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[OPEN]

A small Rab GTPase orchestrates salicylic acid-mediated defense responses by recruiting Rab effector proteins that normally suppress these defense responses in uninfected plants. 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.


INTRODUCTION
Plants have a complex and sophisticated defense strategy to protect themselves from pathogenic organisms, which includes premade physical barriers as well as the accumulation of newly synthesized defense compounds. The perception of microbial pathogens occurs initially through the interaction of plasmamembrane localized receptor proteins of the host cell with pathogenassociated molecular pattern components (Liew et al., 2005). The detection of pathogen-associated molecular pattern components by these surface-localized receptors initiates defense signaling pathways that result in cellular responses to limit microbial growth and the induction of host disease resistance (Liew et al., 2005). One key feature of this system is a need for the trafficking and secretion of various molecules, including the secretion of antimicrobial proteins (van Loon et al., 2006) and reinforcement of the plant cell wall with callose, as first discovered in 1863 by de Bary (Nishimura et al., 2003). These trafficking processes must be carefully regulated to ensure that an effective defense response is achieved.
The Rab family of GTPase proteins are important regulators of membrane trafficking and cycle between an active GTP-bound form and an inactive GDP-bound form (Stenmark et al., 1994). In their active, GTP-bound forms, Rab GTPases recruit Rab effector proteins, and it is through these protein-protein interactions that they carry out their regulatory functions (Pereira-Leal and Seabra, 2001;Zerial and McBride, 2001;Vernoud et al., 2003). We previously showed that one member of the Arabidopsis thaliana RabA family, RabA4B, localizes to vesicles that emerge from the trans-Golgi network (TGN) and that this Rab GTPase plays important roles during the regulation of polarized membrane trafficking in root hair cells (Preuss et al., 2004;Kang et al., 2011). RabA4B specifically interacts with the lipid kinase isoforms phosphatidylinositol 4-kinase b1 (PI4Kb1) and PI4Kb2, and these proteins coordinate to regulate polarized expansion in root hairs (Preuss et al., 2006). However, RabA4B is widely expressed in numerous plant organs, including leaves (Preuss et al., 2004), and likely has additional functions outside of polarized cell growth. Rab GTPases also have been shown to interact with multiple effector proteins, and these effectors are diverse in both form and function (Grosshans et al., 2006).
To identify additional potential RabA4B effector proteins, a yeast two-hybrid (Y2H) screen for novel RabA4B effectors identified PUB13, a member of the Plant U-Box (PUB) family of E3 ligase proteins (Mudgil et al., 2004). While most PUB family members have not been extensively characterized, current evidence indicates that PUB proteins function in plant responses to various external cues. Arabidopsis PUB44 is required for delaying premature leaf senescence (Raab et al., 2009), while another Arabidopsis PUB family member, CHIP, is important for sensitivity to temperature stress (Yan et al., 2003). The oilseed rape (Brassica napus) ARM REPEAT CONTAINING1 (ARC1) is a positive regulator of self-incompatibility. ARC1 has verified E3 ligase activity and is proposed to degrade stigmatic factors that are important for early events in pollen germination, thus leading to the rejection of self-pollen (Stone et al., 2003). S-receptor kinase is a known interacting partner that mediates the selfincompatibility response in oilseed rape. S-receptor kinase has been shown to interact and phosphorylate ARC1, which, in turn, acts as a positive regulator of the self-incompatibility response (Samuel et al., 2008). Loss of Spl11, a PUB homolog in rice (Oryza sativa), regulates flowering via interaction with SPL11-INTERACTING PROTEIN1, a Signal Transduction and Activation of RNA family member (Vega-Sánchez et al., 2008).
The closest Arabidopsis homolog to oilseed rape ARC1 is PUB17 (Yang et al., 2006). Unlike oilseed rape, Arabidopsis is self-compatible, and the function of PUB17 is unrelated to pollenstigma interactions. Instead, PUB17 is important for the regulation of cell death during pathogen infection. Loss of PUB17 causes decreased resistance to the plant pathogen Pseudomonas syringae pv tomato DC3000 (Pst DC3000; Yang et al., 2006). Therefore, PUB17 is predicted to activate plant defense responses. Other PUB family members, PUB22, PUB23, and PUB24, function to downregulate defense responses (Trujillo et al., 2008). Similarly, loss of Spl11 causes spontaneous lesion formation in the absence of infection and increased resistance to pathogens (Zeng et al., 2004). The increased resistance observed in these PUB mutants is thought to be due to the constitutive activation of defense responses, and Spl11 is proposed to be a negative regulator of cell death and defense. PUB13 was also recently shown to be involved in the regulation of salicylic acid (SA)-mediated resistance via the targeted degradation of the receptor protein FLAGELLIN SENSITIVE2 (FLS2) upon elicitation with flagellin (Lu et al., 2011;Li et al., 2012aLi et al., , 2012bLiu et al., 2012). These data show the importance of PUB proteins for responding to diverse extracellular signals in plants.
Here, we present data describing a functional connection between PUB13, RabA4B, and two other RabA4B-recruited effector proteins, PI4Kb1 and PI4Kb2, during SA-mediated defense responses to the pathogen Pst DC3000. We show by Y2H analysis that PUB13 specifically associates with RabA4B in a GTPdependent manner through N-terminal U-box N-terminal domain (UND) and U-box domains. In vivo, a functional fluorescent fusion of PUB13 primarily localizes to internal subcellular compartments, and these EYFP-PUB13-labeled membranes display significant colocalization with the Golgi-localized syntaxin 32 marker protein ECFP-SYP32 (Geldner et al., 2009) and ECFP-RabA4B on Golgiassociated TGN compartments. Consistent with previous observations (Lu et al., 2011;Li et al., 2012a), pub13 mutants were significantly smaller than wild-type plants and displayed an increased resistance to Pst DC3000. We describe similar reduced stature and increased pathogen resistance in pi4kb1/pi4kb2 double mutant plants, and both mutant pub13 and pi4kb1/pi4kb2 double mutant plants display constitutively induced expression of SAdependent PATHOGENESIS RELATED (PR) genes and increased levels of ectopic deposition of callose in leaves. We also demonstrate that PUB13 interacts directly with phosphoinositides, including phosphatidylinositol 4-phosphate (PI-4P), through C-terminal sequences. We propose that RabA4B orchestrates the regulation of SA-mediated defense response by recruiting and trafficking of the effectors PUB13, PI4Kb1, and PI4Kb2, which normally function to suppress the induction of SA-mediated defense responses in uninfected plants.

PUB13 Is a RabA4B-Interacting Protein
In order to identify additional RabA4B effector proteins, we used a Y2H screen to find proteins that interact with constitutively active (GTP-bound) RabA4B. Out of 127 sequenced yeast colonies containing putative RabA4B-interacting proteins, we identified one clone that contained the N-terminal region of PUB13 (amino acids 1 to 551, termed PUB13 D552-660). Rab GTPases interact with their effector proteins in a manner regulated by GTP binding and hydrolysis, with Rab effector protein interactions being stabilized when the Rab GTPase is in a GTP-bound state and this interaction being lost upon hydrolysis of GTP to GDP (Hutagalung and Novick, 2011). To determine if the interaction between PUB13 and RabA4B was nucleotide-dependent and specific to RabA4B, we also tested the interaction of PUB13 with RabG3C, RabF2A, and the small GTPase ROP1, which localize to plant vacuolar, endosomal, and plasma membrane compartments, respectively (Ueda et al., 2001;Gu et al., 2005;Haas et al., 2007). We found that PUB13 interacts with the active, GTP-bound form of RabA4B but not with a GDP-locked mutant form ( Figure  1A). Additionally, PUB13 did not interact with other members of the Rab family or with ROP1 ( Figure 1A). From these data, we concluded that PUB13 is a potential effector of RabA4B.
We were interested to determine which domains of PUB13 are responsible for interaction with RabA4B. PUB13 contains three domains, the UND (amino acids 1 to 255), a U-box domain (amino acids 256 to 328), and a series of six armadillo (ARM) repeats (amino acids 383 to 587) ( Figure 1B). We tested these domains singly and in combination and determined that the UND-U-box domain combination is necessary for interaction with active RabA4B ( Figure 1C). We were unable to detect Y2H interaction between full-length PUB13 and RabA4B; however, this may be a reflection of the experimental method, as the full length of the RabA4B effector protein PI4Kb1 also fails to interact with RabA4B in a Y2H assay, even though full-length PI4Kb1 interacts with RabA4B in a biochemical assay (Preuss et al., 2006).
To determine whether PUB13 is capable of interacting directly with RabA4B, we performed a far-western protein gel blot experiment, where different dilutions of recombinant PUB13-UND-U-box protein were spotted onto a nitrocellulose membrane and then hybridized with RabA4B bound to a nonhydrolyzable form of GTP (RabA4B-GTPgS). Recruitment of RabA4B was detected with anti-RabA4B antibody (Figure 2A). To confirm the interaction between RabA4B and PUB13 in planta, we performed bimolecular fluorescence complementation (BiFC; Hu et al., 2002), where the constitutively active form of RabA4B was fused to the N-terminal portion of yellow fluorescent protein (YFP) (nYFP-RabA4B) or to the C-terminal portion of YFP (RabA4B-cYFP), while PUB13 was fused to the C-terminal portion of YFP (PUB13-cYFP) and to the N-terminal portion of YFP (nYFP-PUB13). Then, nYFP and cYFP fusions proteins were transiently expressed in Nicotiana benthamiana leaf epidermal cells. The relative level of reconstituted BiFC fluorescence was detected by confocal microscopy in leaves prestained with propidium iodide to detect cell peripheries ( Figure 2B). While we did observe that minimal levels reconstituted BiFC fluorescence when nYFP alone was coexpressed with PUB13-cYFP ( Figure 2B, panel 2) or when nYFP-RabA4b was coexpressed with cYFP alone ( Figure 2B, panel 3), the level of reconstituted BiFC fluorescence was significantly enhanced when RabA4B and PUB13 were both present ( Figure 2B, panels 1 and 4). Reconstituted BiFC fluorescence could not be detected in cells that expressed only nYFP-RabA4B or cYFP-PUB13 ( Figure 2B, panels 5 and 6). To further characterize the levels of reconstituted BiFC fluorescence, fluorescence in leaves expressing these nYFP and cYFP constructs was quantified. The mean YFP fluorescence intensity of the epidermis increased to approximately five times the initial value in cells when both RabA4B and PUB13 were expressed ( Figure 2C, black bars). These results, taken together with our Y2H (Figure 1) and in vitro far-western protein gel blot (Figure 2A) results, support the conclusion that PUB13 can interact physically with RabA4B, likely through interaction with the UND-U-box domains, both in vitro and in vivo.

PUB13 Localizes to Golgi and TGN Compartments
We examined whether RabA4b and PUB13 colocalized to the same subcellular membranes in Arabidopsis cells. To determine (A) Y2H interaction of PUB13D552-660 with active GTP-bound RabA4B (T), but not inactive GDP-bound RabA4B (D), was detected on high-stringency medium (2HisTrpLeu + 3-aminotriazole [2HTL+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 (2LT) or tryptophan (2T), 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 (2HisTrpLeu+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. the distribution of PUB13, we examined the subcellular localization of a functional EYFP-PUB13 fluorescent fusion protein that was stably transformed into Arabidopsis seedlings, which were then crossed with plants harboring either enhanced cyan fluorescent protein (ECFP)-conjugated RabA4B or the Golgi-localized red fluorescent protein (RFP)-SYP32 (Geldner et al., 2009). In these lines, the EYFP-PUB13 fusion protein selectively labeled mobile punctate structures (Supplemental Movie 1) in both leaf epidermal (Figures 3A to 3C) and root epidermal (Figures 3D to 3F) cells. While the YFP-PUB13-labeled structures displayed significant levels of colocalization (63.6%) with CFP-RabA4B in leaf tissues ( Figure 3C), the degree of colocalization (29.14%) of these two fusion proteins in root tissues was significantly lower ( Figure 3F). EYFP-PUB13 also displayed similar levels of colocalization with the RFP-SYP32-labeled Golgi compartments (63.41%) in leaf epidermal cells ( Figures 3G to 3I). However, unlike ECFP-RabA4B, EYFP-PUB13 colocalization levels with the RFP-SYP32 marker were maintained (73.47%) in root tissues (Figures 3J to 3L). In growing root hairs, EYFP-PUB13 labeled mobile Golgi-like structures, but no significant accumulation was observed in the (A) Different dilutions of recombinant PUB13 were spotted onto a nitrocellulose membrane and then hybridized with 25 nM RabA4B-GTPgS. 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 6 SD, n = 50). Asterisks indicate significant differences (P < 0.001) calculated by Student's t test.
tips of these cells (Supplemental Movie 1). Taken together, these results indicate that EYFP-PUB13 labeled compartments that partially colocalize both with Golgi-localized RFP-SYP32 compartments and with a subset of CFP-RabA4B-labeled compartments. However, this colocalization is restricted to Golgi/TGN subcellular membranes and is not observed in TGN-derived RabA4B-labeled secretory vesicles in tip-growing root hair cells.

Mutant pub13 and pi4kb1/pi4kb2 Double Mutant Plants Display Similar Growth Defects
To gain a better understanding of PUB13 function during plant growth and development, we obtained a SALK T-DNA insertion line for PUB13. We performed RT-PCR analysis on wild-type Columbia-0 (Col-0) plants and homozygous PUB13 mutant (pub13) plants. The pub13 plants showed disrupted expression of PUB13 transcript as compared with wild-type (Col-0) plants (Supplemental Figure 1A). We observed that plants homozygous for the T-DNA insertion in PUB13 displayed defects in growth and development and were significantly smaller than wild-type plants ( Figure 4A), confirming observations reported previously by Li et al. (2012a). Growth defects observed for pub13 could be complemented by expressing EYFP-PUB13 fusion protein (Supplemental Figure 1B). Intriguingly, we noted that the aerial growth phenotypes, which we and others (Li et al., 2012a) described for the pub13 mutant, were strikingly similar to those described previously for pi4kb1/pi4kb2 double knockout mutants (Preuss et al., 2006), in which the expression of two other RabA4B effector proteins, PI4Kb1 and PI4Kb2, was eliminated. However, while both pub13 and pi4kb1/pi4kb2 plants displayed similar aerial growth defects ( Figure 4A), pi4kb1/pi4kb2 double mutants also displayed defects in tip-growing root hairs ( Figures 4B and 4D), but root hairs of pub13 mutant plants were normal (Figures 4B and 4C).
In order to determine whether the growth phenotypes observed for pi4kb1/pi4kb2 double mutants were actually due to the loss of PI4K activity, we tested whether a catalytically inactive form of PI4Kb1 (PI4Kb1 D972A; Strahl et al., 2005) was able to restore aerial and root hair growth phenotypes in pi4kb1/pi4kb2 plants (Supplemental Figure 2). This mutation is based on an invariant residue present in the catalytic domain of all PI4Ks and is required for Pik1p activity in yeast (Strahl et al., 2005). We found that the expression of catalytically inactive PI4Kb1 was unable to rescue either aerial (Supplemental Figure 2A) or root hair growth defects (Supplemental Figures 2B to 2D). Significantly, expression of the catalytically inactive PI4Kb1 D972A mutant did not rescue the constitutively activated expression of the PR1 gene in pi4kb1/pi4kb2 plants (Supplemental Figure 2E). Furthermore, PI4Kb1 and PI4Kb2 interact with RabA4B through a novelhomology domain that is spatially distinct from the catalytic site, and these interactions occur independently of the presence of the catalytic domain (Preuss et al., 2006). These results suggest that these phenotypes are the result o loss of PI4Kb1/2associated PI4K activity and not simply due to a structural deficiency caused by the lack of accumulation of these proteins.

RabA4B Effector Mutants Display Enhanced Resistance to Pst DC3000
Previous characterization of pub13 mutant plants showed that they had increased resistance to biotrophic and necrotrophic pathogens (Lu et al., 2011;Li et al., 2012a). Since similar growth defects were observed for pi4kb1/pi4kb2 double mutants, we were interested in determining whether they also display increased pathogen resistance. Therefore, we compared how loss of either PUB13 or PI4Kb1/b2 affected the ability of Pst DC3000 to grow and induce disease symptoms in these mutant plants. The rate of bacterial growth inside leaves was significantly lower for pub13 and pi4kb1/ pi4kb2 compared with wild-type plants ( Figure 5; n = 30). In both pub13 and pi4kb1/pi4kb2 plants, scoring of disease symptoms was similar, with a limited percentage of plants (50%) showing clear symptoms of bacterial disease, while over 95% of the wild-type plants manifested serious signs of bacterial infection. Disease signs were scored based on the emergence of water-soaked areas (starting from 3 d after inoculation), which then developed into necrosis, chlorosis, and, in the most extreme case, wilting of the whole plant. Purpling of the leaf tissues (possibly due to the accumulation of anthocyanins) was sporadically observed. These results indicate that PUB13 and the two lipid kinases, PI4Kb1 and PI4Kb2, both play important roles during the initiation and establishment of defense responses to Pst DC3000.

RabA4B-Interacting Protein Mutants Display Callose Deposition in the Absence of Pathogens
One physiological outcome of pathogen-related signaling is the induction of ectopic callose deposition in the cell walls of infected 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 pi4kb1/pi4kb2 (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 pi4kb1/pi4kb2 homozygous mutant plants (A); wild-type (Col-0) root hairs (B); pub13 homozygous mutant root hairs (C); and pi4kb1/pi4kb2 homozygous mutant root hairs (D). Bars = 200 mm.
plants (DebRoy et al., 2004). To assess the presence of ectopic callose deposition in pub13 and pi4kb1/pi4b2 mutant plants, we collected rosette leaves from uninfected 4-week-old plants and stained for callose using decolorized aniline blue ( Figure 6). Leaves from both pub13 and pi4kb1/pi4kb2, but not wild-type plants, displayed numerous fluorescent patches indicating an enhanced deposition of callose in these tissues (Figures 6A to 6C and 6G).
In Arabidopsis, the perception of the bacterial component flagellin by the membrane receptor FLAGELLIN-SENSITIVE2 (FLS2) leads to the deposition of callose in the cell wall (Bardoel et al., 2011;Lu et al., 2011), and we wondered whether PUB13, PI4Kb1, and PI4Kb2 are also involved in the signaling of callose deposition. To test this hypothesis, we stimulated leaves of wildtype, pub13, and pi4kb1/pi4kb2 plants with flagellin fragment 22 The data shown are means 6 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 pi4kb1/pi4kb2 plants. P values were calculated by Student's t test.
(flg22), which revealed little difference among the three genotypes (Figures 6D to 6F and 6G), suggesting that pub13 and pi4kb1/ pi4kb2 mutant plants have normal responses to flagellin and a functional FLS2 receptor and signaling pathway.

Mutation of RabA4B-Interacting Proteins Constitutively
Activates SA-Dependent PR Gene Expression One possible explanation for the decreased growth rate, increased resistance, and increased callose deposition observed in pub13 and pi4kb1/pi4kb2 double mutant plants when exposed to pathogens is the constitutive activation of plant defenses in the absence of pathogens (Bowling et al., 1994). Expression of PR genes is an early occurrence in plant defenses (van Loon et al., 1994;van Loon and van Strien, 1999). We compared the relative expression levels of PR genes (PR1 to PR5) in pub13, pi4kb1/pi4kb2, and wild-type plants in the absence of pathogen challenge. Using RT-PCR, we found that PR1, PR2, and PR5 expression levels were higher in 4-week-old pub13 and pi4kb1/pi4kb2 plants compared with wildtype plants ( Figure 7A). These results were confirmed by using quantitative RT-PCR (qRT-PCR), and we concluded that PR1, PR2, and PR5 expression levels were increased between 30-and 50-fold in pub13 and pi4kb1/pi4kb2 plants, whereas the expression of PR3 and PR4 was either unchanged or only slightly induced ( Figure 7B). Expression of PR1, PR2, and PR5 has been shown to be regulated by the plant hormone SA (Uknes et al., 1992).
We inferred that it is likely that in pub13 and pi4kb1/pi4kb2 plants, SA-mediated defense genes are constitutively expressed even in the absence of pathogen challenge. In the case of pub13 mutant plants, constitutive activation of PR1 gene expression was specific to the loss of PUB13 activity and normal PR1 gene expression was restored in stably transformed pub13 expressing either EYFP-PUB13 or untagged PUB13 genes (Supplemental Figure 3). In constitutive-expresser of PR genes1 mutants, the overexpression of SA-dependent genes has been efficiently reversed by mutating SALICYLIC ACID INDUCTION-DEFICIENT2 (SID2) (Korasick et al., 2010). To test whether the constitutive expression of PR genes we observed in pub13 plants was SAdependent, we created homozygous pub13/sid2 double mutants and compared the expression levels of the PR1 to PR5 genes in pub13, sid2, and wild-type plants. As expected, wild-type and sid2 plants expressed low levels of the SA-dependent genes compared with the pub13 mutant ( Figure 7B). At the same time, the loss of SID2 function in the pub13/sid2 plants was sufficient to block the induction of PR1, PR2, and PR5.
If both pub13 and pi4kb1/pi4kb2 mutant plants constitutively express SA-dependent PR genes, have these plants lost the ability to respond to changes in SA levels? To address this question, we examined whether pub13 and pi4kb1/pi4kb2 mutants retained the ability to respond to exogenously added SA. In both pub13 and pi4kb1/pi4kb2 plants, PR1 gene expression was observed to increase after 24 h of treatment with exogenous SA. While wildtype plants showed a 56-fold increase in PR1 expression at 24 h after SA treatment, pub13 and pi4kb1/pi4kb2 plants displayed constitutively increased PR gene expression. These plants retained the ability to respond to SA, as PR1 gene expression increased by 1.6-and 3.2-fold at the same time point, respectively ( Figure 7C). These results indicate that the induction of PR genes in pub13 and pi4kb1/pi4kb2 plants can still be regulated by SA levels.

Loss of FLS2 Function Suppresses the Constitutive SA-Mediated Defense Responses in pub13 Mutants
A key molecule in the innate immunity of Arabidopsis to Pst DC3000 is the receptor protein FLS2, which localizes on the plasma membrane and associates with bacterial flg22 . Upon binding the flg22 peptide, FLS2 initiates the immune signaling pathways associated with the regulation of PR genes (A) to (F) Leaves of wild-type (Col-0), pub13, and pi4kb1/pi4kb2 plants were stained to show basal levels of deposition of callose ( [A], [B], and [C], respectively); the same genotypes also were treated with 5 mM flg22 ([D], [E], and [F], respectively). Arrows indicate deposited callose. Bar = 100 mm. (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 pi4kb1/pi4kb2 plants. P values were calculated by Student's t test. (Gomez-Gomez and Boller, 2000;Gómez-Gómez et al., 2001). The loss of FLS2 results in insensitivity to flg22, leading to an increased sensitivity to pathogenic microorganisms such as Pst DC3000. Both pub13 and pi4kb1/pi4kb2 mutant plants were sensitive to added flg22 (Figure 6), which shows that FLS2mediated pathogen-related signaling still occurs in these plants.
To test for a genetic interaction between PUB13 and FLS2, we generated pub13/fls2 plants and quantified the expression of PR1 to PR5 genes in these plants (Figure 8). We found that the simultaneous loss of PUB13 and FLS2 genes in the pub13/fls2 double mutant attenuates the PR gene overexpression observed in the pub13 mutant alone, reducing the overexpression observed when compared with pub13 ( Figure 8B). Thus, the increased expression of PR genes in pub13 plants may be mediated, in part, by upregulation of the FLS2 pathway. Furthermore, the growth defects observed in pub13 mutants were largely restored to wild-type levels in the pub13/fls2 double mutants ( Figures 8C to 8E).

PUB13 Interacts with Phosphoinositides
As part of the regulation of the membrane trafficking between the Golgi and the plasma membrane, RabA4B recruits the closely related PI4Kb1 and PI4Kb2 lipid kinases, and the generation and turnover of the phosphoinositide PI-4P on these membrane compartments play important roles during the regulation of polarized membrane trafficking in Arabidopsis (Preuss et al., 2006). In animals, Rab5 recruits lipid kinases to generate phosphatidylinositol (C) qRT-PCR analysis of PR1 gene expression in wild-type (Col-0), pub13, and pi4kb1/pi4kb2 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 pi4kb1/pi4kb2 plants. P values were calculated by Student's t test.
3-phosphate (PI-3P) in endosomal membranes (Christoforidis et al., 1999b), and a number of Rab5-interacting proteins are recruited in a dual Rab5-and PI-3P-dependent manner (Simonsen et al., 1998;Nielsen et al., 2000;Zerial and McBride, 2001). We decided to test whether the recruitment of PUB13 by RabA4B might also involve coordinated recruitment through interaction with both RabA4B and phosphoinositides. PUB13 was subdivided into two parts corresponding to the C-terminal ARM-repeat domain and the N-terminal UND-U-box domain. Glutathione S-transferase (GST)-tagged versions of the UND-U-box and ARM domains were expressed and purified from Escherichia coli, and each of these was tested in a lipid blot assay; we used the Pleckstrin homology (PH) domain of the human protein FOUR-PHOSPHATE-ADAPTOR PROTEIN1 (FAPP1) as a positive control for the binding specificity ( Figure 9A). We did not observe any interaction between the UND-U-box domain and lipids ( Figure  9B), but surprisingly, we found that the ARM-repeat domain interacts with several different phosphoinositides, especially with PI-4P and phosphatidylinositol 3,4,5-triphosphate ( Figure 9C). Similar results were observed in an in vitro liposome binding assay, where the PUB13-ARM domain displayed binding to PI-3P, PI-4P, PI-3,5-diphosphate, and PI-4,5-triphosphate ( Figure  9D). We did not observe any interaction between the UND-Ubox domain and phosphatidylinositol-containing liposomes ( Figure 9D).

DISCUSSION
PUB13 has been previously implicated in the regulation of SAdependent innate immunity responses in Arabidopsis (Lu et al., 2011;Li et al., 2012a). In this study, we have determined that PUB13 is a RabA4B-interacting protein and show that two previously characterized RabA4B-interacting proteins, the PI4Kb1 and PI4Kb2 lipid kinases, also are involved in the regulation of SAmediated defense responses in Arabidopsis. Furthermore, we observed that PUB13 interacts with phosphoinositides, including PI-4P, through its C-terminal ARM domain. These results demonstrate that RabA4B-dependent trafficking pathways play important roles in the regulation of SA-dependent innate immunity responses in Arabidopsis.
by Lu et al. (2011), the degradation of FLS2 requires the specific E3 ligase activity of PUB13, and its homolog PUB12, to interact with the FLS2 coreceptor BRASSINOSTEROID INSENSITIVE (BRI1)-ASSOCIATED RECEPTOR KINASE1 (BAK1). BAK1 interacts directly with FLS2 in a flg22-dependent fashion, and formation of the FLS2-BAK1-PUB13/12 complex results in turnover of the FLS2 receptor (Lu et al., 2011). Our results support the hypothesis that PUB13 is a negative regulator of SA-dependent gene expression and defense response and further identify that PUB13 is recruited to Golgi-associated TGN compartments, at least in part through its interaction with RabA4B ( Figure 3). Consistent with an important role for RabA4B in regulating FLS2-mediated defense responses, we show that two other previously identified RabA4B-interacting proteins, PI4Kb1 and PI4Kb2, also are involved in negatively regulating the expression of SA-dependent PR genes.
Importantly, we demonstrated that the constitutive activation of SA-mediated defense responses was significantly reduced in pub13/fls2 plants ( Figure 8) and that overall growth defects observed in pub13 mutants were largely restored to the wild type in the pub13/fls2 double mutant background ( Figures 8A, 8D, and 8E; Supplemental Figure 1B). These findings indicate that the enhanced expression of PR genes and associated growth phenotypes observed in pub13 are, at least in part, directly dependent upon signaling through the FLS2 receptor protein. The FLS2 receptor normally functions to bind and detect the bacterial flagellin protein (Gomez-Gomez and Boller, 2000; Gómez-Gómez et al., 2001). A common defense response initiated by flagellin or flg22 treatment is callose deposition. Our results indicated that both pub13 and pi4kb1/pi4b2 mutant plants had positive callose staining after treatment with flg22 ( Figure 6). These results support the idea that the FLS2 receptor is present and functional in both pub13 and pi4kb1/pi3kb2 mutant plants.
More recently, FLS2 has been shown to continuously cycle between the plasma membrane and internal compartments (Beck et al., 2012), and these studies have indicated that the recycling of internalized FLS2 undergoes two distinct endocytic pathways, according to the activation status of the receptor (nonelicited versus elicited). These results imply that the propensity of flg22-FLS2 and nonelicited FLS2 complexes to activate defense-related signaling pathways is associated with altered subcellular sorting of the FLS2 receptor either for degradation or recycling back to the plasma membrane. This would be compatible with our observation that PUB13 localizes to an internal Golgi-associated TGN compartment (Figure 3), where its interaction with FLS2 may be associated with productive sorting of this receptor protein into either the degradative or recycling membrane trafficking pathway. Interestingly, the expression of mutant forms of different members of the RabA4 GTPase family was recently shown to affect endocytosis, sorting, and recycling events associated with TGN and multivesicular body compartments (Choi et al., 2013). Additionally, RabA4C, which shares a high degree of sequence similarity with RabA4B, has been implicated in regulation of the callose synthase POWDERY MILDEW RESISTANT4, further highlighting an important role for RabA GTPases during plant defense signaling (Ellinger et al., 2014). Combining these earlier results with our observations, we propose a working model for the function of RabA4B-recruited PI4Kb1 (and PI4Kb2) and PUB13 during the subcellular trafficking of FLS2 ( Figure 10). In our working model, we suggest that RabA4B-recruited PUB13 and PI4Kb1/b2 proteins function in the normal sorting of the internalized FLS2 protein ( Figure 10) and that loss of these protein functions would result in altered subcellular dynamics of either empty FLS2 protein or flg22-FLS2 complexes ( Figure 10). In particular, we suggest in our model that the RabA4B recruitment of PI4K activities assists in the enrichment of PI-4P on Golgi-associated TGN compartments and that this in turn assists in the recruitment of other proteins, such as PUB13, that participate in the appropriate sorting and activity of FLS2 and other potential elements associated with SA-dependent defense signaling ( Figure 10).
Rab GTPases regulate aspects of membrane trafficking through the GTP-dependent recruitment of cytosolic proteins (Zerial and McBride, 2001;Grosshans et al., 2006;Hutagalung and Novick, 2011). RabA4B was shown previously to recruit two closely related lipid kinases, PI4Kb1 and PI4Kb2 (Preuss et al., 2006), and the generation and turnover of the phosphoinositide PI-4P was shown to play important roles during polarized membrane trafficking in tipgrowing cells (Thole and Nielsen, 2008;Szumlanski and Nielsen, 2009). PUB13, like PI4Kb1 and PI4Kb2, specifically interacts with RabA4B in its active, GTP-bound state ( Figure 1A), and a functional EYFP-PUB13 fluorescent fusion displays significant colocalization with RabA4B on Golgi-associated membrane compartments in both root and leaf epidermal cells (Figures 3C and 3F). Significantly, however, EYFP-PUB13 was not observed to accumulate on apical vesicle populations in tip-growing root hair cells (Supplemental Movie 1), as has been shown for both RabA4B and PI4Kb1 (Preuss et al., 2004(Preuss et al., , 2006. These results are consistent with the observation that T-DNA insertional mutants of PUB13 failed to display root hair growth defects ( Figure 4) and indicate that, unlike PI4Kb1 and PI4Kb2, PUB13 does not appear to play essential roles during polarized membrane trafficking during tip-restricted expansion in these cells. Instead, colocalization between PUB13 and RabA4B appears to be restricted to Golgi-associated membranes (Figures 3I and 3L) also labeled by the Golgi-localized SYP32 marker (Geldner et al., 2009). RabA4B labels budding vesicle profiles on Golgiassociated TGN compartments as well as independent vesicles derived from these membranes (Kang et al., 2011). We interpret these results as an indication that RabA4B interaction with PUB13 is limited to Golgi-associated TGN elements and that, unlike PI4Kb1, the association between RabA4B and PUB13 is not maintained once RabA4B vesicles have finished budding from these Golgi-associated TGN compartments (Figure 10).
In some instances, the efficient recruitment of Rab-GTPaseinteracting proteins to a particular membrane compartment involves direct interaction with both the Rab-GTPase and the membranes of that compartment through lipid-interacting domains that can recognize one or more classes of phosphoinositide lipids (Simonsen et al., 1998;Zerial and McBride, 2001). During endosomal trafficking in animals, the Rab-GTPase Rab5 recruits PI3Ks to enrich endosomal membranes for PI-3P (Christoforidis et al., 1999b), which in turn facilitates the recruitment of a number of Rab5-interacting proteins that also contain the PI-3P binding domains, such as EARLY ENDOSOME ANTIGEN1, Rabenosyn-5, and Rabankyrin-5 (Simonsen et al., 1998;Christoforidis et al., 1999a;Nielsen et al., 2000;Schnatwinkel et al., 2004;Stenmark, 2009). PI-3P also has been shown to be enriched in plant compartments associated with endocytosis, although whether these represent early or late endocytic compartments remains unclear (Voigt et al., 2005;Robinson et al., 2008). Intriguingly, in plants, early endosomal compartments involved in the recycling of internalized plasma membrane proteins such as BRI1, PIN-FORMED proteins, and FLS2 appear to be associated with compartments labeled by RabA-GTPase family members (Feraru et al., 2012;Choi et al., 2013). These RabA-labeled membranes, which emerge from Golgi-associated TGN compartments (Kang et al., 2011), also appear to play significant roles in the polarized secretion of new cell wall components, both in tip-growing cells and in cells undergoing cytokinesis, suggesting that at least in some cases there may be significant overlap between membrane trafficking compartments involved in endocytic recycling and in polarized secretion in plants. Intriguingly, RabA4B, which was shown previously to regulate polarized membrane-trafficking events in tip-growing cells (Preuss et al., 2004;Szumlanski and Nielsen, 2009), recruits the lipid kinases PI4Kb1 and PI4Kb2, and the production and regulation of PI-4P levels in these membranes play important roles during the regulation of membrane-trafficking events (Walch-Solimena and Novick, 1999;Audhya et al., 2000;Li et al., 2002;Godi et al., 2004;Preuss et al., 2006;Thole and Nielsen, 2008). Our data showed that the ARM repeat domain of PUB13 shows significant binding to phosphoinositides, in particular to PI-4P. While PUB13 did display binding to some other phosphatidylinositol isoforms, phosphatidylinositol 4,5-bisphosphate is thought to be mainly localized to the plasma membrane (Mueller-Roeber and Pical, 2002;Lee et al., 2007) while phosphatidylinositol 3,4-bisphosphate and phosphatidylinositol 3,4,5-triphosphate have not been detected in plants (Mueller-Roeber and Pical, 2002;Furt et al., 2010), excluding the possibility that either of these lipids interacts with PUB13 at the TGN. PI-4P was shown to localize to the TGN membrane . Therefore, it is likely that PUB13 interacts specifically with PI-4P in vivo. These results would be consistent with a model in which RabA4B recruitment of PI4Kb1 (and PI4Kb2) lipid kinases results in the localized accumulation of PI-4P, which creates a membrane microenvironment suitable for the recruitment of PUB13. Intriguingly, the ENHANCED DISEASE RESISTANCE2 (EDR2) protein also has been shown to contain a PH domain that binds specifically to PI-4P, and edr2 mutant plants also display constitutively activated SA defense responses (Vorwerk et al., 2007). This raises the interesting possibility that the ability to bind PI-4P may be common to multiple proteins involved in the appropriate regulation of SA-mediated defense responses in Arabidopsis.
Our data provide additional insight into the functions of some of the domains of PUB13. Although the UND in Brassica ARC1 was used previously to isolate its interacting protein (Samuel et al., 2009), the UND in PUB13 is an uncharacterized plant-specific domain unique to PUB proteins (Mudgil et al., 2004). The U-box domain has been characterized as a putative E3 binding domain (Aravind and Koonin, 2000;Hatakeyama et al., 2001;Ohi et al., 2003), which is necessary for the ubiquitination of target proteins (Aravind and Koonin, 2000;Hatakeyama et al., 2001). Recently, the SPL11 U-box domain was shown to be important for the interaction with E2 (Bae and Kim, 2013), and in this class of proteins, this domain has been shown to be essential to trigger disease resistance in tobacco (Nicotiana tabacum) and tomato (Solanum lycopersicum) . We found that the UND-U-box domain was important for the interaction between PUB13 and RabA4B, perhaps supporting a role for UND in mediating protein-protein interactions. ARM repeats have been described previously as protein-protein interaction domains (Coates, 2003). In this study, we found that the ARM repeat domain of PUB13 has a propensity to interact with phosphoinositides. Structural studies of the ARM repeat domain have indicated significant structural similarity to Epsin NH 2 -Terminal (ENTH) domains (Hyman et al., 2000). ENTH domains have been widely characterized as having phosphatidylinositol binding capacity (De Camilli et al., 2002) and participate in vesicle-trafficking events (Itoh et al., 2001). We propose that PI-4P generation by PI4Kb1/b2 might result in the generation of a membrane microenvironment suitable for the recruitment of other RabA4B-interacting proteins. Targeting of PUB13 and possibly other cytosolic proteins to this PI-4P-enriched domain may play important roles in the sorting of signaling components, such as the FLS2 receptor, between degradation in late endosomal and vacuolar compartments or recycling back to the plasma membrane ( Figure 10) Salomon and Robatzek, 2006;Göhre et al., 2008;Beck et al., 2012).

Construction of Clones and Y2H Interaction Assays
The yeast strain AH109 (Clontech Laboratories) was used for testing protein interactions. GTP-bound and GDP-bound forms of RabA4B, RabG3C, RabF2A, and ROP1 were generated and cloned into the pGBKT7 bait vector as described (Preuss et al., 2006). Sequencing of the original PUB13 clone identified in the Y2H assay determined that it corresponded to amino acids 1 to 551. Protein domains of PUB13 were cloned using primers specific to the UND ( The pGBKT7-RabA4B clones and pGADT7-PUB13 domain clones were transformed into yeast using a LiAc-mediated transformation method (Gietz and Woods, 1994). Two-milliliter liquid cultures, selecting for the presence of plasmid(s) of interest, were grown for 2 d at 30°C at 260 rpm. Cultures were diluted to an OD 600 of 0.02, and 10 mL droplets were placed onto selective and nonselective media. Plates were scanned after 7 d of growth at 30°C.

Protein-Protein Direct Interaction Assay
A far-western protein gel blot assay (Walsh et al., 2012) was conducted to assess the interaction between PUB13 and RabA4B. Briefly, the GSTtagged PUB13 was expressed and purified from BL21DE Escherichia coli cells, and a serial dilution of the protein was applied to a precut, dry nitrocellulose membrane. GST alone was added to the membrane as a negative control. The membrane was shaken in 10 mL of blocking buffer at room temperature. After 1 h, the blocking buffer was removed and replaced with protein binding buffer containing GTP-bound (GTPgS) RabA4B, and the membrane was incubated at 4°C. After 6 h of incubation on a rotary shaker, the membrane was washed three times with blocking buffer and placed in 10 mL of blocking buffer containing anti-RabA4B antibody (1:10,000 dilution) overnight at 4°C on an orbital shaker. The primary antibody solution was removed, and the membrane was washed three times with blocking buffer and incubated for 3 h with 10 mL of blocking buffer containing anti-rabbit horseradish peroxidase secondary antibody (1:2000 dilution). After incubation with the secondary antibody, the membrane was washed once with blocking buffer and twice with protein dilution buffer. The blot was activated by the addition of 1 mL of each component of the chemiluminescent substrate.

BiFC and Quantification
The interaction between PUB13 and the constitutively active form (GTPbound) of RabA4B was verified in a structural complementation assay using pSAT1-and pSAT4-derived plasmids (Citovsky et al., 2006) for transient expression of the fusion proteins in Nicotiana benthamiana (Sparkes et al., 2006). The C-terminal end of PUB13 was fused to the N-terminal moiety or to the C-terminal moiety of YFP (nYFP or cYFP, respectively), while for RabA4B, the fluorescent tags were added to the N-terminal end of the protein to prevent potential interference with the mutation in the C-terminal CAAX box. Two days after injection with Agrobacterium tumefaciens GV3101, leaf tissues were treated with propidium iodide (10 mg/mL solution) to stain the outline of the cells and viewed with an Olympus spinning disc confocal microscope equipped with two laser lines and a set of filters capable of distinguishing between YFP and RFP. The samples were observed under 203 magnification (Olympus UPLSAPO20X; NA = 0.70). To excite the YFP, we used a 515-nm/50-mW diode laser; to excite phosphatidylinositol, we used a 561nm/50-mW diode laser. Quantitation of the fluorescence intensities from BiFC experiments was performed on representative images from five independent leaves transiently expressing nYFP and cYFP constructs. To determine relative BiFC fluorescence, 10 regions of 2 3 10 23 mm 2 were selected for each image, and the relative fluorescence intensity of these regions was determined using ImageJ. These values were averaged to obtain the fluorescence intensity for each of the five independent images examined. BiFC values for each leaf were normalized by comparison with corresponding phosphatidylinositol fluorescence values.

Characterization of PUB13 Insertion Mutant and SID2 Deletion Mutant
The PUB13 insertion mutant was obtained from the SALK T-DNA collection (SALK_093164; ABRC). The T-DNA insertion site in PUB13 was confirmed by sequencing the T-DNA junction using primers specific to PUB13 (59-ATCGGGTACCATGGAGGAAGAGAAAGCTTCTG-39 and 59-ATCGGAGCTCTTAAGTATCTGCAGCTTCTGTGG-39) and to the left border of the T-DNA (59-GCGTGGACCGCTTGCTGCAACT-39). The insertion site consists of a head-to-head insertion of two copies of the T-DNA. Sequencing of both sides of the T-DNA insert confirmed that it is located in the third exon of PUB13 at base pair 1045 of the genomic sequence. The sid2 mutant seeds were obtained from Antje Heese's laboratory (University of Missouri). The mutation consists of a deletion and was detected using primers flanking the deletion site (59-TTCTTCATGCAGGGGAGGAG-39, 59-CAACCACCTGGTG-CACCAGC-39, and 59-AAGCAAAATGTTTGAGTCAGCA-39). Double mutant pub13/sid2 plants were obtained by crossing the two parental lines together and genotyping for double homozygous mutant plants in the F2 generation.

Generation of Transgenic Lines Coexpressing EYFP-PUB13/ECFP-RabA4B and EYFP-PUB13/RFP-SYP32
The full-length PUB13 domain was amplified from wild-type cDNA using primers specific to PUB13 (forward primer, 59-CATGCCATGGAGGAAGA-GAAAGCTTCTGCT-39; reverse primer, 59-CCGCTCGAGTTAAGTATCTG-CAGCTTCTGT-39). The insert was cloned into the plant transformation vector pCAMBIA (CAMBIA) containing a 35S promoter element used to drive the expression of EYFP, thus creating pCAM-35S-EYFP-PUB13 vector. Wildtype Arabidopsis thaliana (Col-0) plants were transformed by floral dipping (Clough and Bent, 1998). Independent transformants were selected by growth on hygromycin and then transplanted to soil. Independent T3 lines were used for imaging. Homozygous transgenic 35S-EYFP-PUB13 plants were crossed with the transgenic line expressing CFP-RabA4B. The F1 progeny were selected on one-quarter-strength Murashige and Skoog plates containing hygromycin (50 mg/mL). Resistant F1 seedlings were analyzed by confocal microscopy for the expression of the EYFP and CFP fusion proteins. Homozygous transgenic 35S-EYFP-PUB13 plants were crossed with the transgenic line generated by Geldner et al. (2009) expressing the marker RFP-SYP32. The F1 progeny were selected on one-quarter-strength Murashige and Skoog plates containing hygromycin (50 mg/mL) and kanamycin (30 mg/mL), and the double resistant F1 seedlings were analyzed by confocal microscopy for the expression of the EYFP and RFP fusion proteins.

RT-PCR Analysis of PUB13 Expression
Leaves of wild-type (Col-0) and PUB13 homozygous mutant (pub13) plants were collected and frozen in liquid nitrogen. Total RNA was extracted from each sample using the Qiagen RNeasy Kit. Total cDNA was synthesized from each RNA sample using the Invitrogen SuperScript III First-Strand Synthesis System for RT-PCR and the oligo(dT) primer according to kit instructions. Reactions with primers specific to UBIQUITIN10 (59-TGCCGGAAAACAATTGGAGGATGG-39 and 59-CGTACTTTGGCCGACTACAACATC-39) were used to equalize the amounts of cDNA from the different sources. PCR was performed with GoTaq Green (Promega) using primers specific to PUB13 (59-CCGGAATTCC-GATCTGCAGCTGGGGAAATC-39 and 59-CCGCTCGAGTTAAGTATCTG-CAGCTTCTGT-39). The amplification program consisted of 30 s at 94°C, 30 s at 56°C, and 1 min at 68°C for 28 cycles, followed by a 10-min extension at 68°C. PCR products were visualized on an agarose gel using ethidium bromide.

PR Gene Expression Analysis by RT-PCR
Plants were grown in soil for 4 weeks, and rosettes from 4-week-old plants were collected and frozen in liquid nitrogen. Total RNA was extracted from each sample using the Qiagen RNeasy Kit. The cDNA was synthesized from each RNA sample using the Invitrogen SuperScript III First-Strand Synthesis System for RT-PCR using oligo(dT) primer according to kit instructions. Reactions with primers specific to UBIQUITIN10 (59-GATCTTTGCCGGAAAACAATTGGAGGATGGT-39 and 59-CGACTTGTCAT-TAGAAAGAAAGAGATAACAGG-39) were used to equalize amounts of cDNA from the different sources. PCR was performed with GoTaq Green (Promega) using primers specific to PR1 The amplification program consisted of 30 s at 94°C, 30 s at 59°C, and 1 min at 68°C for 28 cycles, followed by a 10-min extension at 68°C. PCR products were visualized on an agarose gel using ethidium bromide.

qRT-PCR Analysis of PR Gene Expression
Plants were grown in soil for 4 weeks, and rosettes were harvested and frozen in liquid nitrogen. Total RNA was extracted from each sample using the Qiagen RNeasy Kit. cDNA was synthesized from each RNA sample using the Invitrogen SuperScript III First-Strand Synthesis System for RT-PCR using oligo(dT) primer according to kit instructions. PCR was performed using SybrGreen (Bio-Rad). Reactions were run in a Bio-Rad iCycler using iQ SYBR Green Supermix (Bio-Rad) with primers specific to UBIQUITIN10 (59-GGAAAGCTCCGACACCATTG-39 and 59-CAAGCTGCTTTCCAGCGAAG-39) serving as a standard. Reactions without reverse transcriptase were used to detect the presence of contaminating genomic DNA. Relative gene expression levels were obtained using primers specific to PR1 The cycle consisted of three repeats of 95°C for 30 s, a 95°C hold for 3 min, followed by 50 repeats of the following amplification cycle: 95°C for 15 s, 62°C for 30 s, and 72°C for 30 s. After a final hold at 95°C for 1 min, the samples were cooled to 55°C and incrementally raised to 95°C for a melt-curve analysis. Reactions were tested in triplicate, and similar results were obtained in three separate experiments.

Aniline Blue Staining of Leaves for Detection of Callose Deposition
To visualize the induction of callose deposition, leaves from 4-week-old plants were injected with 1 µM flg22 and stained after 12 h as described before (Korasick et al., 2010). Briefly, the leaf tissue was fixed overnight in 1% glutaraldehyde, 5 mM citric acid, and 90 mM Na 2 HPO 4 , pH 7.4, cleared, and dehydrated with 100% ethanol. Cleared tissue was transferred sequentially in 50% ethanol and in 67 mM K 2 HPO 4 , pH 12, and then stained for 1 h at room temperature in 0.01% aniline blue in 67 mM K 2 HPO 4 , pH 12. Leaves injected with water alone were used as a negative control. Stained material was mounted in 70% glycerol and 30% stain and examined using UV epifluorescence (Nikon E600) with fluorescence optics and appropriate filters.
Challenge Experiments with Pseudomonas syringae pv tomato DC3000 The phytopathological assay was performed according to the protocol described by Ishiga et al. (2011). Briefly, surface-sterilized seeds were grown in a Petri dish (100 mm 3 25 mm) under a 12-h-light/12-h-dark photoperiod in environmentally controlled chambers. Two-week-old plants were flood-inoculated with 40 mL of a suspension of Pst DC3000 cells (1 3 10 5 starting concentration), and the rate of bacterial growth inside leaves was measured 0, 2, and 4 d after inoculation. Plants floodinoculated with water alone were used as a negative control. The numerical value of bacterial growth was normalized to colony-forming units/ mg using the total weight of the tissues used for the assay.

Characterization of pub13/fls2 Mutant Plants
The fls2 mutant seeds were obtained from TAIR. The mutation consists of a T-DNA insertion in the FLS2 gene (SALK_093905) and was detected using primers flanking the deletion site (59-CATAAGCTTGAACCTTTCAAGG-AACAG-39 and 59-TATCCAGAATGAAGATAATCGATTCCGC-39). Double mutant pub13/fls2 plants were obtained by crossing the two parental lines together and genotyping for double homozygous mutant plants in the F2 generation. The number of leaves in 24-d-old plants, the length of the stem, and the leaf area of pub13/fls2 plants were measured and compared with wild-type (Col-0), pub13, and fls2 plants. For the leaf area, we measured five leaves for each plant (from leaf 4 to leaf 8) on a total of six plants for each genotype; the surface area was calculated with ImageJ.

Lipid-Protein Binding Assay
The UND-U-box and ARM domains of PUB13 were cloned in pGEX4T-1 for expression in BL21DE E. coli cells, and the purified proteins were used in a lipid binding assay. The lipid binding activity was assessed on PIPs-P6001 strips (Echelon Biosciences). Each strip was placed in a Petri dish containing PBST with albumin (3%) and shaken in darkness at room temperature; after 1 h, the blocking buffer was removed and 10 mL of PBST containing the UND-U-box or ARM protein (0.20 pmol) was added to the Petri dish. The protein was removed after overnight shaking at 4°C, and the membrane was washed three times with PBST (10 min for each wash) before adding 10 mL of PBST containing anti-GST antibody from goat (1:2000 dilution; GE Healthcare). The lipid filter was then incubated on a shaker for 1 h at room temperature and washed three times before adding 10 mL of PBST containing anti-goat IgG antibody from donkey, horseradish peroxidase conjugate (1:7500 dilution; GE Healthcare). The lipid strip was incubated on a shaker at room temperature and after 45 min washed twice with 10 mL of PBST and rinsed in 10 mL of PBS for 10 min. The blot was activated by the addition of 1 mL of each component of the chemiluminescent substrate (Amersham ECL Prime Western Blotting Detection Reagent) and exposed to film for 10 s. For PolyPIPosome experiments, 20 mL of l mM PolyPIPosomes (65% phosphatidylcholine, 29% phosphatidylethanolamine, 1% biotinylated phosphatidylethanolamine, and 5% phosphoinositides; Echelon Biosciences) was mixed with 30 pmol of purified proteins in 1 mL of binding buffer (50 mM Tris-HCl, pH 7.6, 150 mM NaCl, and 0.05% Nonidet P40) and rotated for 90 min at room temperature. Neutravidin beads (20 mL) were added to pull down the lipid-protein mixtures for 1 h at room temperature. The mixture was subsequently centrifuged (2 min, 2500g), and supernatant was recovered for input analysis. Following three washes with 1 mL of binding buffer as detailed above, the pellet was resuspended with 23 Laemmli sample buffer followed by analysis by SDS-PAGE and protein gel blotting with antibodies specific to GST (1:1500 dilution), as detailed above.

Microscopic Observations of Fluorescent Proteins
The samples were observed with an Olympus spinning disc confocal microscope equipped with laser lines and a set of filters capable of distinguishing CFP, YFP, and RFP. The samples were observed under 103 (Olympus UPLFLN10X2; NA = 0.3) or 203 (Olympus UPLSAPO20X; NA = 0.70) magnification. To excite the CFP, we used a 445-nm/40-mW diode laser; to excite the YFP, we used a 515-nm/50-mW diode laser; to excite the RFP, we used a 561-nm/50-mW diode laser.

Complementation of pub13 Mutant Plants
Full-length PUB13 was amplified from wild-type leaf cDNA and cloned into the pCAM-35S vector and into the pCAM-35S-EYFP vector for stable expression in Arabidopsis. Homozygous pub13 mutant plants were transformed by floral dipping (Clough and Bent, 1998). Independent transformants were selected by growth on hygromycin and were transplanted to soil. Plants were verified as homozygous pub13 by PCR using primers specific to the genomic clone of PUB13 (59-GCTTTGTTTTGCTATGGTAGTGTTTGAG-39 and 59-ATCGGAGCTCTTAAGTATCTGCAGCTTCTGTGG-39) and to the left border of the T-DNA (59-GCGTGGACCGCTTGCTGCAACT-39). The presence of 35S-EYFP-PUB13 was confirmed using primers specific to EYFP (59-CGCCACCATGGTGAGCAAGGGCGAGGAGCT-39) and to PUB13 (59-GCGGATCCTTGGTCCTCGGGGTTTCCGTA-39). The presence of 35S-PUB13 was confirmed using primers specific to the 35S promoter (59-CCGGAATTCATGGAGTCAAAGATTCAAATA-39) and to PUB13 (59-GCGGATCCTTGGTCCTCGGGGTTTCCGTA-39). The expression of PR1 in individual complementation plants was tested by qRT-PCR as described above.
In Figure 1C, an image of a yeast two-hybrid drop assay was inadvertently duplicated (U-box domain interaction with GTP-bound RabA4b in -T and -LT growth conditions; circled in red in the original figure). In examining how this duplication occurred, we determined that images in the original figure also combined spots from two sets of replicate plates. Although these two sets of plates were spotted at the same time, used the same yeast strains, and presented virtually identical results, we now present images from only one set of plates in the corrected version of Figure 1C (all images replaced in this manner are underlined in red). In addition, we also now provide white borders between individual yeast spots in Figures 1A and 1C to indicate that they have been arranged in order to present this data in a more efficient and logical manner. Finally, "2HisTrpLeu" and "2HTL" were replaced with "2HisLeuTrp" and "2HLT," respectively, in the figure legend to match the figure.  (A) Y2H interaction of PUB13D552-660 with active GTP-bound RabA4B (T), but not inactive GDP-bound RabA4B (D), was detected on high-stringency medium (2HisTrpLeu 1 3-aminotriazole [2HTL13-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 (2LT) or tryptophan (2T), 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 (2HisTrpLeu13-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. (Inadvertently duplicated spots are circled in red).