrupt the actin cytoskeleton. After over-night incubation with dexamethasone and/or LatB, protoplasts were labeled with FM4-64 and viewed after 0.5, 1 and 2 h. Arrows point to some of the FM4-64-stained endosomes. Protoplasts from two independent transgenic dex:HopW1 lines were used with similar results. (B) Endosomes were quantified in protoplasts from the indicated plants that were treated and stained as in (A). At least 20 protoplasts per treatment, per time-point, from three biological repeats (independent experiments) were analyzed. Bars indicate SEM. Different letters indicate significantly different numbers of endosomes (P<0.0001) between wild type versus dex:hopW1 or LatB-treated wild type, determined using ANOVA/Tukey's test.

Figure 7. PtoDC3000/HopW1 infection inhibits endocytosis.

(A) Examples of microscopic images of infected tissue in which endosomes are visualized using FM4-64. Cotyledons of Arabidopsis Col seedlings grown on MS plates were infected with PtoDC3000 carrying either empty vector (pME6012) or vector with the HopW1 gene at OD₆₀₀ = 0.01. 100 µM LatB was used as an actin cytoskeleton-disrupting control. After infections and treatments for the indicated times, cotyledons were labeled for 1 h with FM4-64 and viewed. Arrows indicate some of the FM4-64-labeled endosomes. (B) Quantitation of the data in (A). Endosomes per cell were manually counted in at least 10 images per treatment, per time-point, from two or three biological repeats. Bars indicate SEM. Different letters indicate significantly different numbers of endosomes for given treatments, as determined by ANOVA/Tukey's test (P<0.05).

Figure 8. A schematic diagram of HopW1's effects on plant cells.

Model showing that the P. syringae effector HopW1, secreted through the type three secretion system (TTSS) disrupts actin-dependent events that may be important for defense. Disruption of actin-dependent endocytosis may reduce active signaling from endosomes. It is also possible that disruption of actin-dependent vacuolar trafficking suppresses defenses by preventing efficient delivery of defense proteins to the vacuole or that some other actin-dependent defense process is disrupted. Proteins and possibly other molecules that are targeted to vacuoles have previously been implicated in various defense-related responses to infection , .