Heat-induced SUMOylation differentially affects bacterial effectors in plant cells

Abstract

Bacterial pathogens deliver effectors into host cells to suppress immunity. How host cells target these effectors is critical in pathogen-host interactions. SUMOylation, an important type of posttranslational modification in eukaryotic cells, plays a critical role in immunity, but its effect on bacterial effectors remains unclear in plant cells. In this study, using bioinformatic and biochemical approaches, we found that at least 16 effectors from the bacterial pathogen Pseudomonas syringae pv. tomato DC3000 are SUMOylated by the enzyme cascade from Arabidopsis thaliana. Mutation of SUMOylation sites on the effector HopB1 enhances its function in the induction of plant cell death via stability attenuation of a plant receptor kinase BRASSINOSTEROID INSENSITIVE 1 (BRI1)-ASSOCIATED RECEPTOR KINASE 1. By contrast, SUMOylation is essential for the function of another effector, HopG1, in the inhibition of mitochondria activity and jasmonic acid signaling. SUMOylation of both HopB1 and HopG1 is increased by heat treatment, and this modification modulates the functions of these 2 effectors in different ways in the regulation of plant survival rates, gene expression, and bacterial infection under high temperatures. Therefore, the current work on the SUMOylation of effectors in plant cells improves our understanding of the function of dynamic protein modifications in plant-pathogen interactions in response to environmental conditions.

Figure 1.

Multiple effectors of P. syringae are SUMOylation substrates. A) Schematic design of screening of SUMOylated P. syringae effectors. Thirty-six effectors were predicted using the GPS-SUMO software and 30 of them harbor potential SUMOylation sites; 17 effectors with predominant consensus and SUMOylation sites were selected for in vitro detection. B) Summary of effector SUMOylation in an E. coli reconstitution system. In 17 selected effectors, 9 of them were SUMOylated at 25 °C, and 16 of them were SUMOylated at 37 °C. C) The immunoblots for in vitro SUMOylation of 16 effectors at 25 °C and 37 °C. Full-length effectors were used in the assay except AvrE1. Because AvrE1 is an effector with a high molecular weight and was difficult to express, its C-terminal region (822 to 1,795 aa) with a predicted SUMOylation site was used in the detection. The FLAG-tagged effectors were coexpressed in a reconstitution system in E. coli cells expressing Arabidopsis SUMO E1, SUMO1GG, with or without SUMO E2. The E. coli cells were incubated at 25 °C or 37 °C for 16 h, and total proteins were subjected to SDS–PAGE and immunoblots using an anti-FLAG antibody. The unmodified proteins are indicated with black triangles and the SUMOylated proteins are indicated by asterisks. The data are representatives of 3 biologically independent experiments (independent plasmid transformation, protein expression, and immunoblotting).

Figure 2.

Suppression of HopB1 SUMOylation enhances cell death. A) Prediction of potential SUMOylation sites of HopB1 by the GPS-SUMO software. The positions of 2 predicted lysine residues K66 and K73 on HopB1 are indicated. B) Identification of SUMOylation sites on HopB1 in the E. coli reconstitution system at 37 °C. SUMOylation of the WT and mutant (K66R, K73R, and K66/73R) versions of HopB1-FLAG was detected via immunoblots using an anti-FLAG antibody. The representative immunoblot from 3 biologically independent experiments (independent plasmid transformation, protein expression, and immunoblotting) is shown. The unmodified forms are indicated with triangles, and the SUMOylated forms are indicated with asterisks. C) Detection of HopB1 SUMOylation in plant cells. The WT or 2KR (K66/73R) version of HopB1-GFP (free GFP was used in the control sample) was coexpressed with Myc-SUMO1GG in Arabidopsis protoplasts. After incubation at 22 °C or 32 °C for 16 h, the cells were collected for IP using anti-GFP beads. The proteins in input and IP samples were detected in immunoblots with anti-GFP and anti-Myc antibodies. The representative immunoblots from 3 biologically independent experiments (independent plant growth, protoplast transformation, and IP assay) are shown. D) The effect of 2KR mutation on the phenotypes of HopB1-overexpressing plants. The representative developmental phenotypes of 3-wk-old plants, namely WT (untransformed), Vector (UBQ:GFP), and 2 independent lines of UBQ:HopB1(WT)-GFP and UBQ:HopB1(2KR)-GFP, from 3 biologically independent experiments (independent plant growth and phenotype observation), are shown. E) The effect of 2KR mutation of HopB1 on Pst DC3000 infection. The 5 × 10⁸ CFU/mL suspensions of indicated Pst DC3000 strains were sprayed onto 3-wk-old Arabidopsis at 22 °C. Leaf discs from 0 or 3 d after inoculation were crushed in water, and colony numbers were counted. The data are means ± Sd from 4 replicates in an experiment; 3 biologically independent experiments showed similar patterns. F) The Venn diagram for the numbers of upregulated genes in the WT and 2KR versions of HopB1-overexpressing plants. G) GO enrichment analysis for genes affected by HopB1 SUMOylation. The biological process enrichment for the genes (1,447) specifically upregulated in the HopB1(2KR)-overexpressing plants is shown. P < 0.05. H) The effect of SUMOylation on HopB1 function in the expression control of plant immunity genes. The relative transcript levels of indicated genes in the WT, Vector, 2 independent HopB1(WT)- and HopB1(2KR)-overexpressing plants were detected using RT-qPCR. ACTIN2 was used as a reference gene. The relative expression level in WT was set to 1. The data are means ± Sd from 3 repeats in an experiment; the patterns were consistent in 3 biologically independent experiments (independent plant growth, RNA preparation, and RT-qPCR). I) The effect of SUMOylation on HopB1 function in regulating the protein stability of BAK1. Proteins of 3-wk-old plants of indicated genotypes were extracted in a buffer with phosphatase inhibitors. The immunoblots with anti-BAK1 or anti-GFP antibodies from 3 biologically independent experiments (independent plant growth, protein extraction, and immunoblotting) are shown. Rubisco levels in Coomassie blue staining are shown as loading controls. J) The quantitative data of transcript and protein levels of BAK1 in the indicated plants. The transcript levels of BAK1 in 3-wk-old plants were measured using RT-qPCR. To detect the protein levels of BAK1, the immunoblot and Rubisco signals were quantified by ImageJ, and the relative protein levels were calculated from signal ratios (BAK1/Rubisco). The relative interaction intensity in the WT sample was set to 1. The data are means ± Sd from 3 biologically independent experiments (independent plant growth and RNA/protein analysis). Significance analysis in E), H), and J) was performed with 1-way ANOVA followed by Tukey's multiple comparison tests (P < 0.05).

Figure 3.

SUMOylation is essential for the function of HopG1 in plant cells. A) Prediction of SUMOylation sites of HopG1 by the GPS-SUMO software. The positions of 3 predicted lysine residues K24, K180, and K221 on HopG1 are indicated. B) Identification of SUMOylation sites on HopG1 in the E. coli reconstitution system at 37 °C. SUMOylation of the WT and double mutant (K24/180R, K180/221R, and K24/221R) versions of HopG1-FLAG was detected in immunoblots with an anti-FLAG antibody. The data of single-site mutants are included in Supplementary Fig. S6. The representative immunoblot from 3 biologically independent experiments (independent plasmid transformation, protein expression, and immunoblotting) is shown. The SUMOylated forms are labeled with asterisks, and unmodified forms are labeled with triangles. C) Detection of HopG1 SUMOylation in plant cells with different temperatures. The WT or 2KR (K24/221R) version of HopG1-GFP (free GFP was used in the control sample) was coexpressed with Myc-SUMO1GG in protoplasts. After the incubation at 22 °C or 32 °C for 16 h, the protoplasts were harvested for IP with anti-GFP beads. The proteins in IP and input samples were detected in immunoblots with anti-Myc and anti-GFP antibodies. The representative immunoblots from 3 biologically independent experiments (independent plant growth, protoplast transformation, and IP assay) are shown. D) The effect of SUMOylation on the function of HopG1 in the regulation of plant development. The representative phenotypes of 10-d-old seedlings on the medium with 10 μm estradiol (EST, for HopG1 overexpression induction) or DMSO (control) from 3 biologically independent experiments (independent plant growth and phenotype observation) are shown. Bars: 1 cm. E) The function of HopG1 SUMOylation in Pst DC3000 infection. The 5 × 10⁸ CFU/mL suspensions of indicated Pst DC3000 strains were sprayed onto 3-wk-old plants at 22 °C. The plant phenotypes were recorded 3 d after inoculation, and the representative image of 3 biologically independent experiments (independent plant growth and bacteria inoculation) is shown. F) The Venn diagram for the numbers of upregulated genes in plants overexpressing the WT and 2KR versions of HopG1. The 3-wk-old seedlings on the medium with 10 μm estradiol were used for RNA-seq. G) GO analysis of genes affected by HopG1 SUMOylation. The biological process enrichment of the genes (978) specifically upregulated in the HopG1(WT)-induced overexpressing plants is shown. P < 0.05. H) The effect of SUMOylation on HopG1 functions in the expression control of JA signaling genes. The relative transcript levels of indicated genes in the WT, Vector, and 2 independent lines of HopG1(WT) and HopG1(2KR) estradiol-induced overexpressing plants were measured using RT-qPCR. ACTIN2 was used as an internal control. The relative expression level in WT was set to 1. The data are means ± Sd from 3 repeats of an experiment; the results from 3 biologically independent experiments (independent plant growth, RNA preparation, and RT-qPCR) showed similar patterns. I) Regulation of hydrogen peroxide (H₂O₂) contents mediated by HopG1 SUMOylation. The indicated 10-d-old seedlings grown on the medium with 10 μm estradiol were prepared for H₂O₂ detection. The data are means ± Sd from 3 biologically independent experiments (independent plant growth and detection). J) Regulation of active mitochondria numbers by HopG1 SUMOylation. The indicated 10-d-old seedlings grown on the medium with 10 μm estradiol were used for Janus green B staining, and the numbers of stained mitochondria in 6 cells in similar root meristem regions are shown. The data are means ± Sd from 20 roots in an experiment; 3 biologically independent experiments (independent plant growth and detection) showed similar patterns. Significance analysis in H), I), and J) was performed with 1-way ANOVA followed by Tukey's multiple comparison tests (P < 0.05).

Figure 4.

SUMOylation differently modulates HopB1 and HopG1 at high temperatures. A, B) The effect of HopB1 SUMOylation on plant survival rates under heat treatments. Ten-day-old seedlings grown on the regular medium were incubated at 37 °C for 5 d and then were moved to 22 °C for 3 d. The representative phenotypes from 3 biologically independent experiments are shown in A). At least 15 seedlings for each genotype were used in a replicated plate. The survival rates in B) are means ± Sd from 3 biologically independent experiments (independent plant growth and independent heat treatment). C, D) The effect of HopG1 SUMOylation on plant survival rates under heat treatments. Ten-day-old seedlings on the medium with 10 μm estradiol were incubated at 37 °C for 3 d and then were moved to 22 °C for 3 d. The representative phenotypes from 3 biologically independent experiments are shown in C). At least 15 seedlings for each genotype were used in a replicated plate. The survival rates in D) are means ± Sd from 3 biologically independent experiments (independent plant growth and independent heat treatment). E, F) Regulation of immunity gene expression mediated by SUMOylation of HopB1 or HopG1 under high temperatures. The indicated 10-d-old seedlings were treated at 37 °C for 6 h before RNA extraction for RT-qPCR analysis of PR1E). The indicated 10-d-old seedlings grown on the medium with 10 μm estradiol were treated at 37 °C for 6 h before RNA extraction for RT-qPCR analysis of JAZ1F). CK, control condition; HT, heat treatment. ACTIN2 was used as an internal control. The relative expression level in WT was set to 1. The data are means ± Sd from 3 repeats of an experiment; the results from 3 biologically independent experiments (independent plant growth/treatment, independent RNA preparation, and independent qRT-PCR) showed similar patterns. G, H) The effect of SUMOylation of HopB1 or HopG1 on Pst DC3000 infection under high temperatures. The Pst DC3000 strains with indicated genotypes were used for inoculation on 3-wk-old Arabidopsis plants in a spray assay at 22 °C. After 1-d incubation at 22 °C, the inoculated plants were moved to 22 °C or 28 °C for 2 d. Leaf discs were crushed in water, and colony numbers were counted. The data are means ± Sd from 4 replicates in an experiment; 3 biologically independent experiments (independent plant growth, independent bacteria inoculation, and independent temperature treatment) showed similar patterns. I) The proposed model for the function of heat-induced SUMOylation of HopB1 and HopG1 in plant cells during Pst DC3000 infection. HT, high temperature; S, SUMO; dots in the right graph indicate mitochondria (blue for active and gray for inactive). Significance analysis in B), D), E), F), G), and H) was performed with 1-way ANOVA followed by Tukey's multiple comparison tests (P < 0.05).