Recruitment of PLANT U-BOX13 and the PI4Kβ1/β2 phosphatidylinositol-4 kinases by the small GTPase RabA4B plays important roles during salicylic acid-mediated plant defense signaling in Arabidopsis

Abstract

Protection against microbial pathogens involves the activation of cellular immune responses in eukaryotes, and this cellular immunity likely involves changes in subcellular membrane trafficking. In eukaryotes, members of the Rab GTPase family of small monomeric regulatory GTPases play prominent roles in the regulation of membrane trafficking. We previously showed that RabA4B is recruited to vesicles that emerge from trans-Golgi network (TGN) compartments and regulates polarized membrane trafficking in plant cells. As part of this regulation, RabA4B recruits the closely related phosphatidylinositol 4-kinase (PI4K) PI4Kβ1 and PI4Kβ2 lipid kinases. Here, we identify a second Arabidopsis thaliana RabA4B-interacting protein, PLANT U-BOX13 (PUB13), which has recently been identified to play important roles in salicylic acid (SA)-mediated defense signaling. We show that PUB13 interacts with RabA4B through N-terminal domains and with phosphatidylinositol 4-phosphate (PI-4P) through a C-terminal armadillo domain. Furthermore, we demonstrate that a functional fluorescent PUB13 fusion protein (YFP-PUB13) localizes to TGN and Golgi compartments and that PUB13, PI4Kβ1, and PI4Kβ2 are negative regulators of SA-mediated induction of pathogenesis-related gene expression. Taken together, these results highlight a role for RabA4B and PI-4P in SA-dependent defense responses.

Figure 1.

PUB13 Interacts Specifically with the Active Form of RabA4B. (A) Y2H interaction of PUB13Δ552-660 with active GTP-bound RabA4B (T), but not inactive GDP-bound RabA4B (D), was detected on high-stringency medium (−HisTrpLeu + 3-aminotriazole [−HTL+3-AT]). No interaction was observed with RabF2A, RabG3C, and ROP1. The presence of prey and/or bait vectors was monitored by growth in the absence of leucine and tryptophan (−LT) or tryptophan (−T), respectively. (B) UND, U-box, and ARM domains are indicated. Deletion fragments of PUB13 were constructed to determine the binding site of RabA4B. (C) Y2H interaction was seen between active RabA4B and PUB13 fragments on selective medium (−HisTrpLeu+3-AT). The interaction between RabA4B and PUB13 requires the presence of UND and U-box domains. No interaction was observed between RabA4B and the ARM domain. Surprisingly, full-length PUB13 did not interact with RabA4B in the Y2H assay.

Figure 2.

The Physical Interaction between PUB13 and RabA4B Was Confirmed in a Far Protein Gel Blot and BiFC. (A) Different dilutions of recombinant PUB13 were spotted onto a nitrocellulose membrane and then hybridized with 25 nM RabA4B-GTPγS. RabA4B was recruited by PUB13, and the recruitment was detected using an anti-RabA4B antibody. (B) BiFC assays using RabA4B and PUB13 to reconstruct the YFP signal. A strong interaction was detected in leaves expressing nYFP-RabA4B and PUB13-cYFP (panel 1). BiFC fluorescence intensity was weaker in leaves expressing nYFP-PUB13 and RabA4B-cYFP fusion domains (panel 4). No significant fluorescence was detected when only one of the fusion proteins was expressed with the complementary portion of YFP (panels 2, 3, 5, and 6). Panels are as follows: 1, nYFP-RabA4B with PUB13-cYFP; 2, nYFP with PUB13-cYFP; 3, nYFP-RabA4B with cYFP; 4, nYFP-PUB13 with RabA4B-cYFP; 5, nYFP-PUB13 with cYFP; 6, nYFP with RabA4B-cYFP. Leaves were treated with propidium iodide to stain the outline of the cell. (C) BiFC (black bars) and propidium iodide (gray bars) fluorescence intensities were quantified as mean values of YFP and RFP fluorescence intensity, respectively (mean ± sd, n = 50). Asterisks indicate significant differences (P < 0.001) calculated by Student’s t test.

Figure 3.

PUB13 Localizes on the TGN and Golgi. Leaf tissues coexpressing YFP-fluorescent PUB13 ([A] and [G]) together with CFP-RabA4B (B) or together with RFP-SYP32 (H), and merged images ([C] and [I]), are shown. Note the colocalization (yellow signal) on TGN and Golgi (the percentage of overlapping fluorescent signals). Root epidermal cells coexpressing YFP-PUB13 ([D] and [J]) together with CFP-RabA4B (E) or together with RFP-SYP32 (K), and merged images ([F] and [L]), are shown. Note the colocalization (yellow signal) on TGN and Golgi (the percentage of overlapping fluorescent signals). Bars = 15 μm.

Figure 4.

PUB13 Function Is Essential for Normal Aerial Growth in Arabidopsis but Not for Root Hair Development. The pub13 homozygous mutant plants were significantly smaller than wild-type (Col-0) plants (A). Despite the close similarity of the aerial parts between pub13 and pi4kβ1/pi4kβ2 (A) mutants, pub13 plants did not show impaired root hair development (C). Panels are as follows: aerial part of wild-type (Col-0), pub13 homozygous mutant, and pi4kβ1/pi4kβ2 homozygous mutant plants (A); wild-type (Col-0) root hairs (B); pub13 homozygous mutant root hairs (C); and pi4kβ1/pi4kβ2 homozygous mutant root hairs (D). Bars = 200 μm.

Figure 5.

pub13 and pi4kβ1/pi4kβ2 Mutant Plants Display Enhanced Resistance to Pst DC3000. Wild-type (Col-0), pub13, and pi4kβ1/pi4kβ2 plants were flood-inoculated with a suspension of bacterial cells (1 × 10⁵ colony‑forming units [CFU]). (A) Bacterial growth on leaf tissue. (B) and (C) Chlorotic symptoms in plants 4 d after inoculation. The data shown are means ± sd from 30 leaves, with three replicates for each genotype. Asterisks indicate significant differences at P < 0.1 (*), P < 0.05 (**), and P < 0.01 (***) between wild-type (Col-0) and pub13 plants or between wild-type (Col-0) and pi4kβ1/pi4kβ2 plants. P values were calculated by Student’s t test.

Figure 6.

pub13 and pi4kβ1/pi4kβ2 Mutant Plants Are No Longer Able to Specifically Induce Callose Deposition. (A) to (F) Leaves of wild-type (Col-0), pub13, and pi4kβ1/pi4kβ2 plants were stained to show basal levels of deposition of callose ([A], [B], and [C], respectively); the same genotypes also were treated with 5 μM flg22 ([D], [E], and [F], respectively). Arrows indicate deposited callose. Bar = 100 μm. (G) Average numbers of callose deposits per field of view. The results are from three independent experiments, with sd indicated by error bars (n = 15). Asterisks indicate significant differences at P < 0.05 (**) and P < 0.01 (***) between wild-type (Col-0) and pub13 plants or between wild-type (Col-0) and pi4kβ1/pi4kβ2 plants. P values were calculated by Student’s t test.

Figure 7.

PUB13 and PI4Kβ1/PI4Kβ2 Are Involved in the Expression of the SA-Mediated Defense Response. (A) and (B) RT-PCR (A) and qRT-PCR (B) analyses of PR gene expression in wild-type (Col-0), pub13, and pi4kβ1/pi4kβ2 plants. The mutant genotypes showed a constitutive expression of SA-mediated PR1, PR2, and PR5 genes. (C) qRT-PCR analysis of PR1 gene expression in wild-type (Col-0), pub13, and pi4kβ1/pi4kβ2 plants upon 2 mM SA treatment. The results are from four technical replicates and three independent experiments, with sd indicated by error bars. Asterisks indicate significant differences at P < 0.05 (**) and P < 0.01(***) between wild-type (Col-0) and pub13 plants or between wild-type (Col-0) and pi4kβ1/pi4kβ2 plants. P values were calculated by Student’s t test.

Figure 8.

The Loss of FLS2 Restores the Wild-Type Phenotype in pub13 Mutant Plants. (A) Aerial view of 24-d-old wild-type (Col-0), pub13, fls2, and pub13/fls2 plants. (B) qRT-PCR analyses of PR gene expression in wild-type (Col-0), pub13, fls2, and pub13/fls2 plants. (C) Number of leaves in 4-week-old plants (n = 30 for each genotype). (D) Leaf surface area calculated for 30 leaves harvested from six individual plants. (E) Stem height in 4-week-old plants (n = 30 for each genotype). The data are shown as means ± sd.

Figure 9.

The ARM Repeat of PUB13 Is a Lipid Binding Domain. (A) The UND-U-box domain and the ARM domain of PUB13 were tested in a lipid blot assay. The PH domain of the human protein FAPP1 was used as a control for the binding specificity in this assay. (B) and (C) The UND-U-box domain did not show any interaction with phosphatidylinositol-phosphates (B), while the ARM domain interacted with multiple phosphoinositides (C). (D) PolyPIPosome assays with GST-tagged FAPP1-PH, PUB13-UND-U-box, and PUB13-ARM. FAPP1-PH and PUB13-ARM showed binding to PI-4P. LPA, lysophosphatidic acid; LPC, lysophosphatidylcholine; PA, phosphatidic acid; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PI, phosphatidylinositol; PI-3,4P₂, phosphatidylinositol 3,4-bisphosphate; PI-3,4,5P₃, phosphatidylinositol 3,4,5-triphosphate; PI-3,5P₂, phosphatidylinositol 3,5-bisphosphate; PI-4,5P₂, phosphatidylinositol 4,5-bisphosphate; PS, phosphatidylserine; S1P, sphingosine 1-phosphate.

Figure 10.

Model for the RabA4B-Mediated Recruitment of PUB13 and PI4Kβ1/β2 to Regulate the Plant Defense Response. FLS2-BAK1 complexes localize predominantly at the plasma membrane in nonelicited cells, but this complex continuously cycles between plasma membranes and internal TGN/early endosome (TGN/EE) compartments in nonelicited cells. These TGN/EE compartments are enriched for PI-4P through the recruitment of PI4Kβ1 (and PI4Kβ2) by active RabA4B. The presence of RabA4B and PI-4P recruits PUB13 as well as other PI-4P binding proteins to these TGN/EE compartments. In nonelicited cells, RabA4B and RabA4B-recruited PI4Kβ1 (and PI4Kβ2) participate in the rapid recycling of FLS2-BAK1 complexes to the plasma membrane. Upon elicitation with flg22, flg22-FLS2-BAK1 complexes are rapidly internalized and ubiquitinated by PUB13 in TGN/EE compartments containing RabA4B and RabA4B-recruited PI4Kβ1 (and PI4Kβ2). Loss of PI-4P and/or PUB13 on RabA4B-associated TGN/EE compartments interferes with the recycling of FLS2-BAK1 to the plasma membrane (nonelicited cells) or the sorting and turnover of flg22-FLS2-BAK1 complexes in multivesicular body (MVB)/vacuole compartments (elicited cells).