Exportin-4 coordinates nuclear shuttling of TOPLESS family transcription corepressors to regulate plant immunity

Exportin-4 defines a nuclear transport pathway that modulates SA-dependent immune amplification in effector-triggered immunity through fine-tuning nuclear levels of TPL/TPR transcription corepressors.

Abstract

The regulated nucleocytoplasmic exchange of macromolecules is essential for the eukaryotic cell. However, nuclear transport pathways defined by different nuclear transport receptors (NTRs), including importins and exportins, and their significance in activating distinct stress responses are poorly understood in plants. Here, we exploited a CRISPR/Cas9-based genetic screen to search for modifiers of CONSTITUTIVE EXPRESSION OF PATHOGENESIS-RELATED GENE 5 (cpr5), an Arabidopsis thaliana nucleoporin mutant that activates autoimmune responses that partially mimic effector-triggered immunity (ETI). We identified an NTR gene, Exportin-4 (XPO4), as a genetic interactor of CPR5. The xpo4 cpr5 double mutant activates catastrophic immune responses, which leads to seedling lethality. By leveraging the newly developed proximity-labeling proteomics, we profiled XPO4 substrates and identified TOPLESS (TPL) and TPL-related (TPR) transcription corepressors as XPO4-specific cargo. TPL/TPRs target negative regulators of immunity and are redundantly required for ETI induction. We found that loss-of-XPO4 promotes the nuclear accumulation of TPL/TPRs in the presence of elevated salicylic acid (SA), which contributes to the SA-mediated defense amplification and potentiates immune induction in the cpr5 mutant. We showed that TPL and TPRs are required for the enhanced immune activation observed in xpo4 cpr5 but not for the cpr5 single-mutant phenotype, underscoring the functional interplay between XPO4 and TPL/TPRs and its importance in cpr5-dependent immune induction. We propose that XPO4 coordinates the nuclear accumulation of TPL/TPRs, which plays a role in regulating SA-mediated defense feedback to modulate immune strength downstream of CPR5 during ETI induction.

Introduction

Plant responses to diverse environmental stimuli are highly dependent on the regulated transport of a variety of signaling molecules in and out of the nucleus, which directs transcriptional reprogramming to confer stress resistance. The transport process is mediated by karyopherin proteins, a superfamily of nuclear transport receptors (NTRs) including importins and exportins that selectively bind cargos and facilitate their translocation through the nuclear pore complex (NPC). It has been well established in yeast and humans that importin-α family proteins recognize the classical nuclear localization signal of protein cargo and function as adaptors between cargo and importin-β proteins (Goldfarb et al., 2004). Importin-βs can physically interact with the NPC selective barrier and navigate the cargo/importin-α/importin-β ternary complex across the NPC (Bayliss et al., 2000; Aramburu and Lemke, 2017; Tan et al., 2018). Importin-βs may also recognize and bind cargo directly for nuclear transport independent of importin-α (Ullman et al., 1997; Chook and Suel, 2011), and based on the transport direction, importin-βs are further classified into importins (IMBs) and exportins (XPOs). The small GTPase protein Ran differentially regulates cargo binding of IMBs and XPOs, which determines the transport direction. IMBs bind Ran-GTP in the nucleus to facilitate release of imported cargo, whereas XPOs bind Ran-GTP in the nucleus to promote interaction with cargo for nuclear export (Meier and Somers, 2011).

At least 15 subfamilies of importin-β are predicted to have been established in the last eukaryotic common ancestor (O'Reilly et al., 2011) and were presumably evolved to permit differential cargo transport. The Arabidopsis genome has 18 importin-β genes, representing most of the ancestral and one plant-specific importin-β subfamily (Tamura and Hara-Nishimura, 2014). They appear to play critical roles in various aspects in plant development and stress responses. For example, mutations in six importin-β genes have been reported to affect plant growth and development, including mutations in XPOT/PAUSED that disrupt initiation of the shoot apical meristem and delay root emergence, leaf initiation, and flowering transition (Hunter et al., 2003), mutations in XPO5/HASTY that accelerates the vegetative phase change and reduce fertility (Bollman et al., 2003), and mutations in IMB3, IMB4, XPO1A, and XPO1B which impair gametophytic fertility (Blanvillain et al., 2008; Liu et al., 2019; Xiong et al., 2020a). Four importin-β genes have been reported to be involved in plant stress responses, including IMB1 and Sensitive to ABA and Drought 2(SAD2) that are involved in responses to abscisic acid and drought stress (Verslues et al., 2006; Zhao et al., 2007; Luo et al., 2013), XPO1A that is involved in heat-induced oxidative stress (Wu et al., 2010), and Modifier of suppressor of npr1-1, constitutive 1 (that is involved in innate immune responses (Xu et al., 2011). Moreover, knocking down IMB2 and IMB3 leads to reduced miRNA activity and miRNA biogenesis in the nucleus, respectively (Cui et al., 2016; Zhang et al., 2017).

Despite extensive phenotype characterizations of importin-β mutants, so far, only a few cargos of importin-β proteins have been identified in plants (Zhao et al., 2007; Xu et al., 2011; Zhang et al., 2017; Liu et al., 2019; Panda et al., 2020; Xiong et al., 2020a, 2020b). The shortage of specific cargo spectrum information for individual importin-β proteins and the lack of an efficient cargo profiling method due to the transient nature of NTR-cargo interaction has limited our understanding of the mechanism underlying distinct phenotypes of individual importin-β mutants and of potential specialized functions of each NTR in activating various stress responses, such as immunity.

Effector-triggered immunity (ETI) is a conserved and critical innate immune response, which is activated by the nucleotide-binding leucine-rich repeat (NLR) receptors in plants upon recognition of secreted pathogen proteins called effectors (Jones and Dangl, 2006; Stuart et al., 2013; Zhou and Zhang, 2020). ETI is a robust immune response and usually results in localized programmed cell death (PCD) in the plant cell. Previously, it was reported that the plant-specific membrane nucleoporin CONSTITUTIVE EXPRESSION OF PATHOGENESIS-RELATED GENE 5 (CPR5) functions downstream of NLRs at the NPC and plays a critical role in gating immune activation in Arabidopsis. Loss of CPR5 results in spontaneous PCD and transcriptome features similar to those that are induced by NLR activation and confers effective resistance to pathogens carrying effector proteins even in the absence of cognitive NLR receptors, suggesting induction of typical ETI responses in the cpr5 mutant (Bowling et al., 1997; Boch, 1998; Wang et al., 2014; Gu et al., 2016). CPR5 function compromised NLR activation, which induces NPC conformational changes that promote nuclear accumulation of immune cargos and overactivate E2F transcription factors to induce defense gene expression (Wang et al., 2014; Gu et al., 2016). These and other findings have pointed out a critical role of the NPC and the associated nucleocytoplasmic transport in regulating ETI activation (Palma et al., 2005; Cheng et al., 2009; Garcia and Parker, 2009; Wiermer et al., 2012; Buscaill and Rivas, 2014; Roth et al., 2017). However, how importin-βs and importin-β-dependent nucleocytoplasmic transport are directly involved in regulating ETI induction has not been established.

Here, we designed a CRISPR guide RNAs (gRNA) library to target Arabidopsis NTR genes in the cpr5 mutant background and searched for modifiers of the cpr5 phenotype. We identified an importin-β gene (XPO4) as a genetic interactor of CPR5. We showed that loss of XPO4 dramatically enhanced the PCD and SA-mediated defense gene expression in cpr5 plants, which resulted in seedling lethality. By combining TurboID-based proximity labeling technology, label-free quantitative mass spectrometry, and ratiometric analysis (Branon et al., 2018; Huang et al., 2020; Tang et al., 2020), we established the in vivo cargo spectrum of XPO4 and other exportins and determined TOPLESS (TPL) and TPL-related (TPRs) transcription corepressors as XPO4-specific cargo. TPL/TPRs are known to be involved in multiple stress- and hormone-related signaling pathways (Long et al., 2006; Szemenyei et al., 2008; Pauwels et al., 2010) and are required for the activation of ETI mediated by multiple NLR proteins, including SUPPRESSOR OF npr1, CONSTITUTIVE 1 (SNC1; Zhu et al., 2010). We showed that XPO4 interacts with TPR1 and inhibits its nuclear accumulation in the presence of high levels of salicylic acid (SA) likely through constitutive nuclear export. We also found that xpo4-induced enhancement of the cpr5 phenotype is dependent on TPL and TPRs as the tpl tpr1 tpr4 triple mutant can rescue the detrimental immune activation in xpo4 cpr5 plants. Our data suggest that coordinated XPO4 activity regulates the nuclear level of TPL/TPRs upon SA accumulation, which plays a role in the SA-mediated defense amplification loop to modulate immune strength downstream of CPR5 during ETI induction.

Results

A CRISPR/Cas9-assisted genetic screen identified XPO4 as a genetic interactor of CPR5

To better understand the role of NTR-dependent nuclear transport in cpr5-mediated autoimmune activation, we conducted a targeted genetic screen for modifiers of cpr5. We designed 27 pairs of CRISPR gRNAs that target two different loci in each of the 27 NTR genes (9 importin-αs and 18 importin-βs) in Arabidopsis (Figure 1A and Supplemental File 1). Each pair of gRNAs was cloned individually in a vector carrying a maize codon-optimized Cas9 gene driven by the CaMV 35S constitutive promoter (Xing et al., 2014). To screen cpr5 phenotype modifiers, we pooled the 27 CRISPR vectors and bulk-transformed this vector library into cpr5 mutant plants using Agrobacteria-mediated transformation (Figure 1B).

Figure 1.

A targeted genetic screen for modifiers of cpr5 by CRISPR/Cas9-mediated knockout of NTR genes. (A) Phylogenetic trees of Arabidopsis important-α and important-β family proteins. Neighbor-joining trees were generated with full-length amino acid sequences using MEGA 10.8 with JTT model and 10,000 bootstraps. Scale bars represent amino acid substitutions per site. XPO4 is labeled red. (B) A schematic diagram for the CRISPR/Cas9-assisted genetic screen for modifiers of cpr5. A pair of gRNAs was designed to target two different loci of the same NTR gene. A total of 27 pairs of gRNAs that target 27 NTR genes were individually cloned into a CRISPR vector backbone that contains a zCas9 gene driven by the CaMV 35S constitutive promoter. The resulting vector library was bulk transformed into cpr5 mutant plants, and T1 transformants were identified in soil supplemented with Basta. Modifiers of cpr5 may be identified in T1 or T2 population. Genomic DNA of identified cpr5 modifiers was extracted, and the gRNA cassette of the transgene and the genomic region of the corresponding target gene were amplified and sequenced. (C) Simulated relationship between the number of T1 transformants obtained in the screen and the probability of saturating all constructs in the CRISPR library. The simulation process is based on multinomial distribution in the probability theory and realized by python. The simulation result was plotted by R. Blue solid lines indicate the number of T1 transformants needed to saturate the library with a 90% probability. Red dashed lines indicate that obtaining 151 T1 transformants almost certainly saturates the library. (D) The representative phenotype of a cpr5 mutant plant and a 4-week-old cpr5 enhancer identified from the screen. This particular plant is confirmed to be xpo4-1 cpr5 double mutant by sequencing. The plant in the right box shows an enlarged image of the enhancer mutant. Bar = 2 cm.

We obtained a total of 151 T1 transformants. Under the premise that each transformant carries at least one gRNA construct, this population is expected to have saturated the vector library (Figure 1C). Interestingly, although we aimed to identify modifiers in the T2 population, we were able to find three potential cpr5 enhancers showing dramatic phenotypes in T1 transformants (Figure 1D). Mapping the cpr5 enhancer mutations would be infeasible if they were introduced by random mutagenesis, because cpr5 enhancer lines were unlikely to survive until reproduction due to strong autoimmune induction and severe PCD (Figure 1D). In contrast, the CRISPR/Cas9-dependent mutagenesis allowed us to identify the potential enhancer genes by recovering the gRNA information directly from the enhancer plants. Genomic DNA from the three potential enhancer lines was extracted, and the gRNA cassette in the CRISPR vector was amplified and sequenced. Remarkably, the three independent enhancer lines showing similar phenotypes contained the same gRNAs that target XPO4 (AT3G04490), a homolog of mammalian Exportin-4. These data suggest that potential mutations in XPO4 may enhance the cpr5 phenotype. Sequencing of XPO4 genomic region in the three enhancer lines revealed the same base editing by Cas9 in the first exon of XPO4, which resulted in a premature stop codon at the position of the 16th amino acid (Figure 2A), suggesting a highly reproducible editing pattern for this particular targeting sequence. We named this mutation as xpo4-1.

Figure 2.

xpo4 is a genetic enhancer for the cpr5-dependent autoimmune activation. (A) The gene structure of AtXPO4 and xpo4 mutations. In the xpo4-1 mutant, an adenine is inserted after +23 bp, which leads to a premature stop codon (E16 to stop) in the XPO4 protein. In the xpo4-2 mutant, a T-DNA is inserted in the 5th intron of the XPO4 gene. RT-qPCR using WT and xpo4-2 plants. Ubiquitin 5 was used as the reference gene, and the relative expression level of XPO4 was normalized by that of WT. Data are presented as mean ± standard deviation (sd; n = 2 biological replicates). Student’s t test was performed using WT as control (****P < 0.0001). (B) Three-week-old plants grown in soil (left) and 10-day-old seedlings grown on 1/2 MS media (right). Bars = 2 cm. (C) Measurement of rosette diameter of 4-week-old plants. Data are presented as mean ± sd for eight individual plants. Statistical analysis was performed using two-way ANOVA, ****P < 0.0001, and ns stands for not significant. (D) Principal component analysis (PCA) and hierarchical clustering of the global transcriptome profiling data obtained from RNA-seq (GSE146914) using 10-day-old seedlings with three biological replicates for each genotype. (E) Venn diagram showing overlaps of upregulated DEGs obtained from indicated analyses. All DEGs were selected using cutoffs P <0.05 and fold-change >2 comparing to controls. Most cpr5-dependent up-regulated DEGs are also upregulated in the xpo4-2 cpr5 double mutant (598/789), and the expression of approximately one-third of the overlapping genes (190/598) was further enhanced in xpo4-2 cpr5 compared with cpr5 (190 genes labeled in red). (F) Expression levels of the 190 genes in WT, cpr5, and xpo4-2 cpr5 plants in three biological replicates and their GO enrichment terms. Defense and cell death-related GO terms are labeled in red. TPM, transcripts per million. FDR, false discovery rate.

To validate that the enhancer phenotype is caused by mutations in XPO4, we obtained a T-DNA knockdown mutant for XPO4 (SALK_093159, xpo4-2) and crossed it with the cpr5 mutant. Although the xpo4-2 single mutant did not show apparent developmental phenotypes compared with wild-type (WT) plants, the xpo4-2 cpr5 double mutant displayed severely restricted growth, which eventually led to seedling lethality (Figure 2B), similar to what was observed in the xpo4-1 cpr5 enhancer line (Figure 1D). This deleterious effect was not seen when cpr5 was crossed with mutants of another two exportins, including XPO5 and XPO7 (close homologs of XPO4), suggesting a specific functional connection between XPO4 and CPR5 (Figure 2C and Supplemental Figures 1, A–C).

CPR5 is a transmembrane nucleoporin and physically and genetically interacts with other components of the NPC, such as Nup93a and Nup85, to influence the NPC function (Gu et al., 2016). However, XPO4 did not confer obvious genetic interactions with Nup93a, Nup85, or other nucleoporins that we tested (Supplemental Figure 1D). Therefore, it is unlikely that the genetic interaction between XPO4 and CPR5 is based on functional connections between XPO4 and structural components of the NPC.

XPO4 is required to repress immune gene expression activated by loss-of-CPR5

To better understand the enhancement phenotype in the xpo4 cpr5 mutant, we performed whole-genome transcriptome analyses on 10-day-old seedlings in WT, xpo4-2, cpr5, and xpo4-2 cpr5 background using RNA-seq. Principal component analysis (PCA) and hierarchical clustering illustrated a clear effect of the genetic interaction between XPO4 and CPR5 at the level of global gene expression (Figure 2D). Detailed analysis of differentially expressed genes (DEGs, fold-change >2 and P <0.05, linear model F-test) revealed that mutating XPO4 dramatically altered the cpr5 transcriptome by introducing an additional 1,082 upregulated and 1,446 downregulated DEGs that are independent of DEGs in either the xpo4 or cpr5 single mutant (Supplemental Figure 2 and Supplemental Data Set 1).

Comparative analysis showed that 76% (598/789) of cpr5-dependent upregulated DEGs are also upregulated in the xpo4-2 cpr5 double mutant (Figure 2E, left). However, the expression level of almost 1/3 (190/598) of these overlapping DEGs is significantly enhanced in xpo4-2 cpr5 compared with cpr5 (Figure 2, E, right and 2, F, middle), even though their expression was already highly induced in the cpr5 mutant compared with WT (Figure 2F, left). Gene ontology (GO) analysis revealed that these further induced genes are highly enriched in PCD-related processes and defense, including SA-mediated immune responses (Figure 2F, right). Moreover, xpo4-2 cpr5 also activates expression of additional defense genes that are not induced in the cpr5 single mutant, including those that are significantly enriched in SA responses (Supplemental Figure 2, A and B). These analyses demonstrate that XPO4 plays a critical role in negatively regulating the SA-mediated defense gene expression in the cpr5 mutant.

XPO4 is required for basal resistance and NLR-mediated resistance but not for PTI

To further investigate the functional importance of XPO4 in plant immunity, we tested the defense responses of the xpo4-2 single mutant against the virulent bacterial pathogen Pseudomonas syringae pv. tomato DC3000 (Pst DC3000). Compared with the WT and other exportin mutants, xpo4-2 was the only exportin mutant that displayed an altered defense response in this assay. Surprisingly, we observed significantly higher susceptibility rather than resistance to Pst DC3000 in the xpo4-2 mutant compared with WT and other NTR mutants (Figure 3A and Supplemental Figure 3A). To confirm this result, we used CRISPR to re-generate the xpo4-1 mutation in the WT background. The xpo4-1 single mutant also showed susceptibility to Pst DC3000, at a similar level to xpo4-2 (Figure 3B). Moreover, the observed immune deficiency in the xpo4-2 mutant can be fully complemented by an XPO4 genomic construct. These data suggest that XPO4 is required for basal immune induction in plants. A T-DNA insertion mutant of XPO7, an exportin homolog with the highest protein sequence similarity to XPO4, showed a WT level of resistance against Pst DC3000, and the xpo4-2 xpo7 double mutant did not further enhance the xpo4-mediated susceptibility (Figure 3C), highlighting a specific role of XPO4 in immune induction.

Figure 3.

XPO4 is required for basal and NLR-mediated immunity but not for PTI. (A) Growth of the bacterial pathogen Pst DC3000 on exportin mutants. WT and the npr1 mutant were included as controls. (B) Growth of Pst DC3000 on WT, xpo4-2, xpo4-1, and two independent pXPO4-XPO4-3×HA-TurboID lines in the xpo4-2 background (homozygous T3 plants were used). (C) Growth of Pst DC3000 on WT, xpo4-2, xpo7, and xpo4-2 xpo7 plants. (D) Growth of Pst DC3000 carrying effector AvrRps4 or AvrRpt2 on WT, xpo4-2, and xpo7 plants. (E) Growth of Pst DC3000 ΔhrcC strain on WT, xpo4-2, and xpo7 plants using infiltration or spray inoculation. For pathogen assays in A–E, the fifth and sixth leaves of 3-week-old Arabidopsis plants were inoculated with bacterial suspension culture at OD_600nm = 0.001 (A, B, and D), 0.0001 (C), or 0.04 (E) using infiltration or at OD_600nm = 0.2 (E) using spray. The growth of bacteria was counted at 3 or 4 dpi as indicated. CFU, colony-forming units. Data are presented as mean ± sd for five or six biological replicates. Student’s t tests were performed on 3 or 4 dpi data using Col-0 as control, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. NS stands for not significant. Red dots indicate individual data points. (F) Possible dual regulatory role of XPO4 (labeled in red) in ETI activation. Activation of NLR receptors disrupts CPR5 oligomerization and induces NPC conformational change to activate ETI signaling. Following ETI activation, SA is synthesized and leads to transcriptome reprogramming to substantiate ETI. XPO4 plays a negative regulatory role in the SA-mediated defense amplification during ETI downstream of CPR5 and the NPC. Meanwhile, loss of XPO4 compromises NLR-mediated resistance, suggesting that XPO4 is also required for initiating ETI signaling, likely upstream of CPR5 and the NPC.

To further dissect the immune deficiency in the xpo4 single mutant, we inoculated xpo4-2 plants with avirulent Pst DC3000 carrying the effector protein AvrRps4 or AvrRpt2. These bacterial effectors can be recognized by NLR protein RPS4 (Saucet et al., 2015) and RPS2 (Kunkel et al., 1993), respectively, and induce ETI. We found that RPS4- and RPS2-mediated resistance were significantly compromised in the xpo4-2 mutant (Figure 3D). In contrast, xpo4-2 showed the WT level of resistance against the Pst DC3000 ΔhrcC strain, which lacks the type III secretion system and cannot deliver pathogen effectors into the plant cell (Figure 3E). This result suggests an intact pathogen-associated molecular pattern (PAMP)-triggered immunity (PTI) in the absence of XPO4. Consistently, the xpo4 mutant was sensitive to elf18, a PAMP that induces PTI, and showed stunted growth in the presence of elf18 and WT responses in elf18-induced resistance against Pst DC3000 (Supplemental Figure 3, B–D). These data suggest that XPO4 is required for NLR-mediated resistance but not for PTI.

Loss of XPO4 results in opposite effects in cpr5-induced and pathogen-induced ETI responses, which suggests that XPO4 may play distinct roles in multiple steps during ETI activation. One possible scenario is that XPO4 coordinates the nuclear shuttling of signaling cargos both upstream of CPR5/NPC to initiate ETI induction and downstream of CPR5/NPC to repress the potentiation of ETI by SA-mediated defense amplification (Figure 3F). Here, the cpr5 mutant provides a unique genetic background to unravel the role of XPO4 in ETI downstream of CPR5.

Profiling the cargo spectrum of XPO4 using proximity-labeling proteomics

To further investigate the mechanism of how XPO4 may participate in ETI regulation, we sought to obtain the in vivo cargo information of XPO4. However, capturing the full-range cargo spectrum of an NTR is particularly challenging with the conventional coimmunoprecipitation method due to the transient nature of the NTR-cargo association: NTRs bind their cargos during nuclear transport and release them after arriving at the cargo destination (Meier and Somers, 2011; Sun et al., 2013). As a result, the complete cargo spectrum information for any NTR is not known in plants. Here, we leveraged a recently developed enzyme-catalyzed proximity labeling strategy, namely TurboID (Branon et al., 2018), which uses a promiscuous biotin ligase to efficiently and irreversibly label proteins upon association with the bait. Tagging XPO4 with TurboID allows us to comprehensively probe XPO4 cargos, regardless of their dissociation upon the completion of nuclear transport (Figure 4A).

Figure 4.

Profiling of in vivo XPO4 cargo using proximity-labeling proteomics. (A) Schematic of proximity labeling of XPO4 cargo using XPO4-TurboID. (B) The pXPO4-XPO4-3×HA-TurboID construct used to generate transgenic plants. The XPO4 native promoter (950-bp upstream of the start codon) and genomic DNA (start codon to stop codon) were used. (C) Proximity labeling coupled label-free quantitative mass spectrometry (LFQMS). Rosette leaves of 3-week-old T3 homozygous pXPO4-XPO4-3×HA-TurboID transgenic Arabidopsis (Line 1) were infiltrated with 50 μM free biotin for proximity labeling (H₂O treatment as control). Six hours later, the total protein was extracted and subjected to desalting chromatography to remove free biotin. The protein fraction was collected for AP using streptavidin-coated beads. After AP, 10% of each sample was run on SDS–PAGE and immunoblotted with HRP-conjugated streptavidin and anti-HA antibody. The remaining samples were digested with trypsin and subject to LFQMS. Asterisks represent naturally biotinylated proteins. (D) Volcano plot of peptide intensity data from LFQMS. Proteins enriched in biotin-treated samples (red dots) were selected using PSM >2, fold-change (biotin/mock) >3, and P < 0.05 as cutoffs. (E) Heatmap of normalized and averaged PSM values of proteins that are significantly enriched in biotin-treated samples. The heatmap is ordered by ranking the fold-change (biotin/mock) using averaged PSM. FG nucleoporins are labeled in red.

We generated stable transgenic plants expressing native promoter-driven XPO4 tagged with TurboID (Figure 4B) and used the T3 homozygous plants that complemented the pathogen susceptibility in the xpo4-2 mutant (Figure 3B) for proximity-labeling proteomics. Rosette leaves of 3-week-old pXPO4-XPO4-3×HA-TurboID plants were infiltrated with 50-µM free biotin or water (mock), and 6 h later, the total protein from infiltrated leaves was extracted. Two biological replicates for both biotin- and mock-treated samples were collected. After free biotin removal from the total protein extract using desalting chromatography, biotinylated proteins were affinity-purified using streptavidin-coated beads. Immunoblotting with 10% sample demonstrated that biotin treatment led to significant protein biotinylation, including XPO4 itself, suggesting an efficient and inducible biotin labeling (Figure 4C). The remaining samples were subjected toSDS–PAGE (sodium dodecyl sulfate polyacrylamide gel electrophoresis), trypsin digestion, and label-free quantitative mass spectrometry (MS). Proteins that are significantly enriched in biotin-treated samples were selected by comparing the peptide spectrum level with mock-treated samples using cutoffs fold-change >3 and P < 0.05, and a total of 244 candidates were identified (Figure 4D and Supplemental Data Set 2).

The identified candidates are highly enriched in a specific type of NPC components, phenylalanine–glycine (FG) nucleoporins (Figure 4E, red label). FG Nups interact directly with NTRs and mediate their translocation through the NPC (Bayliss et al., 2000; Tan et al., 2018). The seven identified FG Nups are distributed throughout the NPC (Tamura and Hara-Nishimura, 2013), including the cytosolic filament (Nup214 and Nup98a), the central channel (Nup54), and the nuclear basket (Nup82, Nup50a, Nup50b, and Nup50c), reflecting the entire transportation route of XPO4 across the NPC. This result demonstrates that the proximity labeling by XPO4 occurred in vivo and validates the quality of ourMS profiling.

Beside FG Nups, the top XPO4 cargo candidates ranked by the level of peptide enrichment (biotin versus mock) contain four members of TPL and TPR proteins, including TPL, TPR1, TPR2, and TPR4 (Figure 4E). Notably, it has been reported that TPL, TPR1, and TPR4 are transcription corepressors that redundantly contribute to ETI activation by multiple NLR receptors (Zhu et al., 2010). Identification of TPL and TPRs as XPO4 substrates led us to hypothesize that XPO4 mediates the nuclear export of TPL family proteins and that their protein levels in the nucleus are deregulated in the absence of XPO4 to influence cpr5-dependent immunity.

TPL and TPRs are XPO4-specific cargo

To test the above hypotheses, we first addressed whether TPL and TPRs are specific cargos of XPO4. Because XPO5 and XPO7 did not display discernable genetic interactions with CPR5 (Figure 2C), we generated transgenic lines expressing native promoter-driven and TurboID-tagged XPO5 and XPO7 and performed a second round of proximity-labeling proteomics using XPO4-TurboID, XPO5-TurboID, and XPO7-TurboID transgenic plants in parallel (Supplemental Figure 4A). To capture a comprehensive cargo population in different tissues, 10-day-old whole seedlings were sampled after proximity labeling (Supplemental Figure 4B). Quantitative MS results from XPO4, XPO5, and XPO7 samples were normalized and subjected to pairwise ratiometric analysis using fold-change >2 and P <0.05 (linear model, F-test) as cutoffs. We identified 117, 48, and 66 proteins that are specifically enriched in XPO4, XPO5, and XPO7 samples, respectively, using samples of the other two XPOs as control (Figure 5A, Supplemental Figure 4C and D, and Supplemental Data Set 3).

Figure 5.

TPL family transcription corepressors are XPO4-specific substrates (A) Ratiometric analysis of proximity-labeling proteomic data for XPO4-specific substrates using XPO5 and XPO7 as controls. The proximity labeling was performed using 10-day-old transgenic seedlings expressing native promoter-driven XPO4-TurboID, XPO5-TurboID, and XPO7-TurboID, respectively. For the analysis, two biological replicates were included for XPO4 and XPO5 samples and one for XPO7 sample. XPO4-specific preys (red dots) were selected using cutoffs PSM >1, P <0.05, and fold-change >2 compared with both XPO5 and XPO7. (B) The level of TPL and four TPRs probed by XPO4-TurboID, XPO5-TurboID, and XPO7-TurboID in proximity-labeling proteomics. Normalized PSM values are shown. Student’s t tests were performed using XPO4 samples as control, **P < 0.01 and ***P < 0.001. Statistical analysis for XPO7-TurboID sample is not available as data from only one sample was obtained. (C) Co-IP assays using pXPO4-XPO4-3xHA-TurboID and pXPO7-XPO7-3xHA-TurboID transgenic plants. Total protein was extracted from 10-day-old seedlings and incubated with anti-HA antibody and Protein A beads. Protein sample after IP was immunoblotted with anti-HA and anti-TPR1 antibodies. (D) GO analysis of downregulated DEGs in the xpo4-2 mutant compared to WT. GO terms with P <10⁻⁵ were expanded. RNA-seq (GSE146914) was performed using three biological replicates for each genotype. (E) Expression levels of selected genes involved in JA biosynthesis and signaling pathways in WT and the xpo4-2 mutant. Data were obtained from RNA-seq analysis. p.adj, adjusted P value (linear model, F-test).

Remarkably, all TPL family proteins, including TPL and the four TPRs, were identified as significantly enriched preys of XPO4 using XPO5 and XPO7 samples as control (Figure 5, A and B), suggesting that their nuclear translocation is predominantly mediated by XPO4 rather than XPO5 or XPO7. Indeed, when we performed an immunoprecipitation (IP) assay with anti-HA antibody using pXPO4-XPO4-3xHA-TurboID plants, we detected the co-IPed TPR1 with a custom antibody against TPR1 (Figure 5C), which may also recognize TPL and other TPRs (Niu et al., 2019). In contrast, TPL/TPRs barely interacted with XPO7 in the co-IP experiment using pXPO7-XPO7-3xHA-TurboID plants.

Aside from the reported role of TPL and TPRs in ETI, TPL has long been known as a critical transcription corepressor that associates withJAZ (JASMONATE ZIM-DOMAIN) proteins to repress jasmonic acid (JA) responses (Pauwels et al., 2010). Consistently, the unchallenged xpo4 mutant displayed a clear transcriptome signature with downregulated JA responses compared with WT (Figure 5, D and E). Repressed basal JA responses in the xpo4 mutant support the hypothesis that the nuclear accumulation of TPL, and likely also TPRs, is misregulated in the absence of XPO4.

XPO4 regulates the nuclear accumulation of TPR1 in response to elevated SA

TPL and TPRs activate ETI through repressing the expression of negative regulators of immunity (e.g. DND1 and DND2), and overexpression of TPR1 is sufficient to activate immune responses (Zhu et al., 2010; Niu et al., 2019). However, the xpo4 single mutant did not display an autoimmune phenotype (Figure 2B). To directly measure the nuclear accumulation of TPR1, we extracted nuclear protein from WT and the xpo4 mutant and immunoblotted with the TPR1 antibody. We found that TPR1 is barely detected in the nucleus (Figure 6A). However, a decent level of TPR1 was observed in both the total protein extract and the cytosolic fraction, suggesting that TPR1 is predominantly excluded from the nucleus under normal conditions.

Figure 6.

XPO4 negatively regulates cpr5-dependent immune responses by inhibiting the nuclear accumulation of TPR1 in response to elevated SA. (A) The total protein from 10-day-old WT and xpo4-2 seedlings was extracted for nuclear fractionation. The total protein extract (total), the nuclear fraction (nucleus), and the fraction devoid of nucleus (cytosol; see Materials and methods Section for details) were then subject to immunoblotting with anti-TPR1, anti-Actin, and anti-Histone H3 antibodies. (B) Ten-day-old WT and xpo4-2 seedlings were treated with 150 μM SA for the indicated time before the total protein extract, and the nuclear fraction was subjected to immunoblotting. (C) The relative expression level of PR1, EDS1, EDS5, and ICS1 in 2-week-old seedlings 12 h after 150 µM SA treatment. RT-qPCR was performed, and Actin was used as the reference gene. The gene expression level was normalized to that in plants without SA treatment. Each replicate contains five seedlings. Data are represented as the means ± sd (n = 2 biological replicates). The P values (Student’s t tests) using WT as control are labeled. (D, E) Suppression of the enhanced autoimmune phenotype in xpo4-2 cpr5 (xc) by the tpl tpr1 tpr4 triple mutant (ttt). The cpr5 single mutant (c) phenotype was not further alleviated in the cpr5 tpl tpr1 tpr4 quadruple mutant (cttt). Image (D) and measurement of rosette diameter (E) of 2-week-old plants are shown. Bar = 2 cm. Student’s t test was performed. (F) The relative expression level of PR1, EDS1, ICS1, and EDS5 in 2-week-old seedlings. RT-qPCR was performed, and Actin was used as the reference gene. The gene expression level was normalized to that in WT plants. Each replicate contains five seedlings. Data are represented as the means ± sd (n = 2 biological replicates). Similar results have been obtained twice. *P < 0.05, **P < 0.01, ***P < 0.001.

ETI activation is accompanied by biosynthesis and accumulation of SA, which activates SA-mediated defense amplification to substantiate ETI responses (Zhang and Li, 2019). To test whether XPO4 plays a role in TPR1 nuclear accumulation in the presence of high levels of SA, we treated 10-day-old plate-grown Arabidopsis seedlings with 150 μM SA for 12 h before nuclear fractionation was performed. We found that after SA treatment, the nuclear accumulation of TPR1 became readily detectable and substantially higher in the xpo4 mutant compared with WT (Figure 6B). The enhanced nuclear accumulation of TPR1 in xpo4 plants is evident at different time points after SA treatment. This result indicates that elevated SA can promote the nuclear shuttling of TPR1, and importantly, XPO4 is required to reduce TPR1 nuclear accumulation during this process, most likely by mediating nuclear export of the protein.

The cpr5 mutant accumulates high levels of SA in vivo (Bowling et al., 1997) presumably due to ETI activation; thus, we propose that loss of XPO4 in the cpr5 mutant can lead to hyperaccumulation of nuclear TPR1 in the presence of elevated SA and consequently result in the detrimental immune phenotype in the xpo4 cpr5 double mutant. Note that loss-of-XPO4 potentiates SA-mediated defense gene expression in the cpr5 mutant background (Figure 2F;Supplemental Figure 2B), but this effect cannot be replicated simply by treating the xpo4 mutant with exogenous SA (Figure 6C), suggesting that amplification of SA responses by loss-of-XPO4 requires cpr5-dependent immune induction first and may function downstream of CPR5 and the NPC during ETI signaling (Figure 3F).

TPL and TPRs are required for the enhanced autoimmune phenotype in the xpo4 cpr5 mutant

Although it is impractical to directly measure TPR1 accumulation in the nucleus of xpo4 cpr5 plants due to the difficulty in extracting nuclei from the severely stunted double mutant plants, we assessed the functional importance of TPL and TPRs to the xpo4 cpr5 double mutant phenotype. Because TPL, TPR1, and TPR4 were reported to be functionally redundant for immune activation (Zhu et al., 2010), we crossed the xpo4-2 cpr5 double mutant with the tpl tpr1 tpr4 triple mutant. We found that both the severe phenotype and high levels of SA-mediated defense gene expression in the xpo4 cpr5 double mutant were significantly alleviated in the xpo4 cpr5 tpl tpr1 tpr4 quintuple mutant (Figure 6, D–F), demonstrating that TPL and TPRs are required for the enhanced autoimmune phenotype in the xpo4 cpr5 compared with cpr5. However, the cpr5 single mutant phenotype is not affected in the cpr5 tpl tpr1 tpr4 quadruple mutant, consistent with our hypothesis that TPL and TPRs may function downstream of SA accumulation during ETI signaling to substantiate defense responses initiated by cpr5-dependent signaling. The genetic interaction between XPO4 and TPL/TPRs in the cpr5 background underscores the functional importance of XPO4-mediated TPL/TPR nuclear export during cpr5-dependent immune induction.

Discussion

The Arabidopsis genome encodes seven different exportins (Tamura and Hara-Nishimura, 2014), which may function in parallel to permit differential regulation of cargo export from the nucleus. Some of the exportins are specialized in transporting a single type of molecule, and others may mediate transport of a multitude of cargos; however, the exact cargo spectrum of any plant NTR is not defined, which has greatly hampered our understanding of specialized cargo transport by NTRs and its role in regulating distinct stress responses in plants.

XPO4 is a critical negative regulator of cpr5-dependent autoimmune induction

Here, we found that mutations in Arabidopsis XPO4 dramatically enhanced the autoimmune phenotype in the cpr5 mutant, resulting in severely stunted growth and significantly upregulated SA-mediated defense gene expression (Figure 2). The in vivo cargo profiling using proximity-labeling proteomics identified transcription corepressors of the TPL family that plays a crucial role in immune signaling as XPO4-specific substrates (Figures 4 and 5). XPO4 may mediate constitutive nuclear export of TPL and TPRs, because XPO4-TurboID probed these proteins in plants without SA or pathogen treatment. However, loss-of-XPO4 did not result in detectable accumulation of TPL/TPRs in the nucleus unless a high level of SA is present, indicating that the nucleocytoplasmic exchange of TPL/TPRs is maintained at a low level without stress challenges. It also suggests that SA can promote the nuclear translocation of TPL and TPRs and XPO4 is required for counteracting this process, likely through nuclear export (Figure 6). The involvement of TPL and TPRs in xpo4-induced immune enhancement in cpr5 is supported by significant suppression of the xpo4 cpr5 phenotype by the tpl tpr1 tpr4 triple mutant (Figure 6). However, the triple mutant did not further suppress the cpr5 single mutant phenotype, indicating that XPO4-dependent TPL/TPR nuclear export functions downstream of cpr5-induced initial immune signaling to affect the final immune outcome. In line with this hypothesis, loss-of-XPO4 potentiates SA-mediated defense responses in cpr5 but not WT background, arguing that immune signaling activated by loss-of-CPR5 is a prerequisite for XPO4-regulated amplification of SA responses.

A working model for the role of XPO4 in regulating ETI strength

Based on our findings, we propose a model to describe a negative regulatory role of XPO4 during ETI induction in plants: In the absence of pathogen effectors or activated NLR proteins, CPR5 gates ETI induction at the NPC (Figure 7A). NLR activation compromises CPR5 function and induces an NPC conformational change to initiate ETI signaling partly through noncanonical E2F activities (Wang et al., 2014; Gu et al., 2016). This initial signaling promotes SA synthesis, and the elevated SA level stimulates the nuclear shuttling of TPL/TPRs, which repress the expression of negative defense regulators (Zhu et al., 2010). This completes a feedback loop for SA-mediated defense amplification and substantiates ETI induction (Figure 7B). XPO4 plays a negative regulatory role in the SA-mediated defense amplification by mediating constitutive nuclear export of TPR1. We hypothesize that the transport capacity of XPO4 can be partially reduced or its substrate preference can be altered by NLR activation as a mechanism to promote the defense amplification and modulate the immune strength. However, because the cpr5-induced immune activation bypasses NLRs, it may lack the essential signal initiated by NLRs to repress XPO4 activities, resulting in immune activation that mimics only partial ETI induction in the cpr5 mutant (Figure 7C). Consistent with this hypothesis, in the absence of XPO4, SA-mediated defense responses are significantly enhanced in the cpr5 mutant (Figure 2F;Supplemental Figure 2B) and the autoimmunity and cell death in cpr5 are boosted to a level that results in lethality, a strength that appears closer to the NLR-activated full ETI response.

Figure 7.

Working model: coordinated XPO4 activity determines the nuclear level of TPL/TPRs in response to elevated SA and is part of the SA-dependent defense amplification loop that modulates ETI strength. (A) TPR1 is predominantly excluded from the nucleus before ETI is activated. (B) NLR activation (1) induces a CPR5-mediated NPC conformational change (2) and initiates ETI signaling (e.g. through E2F transcription factors; 3). Following the initial ETI activation and defense gene expression (4), SA is synthesized (5), and the elevated SA promotes the nuclear shuttling of TPR1 (6), which represses the expression of negative defense regulators to substantiate ETI (7). The activity of XPO4 may be altered following NLR activation as a mechanism to facilitate the subsequent nuclear accumulation of TPR1 and potentiate the full ETI induction. This situation is mimicked in the xpo4 cpr5 double mutant. Numbers in parentheses indicate the sequence of defense activation in the model. (C) XPO4 negatively regulates the nuclear accumulation of TPR1 through mediating nuclear export of TPR1, which compromises the SA-mediated defense amplification loop downstream of CPR5 and results in partial ETI activation in the cpr5 mutant. Because the cpr5-induced ETI bypasses NLRs, we hypothesize that it lacks an essential signal initiated by NLRs that can alter the XPO4 activity to inhibit the TPR1 export.

Dual function of XPO4 in ETI

The involvement of XPO4 in ETI induction appears to be more complicated, as we found that the xpo4 single mutant is compromised in pathogen-induced ETI (Figure 3D). This evidence indicates that XPO4 is also essential for ETI signaling, possibly playing a role upstream of CPR5 and the NPC as well (Figure 3F). In other words, the cpr5 mutant provides a unique immune context that allowed us to discover the negative regulatory role of XPO4 in ETI downstream of CPR5, which would have been masked by using just the xpo4 single mutant. We think that the positive contribution of XPO4 to ETI induction can be attributed to other immune-related cargo transported by XPO4. Indeed, aside from TPL/TPRs, we also identified SGT1 (SGT1A and SGT1B) as potential substrates of XPO4 (Figure 4, D and E). SGT1 plays a critical role in NLR protein folding (Austin et al., 2002; Azevedo et al., 2002), and it is also involved in negatively regulating accumulation of multiple NLR proteins to avoid autoimmune activation (Holt et al., 2005; Li et al., 2010). It is possible that XPO4 modulates the nucleocytoplasmic distribution of SGT1, which in turn affects the nuclear level and activity of NLRs (e.g. RPS4 and SNC1) that are required for activation of ETI upstream of CPR5 and the NPC.

Extensive involvement of XPO4 in regulating plant immunity

In addition to ETI, the xpo4 mutant is compromised in basal immunity, and this phenotype is unique to XPO4 in the exportin family. These data implicate an extensive involvement of XPO4 at multiple levels during plant immune activation and perhaps are not surprising considering that XPO4 may mediate nuclear transport of a variety of cargo. It also suggests that XPO4 may have evolved to specialize, at least partially, in regulating plant immune responses and supports the idea of functional specialization of NTRs in plants. Because PTI activation is not affected in the xpo4 mutant (Figure 3E), XPO4 substrates and mechanisms that are involved in regulating basal immunity are not clear and await future research.

Materials and methods

Plant materials and bacterial strains

The xpo4-2 (Salk_093159), xpo5 (Salk_006481), xpo7 (Salk_049002), xpo1a (Salk_028886), xpo1b (Salk_036720), tnpo3-1 (Salk_052896), nup54 (Salk_015252), nup85 (Salk_133369), nup93a (Salk_137710), and nua2 (Salk_069922) mutants were obtained from Arabidopsis Biological Resource Center (ABRC). The tpl tpr1 tpr4 triple mutants were obtained from Dr. Jingbo Jin’s laboratory at the Institute of Botany, Chinese Academy of Sciences. The cpr5 and npr1-2 mutants were obtained from Dr Xinnian Dong’s laboratory at Duke University. The xpo4-2 cpr5, nup54 xpo4-2, nup85 xpo4-2, nup93a xpo4-2, nua-2 xpo4-2, xpo5 cpr5, and xpo7 cpr5 double mutant lines were generated by genetic crossing. The plants were grown in soil or on half-strength Murashige and Skoog(1/2 MS) medium containing 0.8% agar at 22°C under a photoperiod of 16-h light/8-h dark and a light intensity of ∼110 μE/m² provided by full-spectrum fluorescent bulbs. For the disease assays, plants were grown in soil under a 12-h light/12-h dark cycle.

The bacterial strains Pseudomonas syringae pv tomato DC3000 (Pst DC3000), Pst DC3000 (AvrRpt2), Pst DC3000 (AvrRps4), and Pst ΔHrcC were obtained from Jianmin Zhou’s laboratory at Institute of Genetics and Developmental Biology, Chinese Academy of Sciences. The Pst DC3000 strains were grown on King’s medium B (KB) plates with 25 mg/L rifampicin. Pst DC3000 (AvrRpt2), Pst DC3000 (AvrRps4), and Pst ΔhrcC were grown on KB plates with 25 mg/L rifampicin and 50 mg/L kanamycin. All the above strains were grown at 28°C and reinoculated twice on KB plates before used for plant infection.

Phylogenetic analysis

Full-length amino acid sequences of Arabidopsis importin-α and importin-β family proteins were aligned using Multiple Protein Sequence Alignment. Neighbor-joining trees were constructed using MEGAX_10.1.8 with the JTT model and 10,000 bootstraps.

Constructs and transgenic plants

The genomic DNA sequence (from start to stop codon) of XPO4 and its upstream 950 bp (XPO4 promoter) were subcloned into pAB002, a modified pEarleyGate100 vector that contains a multiple cloning site followed by a 3×HA and TurboID tag, to generate the pXPO4-XPO4-3×HA-TurboID construct. Similarly, the genomic sequence of XPO7 and its upstream 1,867 bp (XPO7 promoter) and the coding sequence of XPO5 and its upstream 693 bp (XPO5 promoter) were used to generate pXPO7-XPO7-3×HA-TurboID and pXPO5-XPO5-3×HA-TurboID. These constructs were confirmed by sequencing and introduced into Agrobacterium tumefaciens strain GV3101. Transgenic Arabidopsis plants were generated by floral dipping. The xpo4-2, xpo5, and xpo7 plants were transformed with the corresponding TurboID constructs to obtain transgenic lines, and T3 plants with homozygous transgene insertion were used for experiments.

CRISPR library construction, transformation, and genetic screen

The CRISPR gRNA target sites were predicted using the service provided by www.genome.arizona.edu/crispr/CRISPRsearch. Two gRNAs that target the same NTR gene were cloned into two separate gRNA cassettes in a modified pHSE401 vector carrying a bar gene conferring resistance to the herbicide Basta for plant selection (Dr Qijun Chen’s Lab at College of Biological Sciences, China Agricultural University). The two gRNAs are driven by the U6-26 promoter and U6-29 promoter, respectively, and are both terminated by the U6-26 terminator. The cloning was carried out as previously described (Huang et al., 2020). A total of 27 CRISPR gRNA constructs were made for the 27 NTR genes. All target sequences and cloning primers containing gRNAs are provided in Supplemental Data Set 4. An equal amount of the 27 constructs was aliquoted and mixed to obtain the vector library, which was transformed into Agrobacterium strain GV3101. Bacterial transformants were bulk collected and cultured for cpr5 plant transformation using floral dipping (Clough and Bent, 1998).

Pathogen infections

Pst DC3000 grown on KB plates was resuspended in 10 mM MgSO₄. (1) For infiltration infection, a bacterial suspension of Pst DC3000 (OD_600nm = 0.0001 or 0.001), Pst DC3000 carrying AvrRpt2/AvrRps4 (OD_600nm = 0.001), or Pst ΔhrcC (OD_600nm = 0.04) was syringe-infiltrated into the third and fourth rosette leaves of 3-week-old plants. (2) For elf18-induced immune assay, 1 μM of elf18 or ddH₂O (mock) was infiltrated into the third and fourth rosette leaves of 3-week-old seedlings. Four hours later, Pst DC3000 (OD_600nm = 0.001) was infiltrated into the same leaves. (3) For spray infection, Pst ΔhrcC (OD_600nm = 0.2) was sprayed on the whole plant containing 0.02% Silwet L-77.

For infiltration inoculation, on the day of inoculation and 3 or 4 days past-inoculation, four leaf discs from four independent plants were pooled as one sample. Five to six samples were collected per genotype per treatment. The samples were ground in 500 μL of 10 mM MgSO₄. Serially diluted samples were plated on KB medium containing the appropriate antibiotics and incubated at 28°C. Bacterial colonies were counted 2 days after incubation. The colony-forming units were normalized by leaf area (per cm²). Statistical analyses for bacterial growth data and plotting of the results were conducted using Prism GraphPad (v8.2.0).

elf18 treatment

Col-0, xpo4-2, xpo7, and efr plants were grown in 1/2 MS media with 1% sucrose and 0.4% Phytagel. Seven days later, seedlings with similar size were transferred into a single well of a 48-well plate containing 1.5 mL of 1/2 MS media with or without 1 μM elf18. Five days later, the fresh weight of 12-day-old seedlings (10 plants) was measured using a precision scale. Three biological replicates were performed per genotype per treatment.

RNA extraction, RT-qPCR, and RNA-seq

Total RNA was extracted from about 50 mg samples of 10-day-old seedlings using an RNAprep pure Plant Kit (TIANGEN Biotech, China, Cat. # DP432). First-strand cDNA was synthesized using a PrimeScript™ RT reagent Kit with gDNA Eraser (TaKaRa, USA, Cat. # RR047A). RT–qPCR was performed on a CFX96^TM Real-Time System (Bio-Rad) using SYBR™ Select Master Mix (Thermo Fisher, USA, Cat. # 4472908). Statistical analyses for RT-qPCR and plotting of the results were conducted using Prism GraphPad (v8.2.0).

For RNA-seq, total RNA was extracted from 10-day-old Col-0, xpo4, cpr5, and xpo4cpr5 seedlings grown on 1/2 MS medium containing 1% sucrose and 0.7% agar. First- and second-strand cDNA generation, mRNA-seq library construction, quality control, library sequencing using Illumina Novaseq 6000, mapping of sequence reads to Arabidopsis reference genome (TAIR10), and generation of the total number of uniquely mapped reads for each gene were conducted by Tianjin Novogene Bioinformatic Technology Co., Ltd. For subsequent differential expression analysis in RStudio (version 1.1.4), the DESeq2 package was used for PCA and the determination of DEGs, the factoextra package was used for hierarchical clustering, and the pheatmap package was used for building heatmaps. GO enrichment analysis was conducted using agriGO (v2.0; http://systemsbiology.cau.edu.cn/agriGOv2/index.php) and VirtualPlant (v1.3; http://virtualplant.bio.nyu.edu/cgi-bin/vpweb/).

Proximity labeling

For the first round of proximity labeling proteomics (data shown in Figure 4), 3-week-old seedlings of pXPO4-XPO4-3×HA-TurboID transgenic plants were grown in soil. The third and fourth rosette leaves were infiltrated with 50 μM of biotin or ddH₂O (mock). Six hours later, 0.5 g of treated leaves was harvested as one sample. Samples were placed in liquid nitrogen and stored at −80°C until use. Two biological replicates were collected for each treatment.

For the second round of proximity labeling proteomics (data shown in Figure 5), 10-day-old seedlings of pXPO4-XPO4-3×HA-TurboID, pXPO5-XPO5-3×HA-TurboID, and pXPO7-XPO7-3×HA-TurboID transgenic plants were grown in 1/2 MS media with 1% sucrose, 0.4% Phytagel, and 10 mg/L Bialaphos. The whole seedlings were collected and submerged in 50 μM of biotin solution. Six hours later, 0.5 g of treated plants was harvested in liquid nitrogen and stored at −80°C until use. Two biological replicates were collected for each treatment.

Affinity purification and mass spectrometry

Total protein was extracted from collected samples with buffer (50 mM Tris [pH 7.5], 150 mM NaCl, 0.5% Triton X-100, 0.5% Nonidet P-40, 0.25% Nadeoxycholate, plant protease inhibitor cocktail, and 40 µM MG132). Before affinity purification (AP) with streptavidin-coated magnetic beads (Dynabeads MyOne Streptavidin C1, Invitrogen, USA), the protein extract was subject to a desalting procedure to remove free biotin using HiTrap^TM desalting column (GE Healthcare, USA) on the protein purifier systemAKTA^TM. AP was conducted overnight at 4°C. After AP, samples were washed five times with the buffer before being separated by SDS–PAGE. Each lane was cut into three gel pieces, digested with trypsin, and analyzed by LC–MS/MS.

Proteomic analysis

A detailed proteomic analysis procedure has been described previously (Huang et al., 2020). Briefly, MS/MS spectra from each LC–MS/MS run were searched against the TAIR10 database using Proteome Discoverer (Version 1.4) with the following settings: Full tryptic specificity was required, at most two missed cleavages were allowed, carbamidomethylation was set as a fixed modification, oxidation (M) was set as a variable modification, precursor ion mass tolerance was 20 p.p.m. for all MS acquired in the Orbitrap mass analyzer, and fragment ion mass tolerance was 0.02 Da for all MS2 spectra. A high confidence score filter (FDR < 1%) was used to select the "hit" peptides, and their corresponding MS/MS spectra were manually inspected. For enrichment analysis, protein enrichment areas (label-free quantification - LFQ intensities) were integrated and then normalized byDEP (Differential Enrichment analysis of Proteomics data) package (version 1.8.0) in RStudio (version 1.1.4) as the input for statistical analysis. Candidates that met the following criteria were selected as XPO4 substrates in the first round of proteomic analysis using biotin mock-treated samples as the control: fold-change >3, P < 0.05, and peptide-spectrum match (PSM) >2. During the second round of proteomic analysis, the criteria for identification of XPO4-specific substrates were fold-change >2, P <0.05, and PSM > 1. A slightly lower threshold for fold-change was used because the candidates need to meet these criteria when compared with both XPO5 and XPO7 samples.

Co-IP

One gram of 2-week-old pXPO4-XPO4-3×HA-TurboID (Line 1) and pXPO7-XPO7-3×HA-TurboID transgenic Arabidopsis seedlings were used for in vivo co-IP assay. Total protein was extracted with extraction buffer (50 mM Tris–HCl [pH 7.5], 150 mM NaCl, 0.5% Trition-X 100, 0.5% NP-40, 1 mM PMSF (phenylmethylsulfonyl fluoride), 40 μM MG132, and protease inhibitor cocktail). For anti-HA IP, the total protein extract was incubated with 1 μg anti-HA antibody (12CA5, Roche, USA, Cat#11666606001) and 30 μL protein A magnetic beads for 4 h at 4°C. The beads were washed seven times with washing buffer (50 mM Tris–HCl [pH 7.5], 150 mM NaCl). The bound proteins were eluted by boiling with SDS loading buffer for 5 min at 100°C and separated by SDS–PAGE before being immunoblotted with anti-HA and customized anti-TPR1 antibody, respectively.

Nuclear fractionation

Nuclear fractionation was performed as described before (Liu et al., 2018) with some modifications. Briefly, about 2.0 g of the 10-day-old Arabidopsis seedlings was ground into fine powder in liquid nitrogen and suspended in 1.5-mL Honda buffer (0.4 M Sucrose, 2.5% Ficoll, 5% Dextran T40, 25 mM Tris–HCl, pH7.4, 10 mM MgCl₂, 0.5% Triton X-100, 0.5 mM PMSF,10 mM β-mercaptoethanol, Roche protease inhibitor cocktail, 2 mL/g). The homogenate was sequentially filtered through 100-μm and 40-μm mesh nylon, and 100-μL flow-through was aliquoted as the total protein sample. The rest flow-through was centrifuged at 1,500 g for 5 min at 4°C. The supernatant was collected as the cytosolic fraction. The pellet was washed four times with 5 mL of NIB buffer (20 mM KCl, 20 mM HEPES, pH 7.4, 0.5% Triton X-100, 13.8% hexylene glycol, 0.1% β-mercaptoethanol, 50 μM spermine, 125 μM spermidine, 1 mM PMSF, Roche protease inhibitor cocktail) and centrifuged at 1,500 g for 10 min at 4°C to pellet the nuclei. The nuclear pellet was resuspended in 30 μL of cold glycerol buffer (20 mM Tris–HCl, pH 7.9, 50% glycerol, 75 mM NaCl, 0.5 mM EDTA, 0.85 mM DTT, 0.125 mM PMSF, and Roche protease inhibitor cocktail). Then, 30 μL of prechilled nuclei lysis buffer (10 mM HEPES, pH 7.6, 1 mM DTT, 7.5 mM MgCl₂, 0.2 mM EDTA, 0.3 M NaCl, 1M Urea, 1% NP-40, 0.5 mM PMSF, 10 mM β-mercaptoethanol, and Roche protease inhibitor cocktail) was added and the mixture was vortexed and incubated for 10 min on ice to obtain the lysis of the nucleus as the nucleus fraction. Protein samples were mixed with 4 × SDS loading buffer and were boiled at 100°C for 10 min. We loaded 15 μL (∼1%) of the total and the cytosol sample and 45 μL (∼50%) of the nuclear sample on the SDS–PAGE gel. Anti-TPR1 (obtained from Dr. Jingbo Jin’s laboratory), anti-Actin (Abiocode, USA, Cat. # R3772-1P), and anti-Histone H3 (Abcam, USA, Cat. # ab18521) antibodies were used for immunoblotting.

Primers

All primers used in the study are listed in Supplemental Data Set 4.

Statistical analysis

Student’s t test was used to compare differences in gene expression, bacterial growth, and rosette leaf diameter between genotypes. Two-way ANOVA (analysis of variance)analysis was used to compare the differences between groups that were affected by two factors (e.g. two genotypes or genotype and treatment). Results and parameters of all above statistical tests are provided in Supplemental Data Set 5. Statistical analyses for transcriptomic and proteomic data were based on linear regression model (F-test) and can be found in Supplemental Data Set 1–3.

Accession numbers

All RNA-seq data have been deposited to NCBI Gene Expression Omnibus (GSE146914). All mass spectrometry data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository (Identifier: PXD018065 and PXD018109). All MS dataset names and figures associated with each dataset are listed in Supplemental Data Set 6. Sequence data from this article can be found in the Arabidopsis Genome Initiative or GenBank/EMBL databases under the following accession numbers: XPO4 (AT3G04490), XPO5 (AT3G05040), XPO7 (AT5G06120), CPR5 (AT5G64930), TPL (AT1G15750), TPR1 (AT1G80490), TPR4 (AT3G15880).

Supplemental data

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

Supplemental Figure 1. Characterization of cpr5 xpo and xpo4 nup double mutants.

Supplemental Figure 2. The genetic interaction between CPR5 and XPO4 revealed by transcriptome analyses.

Supplemental Figure 3. Loss of XPO4 does not affect PTI activation.

Supplemental Figure 4. Proximity-labeling proteomics using XPO4, XPO5, and XPO7 as bait.

Supplemental Data Set 1. Statistics of differentially expressed genes from RNA-seq analyses.

Supplemental Data Set 2. Proteins identified by proximity-labeling proteomics using XPO4-TurboID expressed in rosette leaves.

Supplemental Data Set 3. Proteins specifically identified by XPO4-TurboID, XPO5-TurboID, and XPO7-TurboID using proximity-labeling proteomics and ratiometric analysis.

Supplemental Data Set 4. Primers used in this study.

Supplemental Data Set 5. Statistical support.

Supplemental Data Set 6. The list of deposited mass spectrometry datasets and figures associated with each dataset.

Supplemental File 1. Alignments used to generate the phylogeny presented in Figure 1A.

 

F.X. and Y.G. designed the research. F.X., M.J., X.L., Y.T., and K.J. generated constructs, genetic crosses, and transgenic plants used in this study. F.X. performed all the rest experiments. F.X., J.B., and Y.G analyzed the data, and F.X. and Y.G wrote the paper.

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/pages/General-Instructions) is: Yangnan Gu (guyangnan@berkeley.edu).