Bacterial effector targeting of a plant iron sensor facilitates iron acquisition and pathogen colonization

Abstract

Acquisition of nutrients from different species is necessary for pathogen colonization. Iron is an essential mineral nutrient for nearly all organisms, but little is known about how pathogens manipulate plant hosts to acquire iron. Here, we report that AvrRps4, an effector protein delivered by Pseudomonas syringae bacteria to plants, interacts with and targets the plant iron sensor protein BRUTUS (BTS) to facilitate iron uptake and pathogen proliferation in Arabidopsis thaliana. Infection of rps4 and eds1 by P. syringae pv. tomato (Pst) DC3000 expressing AvrRps4 resulted in iron accumulation, especially in the plant apoplast. AvrRps4 alleviates BTS-mediated degradation of bHLH115 and ILR3(IAA-Leucine resistant 3), two iron regulatory proteins. In addition, BTS is important for accumulating immune proteins Enhanced Disease Susceptibility1 (EDS1) at both the transcriptional and protein levels upon Pst (avrRps4) infections. Our findings suggest that AvrRps4 targets BTS to facilitate iron accumulation and BTS contributes to RPS4/EDS1-mediated immune responses.

The bacterial pathogen Pseudomonas syringae secretes an effector protein, AvrRps4, that facilitates colonization by manipulating iron homeostasis in Arabidopsis plants.

 

Introduction

Iron (Fe) is an essential element for almost all organisms. The catalytic function of Fe is required for many key enzymes in cellular processes, such as DNA synthesis and energy production (Aznar et al., 2015; Verbon et al., 2017). Fe deficiency can lead to anemia in humans and chlorosis in plants, while an overload of Fe can be harmful due to induction of the Fenton reaction, which is catalyzed by free Fe²⁺. Hydroxyl radicals produced by the Fenton reaction damage DNA and lead to protein or lipid peroxidation (Verbon et al., 2017). Because of the importance of Fe, many organisms, including pathogens during infection, have evolved sophisticated strategies to acquire or sequester Fe to maintain Fe homeostasis (Oide et al., 2006; Lopez-Berges et al., 2012; Grinter et al., 2018).

Control of Fe homeostasis is also important for hosts to resist pathogen infection (Soares and Weiss, 2015). One vertebrate defense strategy is to withhold Fe by regulating Fe metabolism in order to reduce pathogen proliferation, a process known as ‘nutritional immunity’ (Soares and Weiss, 2015). Fe status also impacts plant immune responses. For example, Fe -starved Arabidopsis thaliana plants exhibited enhanced disease resistance to the bacterial pathogen Dickeya dadantii and fungal pathogen Botrytis cinerea (Kieu et al., 2012). Depending on the pathogen mode of infection, however, plants can increase local Fe accumulation to generate highly toxic hydroxyl radicals by activating the Fenton reaction, thereby reducing infection (Liu et al., 2007).

Seed plants have developed two distinct strategies for Fe acquisition. Eudicot species, such as A. thaliana (Arabidopsis), secrete phenolic compounds that mobilize Fe. The ferric chelate reductase (FCR) Ferric reduction oxidase 2 (FRO2) reduces Fe³⁺ to Fe²⁺, and Fe²⁺ is then transported to cells via the Fe-regulated transporter IRT1 (Eide et al., 1996; Robinson et al., 1999). Multiple bHLH transcription factors (TFs) regulate FRO2 and IRT1. For example, bHLH38, bHLH39, bHLH100, bHLH101, bHLH104, bHLH115, and ILR3 (bHLH105) encode positive regulators of Fe acquisition, and function upstream of FRO2 and IRT1 under Fe-limited conditions (Wang et al., 2013; Verbon et al., 2017). Another bHLH TF Upstream Regulator of IRTI could bind to the promoters of bHLH38, bHLH39, bHLH100, and bHLH101 to control their transcriptions (Kim et al., 2019). bHLH TFs were shown to be negatively regulated by the nucleus-located Fe sensor BRUTUS (BTS), which coordinates with POPEYE (PYE) or PYE-like proteins to prevent Fe overload in Fe-sufficient environments (Selote et al., 2015). Recently, BRUTUS-LIKE1 (BTSL1) and BTSL2 were found to target FIT (Fer-like Fe deficiency-induced TF) for degradation to avoid overload of Fe, and were primarily expressed in root epidermis and cortex cells; however, BTS was expressed in the stele and shoot and thereby maintained Fe homeostasis in leaves (Rodríguez-Celma et al., 2019). In contrast, Graminaceae monocot species employ an Fe acquisition strategy similar to that of microbes by secreting phytosiderophores (PS), such as nicotianamine, to chelate Fe in order to create an Fe–PS complex (Curie et al., 2001). The complex is then transported to root cells via the Yellow Stripe or Yellow Stripe-like family transporters YS1/YSL (Curie et al., 2001; DiDonato et al., 2004).

Pathogens also secrete Fe-scavenging siderophores that facilitate Fe uptake from their hosts. Siderophores are low-molecular weight molecules with high affinity for chelating ferric Fe (Khan et al., 2018). In the rhizosphere, some microbe-secreted siderophores can enhance plant growth and disease resistance (Verbon et al., 2017). Fe chelates from compost microorganisms improved the Fe nutrition of soybean (Glycine max) and oat (Avena sativa) (Chen et al., 1998). Also, sequestering of Fe by beneficial microbe-secreted siderophores limits certain soil-borne pathogens from accessing Fe and causing plant disease (Loper and Buyer, 1991). Competition for Fe is thought to arise between plants and pathogens during infection. Taguchi et al. (2010) demonstrated that the siderophore pyoverdine from Pseudomonas syringae pv. tabaci is required for full virulence on tobacco (Nicotiana tabacum). Pyoverdine biosynthesis-defective mutants were less virulent than the wild-type strain, indicating that siderophores contribute to infection by some pathogens (Khan et al., 2018). Notably, pyoverdine was not required for virulence of many other P. syringae pathovars, such as P. syringae pv. tomato (Pst) DC3000 on tomato (Solanum lycopersicum) plants (Jones and Wildermuth, 2011), suggesting that these bacteria employ other strategies for Fe acquisition. A transcriptomic study of Pst DC3000 during infection of Arabidopsis leaves showed that activation of plant immunity causes reprogramming of bacterial Fe homeostasis-related genes (Nobori et al., 2018). It remains unclear whether immunity modulation of Fe availability by the plant directly limits Pst DC3000 growth.

Different P. syringae bacterial strains deliver suites of virulence factors (called effectors) to plant cells to promote colonization (Toruno et al., 2016; Xin et al., 2018). Some effectors interfere with pathogen-associated molecular pattern-triggered immunity (PTI) mediated by pattern recognition receptors (PRRs) at the plasma membrane. In plant–pathogen co-evolutionary cycles, specific effectors become recognized by intracellular nucleotide-binding domain/leucine-rich repeat (NLR) receptors in certain host genotypes (Zhang et al., 2017), often to induce localized host cell death (a hypersensitive response, HR) and resistance in a process called effector-triggered immunity (ETI) (Cui et al., 2015).

The P. syringae pv. pisi derived a coiled-coil protein effector, AvrRps4, enhanced bacterial infection when delivered by Pst DC3000 (Pst) to Arabidopsis or P. syringae pv. tabaci to Nicotiana benthamiana (Sohn et al., 2009; Sohn et al., 2012). Host virulence mechanisms reported for AvrRps4 in susceptible plants include dampening of PTI and direct disabling of nuclear WRKY TF-DNA binding to suppress transcriptional mobilization of plant defense pathways (Sarris et al., 2015). Hence, it is possible that AvrRps4 has multiple sites of action to aid infection of the host. After AvrRps4 delivery to plant cells, the protein is cleaved by an unknown protease to release a 133 amino acid N-terminal and 88 amino acid C-terminal peptide (Sohn et al., 2009; Halane et al., 2018). Analysis of the AvrRps4 domain functions indicated that the N-terminal domain has a bacterial growth-promoting virulence activity and the N- and C-terminal domains, together, are recognized in ETI (Halane et al., 2018). NLR determinants for AvrRps4 recognition in Arabidopsis accessions Col-0 and Ws-2 are the receptor pair of RESISTANCE TO RALSTONIA SOLANACEARUM1 (RRS1) and RESISTANCE TO PSEUDOMONAS SYRINGAE4 (RPS4) belonging to a major subfamily of toll-Interleukin-1 Receptor (TIR) domain NLRs (abbreviated to TIR-NLRs or TNLs) (Birker et al., 2009; Gassmann et al., 1999; Narusaka et al., 2009; Sarris et al., 2015; Saucet et al., 2015). RRS1 proteins possess a C-terminal domain with a WRKY DNA-binding motif, which serves as a decoy for WRKY TF virulence targets of AvrRps4 and another RRS1/RPS4-recognized bacterial effector, PopP2 (Deslandes et al., 2002; Le Roux et al., 2015; Sarris et al., 2015). AvrRps4- or PopP2-mediated disruption of RRS1-DNA binding leads to RRS1–RPS4 ETI (Le Roux et al., 2015; Ma et al., 2018; Sarris et al., 2015). Like all studied TNL receptors, RRS1–RPS4 confer ETI via the Enhanced Disease Susceptibility1 (EDS1) family of nucleocytoplasmic lipase-like proteins (Feys et al., 2005; Wirthmueller et al., 2007; Heidrich et al., 2011; Cui et al., 2018; Bhandari et al., 2019). AvrRps4 interacted with EDS1, and EDS1 interacted with the TNL RPS4, suggesting that EDS1 serves both as a host target of AvrRps4 and a signaling bridge to downstream ETI pathways (Bhattacharjee et al., 2011; Heidrich et al., 2011; Huh et al., 2017; Halane et al., 2018).

Here, we show that the Fe sensor BTS is targeted by Pst-delivered AvrRps4 and that this changes Fe homeostasis of the host plant. We present further evidence that an EDS1-associated NLR immune complex responds to Fe deficiency in the plant and possibly guards BTS during Pst (avrRps4) infection. As a result, RRS1–RPS4–EDS1 immunity activation restricts Fe accumulation in the apoplast to reduce pathogen infection.

Results

Fe regulatory genes respond to Pst (avrRps4) infection

To investigate AvrRps4 activity during Pst infection, we examined the transcriptome of Col-0 (WT, resistant) and Col eds1 (eds1-2, hypersusceptible) mutant plants at 4 and 8 h after leaf infiltration with Pst (avrRps4) (Bhandari et al., 2019) (Supplemental Data Set 1). Statistically significant differences between the WT and eds1 mutant at the level of individual genes were most prominent at 8 hours post-inoculation (hpi) (|log₂FC|≥1, FDR ≤ 0.05), showing that the initial transcription reprogramming linked with RRS1–RPS4–EDS1-dependent ETI occurs between 4 and 8 hpi (Supplemental Data Set 1). We performed gene ontology (GO)-based analysis of gene expression to learn which host processes are affected by the Pst (avrRps4) infection. The gene set expression test (PAGE, Parametric analysis of gene set enrichment) revealed that genes from GO terms connected to defense responses were upregulated upon Pst (avrRps4) infection at 4 and 8 hpi in both genotypes, albeit to a lower degree in the eds1 mutant (Supplemental Data Set 2). Interestingly, GO terms linked to Fe homeostasis showed signs of downregulation in both WT (Wild type) and eds1 at 4 hpi but were not statistically significant (Supplemental Data Set 2, mean log₂FC< –0.2). Because overall Fe-responsive genes are not clear in plants, to further test this observation, we selected the Arabidopsis genes that have been reported to be specifically induced upon Fe limitation as reference genes, where 295 genes were differentially expressed (Rodríguez-Celma et al., 2013). We then examined their expression trends in our RNA-seq data. The 295 genes were generally repressed in both the WT and eds1 mutant at 4 and 8 hpi, but differences between the two genotypes were significant only at 8 hpi (Tukey’s HSD, α = 0.001; Figure 1A). Therefore, we concluded that expression of the Fe deficiency-stimulated genes drops early in response to Pst (avrRps4) infection (<4 hpi) in an EDS1-independent manner and that the Fe-responsive genes are involved in the Pst (avrRps4)-triggered immune response.

Figure 1.

Fe deficiency-responsive mutants respond to Pst and Pst (avrRps4) infection. A, Infection with Pst (avrRps4) results in lower expression of genes induced under Fe deficiency condition. A total of 295 genes differentially expressed upon Fe limitation (Rodríguez-Celma et al., 2013) showed reduced expression in both wild-type (WT) Col-0 and the eds1-2 mutant at 4 and 8 h after the bacteria infiltration (hpi, OD₆₀₀ = 0.005). In the eds1 mutant, the repression was weaker than that in the WT at 8 hpi. Overlapping letters above the boxplots indicate no statistically significant differences between the samples (Tukey’s HSD, α = 0.001). B, Schematic diagram of the regulation of Fe homeostasis in Arabidopsis. The master Fe sensor protein BTS negatively regulates downstream TF bHLH family genes, including ILR3 (also called bHLH105). Under Fe deficiency, BTS and downstream TFs are induced. BTS modulates downstream signaling pathways to fine-tune Fe uptake. FRO2 and NAS4 are responsible for Fe transportation. Solid lines indicate the direct induction or regulation of downstream targets. Broken line indicates the potential targets of BTS. The regulation pathway of Fe homeostasis is modified from (Tissot et al., 2019). C, Bacterial growth in Fe regulation mutants. The leaves of 4-week-old Fe regulation Arabidopsis mutant plants were infiltrated with 5×10⁴ cfu/mL Pst DC3000 and Pst (avrRps4). Bacterial titers were determined at 4 dpi under standard growth conditions. D, Bacterial growth in the bts-2 mutant. The bts-2 mutant was grown under 26°C and then subjected to pathogen inoculation assays at 23°C (standard growth condition). E, Bacterial growth in bts-2 alleles. The BTS knockdown allele bts-1 and three independent homozygous RNAi lines (#4, #18, and #20) were grown under standard conditions for pathogen inoculation assays. Others are the same as in (C). Different letters indicate significant differences based on a two-factor ANOVA with Tukey’s HSD test (P≤0.05, Data are means ± sd, n = 18 biological replicates).

Pst infection in Fe deficiency response mutants

To explore whether AvrRps4 and/or recognition of AvrRps4 has a role in host Fe regulation during infection, we inoculated various Arabidopsis mutants known to alter Fe regulation with virulent Pst or avirulent Pst (avrRps4). During Fe deficiency, the Fe regulator BTS modulates ILR3 and its heterodimerization with bHLH34, bHLH104, and bHLH115, which further regulates downstream bHLHs (bHLH38/39/100/101) and PYE TFs, leading to the activation of FRO2/IRT1 (Fe acquisition) or NAS4 (Fe transportation) (Figure 1B) (Tissot et al., 2019). Growth of Pst and Pst (avrRps4) was not different between most bHLH TF mutants and Col-0, except for ilr3-2, nas4-1, and fro2 (Figure 1C). Of these three mutants, the ilr3-2 mutant displayed strongly enhanced resistance to both Pst strains compared with Col-0, while fro2 was slightly more susceptible.

Additionally, we tested Pst and Pst (avrRps4) growth phenotypes in the Fe sensor mutant bts-2 (Li et al., 2019). BTS encodes a ubiquitin E3-ligase with an essential role in maintaining Fe homeostasis in Arabidopsis (Selote et al., 2015; Zhang et al., 2015). Surprisingly, although bts-2 displayed enhanced disease susceptibility to Pst (∼10-fold bacteria titer than Col-0), it allowed a ∼100-fold higher Pst (avrRps4) titer than that in Col-0 (Figure 1C). Because bts-2 mutant is temperature sensitive (Supplemental Figure 1A), we also tested the pathogen inoculation assays under standard growth conditions after the plants were grown at 26°C, which showed similar susceptibility to Pst (avrRps4) (Figure 1D). The bts-2 temperature-sensitive growth phenotype was rescued by transforming bts-2 with BTS controlled by its own promoter (Supplemental Figure 1B). Importantly, the disease susceptibility to Pst and Pst (avrRps4) was also rescued (Supplemental Figure 1C). The additional allele bts-1 (Hindt et al., 2017) and three RNAi lines (Zhang et al., 2015) showed reduced expression of BTS but were temperature-insensitive also exhibited enhanced susceptibility to Pst (avrRps4) when compared with Pst (Figure 1E;Supplemental Figure 1A). Therefore, we concluded that BTS is crucial for plants to resist Pst (avrRps4) infection among all the tested Fe-regulatory genes and performed the assays under standard growth condition hereafter. It is worth noting that BTS is not required for the AvrRps4-triggered HR (Supplemental Figure 1D).

To get a sense of the effect of Fe homeostasis on disease resistance, we measured Fe accumulation in ilr3-2 and bts-2 leaves by Perls staining (Long et al., 2010), a technique to visualize Fe³⁺ deposits with Prussian blue. Whereas bts-2 mutant leaves accumulated more Fe³⁺ than Col-0, indicated by a dark blue color, ilr3-2 leaves were deficient in Fe³⁺ accumulation as seen by a pale blue color (Supplemental Figure 1E). Because Pst colonizes plant apoplast, we therefore measured the apoplastic Fe of the plants, which showed that the bts-2 mutant accumulated much more Fe in the apoplast, but the ilr3-2 accumulated much less Fe than Col-0 (Supplemental Figure 1F). These results imply that the elevated levels of apoplastic Fe are beneficial for Pst proliferation. We supplied exogenous Fe into the plant apoplast by infiltration with Pst to the leaves. Indeed, 2 and 5 μM exogenous Fe significantly increased the plant susceptibility to Pst infection (Supplemental Figure 1G). These data suggest that plant Fe homeostasis is critical for pathogen proliferation.

BTS interacts with and exhibits a specific response to AvrRps4

To investigate whether AvrRps4 can manipulate the Fe response to facilitate bacterial proliferation in plants, we focused on the Fe sensor gene BTS as it is a key regulator that fine-tunes Fe acquisition (Figure 1B) (Zhang et al., 2015; Tissot et al., 2019). Because bts-2 was highly susceptible to Pst (avrRps4) (Figure 1, C and D), we tested whether other Pst-delivered effectors sensed by different NLRs enhance bacterial growth in bts-2 plants. Col-0 and bts-2 plants were infiltrated with Pst delivering AvrRps4, AvrB, AvrRpt2, or AvrPphB, recognized by Arabidopsis NLRs RRS1–RPS4, RPM1, RPS2, and RPS5, respectively. The bts-2 mutant showed enhanced susceptibility only to Pst (avrRps4) (Figure 2A), suggesting that RPM1, RPS2, and RPS5 mediated ETI functions normally in bts-2 plants. These results indicate that bts-2 is specifically affected only during Pst (avrRps4) infection.

Figure 2.

AvrRps4 interacts with the Fe response regulator BTS. A, bts-2 mutant plants are susceptible to Pst (avrRps4). The plants were grown under standard conditions for 4 weeks and then were infiltrated with Pst DC3000, Pst (avrRps4), Pst (avrB), Pst (avrRpt2), or Pst (avrPphB) at a concentration of 5×10⁴ cfu/mL. The bacteria titer was determined at 4 dpi. Different letters indicate significant differences based on a two-factor ANOVA with Tukey’s HSD test (P ≤0.05, data are means ± sd, n = 6 biological replicates). B, AvrRps4 interacts with BTS in vivo by co-immunoprecipitation assays. Est:AvrRps4-HA and 35S:BTS-FLAG were transiently expressed in N. benthamiana leaves. After 48 h, the leaves were infiltrated with 100 μM estradiol 6 h before harvesting the samples. 35S:GFP-FLAG was used as a negative control. The immunoblots at the bottom show the expression of the respective proteins. The experiment was repeated three times with similar results. C, BTS and AvrRps4 interact in vivo by split-luciferase complementation assays. 35S:BTS-nLUC and 35S:cLUC-AvrRps4 were transiently expressed in N. benthamiana leaves. 35S:C2-nLUC and 35S:cLUC-S1 were used as a positive control. Luciferase complementation imaging assays were performed at 48 hpi. Right panel shows the protein levels. The full length and cleaved AvrRps4 are indicated by arrows. The experiment was repeated three times with similar results. D, AvrRps4 interacts with the main domains of BTS (1HHE, the single HHE domain; E3, E3 domain) by His pull-down assays. The recombinant MBP-BTS^1HHE, MBP-BTS^E3, and His-AvrRps4 were purified from E. coli and subjected to His pull-down assays. Coomassie brilliant blue (CBB) staining shows the protein abundance. MBP was used as a negative control.

Next, we investigated whether BTS interacts with AvrRps4. In transient immunoprecipitation (IP) assays in N. benthamiana, AvrRps4 interacted with BTS but not GFP (Figure 2B). The split-luciferase complementation imaging (LCI) assays further confirmed that AvrRps4 and BTS associate in vivo (Figure 2C). AvrRps4 is cleaved to release a 11kD C-terminal fragment in plant cells, which is necessary for eliciting RPS4 immunity (Heidrich et al., 2011). By the BiFC, split-LUC, and IP assays, we further demonstrated that the AvrRps4 C terminus but not the N terminus interacted with BTS, and the interaction occurs in the nucleus (Supplemental Figure 2, A–C).

To test whether these two proteins interact directly, we performed maltose-binding protein (MBP) pull-down assays on recombinant proteins expressed in Escherichia coli. BTS has three hemerythrin (HHE) domains and a Really Interesting New Gene (RING) domain (E3) (Selote et al., 2015). The HHE domain binds Fe, and the E3 domain is responsible for degrading proteins such as ILR3 and bHLH115 to mediate the Fe regulatory response (Selote et al., 2015). Full-length recombinant MBP-BTS was not stably expressed in E. coli, so we expressed the 1HHE (BTS^1–250aa) and E3 (BTS^{1100–1255aa}) domains individually and tested them for interaction with full-length AvrRps4. Recombinant His-AvrRps4 interacted with both the 1HHE and E3 domains of BTS, but not with MBP alone (Figure 2D). In a microscale thermophoresis (MST) assay to study the strength of the BTS-AvrRps4 interaction, the BTS E3 domain interacted more strongly with AvrRps4 than the 1HHE domain, as indicated by a lower dissociation constant (K_d) for the BTS E3 domain with AvrRps4 (Supplemental Figure 2D). This result was further confirmed by expressing 3HHE (BTS^1–824aa) or BTS E3 and the AvrRps4 C terminus. The AvrRps4 C terminus interacted much more strongly with BTS E3 than the 3HHE domain (Supplemental Figure 2E). Taken together, these data suggest that BTS directly interacts with AvrRps4 and is thus a potential AvrRps4 virulence target.

AvrRps4 interferes with BTS-mediated bHLH115 and ILR3 degradation

To investigate the biological significance of the AvrRps4–BTS interaction, we assessed whether AvrRps4 interferes with BTS functions in Fe regulation by measuring the transcriptional and post-transcriptional effects of AvrRps4 on two major targets of BTS in Fe regulation, the PYEL TFs bHLH115 and ILR3 (Selote et al., 2015). In Arabidopsis, compared with Pst DC3000-treated samples, Pst (avrRps4) only weakly activated bHLH115 and ILR3 expression at 24 hpi (Figure 3A;Supplemental Figure 3, A and B). As a comparison, however, Pst (avrRps4) significantly activated the expression of AvrRps4-responsive genes FMO1, PR1, and PBS3 in a EDS1 and RRS1a/b-dependent manner (Supplemental Figure 3C). BTS and PYE/PYEL TFs are transcriptionally up-regulated under Fe deficiency (Selote et al., 2015). However, BTS becomes stabilized and targets bHLH115 and ILR3 for 26S proteasome degradation via its E3 ubiquitin ligase activity to fine-tune the Fe deficiency response (Selote et al., 2015). In transfected Arabidopsis Col-0 protoplasts, AvrRps4 co-expression led to increased accumulation of bHLH115 and ILR3, but not bHLH104, a PYEL protein that is not targeted by BTS (Selote et al., 2015), suggesting that AvrRps4 interferes with BTS-mediated degradation of these proteins (Figure 3B).

Figure 3.

AvrRps4 alleviates BTS-mediated bHLH115 and ILR3 degradation. A, AvrRps4 slightly induces bHLH115 and ILR3 transcription. Four-week-old Col-0 seedlings were syringe-infiltrated with Pst DC3000 or Pst (avrRps4) at a concentration of 2.5×10⁶ cfu/mL. The plant leaves were sampled at 24 hpi for RT-qPCR assays. Significant differences were analyzed by Student’s t test (P≤0.05, n =6 biological replicates). Stars indicate the difference from Pst DC3000 treatments. B, bHLH115 and ILR3 but not bHLH104 accumulates in the presence of AvrRps4. The proteins fused with FLAG or T7 epitope were expressed in Arabidopsis protoplasts and were subjected to immunoblotting assays at 16 h after transformation. The band abundance was quantified by Image J. The experiment was repeated three times with similar results. C, BTS destabilizes bHLH115 and ILR3 in N. benthamiana. 35S:BTS-FLAG and 35S:bHLH115-T7 or 35S:ILR3-T7 were transiently co-expressed in N. benthamiana. The protein levels of bHLH115 and ILR3 were examined by immunoblot analysis. The experiment was repeated at least three times with similar results. D, AvrRps4 alleviates BTS-mediated degradation of bHLH115 and ILR3. 35S:BTS-FLAG and 35S:bHLH115-T7 or 35S:ILR3-T7 were co-expressed with Est:AvrRps4-HA in N. benthamiana. After 48 h, the leaves were infiltrated with 100 μM estradiol 6 h before sample harvested. The degradation of bHLH115 and ILR3 was examined by immunoblot analysis. Est:GUS-HA served as a negative control for AvrRps4-HA. The experiment was repeated three times with similar results. ACTIN serves as the loading controls for all the panels. E and F, AvrRps4 facilitated the protein stabilization of bHLH115 and ILR3 in N. benthamiana. The AvrRps4, AvrRps4 N, and C terminus, BTS, bHLH115, and ILR3 proteins were transiently expressed in N. benthamiana. GUS protein served as a negative control. bHLH115 and ILR3 protein levels were examined by immunoblotting. G, Pst (avrRps4) infection stabilized bHLH115 and ILR3. The leaves of 4-week-old plants were inoculated with Pst or Pst (avrRps4) at a concentration of 2.5 × 10⁷ cfu/mL. Mock is 5 mM MgCl₂. The plant leaves were sampled at 6 hpi. 35S:bHLH115-HA-OX and 35S:ILR3-GFP-OX are the bHLH115 and ILR3 overexpression Arabidopsis transgenic plants.

We then transiently co-expressed BTS with bHLH115 or ILR3 in N. benthamiana leaves. BTS reduced bHLH115 and ILR3 accumulation (Figure 3C), consistent with the observation of Selote et al. (2015). Co-expression of BTS and bHLH115 or BTS and ILR3 with AvrRps4 resulted in higher bHLH115 and ILR3 accumulation (Figure 3D), suggesting that AvrRps4 interferes with their degradation by BTS. Importantly, AvrRps4 C terminus is sufficient for the accumulation by expressing these proteins in Arabidopsis protoplasts or N. benthamiana leaves (Figure 3, E and F; Supplemental Figure 3, D and E). Furthermore, the function of AvrRps4 was required, as AvrRps4^KRVY-AAAA and AvrRps4^E175A/E187A, which are mutated at the C-terminus (both of which cannot trigger immunity), impaired the capability of stabilizing bHLH115 and ILR3 (Supplemental Figure 3, F and G) (Sohn et al., 2012). Pst (avrRps4) infection also led to the accumulation of bHLH115 and ILR3 in planta (Figure 3G). In addition, the presence of MG132, an inhibitor of the 26S proteasome, did not further increase AvrRps4-mediated bHLH115 and ILR3 stabilization in the presence of BTS (Supplemental Figure 4). These results suggest that AvrRps4 increases bHLH115 and ILR3 protein levels by countering their targeting by BTS for degradation via the 26S proteasome.

AvrRps4 induces Fe accumulation in plants

As BTS is a potential AvrRps4 virulence target (Figure 2A), and the bts-2 mutant accumulates high levels of Fe (Supplemental Figure 1E) (Zhang et al., 2015), we hypothesized that AvrRps4 assists Pst acquisition of Fe during infection. To test this, we first supplied exogenous Fe into the plant apoplast by infiltration with Pst avrRps4^KRVY-AAAA. The result showed that 2 and 5 μM Fe significantly facilitated Pst (avrRps4^KRVY-AAAA) proliferation in the rrs1a/b mutant plants, which cannot recognize AvrRps4 (Supplemental Figure 5A). We then quantified total Fe content in Arabidopsis leaves using inductively coupled plasma mass spectrometry (ICP-MS) after dip-inoculation with Pst or Pst (avrRps4). Pst- and Pst (avrRps4)-infected Col-0 leaves accumulated higher total Fe than mock-treated leaves (Figure 4A). rps4a and eds1-2 mutants, which are defective in AvrRps4 TNL recognition and signaling, respectively, accumulated Fe with Pst (avrRps4) but not with Pst when they were compared with mock-treated plants (Figure 4A).

Figure 4.

AvrRps4 facilitates Fe accumulation in the plant apoplast. A, AvrRps4 facilitates total Fe accumulation in plants. Col-0, rps4a, and eds1-2 plants were dip-inoculated with Pst DC3000 or Pst (avrRps4) at a concentration of 1×10⁸ cfu/mL. Total Fe was extracted from the plants at 24 hpi and subjected to ICP-MS analysis. Mock is 5 mM MgCl₂ treatment. B, The accumulation of Fe in the Arabidopsis apoplast during Pst DC3000 and Pst (avrRps4) infection. Col-0, rps4a, and eds1-2 plants were dip-inoculated with Pst DC3000 or Pst (avrRps4) at a concentration of 1 × 10⁸ cfu/mL. Fe was extracted from the apoplast at 24 hpi and subjected to ICP-MS analysis. C and D, AvrRps4 facilitates Fe accumulation in the eds1-2 apoplast. pEst:AvrRps4(eds1-2) and pEst:EV(eds1-2) transgenic plants were sprayed with 50 μM estradiol. Total and apoplastic Fe were extracted at 12 after estradiol treatment and subjected to ICP-MS analysis. Significant differences were analyzed by Student’s t test (P ≤0.05, n = 3 biological replicates). E, AvrRps4-induced FCR activity in Arabidopsis. The plants were inoculated with Pst or Pst (avrRps4) at a concentration of 0.5 × 10⁵ cfu/mL, and the FCR activities were measured at 24 hpi. Significant differences were analyzed by Student’s t test (n = 4 biological replicates). F, The FCR activity in pEst:AvrRps4 plants. The plants were treated with estradiol as in (C), and the FCR activity was measured 12 h later. Significant differences were analyzed by Student’s t test (P ≤0.001, data are means ± sd, n =4 biological replicates). *, **, and *** represent P ≤0.05, 0.01, and 0.001 levels, respectively, compared with the mock treatment.

Because Pst colonizes the plant apoplast, Fe accumulation in this compartment might be beneficial for bacterial growth. To test this, we inoculated Col-0, rps4a, and eds1 plants with Pst or Pst (avrRps4) and quantified apoplast Fe contents. Strikingly, Pst (avrRps4) but not Pst infection increased apoplastic Fe accumulation in rps4a and eds1 and not in Col-0 (Figure 4B). These data suggest that Pst (avrRps4) facilitates apoplastic Fe accumulation during infection when defenses are compromised.

To explore this phenomenon, we generated stable transgenic Col eds1-2 lines conditionally expressing AvrRps4 or empty vector (EV) controlled by an estradiol-inducible promoter (pEst) (Supplemental Figure 5B). Upon estradiol treatment, total Fe contents were similar between pEst:AvrRps4 (eds1) and pEst:EV (eds1) plants, whereas apoplastic Fe contents were higher in pEst:AvrRps4 (eds1) than pEst:EV (eds1) plants at 12 h after treatment (Figure 4, C and D). The in situ staining of Fe also suggests that Fe accumulated in the apoplast (indicated by stained cell wall) in eds1-2 plants after infection with Pst (avrRps4) but not Pst DC3000. In contrast, Pst DC3000 and Pst (avrRps4) infection increased total Fe levels indicated by stained plastid and nucleus but not apoplast in both Col-0 and eds1-2 plants (Supplemental Figure 5C). In rpm1 rps2 mutant, which is susceptible to Pst (avrRpt2), infection with Pst (avrRpt2) failed to increase Fe accumulation in the apoplast, although both Pst and Pst (avrRpt2) increased total Fe levels (Supplemental Figure 5, D and E). Dexamethasone (Dex)-induced expression of AvrRpt2 did not lead to an increase of total or apoplastic Fe accumulation in a pDex:AvrRpt2 (rpm1 rps2) transgenic line (Supplemental Figure 5, F and G). These results suggest that the observed increase in apoplastic Fe levels is mediated by AvrRps4 when the TNL immune response is low or absent, and not as a general consequence of Pst bacterial infection.

To further validate the aforesaid observation, we measured the activity of FCR in plant leaves to assess the ferric reduction capacity, which is induced under Fe deprivation and reduces Fe³⁺ to Fe²⁺, a critical step for IRT1 to transport Fe to plant tissue (Verbon et al., 2017). Both Pst and Pst (avrRps4) infection led to elevated FCR activity in Col-0 leaves compared with the mock treatment (Figure 4E). In eds1-2 plants, Pst (avrRps4) infection resulted in much higher FCR activity than Pst (Figure 4E). Higher FCR activity was also observed in the estradiol-treated pEst:AvrRps4 (eds1) plants (Figure 4F). However, Pst (avrRpt2) did not induce higher activity of FCR than Pst in rpm1 rps2 plants (Supplemental Figure 5H). Also, Dex-induced AvrRpt2 did not result in higher FCR activity at 12 h (Supplemental Figure 5I). These results suggest that AvrRps4-induced Fe accumulation in the apoplast is probably due to the elevated Fe metabolism in plants.

The AvrRps4 recognition complex responds to Fe deficiency

Fe deficiency suppresses Col-0 root elongation (Long et al., 2010). Since AvrRps4-induced Fe accumulation in the leaf apoplast occurs in the absence of RPS4a or EDS1 (Figure 4), we investigated root growth responses of the eds1-2 and rps4a mutants under Fe-deficient conditions by growing the plants in the medium containing ferrozine, a small molecule that can chelate Fe (Gibbs, 1976; Long et al., 2010). The rps4a and eds1-2 mutant roots exhibited enhanced tolerance to Fe deficiency, as seen by the limited effect of Fe deprivation on root length compared with Col-0 or rps2 (Figure 5A;Supplemental Figure 6A). An rrs1a/b double mutant, which is fully defective in AvrRps4 recognition (Cui et al., 2018), displayed a similarly weak root response (Figure 5A). We found that bts-2 eds1-2, bts-2 rps4a, and bts-2 rrs1a double mutants had similar root length phenotypes as bts-2 to ferrozine treatment (Supplemental Figure 6, B and C). These data suggest that the AvrRps4 recognition complex, like BTS, impacts Fe deficiency responses.

Figure 5.

AvrRps4 recognition components are intrinsically involved in the BTS-mediated Fe regulatory module. A, Root length of AvrRps4 recognition mutants under Fe deficiency. The mutants were grown on ½ MS for 5 days and then transferred to ½ MS medium supplemented with (0.1 mM Fe-EDTA) or without (300 μM Ferrozine) Fe for 7 days. Ferrozine was used to chelate Fe. The root length of the plants was measured at 12 days. Different letters indicate significant differences based on a two-factor ANOVA with Tukey’s HSD test (P ≤0.05, data are means ± sd, n =4 biological replicates). B–E, The Fe regulation gene expression in EDS1-related immune mutants under Fe deficiency. RT-qPCR was used to evaluate the expression of upstream (FIT and PYE) and downstream (FRO2 and NAS4) Fe regulatory genes in Col-0, bts-2, eds1-2, rps4a, rrs1a, and rrs1a/b. Plants were grown on ½ MS for 6 days and transferred to ½ MS with or without Fe for 4 days. The experiments were repeated at least three times. Data are represented relative to ACTIN2. Significant differences were analyzed by a two-factor ANOVA with Tukey’s HSD test (P≤0.05, Data are means ± SD, n =3 biological replicates). F, BTS is part of the EDS1-mediated immune response. The leaves of 4-week-old Arabidopsis plants grown under standard conditions were syringe inoculated with Pst DC3000 or Pst (avrRps4) at a concentration of 5 × 10⁴ cfu/mL. Bacterial growth curve analysis was performed at 4 dpi. Different letters indicate significant differences based on a two-factor ANOVA with Tukey’s HSD test (P ≤0.05, n = 6 biological replicates).

We performed RT-qPCR (Quantitative reverse transcription PCR) to examine the expression of key Fe regulatory genes in the AvrRps4 recognition and response mutants. Fe deficiency induced the expression of PYE, FRO2, and NAS4 but not FIT in all tested plants; however, overall the gene expression was lower in the tested mutants when compared with Col-0 (Figure 5, B–E). Under Fe deficiency conditions, all the tested mutants except rps4a had lower transcript levels for FIT and/or FRO2. With the exception of eds1-2, these mutants showed decreased PYE and/or NAS4 expression compared with Col-0. These results suggest that RPS4 and EDS1 regulate Fe response pathways.

In pathogen infection assays, the bts-2 rps4a double mutant was as susceptible to Pst (avrRps4) as bts-2 alone (Figure 5F). Susceptibility of the bts-2 eds1-2 double mutant to Pst (avrRps4) was equal to that of eds1-2 alone and greater than the bts-2 single mutant (Figure 5D). These data suggest that BTS is part of the same pathway of RPS4-mediated immune responses.

EDS1 protein levels were diminished in bts-2 mutant

Since the AvrRps4-triggered immune response is compromised in bts-2 (Figure 2A), we tested whether EDS1 transcript and/or protein levels are reduced in the bts-2 mutant. In EDS1 immunity to Pst (avrRps4), EDS1 is upregulated (Bhandari et al., 2019; Feys et al., 2001; Garcia et al., 2010). We observed a similar trend in RT-qPCR experiments in which EDS1 transcripts were upregulated at 2 and 6 hpi with Pst (avrRps4) but not with Pst in Col-0 (Figure 6A). In contrast, in the bts-2 mutant, Pst (avrRps4) infection induced very low levels of EDS1 expression at 2 and 6 hpi (Figure 6A). These data suggest that BTS has a role in the timely upregulation of EDS1 in AvrRps4-triggered immunity.

Figure 6.

BTS stabilizes EDS1. A, EDS1 expression is downregulated in the bts-2 mutant. RT-qPCR was used to evaluate gene expression in pathogen-infected Arabidopsis leaves at 0, 2, and 6 hpi. Col-0 and bts-2 plants were grown under standard conditions, and the leaves of 4-week-old plants were infiltrated with 5 × 10⁷ cfu/mL Pst DC3000 or Pst (avrRps4). Data are represented relative to ACTIN2. Differences from the respective 0 h samples that are statistically significant are indicated. Significant differences were analyzed by a two-factor ANOVA with Tukey’s HSD test (P ≤0.05, n = 3 biological replicates). B, EDS1 protein levels in bts-2 plants during Pst DC3000 and Pst (avrRps4) infection. The plants were grown at 26°C for 3–4 weeks, and then, the leaves of Col-0, eds1-2, and bts-2 plants were syringe infiltrated with 2.5 × 10⁶ cfu/mL Pst DC3000 or Pst (avrRps4). EDS1 protein levels were evaluated by immunoblot analysis using anti-EDS1 antibody at 8 hpi under standard growth conditions. The Ponceau staining(S) of Rubisco serves as a loading control. The experiment was repeated at least three times with similar results. C, BTS stabilizes EDS1 in vivo. 35S:BTS-FLAG and 35S:EDS1-cLUC were transiently co-expressed in N. benthamiana leaves. EDS1 protein levels were examined by immunoblotting. 35S:GFP-FLAG served as the negative control for BTS. The experiments were repeated at least three times with similar results.

We determined whether EDS1 protein accumulation is also influenced by BTS by measuring EDS1 levels on immunoblots of Col-0, eds1-2, and bts-2 leaf extracts prepared at 8 hpi after treatment with buffer (mock), Pst, or Pst (avrRps4). EDS1 levels were reduced in bts-2 compared with Col-0 in both treatments (Figure 6B), which can explain the loss of resistance to Pst (avrRps4) in the bts-2 mutant (Figure 2A). We also co-expressed BTS and EDS1 in N. benthamiana, where we found that BTS was able to promote EDS1 protein levels (Figure 6C). These data demonstrate that BTS likely plays a role in stabilizing EDS1 in planta.

Because EDS1 is required for multiple ETI, we asked whether the BTS-involved immune response is specific to AvrRps4. HopA1 (an effector of P. syringae pv syringae strain 61)-triggered immunity also requires EDS1 (Kim et al., 2009). Although bts-2 plants showed enhanced susceptibility to Pst (hopA1), the bacterial titer was much lower than Pst (avrRps4) (Supplemental Figure 7A). Unlike Pst (avrRps4), Pst (hopA1) infection did not cause Fe accumulation in the plant apoplast (Supplemental Figure 7B). In addition, IP and split-LUC assays showed that HopA1 did not interact with BTS (Supplemental Figure 7, C and D). These data suggest that BTS is unlikely a virulence target of HopA1.

Discussion

Fe is an essential nutrient for pathogens and pathogen proliferation in the plant apoplast disturbs Fe homeostasis in the host plant. It is conceivable that withholding Fe during infection is a strategy employed by plants to counter pathogen proliferation. Reducing the availability of Fe to invading pathogens is a strategy that has been adopted by many higher organisms (Segond et al., 2009; Kieu et al., 2012; Soares and Weiss, 2015). Also, activation of plant ETI suppresses the expression of Fe-related genes in the pathogen (Nobori et al., 2018).

Many phytopathogens deploy siderophores that acquire Fe from their hosts, such as the P. fluorescens siderophore pyoverdine (Trapet et al., 2016). However, it appears that the Pst siderophore pyoverdine is not required for virulence (Jones and Wildermuth, 2011), signifying alternative routes for acquiring Fe. In this study, we found that a bacterial effector protein AvrRps4 targets the Fe sensor BTS to interfere with Fe homeostasis in the plant, uncovering a strategy by which Pst deploys an effector protein to acquire Fe (Figure 7).

Figure 7.

Working model for AvrRps4 in Fe acquisition. In the absence of RPS4/EDS1, bacterial pathogen secretes effector protein AvrRps4 to target the Fe sensor BTS, leading to the alleviation of BTS fine-tuned degradation of Fe-responsive PYEL proteins. Consequently, plants accumulate Fe in the cell and apoplast, thereby facilitating pathogen proliferation. However, in the presence of RPS4/EDS1, BTS is guarded by EDS1/RPS4. AvrRps4 and BTS interaction triggers EDS1/RPS4-mediated immune activation, limiting iron accumulation and restricting pathogen proliferation.

BTS negatively regulates Fe deficiency responses and targets the downstream Fe regulatory proteins bHLH115 and ILR3 for degradation (Selote et al., 2015). In our study, interestingly, AvrRps4 leads to increased accumulation of both bHLH115 and ILR3 by reducing the BTS-mediated degradation (Figure 3, B–D) and increased activity of FCR in the plants (Figure 4, E and F). This might result in elevated Fe uptake and enhanced Fe availability to the pathogen (Figure 4). Pst colonizes the plant apoplast, and the observed increase in apoplastic Fe levels during Pst (avrRps4) infection likely favors Pst proliferation in the apoplast (Figure 4). The apoplastic Fe content increased in response to Pst (avrRps4) treatment in rps4a and eds1-2, but not in Col-0 (Figure 4B), implying that TNL/EDS1 immune activation effectively counters AvrRps4-mediated enrichment of Fe in the apoplast. Notably, rps4a and eds1-2 mutants did not accumulate Fe after Pst infection (Figure 4A), probably because of insensitivity to Fe deficiency in these mutants (Figure 5A;Supplemental Figure 6A). Nevertheless, AvrRps4 still significantly promoted Fe accumulation in plant tissue (Figure 4A), suggesting that TNL recognition and concomitant EDS1-mediated transcriptional reprogramming are essential for counteracting Fe accumulation in plants. AvrRps4-facilitated Fe accumulation in the plant apoplast does not require pathogen infection, as transgenic eds1-2 plants expressing inducible AvrRps4 exhibited higher Fe levels than the controls that did not express AvrRps4 (Figure 4D). Therefore, AvrRps4 contributes to Pst proliferation in eds1-2 and rps4a mutants, at least partially by increasing Fe levels in the plant apoplast. These results uncover a novel role of AvrRps4 in plant–pathogen interactions.

Fe efflux has been shown to induce a ROS burst in the apoplast of wheat (Triticum aestivum) plants during powdery mildew infection, a strategy employed by plants to restrict pathogen invasion (Liu et al., 2007). We think it is likely that AvrRps4-induced Fe accumulation in the plant apoplast provides an important micronutrient for Pst rather than causing a Fenton reaction to generate ROS, because (1) although the bts-2 mutant contains high levels of Fe in the apoplast, it is highly susceptible to Pst and Pst (avrRps4) (Figure 1C;Supplemental Figure 1E) and (2) AvrRps4 facilitates Fe accumulation in the plant apoplast only when immune recognition and signaling are compromised (Figure 4). AvrRps4 is processed in planta. The N-terminus in the absence of the C-terminus of AvrRps4 enhances virulence in Col-0, while delivering the AvrRps4 C terminus together with the N-terminus fully complements AvrRps4-triggered immunity (Halane et al., 2018). Here, we found that the AvrRps4 C terminus interacts with BTS and is required for the accumulation of bHLH115 and ILR3 (Supplemental Figures 2 and 3). As the AvrRps4 N terminus is known to localize to chloroplasts (Li et al., 2014), it unlikely interferes with BTS function in the nucleus. However, ILR3 influences plant photosynthesis and chlorosis via downstream targets (Li et al., 2019). Considering the chlorosis phenotype of bts-2 (Supplemental Figure 1A), whether the AvrRps4 N-terminus targets downstream Fe regulation components for virulence is worth investigating. Alternatively, the chlorosis and temperature-sensitive phenotype may be attributed to high Fe content in the plants.

The RRS1–RPS4–EDS1 immunity also appears to modulate Fe deficiency responses (Figure 5), implying that the RPS4-related immune complex directly or indirectly affects plant Fe homeostasis. Alternatively, the RPS4-related immune complex might monitor virulence factors that target plant Fe regulatory pathways. We hypothesize that, during evolution, Pst deployed AvrRps4 to interfere with Fe regulation pathways. Thus, disturbance of BTS could trigger the EDS1–RPS4/RRS1-mediated immune response (Figure 2A). Additionally, BTS positively regulates EDS1 expression and protein steady state accumulation in ETI (Figure 6), highlighting the role of BTS in TNL/EDS1 immunity. We propose the following scenario. In the absence of RRS1/RPS4 or EDS1, AvrRps4 interferes with BTS degradation of Fe regulators, resulting in increased Fe accumulation in the leaf apoplast. In the presence of RRS1/RPS4 and EDS1, recognition of AvrRps4 leads to the immune activation, resulting in reduced Fe availability in the apoplast (Figure 7).

Many immunity proteins regulate Fe homeostasis. For example, mammalian lipocalin 2 protein limits bacterial growth by sequestrating the Fe-laden siderophore, while lipopolysaccharide (LPS) and flagellin markedly activate lipocalin 2 transcription via their respective toll-like receptors (TLRs), highlighting the important roles of TLRs in nutritional immunity (Flo et al., 2004). Although pathogen-derived effectors have been shown to target diverse biological processes, our work suggests the additional roles of effectors in helping pathogen proliferation. Some effectors have evolved to interfere with plant physiology and metabolism, in some cases to acquire nutrients for the pathogen as does the Xanthomonas oryzae effector PthXo1 (Chen et al., 2010). In response, certain plant NLR proteins might have evolved to recognize these effectors directly or indirectly by associating with plant growth-related regulators and mounting defense responses. Because many Pst strains do not produce AvrRps4, it will be important to investigate how these bacteria acquire Fe from plants during infection.

Materials and methods

Plant material and growth conditions

Arabidopsis thaliana T-DNA mutants ilr3-2 (Salk_004997), pye-1 (Salk_021217), bhlh104-1 (Salk_099496), nas4-1 (Salk_135507), and fro2 (Salk_202892) were obtained from the ABRC (www.Arabidopsis.org). Homozygous seeds for bhlh38 (Salk_108159), bhlh39 (Salk_025676), bhlh100 (Salk_074568), bhlh101 (Salk_011245), bhlh115 (WiscDsLox384C11), rps4a (Salk_057697）, rrs1a (Salk_061602), the eds1-2 null allele, rpm1 rps2, and Dex:AvrRpt2 (rpm1 rps2) in the Col-0 background were used. rrs1a/b was obtained by crossing rrs1a and rrs1b (Salk_001360). Homozygous bts-2 was obtained by crossing the mutant of Salk_004748 to Col-0 to remove the mutation allele of At1g67520. An estradiol-inducible promoter-driving AvrRps4 was generated by cloning full-length AvrRps4 into the pER8 vector (a gift from Dr Jian Ye at Institute of Microbiology, Chinese Academy of Sciences), and then, this construct was transformed into eds1-2 plants, and the homozygous line was used in the assays. All plants were grown at 23°C or 26°C under a 10-h light/14-h dark cycle and 4- to 5-week-old plants were used for experiments. Light was provided by white fluorescent bulbs with an intensity of ∼120 μmol m⁻² s^-1. For plate grown plants, seeds were surface sterilized using 70% ethanol for 3 min followed by 40% bleach for 8 min and rinsed four times with sterile water. Then, the seeds were stratified at 4°C for 3 days in dark before being planted on media.

Fe-sufficient medium was composed of half-strength Murashige and Skoog (½ MS) medium, 0.05% MES, 1% sucrose 0.4% agar, and 0.1 mM Fe(II)-EDTA (Ethylene diamine tetraacetic acid) to substitute Fe sulfate. Fe-deficient medium was supplied with 300 μM ferrozine (an Fe chelator) to chelate Fe sulfate in the MS medium. The pH was adjusted to 5.8. Plants were grown at 23°C under a 14-h light/10-h dark cycle in environmentally controlled plant growth chambers. Light was provided by white fluorescent bulbs with an intensity of ∼120 μmol m⁻² s^-1.

RT-qPCR analysis

Four-week-old plants were syringe infiltrated with 5 × 10⁷ cfu/mL Pst or Pst (avrRps4). The infected leaves were sampled, and the total RNA was extracted by the Trizol method (Invitrogen). The cDNA was synthesized with 500 ng of total RNA in a 10 μL HiScript II Q RT SuperMix with genomic DNA wipe (Vazyme Biotech Co.,Ltd, R223) for RT-qPCR assays. The gene transcription was quantified on CFX96^TM Real-time System (Bio-RAD, USA) with the ChamQ SYBR qPCR Master Mix (Vazyme Biotech Co.,Ltd, Q311). ACTIN2 was used as the reference gene. All primers used are listed in Supplemental Table 1.

Analysis of bacterial growth

Pst and Pst (avrRps4) were grown on NYGA plates at 28°C for 48 h. The bacteria were collected and re-suspended in 5 mM MgCl₂. The leaves of 4-week-old plants were syringe infiltrated with 5 × 10⁴ cfu/ml Pst or Pst (avrRps4). Bacterial populations were determined by growth curve analysis at 3-day post-inoculation as described by Luo et al. (2017).

Root length analysis

For the analysis of root response to Fe deficiency, the seeds of each genotype were grown on +Fe medium for 5 days and then transferred to either +Fe or –Fe medium for 7 days (5 days +Fe, 7 days –Fe). The plants were photographed at 12 days after germination. ImageJ software (http://rsb.info.nih.gov/ij/) was used for root length measurements (Selote et al., 2015).

Perls staining for Fe in plants

For Fe staining, the leaves of 4-week-old plants grown in soil were detached and incubated with Perls fixing solution (methanol/chloroform/glacial acetic acid, 6:3:1) for 1–2 h under vacuum (500 mbar). Then, the tissues were further vacuum infiltrated with Perls staining solution (equal volumes of 4% [v/v] HCl and 4% [w/v] K-ferrocyanide) for 30 min. Leaves were then incubated for another 1 h in Perls staining solution at 37°C and washed three times with distilled water according to the method of Long et al. (2010). The stained leaves were photographed using a stereomicroscope with a CCD camera (Olympus BX51, Tokyo, Japan).

Fe content determination

Four-week-old plants were dipped with 1 × 10⁸ cfu/ml (0.025% Silwet L-77) Pst DC3000 or Pst DC3000 (avrRps4). Leaves were incubated for 4 min in buffer containing 5 mM CaSO₄ and 10 mM EDTA. The leaves were further rinsed with deionized water to remove trace surface metals. Leaves were dried at 80°C for 2 days, and 50–100 mg tissue was used for acid-based digestion. Briefly, all samples were digested with 5 mL concentrated nitric acid and 2 mL 30% hydrogen peroxide in microwave digestion system (ETHOS UP, Milestone, Italy). The digested samples were diluted to 20 mL with ddH₂O and analyzed by inductively coupled plasma-mass spectrometry (ICP-MS, Hitachi, Japan) for Fe analysis as described by Morrissey et al. (2009).

Apoplastic Fe extraction

Arabidopsis apoplastic washing fluid (AWF) extraction assays were performed as previously described by O'Leary et al. (2014), with slight modifications. Briefly, 4-week-old plants grown in soil were used for this experiment. Fully expanded rosette leaves were collected, weighed, and incubated for 4 min in buffer containing 5 mM CaSO₄ and 10 mM EDTA. The freshly excised leaves were rinsed with distilled water and maintained for 1 min under vacuum pumping at 7.5 psi in distilled water containing 0.004% silwet-77. The leaves were then rinsed with distilled water to remove the silwet-77. The apoplastic fluid was collected by centrifugation at 500g for 7 min at 4°C. The recovered AWF was pipetted into fresh 1.5-mL tubes and centrifuged at 15,000g for 5 min at 4°C. The supernatant was transferred to a fresh centrifuge tube for Fe content measurements by an ICP-MS (Inductively coupled plasma mass spectrometry).

IP and immunoblot analysis

The indicated constructs were expressed in N. benthamiana leaves by Agrobacteria-mediated transient expression. At about 48 h, the infiltrated leaves were sampled and total proteins were extracted with extraction buffer (50 mM Tris–HCl, 150 mM NaCl, 0.1% Triton, 0.2% NP-40, 6 mM 2-mercapto-Ethanol, and proteinase inhibitor cocktail (Roche), pH 7.5). The anti-FLAG IP was performed by incubating the proteins with 30 μL anti-FLAG ^® M2 Affinity Gel (Sigma-Aldrich, catalog # A2220) for 2 h on an end-over-end shaker at 4°C. The gels were washed five times with extraction buffer. The associated proteins on the gels were eluted with Laemmli buffer (0.0625 M Tris–HCl, 10% glycerol, 2% SDS (Sodium dodecyl sulfate), 0.0025% bromophenol blue, 5% 2-mercaptoethanol, pH 6.8). The anti-HA IP was performed by incubating the proteins with 40 μL anti-HA magnetic beads (Bimake, catalog # B26201, lot # 620027) in TBS buffer (50 mM Tris–HCl, 150 mM NaCl, proteinase inhibitor cocktail (Roche), pH7.4) overnight at 4°C on an end-over-end shaker. Magnetic beads-based IP was carried out according to the manufacturer’s instruction. After washing with PBST buffer (136.89 mM NaCl, 2.67 mM KCl, 8.1 mM Na₂HPO₄, 1.76 mM KH₂PO₄, 0.5% Tween-20) for five times, the beads were incubated with Laemmli buffer. The eluted proteins were separated by SDS–PAGE and revealed by immunoblot analysis using anti-FLAG (Sigma-Aldrich, catalog # F3165, lot # SLBQ7119V, 1:10000), anti-HA (Cell Signaling Technology, catalog # 3724S, 1:1000), anti-GFP (Transgen Biotech, catalog # HT801, lot # O11025, 1:5000), and anti-T7 (Abcam, catalog # ab9115, lot # GR3231183-2, 1:2,000) antibodies, respectively.

Split-LCI assay

The split-LCI assay was performed as described by Chen et al. (2008). Agrobacterium tumefaciens (strain EHA105) carrying the indicated nLUC and cLUC constructs was mixed and infiltrated into 4-week-old N. benthamiana leaves using a needleless syringe. Two days after infiltration, the leaves were rubbed with 0.5 mM luciferin and kept in the dark for 5 min to quench the fluorescence. The luminescence was recorded for 5–10 min by an in vivo Plant Imaging System with a CCD camera (NightShade LB985, Berthold technologies). The proteins were separated by SDS–PAGE and revealed by immunoblot analysis using cLUC (Sigma-Aldrich, catalog # L2164, lot # 076M4860V, 1:2000) and nLUC (Sigma-Aldrich, catalog # L0159, lot # 067M4869V, 1:10,000) antibodies, respectively.

RNA sequencing and bioinformatics analysis

RNAseq data for 4 hpi presented in this study were generated alongside the expression datasets for 8 and 24 hpi described in Bhandari et al. (2019). Plant genotypes, growth and infection conditions, RNA extraction, sequencing library preparation, and bioinformatic data analysis are detailed in Bhandari et al. (2019). Briefly, 4-week-old plants of Col-0 and eds1-2 grown under short-day conditions were syringe-infiltrated with Pst DC3000 (avrRps4) at OD₆₀₀=0.005. Fold change expression values at 4, 8, and 24 hpi relative to 0 hpi in Col-0 and eds1-2 are provided in Supplemental Data Set 1. Read data are available under GEO accession number GSE148304.

Parametric Analysis of Gene Set Enrichment (PAGE) method at AgriGO (Du et al., 2010) was applied to all expressed genes (Supplemental Data Set 1) using FDR correction for multiple testing and GO terms with at least 10 genes. A total of 295 Fe deficiency-induced genes were selected from an earlier analysis of Fe deficiency-induced changes in the Arabidopsis transcriptome (Rodriguez-Celma et al., 2013) (fold induction ≥2, t-test P value ≤0.05, Hub1+Hub2+Rib-leaf+LeafC1).

Protein transient expression in Arabidopsis protoplasts

The Fe regulatory genes bHLH104, bHLH115, and ILR3 were cloned and fused with a T7 epitope. The genes were then introduced into the pUC19 vector at Kpn I/BstB I restriction enzyme digestion sites. Similarly, the CDS of GFP and AvrRps4 was fused with a FLAG epitope and introduced into the pUC19 vector. All the ligations were performed using a One-step Cloning Kit (Vazyme Biotech Co., Ltd, C112). After sequencing, the confirmed constructs were transformed into Arabidopsis protoplasts. About 16 h later, the proteins were extracted using Laemmli buffer. The protein levels of FLAG- and T7-tagged proteins were detected using anti-FLAG (Sigma-Aldrich, catalog # F3165, lot # SLBQ7119V, 1:10000), anti-T7 (Abcam, catalog # ab9115, lot # GR3231183-2, 1:2000), and anti-actin (EASYBIO, catalog # BE0027, lot # 80790112, 1:5000) antibodies, respectively.

FCR activity assay

FCR activity assays were performed as previously described by Yi and Guerinot (1996) with slight modifications. Briefly, to measure the leaf FCR activity, 4-week-old plants grown in soil were used for this experiment. Five-millimeter diameter leaf disks were collected and weighed, and approximately 100 mg fresh weight tissue was incubated in tubes for 5 min in buffer containing 0.2 mM CaSO₄; then, 2 mL of assay solution (5 mM MES [pH adjusted to 5.5 with KOH],0.1 mM Fe(III)-EDTA, and 0.3 mM ferrozine, 10 mMCaSO₄) was added to the tubes. Then, the leaf disks were vacuum infiltrated for 10 min, and incubated in a shaking water bath at 50 rpm and 23°C in darkness. After incubation, the absorbance of the assay solution was determined in a spectrophotometer at 562 nm against an identical assay solution without any leaves (blank). Values were standardized using a blank assay solution without leaves. The concentration of purple-colored Fe(II)–ferrozine complex formation was calculated using the molar extinction coefficient of 28.6 mM⁻¹cm⁻¹ and leaf fresh weights to calculate Fe reduction activity.

BiFC assay

The pSAT1-nYFP and pSAT1-cYFP plasmids were used for the BiFC assay. The CDS of AvrRps4, AvrRps4^N_, and AvrRps4^C were cloned into pSAT1-cYFP at EcoRI/SmaI sites, and the CDS of BTS was cloned into pSAT1-nYFP at SalI/BamHI sites. A. tumefaciens (strain C58C1) carrying the indicated constructs was mixed and infiltrated into the leaves of 4-week-old N. benthamiana plants. Infiltrated leaves were observed 36–48 h later using a confocal laser scanning microscope (Leica Model TCS SP8).

Histochemical staining of Fe in plants

The histochemical staining of in situ Fe in Arabidopsis leaves was described previously (Roschzttardtz et al., 2009). The leaves of 4-week-old plants were syringe inoculated with 2.5 × 10⁶ cfu/mL Pst DC3000 or Pst (avrRps4) and sampled at 24 hpi. The leaves were fixed overnight in a fixing solution of 4% paraformaldehyde, 1% glutaraldehyde, and 0.1 M sodium phosphate buffer (pH 7.4) at 4°C. Then, the tissues were washed four times in 0.1 M sodium phosphate buffer (pH 7.4) and dehydrated in successive baths of 30, 50, 70, 90, 95, and 100% ethanol before embedding in the Technovit 7100. For light microscopy, transverse sections (0.5 μm) were obtained with a Leica RM2265 microtome. The sections were deposited on glass slides (Leica HI1210) and incubated for 45 min with Perls stain solution. The slides were washed twice for 5 min with distilled water and twice with 0.1 M sodium phosphate buffer (pH 7.4). For the DAB intensification reaction, the slides were incubated for 30 min in the dark at room temperature in the intensification solution (0.025% (w/v) DAB tetrahydrochloride [Sigma-Aldrich], 0.005% (v/v) H₂O₂ and 0.005% (w/v) CoCl₂, 0.1 M sodium phosphate buffer (pH 7.4)). The sections were washed three times for microscope observation.

MST assays

Binding affinity of recombinant MBP-BTS^1HHE and MBP-BTS^E3 to HIS-AvrRps4 and HIS-EDS1 was measured based on MST using a Monolith NT.Label Free instrument (Nano Temper Technologies GMBH, Germany), which detects changes in size, charge, and conformation caused by ligand/protein binding. The 10-μM labeled BTS^1HHE or BTS^E3 was displaced by a buffer containing 50 mM Tris–HCl (pH 7.4), 150 mM NaCl, 10 mM MgCl₂, and 0.05% (V/V) Tween-20. A range of concentrations of AvrRps4 (from 0.4 μM to 6.25 μM) or EDS1 (from 1.56 μM to 25 μM) in the assay buffer (50 mM Tris–HCl, 150 mM NaCl, 10 mM MgCl₂, 0.05% Tween-20, pH 7.8) was incubated with labeled protein (1:1, v/v) for 10 min. The sample was loaded into the NT. Label Free standard capillaries were measured with 20% LED power and 80% MST power. The KD Fit function of the Nano Temper Analysis Software (Version 1.5.41) was used to fit the curve and calculate the value of the dissociation constant (K_d).

Statistical analysis

All data are shown as mean ± standard deviation (sd) from no less than three biological repeats. Two-tailed Student’s t test with SPSS 18.0 was used for comparing means between two samples (*, **, and *** represent P ≤ 0.05, 0.01, and 0.001 levels, respectively). One or two-way ANOVA (Analysis of variance) with SPSS 18.0 was used for testing the significance of the difference among different group means (different letters indicate significant differences, P ≤ 0.05). Detailed statistical reports are shown in Supplemental Data Set 3.

Accession numbers

Sequence data from this article can be found in the Arabidopsis Genome Initiative or GenBank/EMBL databases under the following accession numbers: BTS (At3g18290), ILR3 (AT5g54680), bHLH115 (At1g51070), bHLH104 (At4g14410), EDS1 (AT3g48090), FIT (AT2G28160), PYE (AT3G47640), FRO2 (At1g01580), and NAS4 (At1g56430).

Supplemental data

The following materials are available in the online version of this article.

Supplemental Figure 1. bts-2 mutant plants are more susceptible than WT Col-0 to Pst (avrRps4).

Supplemental Figure 2. BTS interacts with the AvrRps4 C terminus.

Supplemental Figure 3. AvrRps4 stabilizes bHLH115 and ILR3 at protein level.

Supplemental Figure 4. MG132 did not markedly enhance AvrRps4-mediated bHLH115 and ILR3 stabilization.

Supplemental Figure 5. AvrRps4 specifically induced iron accumulation in the apoplast of eds1-2 plants.

Supplemental Figure 6. AvrRps4 recognition mutants displayed similar root growth responses to the bts mutant under iron deficiency.

Supplemental Figure 7. HopA1 does not target BTS for virulence.

Supplemental Data Set 1. The transcriptome of Col-0 and eds1 plants at 4 and 8 h after infiltration with Pst (avrRps4).

Supplemental Data Set 2. Iron metabolism-related Gene ontology (GO) term enrichments of WT (Col-0) and eds1 after Pst (avrRps4) infection.

Supplemental Data Set 3. Results of statistical analyses.

Supplemental Table 1. Primers used in this study.

 

J.L., G.C., and J.E.P. conceived and designed the experiments; Y.X., N.X., and X.S. performed most of the experiments; X.L. and Y.W. helped the data analysis; D.D.B. and D.L. performed the RNA-seq analysis; J.C. and H.W. helped with the bioinformatics and iron content assays; Y.X., J.L., G.C., J.E.P., and D.D.B. wrote the article.

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (https://academic.oup.com/plcell) is: Jun Liu (Junliu@im.ac.cn).

Funding

This study was supported by the Chinese Academy of Sciences (Strategic Priority Research Program Grant No. XDB11020300), Natural Science Foundation of China (31570252), the startup fund of “One Hundred Talents” program of the Chinese Academy of Sciences and by the grants from the State Key Laboratory of Plant Genomics (Grant No. O8KF021011) to J.L. G.C. was supported by grants from the National Institutes of Health (R35GM136402) and the USDA-NIFA (2015-67013-23082). D.D.B., D.L., X.S., and J.E.P. were supported by the Max-Planck Society, Germany’s Excellence Strategy CEPLAS (EXC-2048/1, Project 390686111) and Deutsche Forschungsgemeinschaft (DFG; German Research Foundation) CRC-680 project B10 and CRC-1403414786233.

Conflict of interest statement. None declared.