Organelle-targeted biosensors reveal distinct oxidative events during pattern-triggered immune responses

Abstract

Reactive oxygen species are produced in response to pathogens and pathogen-associated molecular patterns, as exemplified by the rapid extracellular oxidative burst dependent on the NADPH oxidase isoform RESPIRATORY BURST OXIDASE HOMOLOG D (RBOHD) in Arabidopsis (Arabidopsis thaliana). We used the H2O2 biosensor roGFP2-Orp1 and the glutathione redox state biosensor GRX1-roGFP2 targeted to various organelles to reveal unsuspected oxidative events during the pattern-triggered immune response to flagellin (flg22) and after inoculation with Pseudomonas syringae. roGFP2-Orp1 was oxidized in a biphasic manner 1 and 6 h after treatment, with a more intense and faster response in the cytosol compared to chloroplasts, mitochondria, and peroxisomes. Peroxisomal and cytosolic GRX1-roGFP2 were also oxidized in a biphasic manner. Interestingly, our results suggested that bacterial effectors partially suppress the second phase of roGFP2-Orp1 oxidation in the cytosol. Pharmacological and genetic analyses indicated that the pathogen-associated molecular pattern-induced cytosolic oxidation required the BRI1-ASSOCIATED RECEPTOR KINASE (BAK1) and BOTRYTIS-INDUCED KINASE 1 (BIK1) signaling components involved in the immune response but was largely independent of NADPH oxidases RBOHD and RESPIRATORY BURST OXIDASE HOMOLOG F (RBOHF) and apoplastic peroxidases peroxidase 33 (PRX33) and peroxidase 34 (PRX34). The initial apoplastic oxidative burst measured with luminol was followed by a second oxidation burst, both of which preceded the two waves of cytosolic oxidation. In contrast to the cytosolic oxidation, these bursts were RBOHD-dependent. Our results reveal complex oxidative sources and dynamics during the pattern-triggered immune response, including that cytosolic oxidation is largely independent of the preceding extracellular oxidation events.

Figure 1

In vivo characterization of roGFP2-Orp1 targeted to different organelles. A, Subcellular localization of roGFP2-Orp1 targeted to the cytosol (Cyt), nuclei (Nuc), chloroplasts (Chl), mitochondria (Mit), peroxisomes (Per), and apoplast (Apo) in guard cells. Representative images of the fluorescence emission through a long-pass filter with a cutoff wavelength at 515 nm after excitation at 470 ± 20 nm. roGFP2-Orp1 and chloroplast fluorescence are depicted in green and magenta, respectively. Scale bars represent 10 µm. B, Initial roGFP2-Orp1 oxidation state in cytosol, nuclei, chloroplasts, mitochondria, peroxisomes, and apoplast. Horizontal lines on violin plots show the median and quartile values. Data are means of at least two independent experiments (n ≥ 40). Different letters indicate significant differences at P < 0.001 based on Tukey's HSD test. C, Initial GRX1-roGFP2 oxidation state in cytosol and nuclei, chloroplasts, mitochondria, and peroxisomes. The oxidation state of roGFP2-Orp1 and GRX1-roGFP2 (ratio 400/485 nm) in untreated condition was measured by multiwell fluorimetry (excitation at 400 ± 8 and 485 ± 8 nm; emission, 525 ± 20 nm) on leaf discs from rosette leaves of 5-week-old plants. Data are means of at least two independent experiments (n ≥ 40). Different letters indicate significant differences at P < 0.001 based on Tukey's HSD test. Horizontal lines on violin plots show the median and quartile values. D–I, In vivo characterization of organelle-targeted roGFP2-Orp1 oxidation and reduction kinetics in response to H₂O₂ and DTT. Leaf discs of plant expressing roGFP2-Orp1 targeted to the cytosol (D), nuclei (E), mitochondria (F), chloroplasts (G), peroxisomes (H), and apoplast (I) were exposed at t = 0 min to control solution, 100 mM H₂O₂ or 50 mM DTT. The 400/485 nm fluorescence ratio (R) was measured over time by multiwell fluorimetry (excitation at 400 ± 8 and 485 ± 8 nm; emission, 525 ± 20 nm) and expressed relative to the mean initial ratio (R_i) before treatment (R/R_i). Dynamic range (δ) of probes in each subcellular compartment was calculated from the 400/485 nm excitation ratios for the oxidized and reduced probe. Data are means ± Se from a representative experiment (n ≥ 4). Two-way ANOVA using repeated measures for time and Tukey's multiple comparisons analyses show H₂O₂ and DTT significantly (P < 0.01) affect R/R_i in all compartments except H₂O₂ in the apoplast (Supplemental Table S2).

Figure 2

PAMP-induced roGFP2-Orp1 and GRX1-roGFP2 redox dynamics in different subcellular compartments. Dose–response kinetics of cytosolic/nuclear roGFP2-Orp1 (A) and GRX1-roGFP2 (B) oxidation in leaves in response to the PAMP flg22. Leaf discs were exposed at t = 0 min to control solution or different concentrations of flg22. The 400/485 nm fluorescence ratio (R) was measured over time by multiwell fluorimetry and expressed relative to the mean initial ratio (R_i) before treatment. Data are means ± Se from a representative experiment (n = 6). The experiments have been repeated at least twice with similar results. Kinetics of roGFP2-Orp1 oxidation in the cytosol (C), nuclei (D), mitochondria (E), chloroplasts (F), and peroxisomes (G) and GRX1-roGFP2 oxidation in the cytosol and nuclei (H), mitochondria (I), chloroplasts (J), and peroxisomes (K) in response to PAMPs. Leaf discs (C–K) were exposed at t = 0 min to control solution, 1 µM elf18, or 1 µM flg22. The 400/485 nm fluorescence ratio (R) was measured over time by multiwell fluorimetry and expressed relative to the mean initial ratio (R_i) before treatment. Data are means ± Se from two independent experiments (n ≥ 6, C–F, H, I, K) or a representative experiment (n ≥ 5, G and J). Two-way ANOVA using repeated measures for time and Tukey's multiple comparisons analyses are shown in Supplemental Table S2. All treatments were significantly different from control (P < 0.05) except: B (0.1 µM flg22); F, G (flg22); I, The Rred/Ri and Rox/Ri fluorescence ratios for fully reduced/fully oxidized probes in each compartment (Figure 1) were: cytosol 0.89/5.05; nuclei 0.89/4.76; mitochondria 0.61/3.35; chloroplasts 0.53/2.88, and peroxisomes 0.43/1.79.

Figure 3

Initial oxidation state and H₂O₂-induced oxidation of roGFP2-Orp1 in mutants of PTI regulators, NADPH oxidases, and apoplastic peroxidases. Initial cytosolic/nuclear roGFP2-Orp1 oxidation state in rbohD and rbohF (A), bak1-5 and bik1 (B), prx33-3 and prx34-2 (C), prx4-2 and prx71-1 (D) and vtc2-4 (E) mutants. The oxidation state of cytosolic/nuclear roGFP2-Orp1 (ratio 400/485 nm) in untreated conditions was measured by multiwell fluorimetry on leaf discs. Data are means ± Se (Figures H–J) or violin plots with horizontal lines showing the median and quartile values (Figures C–E) from at least three independent experiments (n ≥ 40). Different letters indicate significant differences at P < 0.001 (A, B, and E), P < 0.05 (C), and P < 0.01 (D) based on Tukey's HSD test. Kinetics of cytosolic/nuclear roGFP2-Orp1 oxidation in leaves of rbohD and rbohF (F), bak1-5 and bik1 (G), prx33-3 and prx34-2 (H), prx4-2 and prx71-1 (I), and vtc2-4 (J) mutants in response to exogenous H₂O₂. Leaf discs were exposed at t = 0 min to control solution or 1 mM H₂O₂, the 400/485 nm fluorescence ratio (R) was measured over time by multiwell fluorimetry and expressed relative to the mean initial ratio (R_i) before treatment. Data are means ± Se from three independent experiments (n ≥ 15, F–I) or a representative experiment (n ≥ 6, J). Two-way ANOVA using repeated measures for time and Tukey's multiple comparisons analyses are shown in Supplemental Table S2. H₂O₂ significantly (P < 0.05) increased probe oxidation in all cases. rbohF was significantly (P < 0.05) more oxidized than Col-0. The Rred/Ri and Rox/Ri fluorescence ratios for fully reduced/fully oxidized probes in the cytosol/nuclei (Supplemental Figure S2) were 0.85/5.70.

Figure 4

flg22-induced intracellular roGFP2-Orp1 redox dynamics in mutants affecting PTI-mediated ROS production in the apoplast. Kinetics of cytosolic/nuclear roGFP2-Orp1 oxidation in leaves of bak1-5 and bik1 (A), rbohD and rbohF (B), prx33-3 and prx34-2 (C), prx4-2 and prx71-1 (D) and vtc2-4 (E), mutants in response to the PAMP flg22. Leaf discs were exposed at t = 0 min to control solution or 1 µM flg22. The ratio 400/485 nm (R) was measured over time by multiwell fluorimetry and expressed relative to the mean initial ratio (R_i) before treatment (R/R_i). Data are means ± Se from three independent experiments (n ≥ 15, B–D). In (A and E), a representative experiment is shown (n = 6). F, Effect of DPI on flg22-induced oxidation of roGFP2-Orp1. After 2 h of pretreatment with control solution or 20 µM DPI, leaf discs from Col-0 were exposed at t = 0 min to control solution or 1 µM flg22. The 400/485 nm fluorescence ratio (R) was measured over time by multiwell fluorimetry and expressed relative to the mean initial ratio (R_i) before flg22 treatment. Data are means ± Se of three independent experiments (n ≥ 10). G, PAMP-induced apoplastic ROS production detected by luminol assay in Col-0 WT, rbohD, and rbohF mutants. The luminescence was measured over time after treatment with control solution or 1 µM flg22 at t = 0 min. Data are means ± Se (n = 6) from a representative experiment. In (A–G), two-way ANOVA using repeated measures for time and Tukey's multiple comparisons analyses are shown in Supplemental Table S2. flg22 significantly (P < 0.05) increased probe oxidation in all mutants except bak1 (A). flg22-treated bik1 and vtc2-4 flg22 were significantly different from flg22-treated Col-0 (A and E). flg22-induced oxidation was significantly decreased by DPI (F). The 400/485 nm fluorescence ratios for fully reduced/fully oxidized probes in the cytosol/nuclei (Figure 1) were 0.85/5.70.

Figure 5

Pseudomonas syringae bacteria and flg22 induce a biphasic roGFP2-Orp1 and GRX1-roGFP2 oxidation in the cytosol. Oxidation kinetics of roGFP2-Orp1 targeted to the cytosol/nuclei (A), mitochondria (B), chloroplasts (C), and peroxisomes (D), and GRX1-roGFP2 targeted to the cytosol/nuclei (E), mitochondria (F), chloroplasts (G), and peroxisomes (H) in response to WT Pst DC3000 and disarmed Pst hrpA bacteria. Leaf discs were exposed at t = 0 min to mock (10 mM MgCl₂), 10⁸ cfu mL⁻¹Pst DC3000 (Pst), or 10⁸ cfu mL⁻¹Pst hrpA bacteria. Long-term oxidation kinetics of roGFP2-Orp1 targeted to the cytosol (I), mitochondria (J), chloroplasts (K), and peroxisomes (L) in response to PAMP. Leaf discs were exposed at t = 0 min to control solution or 1 µM flg22. In (A–L), the 400/485 nm fluorescence ratio (R) was measured over time by multiwell fluorimetry and expressed relative to the mean initial ratio (R_i) before treatment. Data are means ± Se from two independent experiments (n ≥ 8, C, D, F, and G) or a representative experiment (n ≥ 5, A, B, E, and H–L). Two-way ANOVA using repeated measures for time and Tukey's multiple comparisons analyses are shown in Supplemental Table S2. Pst DC3000 and Pst hrpA significantly (P < 0.05) increased roGFP2-Orp1 oxidation in cytosol (A) and nuclei (B) and GRX1-roGFP2 oxidation in cytosol/nuclei (E) and peroxisomes (H). Pst DC3000 and Pst hrpA were significantly (P < 0.05) different in roGFP2-Orp1 oxidation in cytosol (A). flg22 significantly (P < 0.05) increased roGFP2-Orp1 oxidation in cytosol (H), chloroplasts (K), and peroxisomes (L). The 400/485 nm fluorescence ratios for fully reduced/fully oxidized probes in each compartment (Figure 1) were: cytosol 0.89/5.08; nuclei 0.98/4.76; mitochondria 0.62/3.35; chloroplasts 0.53/2.88, and peroxisomes 0.43/1.79.

Figure 6

The second PAMP- and bacteria-triggered roGFP2-Orp1 oxidation event in the cytosol is not affected in rbohD, rbohF, prx33-3, and prx34-2 mutants. A, Long-term kinetics of PAMP-induced apoplastic ROS production detected by luminol assay in Col-0 WT, rbohD, and rbohF mutants. The luminescence was measured over time after treatment with control solution or 1 µM flg22 at t = 0 min. Data are means ± Se (n = 5) from a representative experiment. Long-term kinetics of roGFP2-Orp1 oxidation in leaves of rbohD and rbohF (B) and prx33-3 and prx34-2 (C) mutants in response to flg22. Leaf discs were exposed at t = 0 min to control solution or 1 µM flg22. Long-term kinetics of roGFP2-Orp1 oxidation in leaves of rbohD and rbohF (D), prx33-3 and prx34-2 (E), and bak1-5 and bik1 (F) mutants in response to WT Pst DC3000 bacteria. Leaf discs were exposed at t = 0 min to Mock control (10 mM MgCl₂) or 10⁸ cfu mL⁻¹Pst DC3000 (Pst) bacteria. In (B–F), the 400/485 nm fluorescence ratio (R) was measured over time by multiwell fluorimetry and expressed relative to the mean initial ratio (R_i) before treatment. Data are means ± Se from two independent experiments (n ≥ 7, B–C, E) or a representative experiment (n ≥ 4, D and F). Two-way ANOVA using repeated measures for time and Tukey's multiple comparisons analyses are shown in Supplemental Table S2. Pst-treated rbohD, rbohF, prx33-3, and prx34-2 mutants were not significantly different to Pst-treated Col-0 (A–E). Pst-treated bak1 was not significantly different to mock bak1-5. Pst-treated bak1-5 and bik1 were significantly different (P < 0.05) to Pst-treated Col-0. Pst-treated rbohD, rbohF, prx33-3, and prx34-2 mutants were not significantly different to Pst-treated Col-0 (A–E). Pst-treated bak1 was not significantly different to mock bak1-5. Pst-treated bak1-5 and bik1 were significantly different (P < 0.05) to Pst-treated Col-0. The 400/485 nm fluorescence ratios for fully reduced/fully oxidized probes in the cytosol/nuclei (Figure 1) were 0.85/5.70.

Figure 7

A model of apoplastic and cytosolic oxidative events during the immune response. PAMPs and virulent Pseudomonas syringae cause distinct oxidative events focused, respectively, on the apoplast and cytosol. The apoplastic oxidation has a very rapid large burst followed by a smaller burst at 4 h (A). Both are dependent upon the RBOHD NADPH oxidase isoform and associated with relatively small cytosolic oxidation. In contrast, oxidation of the cytosol-localized biosensors roGFP2-Orp1 and GRX1-roGFP2 starts after the initial apoplastic burst and is followed by a large oxidation event peaking at 6 h (A). This oxidation is independent of RBOHD but dependent on BAK1, the flagellin co-receptor. The second phase is also dependent on the BIK1 kinase and is suppressed by Pseudomonas syringae effectors. roGFP2-Orp1 is H₂O₂-specific (B), so oxidation could indicate increased H₂O₂ production in the cytosol. GRX1-roGFP2 oxidation may follow because of H₂O₂-induced thiol (glutathione) oxidation. BIK1 or BAK1 could activate an unidentified cytosolic H₂O₂ production mechanism or decrease H₂O₂ scavenging capacity (B, red circle). Alternatively, BIK1 or BAK1 could mediate decreased activity of thiol/glutaredoxin-based reduction of the biosensors (B, red lines). AP, H₂O₂ permeable aquaporin; NOX, NADPH oxidase.