The Phytophthora sojae nuclear effector PsAvh110 targets a host transcriptional complex to modulate plant immunity

Abstract

Plants have evolved sophisticated immune networks to restrict pathogen colonization. In response, pathogens deploy numerous virulent effectors to circumvent plant immune responses. However, the molecular mechanisms by which pathogen-derived effectors suppress plant defenses remain elusive. Here, we report that the nucleus-localized RxLR effector PsAvh110 from the pathogen Phytophthora sojae, causing soybean (Glycine max) stem and root rot, modulates the activity of a transcriptional complex to suppress plant immunity. Soybean like-heterochromatin protein 1-2 (GmLHP1-2) and plant homeodomain finger protein 6 (GmPHD6) form a transcriptional complex with transcriptional activity that positively regulates plant immunity against Phytophthora infection. To suppress plant immunity, the nuclear effector PsAvh110 disrupts the assembly of the GmLHP1-2/GmPHD6 complex via specifically binding to GmLHP1-2, thus blocking its transcriptional activity. We further show that PsAvh110 represses the expression of a subset of immune-associated genes, including BRI1-associated receptor kinase 1-3 (GmBAK1-3) and pathogenesis-related protein 1 (GmPR1), via G-rich elements in gene promoters. Importantly, PsAvh110 is a conserved effector in different Phytophthora species, suggesting that the PsAvh110 regulatory mechanism might be widely utilized in the genus to manipulate plant immunity. Thus, our study reveals a regulatory mechanism by which pathogen effectors target a transcriptional complex to reprogram transcription.

The Phytophthora sojae nuclear effector PsAvh110 suppresses the transcriptional activity of the GmLHP1-2/GmPHD6 complex by disrupting its formation, resulting in reduced expression of immunity genes.

IN A NUTSHELL.

Background: To prevent infection, plants have evolved two layers of immunity, pattern-triggered immunity (PTI) and effector-triggered immunity (ETI). Activation of PTI and/or ETI triggers a series of plant immune responses, including massive transcriptional reprogramming of immune genes regulated by nuclear transcriptional complexes. Among them, BRI1-associated receptor kinase 1-3 (BAK1-3) and pathogenesis-related protein 1 (PR1), which encode two key immune regulators in salicylic acid (SA) pathway and PTI, undergo transcriptional regulation. Pathogens deploy numerous effectors to suppress plant immunity and the Phytophthora effectors directly involved in regulating the function of this nuclear transcriptional complex remain elusive.

Question: We hoped to understand the molecular mechanisms underlying the regulation of plant immune gene transcription by this nuclear transcriptional complex, and to investigate how Phytophthora effectors manipulate the expression of plant immune genes.

Findings: PsAvh110, a nuclear-localized effector from Phytophthora sojae, suppresses soybean (Glycine max) immunity to promote pathogen infection. By transcriptome analysis, we found that PsAvh110 suppressed a subset of soybean immune-related genes. PsAvh110 disrupts the assembly of a nuclear transcriptional complex containing like-heterochromatin protein 1-2 (GmLHP1-2) and plant homeodomain finger protein 6 (GmPHD6) via binding to GmLHP1-2, thus inhibiting the complex’s transcriptional activity. Interestingly, GmLHP1-2/GmPHD6 binds to G-rich elements (GREs) in the promoters of PsAvh110-suppressing immune genes, including GmBAK1-3 and GmPR1, to activate their expression. To counteract soybean immunity, P. sojae utilizes the effector PsAvh110 to compete with GmPHD6 for binding to GmLHP1-2, thereby interfering with the formation of GmLHP1-2/GmPHD6 complex. Therefore, we revealed that the nuclear complex plays a crucial role in regulating gene activation through binding to the GREs, which could be targeted by Phytophthora effector for suppressing plant immunity.

Next steps: Since GREs play important roles in regulating immune gene expression, our future work will investigate whether we can engineer durable disease-resistant crops by editing these elements in the promoters of immune genes using gene-editing technology.

Introduction

Plants encounter various attacks by microbial pathogens throughout their lives and have evolved an innate immune system to defend themselves against microbial infections (Jones and Dangl, 2006). During plant–pathogen interactions, plants deploy plasma membrane (PM)-resident pattern recognition receptors (PRRs), including receptor kinases (RKs) and receptor proteins (RPs), to sense microbe- or host-derived damage-associated molecular patterns (MAMPs/DAMPs), and trigger the first layer of immune responses, termed pattern-triggered immunity (PTI) (Couto and Zipfel, 2016; Yu et al., 2017; Albert et al., 2020; Zhou and Zhang, 2020). To successfully colonize plants, pathogenic microbes secrete numerous virulence factors, such as effectors, into plant cells or the apoplast to suppress PTI by targeting and regulating key immune regulators, leading to effector-triggered susceptibility (ETS) (Cui et al., 2015; Kourelis and van der Hoorn, 2018). In turn, plants have evolved intracellular conserved nucleotide-binding domain (NB) leucine-rich repeat (LRR)-containing receptors (NLRs) to directly or indirectly recognize such virulence effectors and initiate effector-triggered immunity (ETI) (Cui et al., 2015). Recent studies showed that the crosstalk between PTI and ETI can activate strong plant defense against pathogen infections (Ngou et al., 2021; Yuan et al., 2021a, 2021b). Activation of PTI and/or ETI triggers a series of immune responses, such as a transient reactive oxygen species (ROS) burst, rapid Ca²⁺ influx, activation of receptor-like cytoplasmic kinases (RLCKs), and mitogen-activated protein kinase (MAPK) cascades, and massive transcriptional reprogramming of immune-related genes (Yu et al., 2017; Saijo et al., 2018; DeFalco and Zipfel, 2021; Escocard de Azevedo Manhães et al., 2021).

Somatic embryogenesis RKs (SERKs) belong to a small group of RKs that function as coreceptors regulating diverse signaling pathways (Liu et al., 2020). Among the five SERKs in Arabidopsis (Arabidopsis thaliana), SERK3, also known as BRI1-ASSOCIATED RECEPTOR KINASE 1 (BAK1), has been well-studied. BAK1 forms different PRR complexes with other receptors, such as FLAGELLIN-SENSITIVE 2 (FLS2) (Gomez-Gomez and Boller, 2000), Elongation factor-Tu receptor (EFR) (Zipfel et al., 2006), PEP1 RECEPTOR 1 (PEPR1), and PEPR2 (Krol et al., 2010), to regulate different signal outputs including plant cell differentiation, growth and development, and the immune response (DeFalco and Zipfel, 2021; Kong et al., 2021). BAK1 was originally discovered as a key component of brassinosteroid (BR) signaling (Li et al., 2002). BAK1 is a coreceptor for BRASSINOSTEROID INSENSITIVE 1 (BRI1) for the perception of the plant hormone BRs and regulate plant growth and development (Li et al., 2002). Furthermore, extensive studies have shown that BAK1 also functions as a coreceptor in RK- and RP-mediated immune signaling (Yasuda et al., 2017). BAK1 is involved in plant defense against various oomycetes, fungi, viruses, and insects, although the underlying mechanisms remain unclear (Chaparro-Garcia et al., 2011; Wang et al., 2018). To subvert BAK1-mediated immune responses, pathogens have evolved different virulence effectors to directly or indirectly modulate BAK1 by regulating its protein stability, kinase activity, and interactions with the corresponding PRRs (Zhou et al., 2014; Li et al., 2016b; Irieda et al., 2019).

Increasing evidence has shown that profound and dynamic transcriptional reprogramming plays an important role in fine-tuning various PTI outputs (Moore et al., 2011; Tsuda and Somssich, 2015; Li et al., 2016a). With an increasing number of identified patterns (MAMPs/DAMPs) and their corresponding cognate PRRs, recent studies showed that seven patterns can trigger rapid and congruent transcriptional outputs at early time points of infection (Bjornson et al., 2021; Winkelmüller et al., 2021). Transcriptome analysis revealed that different patterns induce the expression of a large set of common genes and a small number of pattern-specific response genes in Arabidopsis (Bjornson et al., 2021). Similarly, one study showed that flg22, a 22-amino-acid epitope derived from bacterial flagellin, can induce extensive transcriptional reprogramming and a set of common genes in four Brassicaceae species including Arabidopsis (Winkelmüller et al., 2021).

Notably, the transcripts of some immune signaling components and PRR complexes are activated upon pattern perception. For instance, immune regulatory genes encoding NON-HOST 1 (NHO1), Flg22-INDUCED RECEPTOR-LIKE KINASE 1 (FRK1), and a WRKY transcription factor (WRKY29) are greatly induced upon flg22 perception (Asai et al., 2002; He et al., 2006). Interestingly, plant bacterial pathogens secrete a set of effectors to suppress FRK1 and NHO1 transcription induced by flg22, resulting in the dampening of plant immunity (Li et al., 2005; Zheng et al., 2014). In addition, several redundant RxLR effectors from the oomycete Phytophthora infestans can suppress flg22-induced FRK1 expression, whose regulatory mechanism remains unclear (Zheng et al., 2014). Although the transcripts encoding components of PRR complexes, such as BAK1, CHITIN ELICITOR RECEPTOR KINASE 1, FLS2, and EFR, SUPPRESSOR of BIR1-1, and BOTRYTIS-INDUCED KINASE 1, were induced by multiple patterns (Boutrot et al., 2010; Zou et al., 2018; Bjornson et al., 2021; Winkelmüller et al., 2021), whether plant pathogens can deploy virulent effectors to modulate their expression needs to be investigated.

Phytophthora, a genus of oomycetes, contains more than 100 species that are identified as plant pathogens causing immeasurable damage to crops and forests (Beakes et al., 2012). Soybean (Glycine max) stem and root rot, caused by Phytophthora sojae, significantly damage global soybean production (Beakes et al., 2012). During the early infection stage, P. sojae secretes numerous apoplast and cytoplasmic effectors to activate or suppress plant immunity (Wang et al., 2019). To date, three groups of P. sojae intracellular effectors have been identified. They are termed RxLR, CRN, and CHxC effectors, respectively, based on the motifs they harbor: RxLR (Arg-x-Leu-Arg), CRN (Crinkler or crinkling- and necrosis-inducing protein), or CHXC (Cys-His-x-Cys) (Bozkurt et al., 2012; Tabima and Grunwald, 2019). RxLR effectors are widely identified in Phytophthora species, which deploy different molecular mechanisms to manipulate plant immunity. For instance, the two P. sojae RxLR effectors PSR1 and PSR2 suppress gene silencing to interfere with plant immunity (Qiao et al., 2015). In addition, several other P. sojae RxLR effectors have been shown to interfere with phytohormone signaling, alternative splicing, epigenetic modification, protein stability, MAPK signaling, autophagy, and protein secretion during plant immune responses (Huang et al., 2017; Kong et al., 2017; Li et al., 2018; Guo et al., 2019; Wang et al., 2019; Yang et al., 2019). Whether RxLR effectors modulate transcriptional reprogramming downstream of PRR complexes remains largely unknown.

The plant homeodomain (PHD) finger is an evolutionarily conserved domain that is ubiquitous among chromatin-binding modules in eukaryotes and regulates transcription and chromatin dynamics (Musselman and Kutateladze, 2011). PHD proteins are characterized as epigenetic effectors, since they can form diverse chromatin-binding modules to recognize different types of histone modification marks, including unmethylated, methylated, and acetylated lysine residues via various consensus sequences (Li and Li, 2012). In humans, more than 100 PHD finger-containing proteins have been identified, some of which can bind to the N-terminal tail of histone H3 at either unmodified or methylated lysine 4 (H3K4) (Jain et al., 2020). Similarly, in the model plant Arabidopsis, increasing evidence has shown that PHD domain proteins function as histone readers recognizing the active mark H3K4me1/2/3 (mono to trimethylation) to regulate gene expression during plant growth and development. For example, Arabidopsis EARLY BOLTING IN SHORT DAYS (EBS) and SHORT LIFE (SHL), both harboring PHD-domains, bind to H3K4me3 to regulate floral phase transition and floral repression, respectively (Qian et al., 2018; Yang et al., 2018). Interestingly, both EBS and SHL also possess a bromo-adjacent homology (BAH) domain, which can recognize the repressive mark H3K27me3 to suppress the expression of distinct floral genes. Additionally, the PHD proteins ASI1-immunoprecipitated protein 3 (AIPP2) and paralog of AIPP2 (PAIPP2) read unmodified H3K4 residues to regulate the expression of a subset of developmental genes in Arabidopsis (Zhang et al., 2020). Of note, PHD proteins can regulate transcriptional activation or transcriptional repression by forming complexes with diverse epigenetic enzymes or readers (Musselman and Kutateladze, 2011). For instance, the Arabidopsis PHD proteins ALFIN1-LIKE PROTEIN 6 (AL6) and its homolog AL7 form a complex with LIKE-HETEROCHROMATIN PROTEIN 1 (LHP1), an H3K27me3 reader of Polycomb Repressive Complex 1 (PRC1), to repress seed development genes by balancing the levels of H3K4me3 and H3K27me3 during plant growth phase transitions, although the mechanism remains unclear (Molitor et al., 2014). The Arabidopsis genome only encodes one LHP1 (also named TERMINAL FLOWER 2), which plays a crucial role in regulating flowering, as well as the plant immunity response to bacterial infections (Feng and Lu, 2017). In addition, soybean GmPHD6 might form a complex with GmLHP1 to activate the expression of salt-tolerance genes via binding to G-rich (GTGGNG/GNGGTG) elements in the promoter of their target genes, although it remains to be investigated whether H3K4me3 levels are changed (Wei et al., 2017b). In contrast to the well-studied function of PHD proteins and LHP1 in the regulation of plant growth and development, their roles in plant biotic stress responses remain largely obscure.

In this study, we report that the P. sojae nucleus-localized RxLR effector PsAvh110 can modulate the promoter activity of immune-associated genes via targeting the heterochromatin complex. PsAvh110 specifically localized to the plant nucleus, which is required for its virulence. The expression of PsAvh110 was induced around 200-fold at the early infection stage. Transcriptome analysis showed that a subset of soybean immune response genes is dysregulated upon infection with a P. sojae Psavh110 mutant, including increased expression of GmBAK1-3 and PATHOGENESIS-RELATED PROTEIN 1 (GmPR1), which encode key components of PRR signaling and salicylic acid (SA) signaling pathways, respectively. We employed a yeast-two-hybrid (Y2H) screen to identify one soybean protein with similarity to Arabidopsis LHP1 (hereafter denoted as GmLHP1-2) as a PsAvh110 interactor. LHP1 is conserved from yeasts to plants and metazoans, with the prototypic human HP1 involved in chromatin dynamics and transcription (Ayyanathan et al., 2003; Lechner et al., 2005). The soybean genome encodes four LHP1s (named GmLHP1-1, GmLHP1-2, GmLHP1-3, and GmLHP1-4), which form transcriptional activation complexes with GmPHD6 (Wei et al., 2017a). We provide evidence showing that the GmLHP1-2/GmPHD6 complex binds to the G-rich elements (GREs) present at the promoters of immune-associated genes, including GmBAK1-3 and GmPR1, to activate their transcription. Conversely, PsAvh110 competes with GmPHD6 for GmLHP1-2 binding to interfere with the transcriptional activity of the complex, thereby suppressing the expression of GmBAK1-3 and GmPR1. Overall, this work reveals a molecular mechanism by which a pathogen effector modulates plant PTI and SA signaling pathways via interfering with nuclear complexes mediating transcriptional reprogramming.

Results

PsAvh110 is a nucleus-localized effector that is essential for P. sojae virulence

We previously systematically screened a multitude of RxLR effectors of P. sojae involved in plant defense (Wang et al., 2011). To explore the mechanisms of P. sojae nuclear effectors in plant immunity, we investigated a subset of effectors that localize to the plant nucleus. Among them, the effector PsAvh110 specifically localized to the nucleoplasm and nucleolus when co-expressed with a construct encoding the nuclear marker histone 2B fused to red fluorescent protein (histone 2B-RFP) in Nicotiana benthamiana epidermal cells (Figure 1A). PsAvh110 contains two nuclear localization signal motifs (NLS1 and NLS2) (Figure 1B; Kosugi et al., 2009). Notably, phylogenetic analysis showed that PsAvh110 is conserved in Phytophthora species including P. parasitica, P. infestans, and Phytophthora capsici, in which all PsAvh110 homologs contain the NLS and W motifs (Figure 1C). To determine the function of PsAvh110 NLS1 and NLS2, we mutated all conserved lysine (K) and arginine (R) residues to alanine (A) residues (hereafter designed as PsAvh110^mNLS) (Figure 1B) and investigated the subcellular localization of the variant protein by confocal microscopy. We observed that PsAvh110^mNLS fails to accumulate in the nucleus, instead mainly localizing in the cytoplasm and PM (Figure 1A). In addition, expression of PsAvh110, but not PsAvh110^mNLS, in N. benthamiana leaves promoted infection by P. capsici isolate LT263, compared with control leaves expressing empty vector (EV) (Figure 1, D and E and Supplemental Figure S1A), suggesting that the nuclear localization of PsAvh110 is essential for its virulence activity.

Figure 1.

PsAvh110 localizes to the plant nucleus and contributes to the virulence of P. sojae. A, PsAvh110 localizes to the nucleus. GFP, GFP-PsAvh110, or GFP-PsAvh110^mNLS were co-expressed with the nuclear marker (H2B-RFP) in N. benthamiana leaves for 36 h before imaging. Fluorescence intensity profiles of GFP and RFP were assessed in the nucleus and PM along the transects shown as white lines. y-axis, GFP intensity (arbitrary units [au]); x-axis, transect length (μm). Scale bar, 10 μm. B, Schematic diagram of PsAvh110 showing its two nuclear localization signals (NLS1 and NLS2) and the mNLS mutant. The lysine (K) and arginine (R) residues within NLS1 and NLS2 are highlighted in red and were mutated to alanine (A) labeled with blue. C, PsAvh110 is a conserved effector in the Phytophthora species. Phylogenetic analysis of PsAvh110 homologs from Phytophthora parasitica (PpPPTG_12652T0), P. infestans (PiPITG_22926), and P. capsici (Pc_22959) is shown. Blue bars indicate the percentage identity of PsAvh110 in different Phytophthora species. The conserved motifs of PsAvh110 and homologs are shown as different shapes with different colors. SP, signal peptide; RxLR, Arg–x–Leu–Arg; W, tryptophan (W) motif. The protein sequences were retrieved from JGI (https://genome.jgi.doe.gov/portal/) for MEGAX phylogenetic analysis using the neighbor-joining method with 1,000 bootstrap replicates. The phylogenetic tree was drawn in iTOL (https://itol.embl.de/). D and E, PsAvh110 promotes Phytophthora infection and nuclear localization is required for its virulence. Nicotiana benthamiana leaves individually expressing Flag-RFP, Flag-Avh110-RFP, or Flag-Avh110^mNLS-RFP were inoculated with P. capsici mycelial plugs at 24 h after Agrobacterium infiltration. D, Infected leaves were photographed under UV light and lesion areas are indicated by white circles at 36 hpi. E, Boxplot showing the average lesion diameter of the infected leaves in (D). The data are shown as an overlay of dot plots with mean ± sd, n=20. Different letters indicate significant differences (P < 0.01; one-way ANOVA). F, The expression of PsAvh110 is highly induced during the early infection stage. Total RNA extracted from P. sojae zoospores and soybean hairy roots infected with zoospores at the indicated time points was subjected to RT-qPCR analysis. Relative PsAvh110 expression levels were normalized to PsActin. Data are shown as means ± sd from three independent replicates, n = 3. G and H, Knockout of PsAvh110 reduces P. sojae virulence. Etiolated soybean hypocotyls were inoculated with zoospore suspensions of P. sojae WT strain P6497, two PsAvh110 knockout transformants (T52 and T119), or a control strain transformed with the EV (CK). The infected soybean hypocotyls were photographed at 48 hpi (G). Relative biomass of P. sojae was determined by qPCR of P. sojae genome DNA normalized to soybean genome DNA at 48 h post inoculation (hpi). The results are shown as an overlay of dot plots with means ± sd, n = 3. Asterisks indicate significant differences (****P < 0.0001; Student’s t test). The experiments were performed three times with similar results. I and J, The PsAvh110^mNLS complementation strains fail to facilitate P. sojae infection in soybean. PsAvh110 was replaced with PsAvh110^mNLS in P. sojae using a CRISPR/Cas9-mediated in situ complementation method. The experiments were performed as in (G and H). The data are shown as an overlay of dot plots with means ± sd from three independent replicates. Asterisks indicate significant differences (****P < 0.0001; Student’s t test).

We next determined the virulence function of PsAvh110 during P. sojae infection. PsAvh110 expression was highly induced by approximately 200-fold at 3, 6, 9, and 12 h post-inoculation (hpi), and 400-fold by 24 hpi (Figure 1F). We generated two P. sojae PsAvh110 mutants (T52 and T119) in the wild-type (WT) strain P6497 using clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated nuclease 9 (Cas9)-mediated gene editing (Supplemental Figure S1B). Genomic PCR and Sanger sequencing confirmed deletion of 261 and 462 bp in the T52 and T119 mutants, respectively (Supplemental Figure S1, B and C). Moreover, we could not detect any PsAvh110 transcripts in the T52 and T119 mutants (Supplemental Figure S1D), confirming that they are PsAvh110 null mutants. The growth rates and colony morphology of PsAvh110 mutants T52 and T119 were similar to those of WT and a control strain (CK, which was transformed with the EV in the WT background) (Supplemental Figure S1, E and F). However, the PsAvh110 mutants exhibited reduced virulence on etiolated soybean hypocotyls compared with the WT and CK strains (Figure 1, G and H). To further determine the role played by PsAvh110 nuclear localization, we replaced the WT PsAvh110 gene with the NLS mutant copy PsAvh110^mNLS in P. sojae using a CRISPR/Cas9-mediated in situ complementation method (Qiu et al., 2021). We obtained three PsAvh110^mNLS replacement transformants (T1–T3) (Supplemental Figure S2, A–C). Furthermore, we also generated PsAvh110 complementation lines (CΔPsAvh110) in the T52 mutant background (Supplemental Figure S2, D–F). All PsAvh110^mNLS mutants and complementation lines showed similar growth rates and colony morphology as T52-CK (a control strain transformed with the EV in T52 background) and WT strains (Supplemental Figure S2, G and H). In contrast to WT and CΔPsAvh110, PsAvh110^mNLS mutants showed significantly reduced disease lesion sizes and P. sojae biomass in etiolated soybean hypocotyls (Figure 1, I and J), indicating that PsAvh110^mNLS cannot restore virulence to the same levels of disease phenotypes as the WT, which further supports the notion that the nuclear localization of PsAvh110 is required for its virulent activity. Together, these results corroborate the idea that PsAvh110 is a nuclear effector and contributes to the virulence of P. sojae.

PsAvh110 suppresses the expression of a subset of immune-associated genes during P. sojae infection

To understand the mechanisms of PsAvh110 virulence function, we carried out a transcriptome deep sequencing (RNA-seq) analysis on soybean hairy roots infected by P. sojae WT P6497 or PsAvh110-knockout mutant T52 at 6 hpi. Compared with WT-infected roots, RNA-seq data revealed the upregulation of 1,287 differentially expressed genes (DEGs) and the downregulation of 1,383 DEGs in T52-infected roots (Supplemental Data Set S1). We defined these upregulated and downregulated DEGs as PsAvh110-regulated genes. Hierarchical clustering analysis of PsAvh110-regulated genes suggested that PsAvh110 suppresses the transcription of a subset of genes (Figure 2A). Enrichment analysis of Gene Ontology (GO) categories indicated that genes related to catalytic activity, transcription regulator activity, cellular metabolism, regulation of biological process immune system, and response to stimulus are enriched among PsAvh110-regulated genes (Figure 2B). Notably, PsAvh110 suppressed the expression of a subset of soybean immune-associated genes such as those encoding GmBAK1-3, GmPR1, GmWRKY71-1 and GmWRKY71-2, N-acetyltransferase family protein (GmGCN5-like), cytochrome P450-related proteins (GmCYP71B21 and GmCYP81D1), disease resistance family protein (GmDRP), glutathione S-transferase family (GmGST), the 14-3-3-like protein GF14 (GmGF14), and CELL WALL-ASSOCIATED RECEPTOR KINASE (GmWAK), which play important roles in plant immunity (Supplemental Figure S3A). Consistent with the RNA-seq data, reverse transcription quantitative PCR (RT-qPCR) confirmed that the expression of these selected immune-associated genes increases in T52-infected roots compared with WT-infected roots (Figure 2C). The soybean genome encodes five BAK1 homologs, named GmBAK1-1/2/3/4/5 (Supplemental Figure S3B). Among them, GmBAK1-1 was shown to function as a co-receptor in regulating soybean immune response against bacterial infection (Wei et al., 2020). In addition, the expression of GmBAK1-3 was induced during P. sojae infection at 3, 6, 12, and 24 hpi (Supplemental Figure S3C). The immune response triggered by the P. sojae-secreted MAMP XEG1 (a glycoside hydrolase 12 protein) depends on NbBAK1, which is also required for N. benthamiana resistance to P. infestans (Chaparro-Garcia et al., 2011; Ma et al., 2015). Moreover, the expression of NbBAK1 was also induced by XEG1 at 1, 2, and 3 hpi in N. benthamiana (Supplemental Figure S3D). Therefore, our data, together with previous studies, indicate that GmBAK1 plays important roles in plant immunity, and PsAvh110 suppresses its transcription to regulate the PTI pathway.

Figure 2.

PsAvh110 suppresses the expression of a subset of immune-associated genes. A, Heatmap representation of gene expression in P. sojae T52-infected and WT-infected soybean hairy roots. The original FPKM values were subjected to data adjustment against reference genes. Rows and hierarchical clustering were generated by the average linkage method using TBtools; red, high expression; blue, low expression. B, GO analysis of PsAvh110-regulated genes. The fold enrichment was calculated based on the transformed values of –Log10(P-value). A list of PsAvh110-regulated genes is shown in Supplemental Data Set S1. C, RT-qPCR analysis of PsAvh110-regulated immune-associated genes in soybean hairy roots infected by P. sojae WT or the T52 mutant at 6 hpi. Relative expression levels were normalized to the soybean control gene GmCYP2; the data are shown as means ± sd (n = 3) from three biological replicates. Asterisks indicate significant differences (***P < 0.001; ****P < 0.0001; Student’s t test). D, Schematic diagram of the effector and reporter, and internal control constructs used in the transfection assays in N. benthamiana protoplasts. proBAK1, plant BAK1 promoter driving the LUC reporter gene; effector constructs contain the 35S promoter (pro35S), RXLR effector genes, or GFP (negative control); the internal control contains the NbActin2 promoter and the RLUC reporter gene. E and F, PsAvh110, but not PsAvh23 or PsAvr3C, inhibits plant BAK1 promoter activity induced by XEG1 in N. benthamiana protoplasts. Nicotiana benthamiana mesophyll protoplasts co-transfected with effector, reporter, and control constructs were treated with or without 200 nM XEG1 for 2 h. The activity of the BAK1 promoter from N. benthamiana (E) or soybean (F) induced by XEG1 was calculated relative to untreated samples and was normalized to the internal control NbActin2. The data are shown as means ± sd from three biological replicates. Asterisks indicate significant differences (**P < 0.01; ***P < 0.001; Student’s t test).

Given the importance of promoters in regulating transcription, we hypothesized that PsAvh110 might modulate the promoter activity of target genes to suppress their expression. To test this hypothesis, we cloned the promoters of NbBAK1 and GmBAK1-3 for generating firefly luciferase (LUC) reporter gene-based reporter constructs (proNbBAK1/proGmBAK1-3:LUC) (Figure 2D). PsAvh110 or GFP (control) was placed under the control of the cauliflower mosaic virus (CaMV) 35S promoter (pro35S:PsAvh110/GFP), an effector construct. The internal control construct contained the NbActin2 promoter driving the Renilla luciferase (RLUC) reporter gene (proNbActin2:RLUC). We thus performed dual-LUC reporter assays by co-expressing the reporter, effector, and internal control constructs in N. benthamiana protoplasts. Upon XEG1 treatment, we determined that LUC activity decreases approximately three-fold when co-expressed with PsAvh110-GFP compared with GFP and another two P. sojae effectors, PsAvh23-GFP and PsAvr3C-GFP serving as controls (Figure 2, E and F), suggesting that PsAvh110 can suppress the transcription of NbBAK1 and GmBAK1-3. We also generated the reporter construct proGmBAK1-3:GmBAK1-3 to ask whether PsAvh110 affects BAK1 protein abundance. However, immunoblotting showed that BAK1 protein levels are only slightly reduced when the encoding construct was co-expressed with PsAvh110 upon XEG1 treatment (Supplemental Figure S3E). Confocal images and immunoblotting analysis showed that GFP, PsAvh110-GFP, PsAvh23-GFP, and PsAvr3C-GFP accumulate in N. benthamiana protoplasts (Supplemental Figure S3, F and G). Together, the results indicate that PsAvh110 suppresses the expression of immune-associated genes by modulating the promoter activity of the target genes.

PsAvh110 targets plant heterochromatin proteins

To further determine how PsAvh110 suppresses the expression of immune genes, we performed a Y2H assay using PsAvh110 as bait to screen a cDNA library prepared from soybean plants infected with P. sojae. We thus identified GmLHP1-2 as a PsAvh110-interacting protein (Figure 3A). The soybean genome encodes four GmLHP1-2 homologs, named GmLHP1-1/2/3/4, which are conserved in plants (Supplemental Figure S4A; Feng and Lu, 2017). Among them, PsAvh110 strongly interacted with GmLHP1-2 compared with GmLHP1-4, but not with GmLHP1-1 or GmLHP1-3 in Y2H (Supplemental Figure S4B). To validate the interaction between PsAvh110 and GmLHP1 homologs, we co-transfected soybean protoplasts with constructs encoding GFP-tagged GmLHP1 homologs and Flag-PsAvh110 or Flag-RFP. Co-immunoprecipitation (Co-IP) assays showed that Flag-PsAvh110 strongly interacts with GmLHP1-2 compared with GmLHP1-1 and GmLHP1-4, but failed to interact with GmLHP1-3 (Figure 3B). Of note, we detected GmLHP1-2 as three different size bands in immunoblots (Figure 3B), suggesting the possibility of post-translational modifications. Furthermore, Co-IP and bimolecular fluorescence complementation (BiFC) assays confirmed that PsAvh110 is also associated with NbLHP1-1/2 (Supplemental Figure S4, C and D). We then observed the localization of PsAvh110 and GmLHP1-2 in plant cells. In the absence of PsAvh110-RFP (only RFP control), GFP-GmLHP1-2 predominantly localized to the nucleolus (92.3% ± 1.2%), with a small percentage of GmLHP1-2 accumulating in the nucleoplasm (7.7% ± 1.2%) in N. benthamiana leaves (Supplemental Figure S4E). However, in the presence of PsAvh110-RFP, the extent of GFP-GmLHP1-2 nucleolar localization decreased (70.7% ± 2.2%), whereas its nucleoplasm accumulation increased (29.3% ± 2.2%) (Supplemental Figure S4E), suggesting that PsAvh110 co-localizes with GmLHP1-2 and partially alters its localization from the nucleolus to the nucleoplasm.

Figure 3.

PsAvh110 interacts with soybean GmLHP1-2, which is required for PsAvh110 virulence activity. A, PsAvh110 interacts with GmLHP1-2 in yeast. Yeast transformants were spotted as 10-fold serial dilutions on SD medium without leucine and tryptophan (SD–2), or without histidine, leucine, adenine, and tryptophan (SD–4) and containing 0.2 mM X-α-gal. Yeast cells were photographed 3–4 days later. pLAW10 and pLAW11 are bait and prey EVs, respectively. Experiments were performed at least three times with similar results. B, PsAvh110 strongly associates with GmLHP1-2 in soybean protoplasts. Protoplasts from soybean seedlings were co-transfected with GmLHP1 homologs with FLAG-PsAvh110 or FLAG-RFP (control) for 12 h. Co-IP assays were carried out with GFP-Trap beads, followed by immunoblotting (IB) with anti-Flag or anti-GFP antibody (top two blots) with input proteins shown (bottom two blots). The molecular weight (kD) is shown to the left. Protein loading is shown by Ponceau S staining (Ponc.) of Rubisco large subunit (RBCL). C, Schematic diagram of PsAvh110 motifs and its variants. PsAvh110 contains two W domains (W1 and W2). The GEGE and GKSE residues present in the W1 and W2 domains are shown in yellow diamonds. The PsAvh110 variants PsAvh110^ΔW1 (lacking the W1 domain), PsAvh110^ΔW2 (lacking W2), and PsAvh110^mW2 (substitution of GKSE residues with AAAA, shown as green diamonds) are shown. D, PsAvh110, but not PsAvh110^mW2, interacts with GmLHP1-2 in an in vitro pull-down assay. Recombinant His-GmLHP1-2 protein immobilized on Ni-NTA resin was incubated with MBP, MBP-Avh110, MBP-Avh110^ΔW1, or MBP-Avh110^mW2. Washed beads were subjected to IB with anti-His or anti-MBP antibodies. E, Y2H assay showing the interaction between GmLHP1-2 and PsAvh110 derivatives. The Y2H assay was performed as described in (A). F, PsAvh110 associates with soybean GmLHP1-2 in planta. Total protein was extracted from N. benthamiana leaves co-expressing GFP-GmLHP1-2 with Flag-PsAvh110-RFP, Flag-PsAvh110^mW2-RFP, or Flag-RFP (control). Co-IP assays were performed as described in (B). G and H, Virulence activity of different PsAvh110 mutants. Nicotiana benthamiana leaves expressing Flag-RFP (control), or different Flag-PsAvh110-RFP derivatives were inoculated with P. capsici mycelial plugs at 36 h after Agrobacterium infiltration. Confocal images show subcellular localization of Flag-RFP (control) and Flag-Avh110-RFP derivatives in N. benthamiana at 36 h after Agrobacterium infiltration (top panel; scale bar, 5 μm). RFP-NLS was used as a nuclear marker. The infected leaves were imaged under UV light at 36 hpi and the diameter of the lesion areas, shown by white circles (bottom panel; scale bar, 1.5 cm), was measured. The data are shown as an overlay of dot plots with means ± sd (n = 22) in (G). Different letters indicate significant differences (P < 0.01; one-way ANOVA). Experiments were performed at least three times with similar results.

The W motif of P. sojae RxLR effectors was shown to be required for their function (Jiang et al., 2008), and PsAvh110 contains two such W motifs (Figures 1C and 3C; Supplemental Figure S4F). We thus generated W1 and W2 motif deletion variants, designated PsAvh110^ΔW1 and PsAvh110^ΔW2, respectively (Figure 3C). Recombinant maltose-binding protein (MBP)-tagged PsAvh110 and PsAvh110^ΔW1 proteins, but not MBP itself (negative control) or PsAvh110^ΔW2, were pulled down by His-tagged GmLHP1-2 (Figure 3D), which was consistent with the Y2H results (Figure 3E). These data suggest that the W2 motif, but not the W1 motif, is required for the interaction between PsAvh110 and GmLHP1-2. Multiple protein sequence alignment showed the relative conservation of Gly–Lys/Glu–Ser–Glu (GE/KSE) residues in the W1 and W2 motifs (Figure 3C and Supplemental Figure S4F), which we changed to alanine (A) in W2 (designated PsAvh110^mW2) to further test its interaction with GmLHP1-2. Notably, pull-down, Y2H, and Co-IP assays showed that PsAvh110^mW2 does not interact with GmLHP1-2 (Figure 3, D–F). Together, these results suggest that the GKSE residues within the W2 domain of PsAvh110 are required for the interaction with GmLHP1-2.

To determine whether the interaction between PsAvh110 and GmLHP1-2 is required for PsAvh110 virulence activity, we expressed constructs encoding the variants PsAvh110^ΔW1, PsAvh110^ΔW2, and PsAvh110^mW2 in N. benthamiana leaves and subsequently inoculated the leaves with P. capsici. Results showed that PsAvh110^ΔW1, which still interacts with GmLHP1-2, promotes P. capsici infection to the same extent as WT PsAvh110 (Figure 3, G and H and Supplemental Figure S4H). In contrast to PsAvh110^ΔW1, PsAvh110^ΔW2 and PsAvh110^mW2 lost virulence activity (Figure 3, G and H and Supplemental Figure S4H), indicating that the interaction between PsAvh110 and GmLHP1-2 is important for the virulence of PsAvh110 to facilitate pathogen infection in plants. Furthermore, the nuclear localization variant PsAvh110^mNLS, which lost its virulence activity (Figure 1, D and E), also failed to interact with GmLHP1-2 (Supplemental Figure S4, I and J).

The C-terminus chromo shadow domain of GmLHP1-2 is required for interaction with both PsAvh110 and GmPHD6

Y2H results showed that GmLHP1-2 interacts with both PsAvh110 and GmPHD6 (Figure 4A). Moreover, Co-IP assays indicated that GmLHP1-2 co-immunoprecipitates with GmPHD6 in soybean protoplasts (Figure 4B), which was consistent with a previous study showing that GmLHP1-2 can form a transcriptional activation complex with GmPHD6 to regulate plant responses to salt stress (Wei et al., 2017a). Given that GmLHP1-2 interacts with both PsAvh110 and GmPHD6, we examined whether PsAvh110 can interfere with the formation of the GmPHD6–GmLHP1-2 complex or form a heterotrimer with GmLHP1-2 and GmPHD6. The Y2H and Co-IP assays revealed that PsAvh110 does not associate with GmPHD6 (Supplemental Figure S5, A and B). However, compared with GmPHD6, PsAvh110 exhibited a stronger interaction with GmLHP1-2 in Y2H; we also observed a reduced association of GmLHP1-2 with GmPHD6 in soybean protoplasts in the presence of PsAvh110 (Figure 4, A and B). These results hint that PsAvh110 may repress or interfere with the formation of the GmPHD6–GmLHP1-2 complex. LHP1 contains an N-terminal chromo domain (CD) and a C-terminal chromo shadow domain (CSD) (Figure 4C; Guan et al., 2011). Previous studies have revealed that the CSD of LHP1 determines its interaction with other proteins (Cowieson et al., 2000; Lechner et al., 2005). Moreover, three residues (Ile437, Leu444, and Leu450) within the CSD are also important for interacting with other proteins (Thiru et al., 2004). To assess which GmLHP1-2 domains are required for its interaction with PsAvh110 and GmPHD6, respectively, we generated four GmLHP1-2 variants: a C-terminal CSD deletion (GmLHP1-2^ΔCSD), a C-terminal truncation without CD domain (GmLHP1-2^CΔCD), a C-terminal truncation with CD and CSD (GmLHP1-2^C), and a site-directed mutant (GmLHP1-2^EHH, with Ile437, Leu444, and Leu450 mutated to Glu [E], His [H], and His [H], respectively) (Figure 4C). Both Co-IP and Y2H assays showed that GmLHP1-2^CΔCD and GmLHP1-2^C, but not GmLHP1-2^ΔCSD or GmLHP1-2^EHH, interact with either PsAvh110 or GmPHD6 (Figure 4, D and E). Notably, PsAvh110, when compared with GmPHD6, displayed a strong interaction with GmLHP1-2 (Figure 4E). Together, these results indicate that the C-terminal CSD is required for GmLHP1-2 to interact with both PsAvh110 and GmPHD6, further leading us to hypothesize that PsAvh110 possibly interferes with the formation of the GmPHD6–GmLHP1-2 complex. Of note, N. benthamiana leaves expressing GmLHP1-2 exhibited more resistance to P. capsici infection, compared with leaves expressing GFP (Supplemental Figure S5, C–E). However, leaves expressing GmLHP1-2^ΔCSD and GmLHP1-2^EHH did not exhibit resistance to P. capsici infection compared with those leaves expressing GmLHP1-2^CΔCD and GmLHP1-2^C (Supplemental Figure S5, C–E). Similarly, leaves expressing GmPHD6 also exhibited more resistance to P. capsici infection than leaves expressing GFP only (Supplemental Figure S5, F–H). Together, these results suggest that both PsAvh110 and GmPHD6 bind to the C-terminal CSD of GmLHP1-2, which plays a pivotal role in plant immunity.

Figure 4.

Both PsAvh110 and GmPHD6 interact with the C terminus of GmLHP1-2. A, Y2H assay showing the interaction between GmLHP1-2 and PsAvh110 or GmPHD6. Ten-fold serial dilutions of yeast cells harboring the indicated constructs were spotted on SD medium without leucine and tryptophan (SD–2), or without histidine, leucine, adenine, and tryptophan (SD–4) supplemented with 0.2 mM X-α-gal. B, PsAvh110 reduces the association between GmLHP1-2 and GmPHD6 in soybean protoplasts. Protoplasts were co-transfected with GFP-GmLHP1-2 and Flag-GmPHD6 with or without PsAvh110-HA for 12 h. Co-IP assays were performed with GFP-Trap beads, followed by IB with anti-Flag or anti-GFP antibody (top two blots) with input proteins shown (bottom three blots). Protein loading is shown by Ponceau S staining (Ponc.) of Rubisco large subunit (RBCL). C, Schematic diagram of GmLHP1-2 and its truncated variants. GmLHP1-2 truncation variants GmLHP1-2^ΔCSD (without the CSD motif), GmLHP1-2^CΔCD (C terminus without the CD motif), GmLHP1-2^C (C terminus with CD and CSD motifs), and GmLHP1-2^EHH (with residues I437,L444,L450 changed to E437,H444,H450) are shown. D, The C-terminal CSD of GmLHP1-2 is required for interacting with both PsAvh110 and GmPHD6 in yeast. The Y2H assay was performed as described in (A). E, Co-IP assay showing the interaction of GmLHP1-2 derivatives with PsAvh110 or GmPHD6. Total protein was extracted from N. benthamiana leaves co-expressing different GmLHP1-2 derivatives with EV (control), PsAvh110, or GmPHD6. Co-IP assays were performed as described in (B).

PsAvh110 competes GmPHD6 for binding to GmLHP1-2

We investigated whether PsAvh110 competitively interferes with the association of GmLHP1-2 with GmPHD6. Co-IP assays showed that the protein levels of Flag-GmPHD6 co-immunoprecipitated by GFP-GmLHP1-2 are notably lower when the encoding constructs were co-expressed with Flag-PsAvh110 relative to Flag-RFP and Flag-PsAvh110^mW2 (Figure 5A). Pull-down assays further confirmed that the abundance of Myc-GmPHD6 co-immunoprecipitated by GST-GmLHP1-2 is lower when the amount of His-PsAvh110 protein increased (Figure 5B), suggesting that PsAvh110 can interfere with the interaction between GmLHP1-2 and GmPHD6 in a dose-depended manner. In addition, microscale thermophoresis (MST) assays showed that the dissociation constant (K_d) of GmLHP1-2 with PsAvh110 is 7.03 µM ± 1.75, whereas the K_d of GmLHP1-2 with GmPHD6 was 44.83 µM ± 8.24 (Figure 5C), suggesting that PsAvh110 displays a higher binding affinity toward GmLHP1-2 than does GmPHD6. This result was consistent with the Co-IP assays that had indicated that PsAvh110 had a stronger interaction with GmLHP1-2, compared with GmPHD6 (Figure 4E). Together, these results indicate that PsAvh110 competes with GmPHD6 to bind to GmLHP1-2, thereby disrupting the formation of the GmPHD6/GmLHP1-2 complex.

Figure 5.

PsAvh110 competes with GmPHD6 for binding to GmLHP1-2 in vitro and in vivo. A, PsAvh110 disrupts the GmPHD6–GmLHP1-2 interaction in a co-IP assay. GFP-GmLHP1-2 and Flag-GmPHD6 were co-transfected with or without Flag-PsAvh110 or Flag-PsAvh110^mW2 in N. benthamiana leaves. Total protein was incubated with GFP-Trap beads for co-IP assays, and immunoprecipitated proteins were analyzed using anti-GFP, or anti-Flag antibodies (top two blots). Input proteins are shown in blots 4 and 5. The black arrows indicate the multiple bands of GFP-GmLHP1-2. Protein loading is indicated by Ponceau S staining of RBCL (bottom panel). B, PsAvh110 competes with GmPHD6 to bind to GmLHP1-2 in vitro in a dose-dependent manner. Recombinant GST-GmLHP1-2 immobilized on glutathione sepharose beads was incubated with His (control), Myc-GmPHD6, and increasing amounts of His-PsAvh110 protein (different gradient dilutions: 1×, 2×, 3×). Protein eluates from washed beads were used for immunoblot with anti-GST, anti-His, or anti-Myc antibodies. C, PsAvh110 displays a higher binding affinity to GmLHP1-2 than GmPHD6 in MST assays. Recombinant purified GmLHP1-2 was labeled with a fluorophore, and purified PsAvh110, GmPHD6, and MBP proteins were used as flow-through analytes for MST assays. Left: dose–response curves of PsAvh110 proteins at gradient concentrations flowing through immobilized and labeled GmLHP1-2; the calculated K_d is 7.03 µM. Middle: dose–response curves of PsAvh110 at gradient concentrations flowing through immobilized and labeled GmLHP1-2; the calculated K_d is 44.83 µM. Right: dose–response curves of PsAvh110 at gradient concentrations flowing through immobilized and labeled GmLHP1-2; no binding affinity was observed between GmLHP1-2 and MBP proteins. The above experiments were performed three times with similar results.

GmLHP1-2 and GmPHD6 positively regulate soybean immunity

To determine the function of GmLHP1-2 and GmPHD6 in soybean immunity, we knocked down the transcript levels of either GmLHP1-2 or GmPHD6 in soybean hairy roots (RNAi-GmLHP1-2 or RNAi-GmPHD6). RT-qPCR showed that the transcript levels of GmLHP1-2 or GmPHD6 are significantly reduced in the corresponding two independent lines, respectively (Figure 6A). We also generated hairy roots knocked down for both GmLHP1-2 and GmPHD6 in two independent lines (RNAi-GmLHP1-2/GmPHD6-L1 and -L2) (Figure 6A). We challenged GmLHP1-2-, GmPHD6-, or GmLHP1-2/GmPHD6-knocked down roots with RFP-labeled P. sojae strain P6497. We determined that the number of oospores produced in GmLHP1-2- and GmPHD6-knock down roots increases (Figure 6, B and C). Moreover, the relative biomass of P. sojae was higher in GmLHP1-2- and GmPHD6-knock down roots, compared with EV roots (Figure 6D). Consistent with the above observations, overexpression of GmLHP1-2 or GmPHD6 in N. benthamiana leaves resulted in enhanced resistance to P. capsici infection (Supplemental Figure S5, C–H), indicating that the function of GmLHP1-2 and GmPHD6 is conserved in plant responses to pathogen infection. Of note, GmLHP1-2/GmPHD6-knock down roots displayed higher disease susceptibility phenotypes in terms of oospores production and relative biomass of P. sojae compared with GmLHP1-2- or GmPHD6-knock down roots (Figure 6, B–D), suggesting that GmLHP1-2 and GmPHD6 likely form a complex to positively regulate plant immunity.

Figure 6.

The GmLHP1-2/GmPHD6 complex positively regulates plant immunity. A, Knockdown efficiency of GmLHP1-2 or GmPHD6 in corresponding hairy roots as indicated by RT-qPCR. Total RNA was extracted from RNAi-GmLHP1-2, RNAi-GmPHD6, or RNAi-GmLHP1-2/GmPHD6 hairy roots. Gene expression was normalized to the soybean internal control GmCYP2. The data are shown as means ± sd from three replicates. Asterisks represent significant differences (***P < 0.001; ****P < 0.0001; Student’s t test). B and C, RNAi-GmLHP1-2, RNAi-GmPHD6, and RNAi-GmLHP1-2/GmPHD6 hairy roots enhance soybean susceptibility to P. sojae infection. Oospore production in RNAi-GmLHP1-2, RNAi-GmPHD6, and RNAi-GmLHP1-2/GmPHD6 hairy roots infected by WT P. sojae labeled with RFP (WT-RFP) was imaged under a microscope (B) and calculated at 48 hpi (C). The data are shown as an overlay of dot plots with means ± sd, n = 20. Different letters indicate significant differences (P < 0.01; one-way ANOVA). Scale bars, 0.2 mm. The experiments were performed at least three times with similar results. D, Relative biomass of P. sojae in RNAi-GmLHP1-2, RNAi-GmPHD6, and RNAi-GmLHP1-2/GmPHD6 hairy roots, as determined by qPCR at 48 hpi of the ratio of P. sojae genome DNA compared with soybean genome DNA. The data are shown as means ± sd, n = 3. Different letters indicate significant differences (P < 0.01; one-way ANOVA). E, Relative expression levels of immune-associated genes in RNAi-GmLHP1-2 and RNAi-GmPHD6 hairy roots infected with P. sojae, as determined by RT-qPCR at 6 hpi. Gene expression levels were normalized to the internal control gene GmCYP2. The data are shown as means ± sd from three biological repeats. Asterisks indicate significant differences (****P < 0.0001; Student’s t test). F, Sequence alignment of the cis-elements within the 1-kb promoter before the translation start codon of the selected immune-associated genes. The sequences were aligned using ClustalW and displayed in ESPript3. The conserved residues of GREs are highlighted in black, and the GREs are highlighted by the black box. The consensus of the GREs was analyzed by WebLogo 3. G, Occupancy of GmPHD6 at different promoter regions of selected immune genes, as assessed by ChIP-qPCR in 35S:GmPHD6-FLAG transgenic hairy roots. IgG was used as the negative control. The ChIP signals were normalized to input, and the data are means ± sem of three replicates. Asterisks indicate significant differences (*P < 0.05; **P < 0.01; ***P < 0.001; Student’s t test). P1–P3 represent the promoter regions examined by ChIP-qPCR. GREs in the promoter are shown as red triangles.

GmLHP1-2 and GmPHD6 regulate PsAvh110-suppressed immune genes

Since PsAvh110 suppressed the expression of a subset of immune-associated genes in soybean hairy roots (Figure 2C and Supplemental Figure S3A), we next investigated whether GmLHP1-2 and GmPHD6 are involved in regulating the expression of these PsAvh110-suppressed genes. Accordingly, we challenged GmLHP1-2- and GmPHD6-knock down roots with P. sojae, before analyzing PsAvh110-suppressed immune genes by RT-qPCR. Compared with EV-transformed roots, GmLHP1-2- and GmPHD6-knock down roots displayed lower expression levels for GmPR1, GmBAK1-3, GmWRKY71-1/2, GmGCN5-like, GmGF14, GmDRP, and GmGST, but not those of GmWAK, GmCYP81D1, or GmCYP71B21 (Figure 6E). These results suggested that GmLHP1-2 and GmPHD6 positively regulate PsAvh110-suppressed immune genes to regulate plant immunity. GmLHP1-2 and GmPHD6 were shown to form a transcriptional activation complex to regulate the expression of salt response genes via the GREs (GTGGNG/GNGGTG) in their gene promoter (Wei et al., 2017a). We assessed the GREs in the promoter regions (1 kb upstream of the start codon) of GmPR1, GmBAK1-3, GmWRKY71-1/2, GmGCN5-like, GmGF14, GmDRP, and GmGST. As shown in Figure 6F and Supplemental Figure S6A, these GmLHP1-2- and GmPHD6-regulated genes contained the GRE (GTGGNG/GNGGTG) in their promoters, indicating the importance of GREs for GmLHP1-2/GmPHD6 to mediate transcription. We further tested whether GmPHD6 can bind to the promoters of these immune-associated genes by chromatin immunoprecipitation followed by qPCR (ChIP-qPCR) with three pairs of primers amplifying different regions of their promoters in soybean transgenic hairy roots expressing 35S:GmPHD6-FLAG (Supplemental Figure S6, B and C). GmPHD6 strongly bound to the P1 region containing the GRE motif in the promoters of GmPR1, GmBAK1-3, GmGCN5-like, GmGF14, and GmGST, although it also exhibited a very lower binding affinity to the P2 and P3 regions in the promoter of these GmLHP1-2/GmPHD6-regulated immune genes (Figure 6G). Notably, GmPHD6 also bound to the P1 and P2 regions containing the GRE motif in the promoters of GmWRKY71-1/2. However, we did not detect any binding for GmPHD6 to P1, P2, or P3 regions of the promoters of GmWAK and GmCYP81D1, which did not contain the GRE motif in their promoters (Figure 6G and Supplemental Figure S6A). In addition, since it was reported that GmPHD6 binds to H3K4me1/2 (Wei et al., 2017b), we determined the H3K4me2 levels at the promoter regions of these immune genes by ChIP-qPCR using. We observed that promoters of GmWRKY71-1/2, GmPR1, GmBAK1-3, GmGCN5-like, and GmGST exhibit high H3K4me2 levels in the absence of pathogen infection, but their H3K4me2 levels only increased slightly after P. sojae infection (Supplemental Figure S6D). Together, the data suggest that the GmLHP1-2/GmPHD6 complex may recognize the GRE motif in the promoter of immune-associated genes to regulate their expression.

GmLHP1-2/GmPHD6 possesses transcriptional activation activity to regulate gene expression

We asked whether the GmLHP1-2/GmPHD6 complex can regulate the expression of immune-associated gene via targeting the GREs in the promoter by establishing a LUC reporter system using the promoter of GmLHP1-2/GmPHD6-regulated immune genes (Figure 6E). To this end, we cloned a 1-kb fragment of selected promoters up to the translation start codon upstream of the LUC gene to yield proGeneX:LUC (GeneX being GmBAK1-3, GmPR1, GmWRKY71-2, GmGCN5-like, GmGF14, GmWAK, or GmCYP71B21) constructs. To eliminate the possible effect of NbLHP1s in transcriptional activation when performing the LUC assay in N. benthamiana, we silenced all three NbLHP1 homologs; RT-qPCR analysis showed that their transcript levels are strongly reduced in the silenced leaves (Supplemental Figure S7A). We measured very low LUC activity in leaves infiltrated with proGmBAK1-3:LUC, proGmPR1:LUC, proGmWRKY71-2:LUC, proGmGCN5-like:LUC, and proGmGF14:LUC, whereas proGmWAK:LUC and proGmCYP71B21:LUC were associated with high LUC activity (Supplemental Figure S7B). These results support the notion that NbLHP1s-silenced leaves can be used for testing the role of GmLHP1-2/GmPHD6 in activating the expression of immune genes. Notably, GmBAK1-3, GmPR1, GmWRKY71-2, GmGCN5-like, and GmGF14 transcription were activated when N. benthamiana leaves co-expressed GFP-GmLHP1-2 and Flag-GmPHD6 compared with control leaves co-expressing GFP and Flag-RFP (Figure 7A). In line with the above observation, leaves co-expressing GFP-GmLHP1-2 and Flag-RFP pair, or the GFP and Flag-GmPHD6 pair, failed to activate the transcription of GmBAK1-3, GmPR1, GmWRKY71-2, GmGCN5-like, or GmGF14 (Figure 7A), indicating that GFP-GmLHP1-2 or Flag-GmPHD6 alone does not have transcriptional activation potential.

Figure 7.

PsAvh110 suppresses the transcriptional activity of the GmLHP1-2/GmPHD6 complex. A, The GmLHP1-2–GmPHD6 complex regulates the promoter activity of GmBAK1-3, GmPR1, GmGCN5-like, GmWRKY71-2, and GmGF14 in NbLHP1s-silenced plants. NbLHP1s-silenced N. benthamiana leaves were infiltrated with Agrobacterium cultures carrying the indicated constructs; relative LUC activity was measured at 48 hpi using a chemiluminescent imaging system. B, Schematic diagram of GREs of the GmBAK1-3 and GmPR1 promoters and their mutants. The DNA probes (GmBAK1-3^GRE and GmPR1^GRE) containing GREs from the GmBAK1-3 and GmPR1 promoters are shown. The GRE motifs of GmBAK1-3^GRE and GmPR1^GRE mutated to red-labeled adenine (A) are labeled GmBAK1-3^mGRE and GmPR1^mGRE. C and D, GmLHP1-2/GmPHD6 activates GmBAK1-3 transcription via the GREs. A 1-kb fragment of the GmBAK1-3 promoter before the start codon containing the GREs (proGmBAK1-3) or the GRE mutation (proGmBAK1-3m) shown in (B) was cloned upstream of the LUC gene to generate proGmBAK1-3/ProGmBAK1-3m:LUC-Nos constructs. C, N. benthamiana leaves were infiltrated with Agrobacterium strains carrying the indicated constructs as shown in the bottom panel, and LUC activity was quantified at 48 hpi (D). Ten individual leaf discs were collected for LUC activity measurement; data are shown as means ± sd (n = 10). Different letters indicate significant differences (P < 0.01, one-way ANOVA). E and F, GmLHP1-2/GmPHD6 activates GmPR1 transcription via the GREs. A 1-kb fragment of the GmPR1 promoter before the start codon containing the GREs (proGmPR1) or the GRE mutation (ProGmPR1m) shown in (B) was cloned upstream of the LUC gene to generate proGmPR1/ProGmPR1m:LUC-Nos. The experiments were performed as in (C and D). Different letters indicate significant differences (P < 0.01, one-way ANOVA). G, GmPHD6 binds to the GREs of the GmBAK1-3 and GmPR1 promoters in a dose-dependent manner. Biotin-labeled probes were incubated with increasing amounts of recombinant purified MBP-GmPHD6 in the absence or presence of 125-fold excess of the corresponding unlabeled probes. The bands corresponding to the DNA–protein complexes (shifted) or free probes are indicated by arrows. Protein loading was determined by Coomassie Brilliant Blue (CBB) staining (bottom panel). H and I, The transcription of GmBAK1-3 activated by GmLHP1-2/GmPHD6 is suppressed by PsAvh110. The experiments were performed as in (C and D). Different letters indicate significant differences (P < 0.01, one-way ANOVA). J and K, The transcription of GmPR1 activated by GmLHP1-2/GmPHD6 is suppressed by PsAvh110. The experiments were performed as described in (C and D). Different letters indicate significant differences (P < 0.01, one-way ANOVA).

GmLHP1-2/GmPHD6 activates the expression of immune genes via binding to GREs in their promoters

We then selected GmBAK1-3 and GmPR1 as example candidate genes to further test whether the GREs in their promoters are required for GmLHP1-2/GmPHD6-mediated transcriptional activation (Figure 7B). We replaced all the nucleosides in the GREs (GTGGNG) to adenines (AAAAAA) and designated the resulting mutant constructs GmBAK1-3^mGRE and GmPR1^mGRE, respectively (Figure 7B). LUC reporter assays showed that the mutated proGmBAK1-3m and proGmPR1m promoters (mutation of GRE to mGRE in the promoter of GmBAK1-3 or GmPR1) exhibit a lower LUC activity than their corresponding intact proGmBAK1-3 and proGmPR1, respectively, when co-expressed with the GmLHP1-2/GmPHD6 complex in NbLHP1s-silenced leaves (Figure 7, C–F). These observations suggest that the GRE motifs play an important role in GmLHP1-2/GmPHD6-mediated transcriptional activation. In all the negative control pairs, including the pairs GFP/RFP, GFP/GmPHD6, and GmLHP1-2/RFP, none had any effect on LUC activity (Figure 7, C–F).

PsAvh110 represses transcription in a GmLHP1-2/GmPHD6-dependent manner

To test whether GmPHD6 or GmLHP1-2 can directly bind to the GRE motifs of GmBAK1-3 and GmPR1, we synthesized probes of 40–50 bp from the promoters of GmBAK1-3 and GmPR1 containing the GRE motif (GmBAK1-3^GRE and GmPR1^GRE) or their corresponding mutants (GmBAK1-3^mGRE and GmPR1^mGRE). We then labeled each probe with biotin to conduct electrophoretic mobility shift assays (EMSAs) (Figure 7B). Notably, EMSA results showed that recombinant purified MBP-GmPHD6 can bind to both GmBAK1-3^GRE and GmPR1^GRE probes, but not the GmBAK1-3^mGRE or GmPR1^mGRE probes (Figure 7G). However, recombinant GmLHP1-2 did not bind to GmBAK1-3^GRE or GmPR1^GRE probes, nor did it have any effect on the binding affinity of GmPHD6 to the GRE motif (Supplemental Figure S7C). These results suggested that the GRE motifs are required for GmPHD6 binding to the promoters of immune-associated genes. In addition, GmPHD6 was able to also bind to the GRE motifs of the GmWRKY71-2^GRE, GmGCN5-like^GRE, and GmGF14^GRE probes, but did not bind to the DNA probe of the GmWAK and GmCYP71B21 promoters, which lack a GRE motif (Supplemental Figure S7D). Together, the results demonstrate that GmLHP1-2/GmPHD6 binds to GRE motifs to modulate promoter activity of immune genes, thereby activating the expression of immune genes.

We next assessed whether PsAvh110 suppresses the expression of immune genes that are under the control of the GmLHP1-2/GmPHD6 complex, again using recombinant purified proteins and EMSAs. We observed that His-PsAvh110 does not bind to, nor did it affect the binding affinity of GmPHD6 to the GmBAK1-3^GRE and GmPR1^GRE probes (Supplemental Figure S7C). However, LUC reporter assays showed that PsAvh110 can repress the activity of the GmBAK1-3 and GmPR1 promoters, which are regulated by the GmLHP1-2/GmPHD6 complex, when infiltrated in NbLHP1s-silenced leaves (Figure 7, H–K and Supplemental Figure S7E). In both experiments, the variant PsAvh110^mW2, which can no longer interact with GmLHP1-2, failed to suppress the promoter activities of GmBAK1-3 and GmPR1 when co-expressed with GmLHP1-2/GmPHD6 in NbLHP1s-silenced leaves (Figure 7, H–K and Supplemental Figure S7E). Together, these data suggested that PsAvh110 suppresses the expression of immune genes by disrupting the formation of the GmLHP1-2/GmPHD6 complex, leading to lower transcriptional activation activity.

Discussion

During infection, pathogens deploy different types of effectors to interfere with plant immunity through diverse regulatory mechanisms (Kourelis and van der Hoorn, 2018; Wang et al., 2019; He et al., 2020a). Here, we showed that the P. sojae nucleus-localized effector PsAvh110 suppresses plant immunity via GmLHP1-2/GmPHD6-mediated transcriptional reprogramming of immune-associated gene expression. The gmLHP1-2/GmPHD6 complex is an important immune regulator whose loss of function increases plant susceptibility to P. sojae infection. In the absence of PsAvh110, GmPHD6 binds to GREs in the promoters of immune genes to recruit GmLHP1-2 and form a transcriptional activation complex upon P. sojae infection, thereby activating the transcription of immune-associated genes (e.g. GmBAK1-3 and GmPR1) and initiating robust plant immunity. PsAvh110, which has a high binding affinity toward GmLHP1-2, blocks the formation of the GmLHP1-2/GmPHD6 complex when present in plant cells, thereby suppressing the transcriptional activity of the complex and leading to reduced plant immunity (Figure 8). Thus, our study sheds light on a previously unidentified infection strategy by which a pathogen suppresses plant immunity by secreting a nucleus-localized effector to interfere with the activity of a transcriptional complex and manipulate plant transcriptional programs.

Figure 8.

A simplified working model of PsAvh110 functions in modulating plant transcriptional reprogramming during P. sojae infection. In the absence of PsAvh110, GmPHD6 binds to the GRE in the promoter of immunity genes to recruit GmLHP1-2 and form a transactivation complex during P. sojae Psavh110 mutant infection, thereby activating the expression of immune-associated genes and mounting robust plant immunity. In the presence of PsAvh110, PsAvh110 has a high binding affinity toward GmLHP1-2, which disrupts the formation of the GmLHP1-2/GmPHD6 complex, thereby suppressing its transcriptional activity, resulting in the suppression of the expression of immune genes and of plant immunity.

The plant defense machinery is very sophisticated and tightly controlled by very complicated immune networks (Yuan et al., 2017; Wang et al., 2019; Kong et al., 2021). For example, plants perceive PAMPs to activate PTI, including accumulating the defense phytohormones SA and jasmonic acid (JA), initiating ROS production, as well as establishing transcriptional reprogramming of defense genes to launch robust immune responses upon infection (Wang et al., 2019; He et al., 2020b). Therefore, this network prevents pathogens from successfully bypassing the defense response. As a counter mechanism, pathogens have evolved diverse effectors to modulate host defense gene expression and facilitate infection. In this study, PsAvh110 offered an interesting example of a conserved effector that targets a host transcriptional complex that is necessary for activating the expression of a subgroup of defense genes containing GREs in their promoters during pathogen attack, thereby suppressing plant immunity. Thus, one effector can directly inhibit a subset of defense genes by regulating upstream transcriptional complexes and the activity of their promoters. PsAvh110 competes with GmPHD6 to bind to the same region of GmLHP1-2, making it difficult for GmLHP1-2 to accumulate mutations that might protect plants against pathogen infections without also losing function. Together, these results suggest that P. sojae has evolved a core effector, PsAvh110, to target a conserved plant chromatin-binding module for suppressing plant immune responses.

HP1 is an evolutionarily conserved protein in most eukaryotes and plays an important role in the maintenance of heterochromatin and euchromatin integrity, replication and repair of DNA, RNA splicing, and transcriptional regulation (Vakoc et al., 2005; Lomberk et al., 2006; Prasanth et al., 2010; Feng and Lu, 2017). In humans, three HP1s (HP1α/β/γ) have different functions in repressing or activating gene expression (Kwon and Workman, 2011; Li et al., 2013; Canzio et al., 2014). For instance, mammalian HP1γ positively regulates heat shock protein 70 (Hsp70) expression in response to heat-shock stress. Similarly, Drosophila melanogaster HP1α and HP1c activate the expression of genes in euchromatic regions (Piacentini et al., 2003; Lin et al., 2008). In plants, the Arabidopsis and soybean genomes encode one LHP1 and four LHP1s (GmLHP1-1/2/3/4), respectively (Turck et al., 2008; Wei et al., 2017a). Arabidopsis LHP1 was shown to regulate flowering time and plant development (Nakahigashi et al., 2005; Exner et al., 2009). We showed here that soybean GmLHP1-2 formed a complex with GmPHD6 to positively regulate plant immunity against P. sojae infection. Similarly, Arabidopsis LHP1 also positively regulates plant immunity against bacterial infection by suppressing SA production (Ramirez-Prado et al., 2019). However, a recent study showed that the GmLHP1-1, the GmLHP1-2 homolog, negatively regulates plant immunity (Zhang et al., 2021), indicating that LHP1 homologs could play dual roles in plant immunity. Our data indicated that PsAvh110 specifically interacts with GmLHP1-2, but not GmLHP1-1, suggesting that plant immune responses are tightly controlled by different regulators and pathogens can deploy virulence effectors, such as PsAvh110, to specifically manipulate the plant immune response.

PHD finger proteins were first discovered in Arabidopsis and maize (Schindler et al., 1993) and were also found in humans, where they are involved in regulating chromatin structure and dynamics via neighboring epigenetic marks and adjacent effectors (Musselman and Kutateladze, 2011). Human PHD finger proteins have been reported to mainly recognize unmodified histone H3 tails, and di- or trimethylated H3 at Lys4 (H3K4me2/3) (Jain et al., 2020). Similarly, a series of studies reported that Arabidopsis PHD domain proteins, such as EBS, SHL, AIPP2, and PAIPP2, recognize unmodified H3K4 and H3K4me2/3 marks (Qian et al., 2018; Yang et al., 2018; Zhang et al., 2020), hinting that the function of PHD domain proteins is conserved across human and plants. In contrast to H3K4me3 as an active mark, H3K27me3 is a repressive mark with opposite roles in balancing chromatin and transcriptional states (Piunti and Shilatifard, 2016). Interestingly, EBS and SHL can also recognize H3K27me3 via their BAH domain to maintain the levels of H3K4me3 and H3K27me3 for the regulation of floral gene expression (Qian et al., 2018; Yang et al., 2018). Several PHD proteins only recognize H3K4me3 through binding to H3K27me3 readers such as LHP1, leading to a switch in binding preference between H3K4me3 and H3K27me3 (Molitor et al., 2014). This binding balance is crucial for plant growth and development, since disruption of these genes caused earlier floral transition (Molitor et al., 2014). In addition to illustrating how the PHD domain proteins and LHP1 recognize two opposite histone marks, our work reveals the underlying mechanism by which soybean GmPHD6 and GmLHP1 form a transcriptional activation complex to regulate immune-associated gene expression during plant–pathogen interactions. We proposed here that the transcriptional activation by the GmPHD6–GmLHP1 complex is independent from their binding activity to H3K4me3 and H3K27me3, because we provided evidence that DNA-binding activity was important for GmPHD6 function in regulating immune gene expression. However, we cannot rule out that the DNA-binding activity of GmPHD6 might affect its binding activity to histone marks as well, such as H3K4me3. Structural and biochemical studies showed that several residues of the Arabidopsis PHD proteins EBS and SHL are required for their binding affinity to histone marks (Qian et al., 2018; Yang et al., 2018). It will be of interest to investigate whether these residues are required for the DNA-binding affinity of GmPHD6, which will increase our understanding of the underlying mechanisms of the GmPHD6/GmLHP1 complex as a transcriptional activation complex. Consequently, the regulatory mechanism by which PsAvh110 attacks the chromatin complex is of broad significance for understanding host–pathogen interactions.

PR1 and BAK1 are inherent components of innate immunity in plants and play very important roles in SA and PRR signaling pathways (Breen et al., 2017; Kong et al., 2021). PR1 family members are highly conserved in plants and have unique anti-oomycete activity (Niderman et al., 1995; Luo et al., 2022). For example, PR1 from tomato and tobacco can inhibit zoospore germination of P. infestans and restrict pathogen colonization in tomato, although the regulatory mechanism remains unclear (Niderman et al., 1995). Intriguingly, a recent study showed that potato (Solanum tuberosum) PR1s can translocate into P. infestans upon infection to target the pathogen AMP-activated protein kinase (AMPK) complex and reduce pathogenicity, further suggesting the importance of PR1 function in anti-oomycete activity (Luo et al., 2022). In addition, PR1 proteins exhibit antifungal and antibacterial activities (Breen et al., 2016). These studies suggest that PR1 proteins confer broad-spectrum resistance to different plant pathogens. To overcome PR1-mediated plant immunity during pathogen infection, our work provides novel insights into the molecular function of PsAvh110 in suppressing PR1 transcription by targeting the heterochromatin complex to reduce plant immunity. Similar to PR1, BAK1 was also shown to regulate plant immunity against different Phytophthora infections. Soybean GmBAK1-1 and N. benthamiana NbBAK1 positively regulate plant resistance to P. sojae and P. infestans infection, respectively (Chaparro-Garcia et al., 2011). BAK1 acts as a coreceptor for multiple PRRs to sense and convert extracellular signals into intracellular signals, which then activate downstream signaling cascades (Kong et al., 2021). Given the importance of BAK1 in plant PRR signaling, plant pathogens have evolved core virulence effectors to target BAK1 through regulating its protein stability, kinase activity, and its interactions with different receptors, thereby leading to the suppression of the PRR-mediated immune responses (Saijo et al., 2018). In contrast to regulating BAK1 protein stability by effectors, we uncovered an unreported mechanism whereby PsAvh110 hijacks GmLHP1 from the GmLHP1-2/GmPHD6 complex, thereby inhibiting GmBAK1 transcription and repressing plant immunity. To date, several MAMPs have been identified from Phytophthora species, such as INF1, nlp20, and XEG1, which can be recognized by the plant PRRs ELICITIN RESPONSE, RECEPTOR LIKE PROTEIN 23, and response to XEG1 (RXEG1), respectively, to relay the downstream signal in a BAK1-dependent manner. Thus, it is possible that PsAvh110 suppresses BAK1 transcription to block or suppress PRR-mediated signaling during the plant–Phytophthora interaction. In addition, PsAvh110 is conserved in other Phytophthora species, thus the regulatory mechanism by which PsAvh110 disrupts the assembly of heterochromatin complexes may be a common strategy utilized by Phytophthora to subvert plant immunity.

Understanding the molecular functions of virulence effectors in modulating plant defense responses will contribute to the framework of elucidating plant–pathogen interactions, and shed light on the putative role of their regulatory targets for disease control. Our study provided clear evidence uncovering that PsAvh110 inhibits the expression of soybean key immune genes GmPR1 and GmBAK1 via disrupting the assembly of the GmLHP1-2/GmPHD6 complex on GREs in promoters. Moreover, we found that the promoters of a subset of immune genes regulated by PsAvh110 contain GREs, such as GmGF14, GmGCN5, and GmWRKY71, which are also involved in plant immunity. Thus, GREs in the promoter of immune genes play a crucial role in regulating transcriptional activation, which is tightly regulated by pathogen virulence effectors during infection. Similarly, a genome-wide association study (GWAS) revealed that a natural polymorphism in the promoter of the rice C₂H₂-type transcription factor gene bsr-d1 (broad-spectrum resistance Digu 1) confers broad-spectrum resistance against rice blast fungus (Li et al., 2017). Therefore, our work, together with the previous study, provides a theoretical basis for breeding durable disease-resistant crops by editing the specific nucleotide in the promoters of immune-associated genes using gene-editing technology.

Materials and methods

Plant and microbe cultivation

Soybean (G. max) Hefeng47 cultivar and N. benthamiana were grown at 24°C–25°C in a growth chamber with a 16-h light/8-h dark photoperiod under the LED lamps at a light intensity of ∼120 − 150 μmol m⁻²s⁻¹. Phytophthora sojae (P6497) and P. capsici (LT263) were cultivated in 25°C incubators on V8 juice medium plates in the dark for 3 days before inoculation. Yeast strain AH109 was incubated at 30°C for 3 days for Y2H assays. Agrobacterium (Agrobacterium tumefaciens) strain GV3101 and Agrobacterium rhizogenes strain K599 were grown on LB medium containing 50 μg/mL rifampicin or streptomycin, respectively, at 30°C overnight before being used for the indicated experiments. The Escherichia coli strain JM109 for vector construction was grown on LB medium at 37°C.

Soybean hairy root transformation and P. sojae infection assays

For silencing GmLHP1-2 and GmPHD6 in soybean hairy roots, 200-bp DNA fragments of the indicated genes (GmLHP1-2 or GmPHD6) without predicted off-targets were designed via the Solanaceae Genomics Network (https://solgenomics.net) and were amplified from soybean Hefeng47 cDNA with primers containing AscI at the 5′ end and BamHI at the 3′ end (Supplemental Data Set S2). The two fragments were individually inserted into the pFGC5941-GFP vector, cloned in-frame with a Chalcone synthase linker from Petunia hybrida. To silence both GmLHP1-2 and GmPHD6 in soybean hairy roots, the same DNA fragments for GmLHP1-2 and GmPHD6 were ligated together and cloned into the pFGC5941-GFP vector. Soybean seeds were cultivated at 25°C the greenhouse for 6 days and the cotyledons were collected. Soybean cotyledons were surface sterilized with 10% (v/v) HClO for 20 min, then soaked in 75% (v/v) ethanol for 1 min, and washed with autoclaved ddH₂O five times. A small wound was made by cutting the lower epidermis of each sterile cotyledon, which was infected with a 10-µL suspension of A. rhizogenes cells (K599) (OD₆₀₀ = 0.5) carrying the indicated construct. The soybean cotyledons were grown on Murashige and Skoog (MS) medium in 25°C incubators without light. After 3 weeks, soybean transgenic hairy roots were screened by fluorescence stereomicroscopy (Nikon, SMZ25) using the GFP filter and further confirmed by RT-qPCR analysis. The hairy roots were inoculated with P. sojae mycelium, and the oospores and biomass were analyzed at 48 hpi.

To generate the transgenic soybean hairy roots expressing pro35S:3xFlag-GmPHD6, the GmPHD6 cDNA was ligated into the pFGC5941-3xFlag vector and was transformed into A. rhizogenes strain K599. The transgenic roots were screened using the herbicide BASTA at a concentration of 3 mg/mL (Bayer; resistance conferred by the pFGC5941 vector). Hairy roots were collected for detection of GmPHD6 protein using anti-Flag antibody (Abmart, Cat No. M20008M) after growth on Basta selection medium plates for 3 weeks.

Y2H assay

A Y2H screen with pLAW10-PsAvh110 was performed as per the Clontech Yeast Protocol Handbook. Specifically, the sequences encoding PsAvh110 without the signal peptide, its variants, or GmPHD6 were individually cloned into the pLAW10 vector as bait. The sequences encoding GmLHP1-2, its variants, or GmPHD6 were individually cloned into the pLAW11 vector. The resulting pLAW10 and pLAW11 constructs were transformed into yeast strain AH109 and selected on synthetic defined (SD) media without Trp, Leu (SD–2), and Trp, Leu, His, Ade (SD–4) containing 0.2 mM X–α–gal. Yeast cells were grown at 30°C for 3–4 days before being photographed.

LUC reporter assays

For dual-LUC reporter assays in N. benthamiana protoplasts, 2-kb promoter fragments of NbBAK1, GmBAK1-3, and NbActin2 were individually ligated into the pHBT-LUC vector. The indicated effectors were ligated into the pHBT-pro35S vector, respectively. The preparation of N. benthamiana mesophyll protoplasts was performed according to a previous method (Yoo et al., 2007). Nicotiana benthamiana protoplasts were co-transfected with the effector, reporter, and control constructs. pHBT-proNbBAK1:LUC or pHBT-proGmBAK1-3:LUC was used as reporter construct. pHBT-pro35S:Avh110-GFP, pHBT-pro35S:Avh23-GFP, or pHBT-pro35S:Avr3c-GFP was used as effector construct. pHBT-proNbActin2:RLUC was used as the internal control construct. Nicotiana benthamiana protoplasts, transfected with the indicated constructs, were incubated at room temperature for 12 h and treated with or without 0.2 μM XEG1 at the indicated time. Protoplasts were collected by centrifugation and used to determine LUC activity with a dual-LUC reporter system (Promega, Cat No. E2920) following the protocol provided by Promega. Statistical analysis was performed using one-way analysis of variance (ANOVA) for multiple comparisons.

For LUC reporter assays in N. benthamiana leaves, 1-kb promoter fragments of the selected genes were individually amplified from soybean genomic DNA and ligated into pCAMBIA1300-LUC. Four- or five-week-old N. benthamiana leaves were infiltrated with Agrobacterium strain GV3101 harboring the indicated constructs resuspended in infiltration buffer (10 mM MgCl₂, 10 mM MES pH 5.6, and 100–200 μM acetosyringone) for 36 h, after which the leaves were sprayed with 1 mM d-luciferin substrate and incubated for 5 min at the dark. The sprayed leaves were photographed using a Tanon-5200 Multi Chemiluminescent Imaging System (Tanon, China). For quantification of LUC activity, N. benthamiana leaf discs were collected at 36 h after infiltration (hpi) and incubated with 1 mM d-luciferin in a 96-well plate, and LUC activity was quantified with a Promega GloMax Navigator microplate luminometer (Promega, USA). GraphPad Prism 8.0 software was used for data analysis and statistical analysis was performed using one-way ANOVA for multiple comparisons.

Co-IP assay

Leaves of 6-week-old N. benthamiana plants were infiltrated with Agrobacterium cultures carrying the indicated constructs. At 36–48 hpi, the infiltrated leaves were collected, frozen in liquid nitrogen, and ground to a powder using a mortar and pestle. The ground powder was mixed in lysis buffer (10 mM Tris–HCl pH 7.5, 150 mM NaCl, 0.5 mM EDTA, 0.5% [v/v] Triton-100, and protease inhibitor cocktail [Sigma-Aldrich]) and the samples were mixed well by vortexing. The samples were centrifuged at 4°C for 15 min at 12,000 g and the supernatant was transferred to a new 1.5-mL tube. For the GFP-Trap-IP assay, 1 mL of supernatant was incubated with 10 μL of GFP-Trap beads (Chromotek, Germany) at 4°C for 3–4 h. The beads were collected by centrifugation at room temperature for 2 min at 500 g and washed four times with 1 mL washing buffer (10 mM Tris–HCl pH 7.5, 150 mM NaCl, and 0.5 mM EDTA) each time. Bound proteins were resuspended in 30–50 μL 1× loading buffer, followed by boiled at 95°C for 5–10 min. Immunoblots were analyzed with anti-Flag (Abmart, Cat No. M20008M) or anti-GFP (Abmart, Cat No. M20004) antibody at a 1:5,000 dilution.

Soybean protoplast isolation and transfection

Soybean protoplasts were isolated as previously described (Yoo et al., 2007). Briefly, newly expanded unifoliate leaves from about 2-week-old soybean seedlings were cut into 0.5–1-mm strips. The leaf strips were floated in an enzyme solution containing 2% (w/v) cellulase R-10 (Yakult Pharmaceutical Industry) and 0.4% (w/v) pectolyase Y-23 (Yakult Pharmaceutical Industry) at 25°C in the light with gentle agitation (40 rpm). After 4–6 h of digestion, an equal volume of W5 solution (154 mM NaCl, 125 mM CaCl₂, 5 mM KCl, and 2 mM MES pH 5.7) was added and the enzyme–protoplast mixture was filtered through a 75-µm nylon mesh into a 50-mL round bottom tube. The protoplast suspension was centrifuged for 2 min at 100 g. After gently removing the supernatant, the protoplasts were resuspended and diluted to a cell density of 2 × 10⁵ cells mL⁻¹ in W5 solution. The protoplasts were kept on ice for 30 min before a quick centrifugation step to collect them and resuspend them at a cell density of 2 × 10⁵ cells mL⁻¹ in MMG solution (0.4 M mannitol, 15 mM MgCl₂, and 4 mM MES pH 5.7). The protoplasts were centrifuged again and the supernatant was gently discarded. The indicated constructs used in this study were transfected into soybean protoplasts via the polyethylene glycol (PEG)-mediated transformation method as previously described (Yoo et al., 2007).

Recombinant protein purification and pull-down assay

The production of recombinant proteins was induced in E. coli BL21 strain carrying the indicated constructs cultured in LB medium (1% [w/v] tryptone, 0.5% [w/v] yeast extract, and 1% [w/v] NaCl) containing 0.1 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) at 18°C for 16–20 h. Recombinant GST-GmLHP1-2 protein was purified with glutathione Sepharose 4B beads (GE Healthcare, Cat No.: 45-000-139) according to the manufacturer’s protocol. MBP-PsAvh110, MBP-PsAvh110^ΔW1, MBP-PsAvh110^mW2, and MBP-GmPHD6 proteins were purified using amylose resin (NEB) according to the manufacturer’s protocol. His-GmLHP1-2 and His-PsAvh110 proteins were purified using Pierce Ni-NTA resin (Qiagen) according to the manufacturer’s protocol. For His pull-down assays, recombinant His-GmLHP1-2 was pre-incubated with pre-washed Ni-NTA resin in 400 μL incubation buffer (20 mM Tris–HCl pH 7.5, 100 mM NaCl, and 0.5% [v/v] Triton X-100) at 4°C for 1 h. The immobilized His-GmLHP1-2 beads were washed twice with washing buffer (20 mM Tris–HCl pH 7.5, 300 mM NaCl, and 0.5% [v/v] Triton X-100), and then incubated with MBP, MBP-PsAvh110, MBP-PsAvh110^ΔW1, or MBP-PsAvh110^mW2 proteins for another 1–3 h. The beads were collected and washed three to five times with washing buffer. Similarly, the GST pull-down assay was performed by glutathione Sepharose 4B beads. Immunoblots were conducted with an anti-MBP (Engibody Biotechnology, Cat No.: AT0030), anti-His antibody (Abmart, Cat No.: M30111), anti-Myc (Engibody Biotechnology, Cat No.: AT0023), and anti-GST antibody (Abmart, Cat No.: M20007) at a 1:5,000 dilution.

Total RNA extraction and RT-qPCR

Total RNA was extracted from N. benthamiana leaves or soybean hairy roots after XEG1 or P. sojae treatment using an E.Z.N.A, Total RNA kit I (Omega Bio-Tek) according to the manufacturer’s protocol. RNA concentration was measured by NanoDrop (Thermo Scientific). One microgram total RNA was reverse transcribed into cDNA using a reverse PCR kit (Takara), and then the cDNA was diluted five times as a template for qPCR, which was performed using iTaq SYBR green supermix (Takara) on an ABI Prism 7500 Fast Real-Time PCR System (Applied Biosystems Inc.). PsActin, NbActin2, and GmCYP2 were used as the internal controls. All primers are listed in Supplemental Data Set S2.

Transcriptome deep sequencing (RNA-seq) analysis

Soybean hairy roots infected by P. sojae WT P6497 or the PsAvh110-knockout mutant T52 at 6 hpi were collected for total RNA extraction. Total RNA was extracted using an E.Z.N.A Total RNA kit I (Omega Bio-Tek) according to the manufacturer’s protocol. Two independent replicates of each sample were collected for constructing the library construction using QuantSeq 3′-mRNA-Seq Library Prep Kit from Illumina. Libraries were run on a BGISEQ-500 sequencer platform with paired end 150 bp reads. RNA-seq reads with low sequencing quality or reads with sequencing adaptors were removed by cutadapt version 1.4.2 (Martin, 2011). The resulting clean reads were then aligned to the soybean reference genome (https://phytozome-next.jgi.doe.gov/info/Gmax_Wm82_a2_v1) using HISAT (Kim et al., 2015) and Bowtie2 (Langmead et al., 2009). Following the alignments, RSEM software (Li and Dewey, 2011) was used to calculate the fragments per kilobase per million (FPKM) mapped reads and identify significant DEGs, using the criteria fold-change ≥ 2 or ≤ –2 and P-value < 0.05 between different samples. GO term enrichment was analyzed using the soybean GO term annotations. The cutoff for significant enrichment was P-value < 0.01 and q (false discovery rate) < 0.05. The fold enrichment was calculated based on the –Log₁₀(P-value).

MST assay

The protein binding affinity of GmLHP1-2 to PsAvh110 or GmPHD6 was measured by MST using a Monolith NT.115 instrument (NanoTemper, Germany) as previously described (Plach et al., 2017; Corbeski et al., 2018). Briefly, recombinant purified His-GmLHP1-2 was diluted in MST buffer (10 mM Tris–HCl pH 7.5, 150 mM NaCl, 10 mM MgCl₂, and 0.05% [v/v] Tween 20) supplied with the Monolith His-tag labeling Kit-RED-tris-NTA second generation dye (NanoTemper, Cat No.: MOL018) at room temperature (20°C) for 30 min. A range of MBP, MBP-PsAvh110, or MBP-GmPHD6 concentrations (from 50 to 0.00153 µM) in MST buffer was incubated with equal volume to the labeled His-GmLHP1-2 protein at room temperature for 5 min. The samples were then loaded onto Monolith NT.115 capillaries and measured with 60% MST power in a Monolith NT.115 instrument (Nano Temper, Germany). Nanotemper analysis software (v.1.2.101) was used to fit the curves and calculate the value of the dissociation constant (K_d).

Electrophoretic mobility shift assay

DNA probes were labeled with digoxigenin-11-ddUTP using a DIG Oligonucleotide 3′-End Labeling Kit (Roche, Cat No.: 3353575910) according to the manufacturer’s protocol. The digoxigenin-labeled DNA probes (2 nM) were incubated with different recombinant purified proteins (MBP-GmPHD6, His-PsAvh110, or His-GmLHP1-2) (1–2 µg for each protein) in 30 µL binding buffer (10 mM Tris–HCl pH 7.6, 25 mM KCl, 2.5 mM MgCl₂, 1 mM EDTA, 1 mM DTT, 0.6 mM BSA, 0.05 mg/mL poly [dI–dC], and 10% [v/v] glycerol) at 25°C for 0.5–1 h. For competition assays, 125-fold unlabeled competitor DNA probes were also added to the reactions. The EMSAs were performed using the DIG Gel Shift Kit (Roche, Cat No.: 3353591910) according to the manufacturer protocol.

For end-labeling with biotin EMSAs, DNA probes labeled with biotin at the 5′ end were synthesized by GeneScript Company. The biotin-labeled DNA probes were incubated with different concentrations of recombinant purified MBP-GmPHD6 (50, 100, or 200 µM) in 20 μL binding buffer (10 mM Tris–HCl pH 7.6, 25 mM KCl, 2.5 mM MgCl₂, 1 mM EDTA, 1 mM DTT, 0.6 mM BSA, 0.05 mg/mL poly [dI–dC], and 10% [v/v] glycerol) at 25°C for 30 min. For competition assays, 125-fold unlabeled competitor DNA probes were also added to the reaction. The EMSA was performed according to the manual from LightShift Chemiluminescent EMSA Kit 20148 (Thermo Fisher Company, USA). Signals were detected using chemiluminescent substrate in the kit according to the manufacturer’s instructions.

ChIP-qPCR assays

Soybean hairy roots expressing 3xFlag-GmPHD6 were used for ChIP experiments as previously described by Kaufmann et al. (2010) with modifications. Briefly, 2–3 g of fresh hairy roots was crosslinked with 1% (w/v) formaldehyde for 5 min under vacuum and then stopped with 125 mM glycine for 2 min. The tissues were ground in liquid nitrogen and resuspended in M1 buffer (0.4 M sucrose, 10 mM Tris–HCl pH 8.0, 5 mM β-mercaptoethanol, 0.1 mM PMSF [Phenylmethanesulfonyl fluoride], and complete protease inhibitor cocktail). After filtering through a mesh (55 µm), the filtrated sample was centrifuged at 2,880 g for 20 min at 4°C. The supernatant was removed, and the nuclear pellet was resuspended in M2 buffer (0.25 M sucrose, 10 mM Tris–HCl pH 8.0, 10 mM MgCl₂, 1% [v/v] Triton X-100, 5 mM β-mercaptoethanol, 0.1 mM PMSF, and complete protease inhibitor cocktail). After centrifugation at 12,000 g for 10 min at 4°C, the supernatant was removed and the nuclear pellet was resuspended in M3 buffer (1.7 M sucrose, 10 mM Tris–HCl pH 8.0, 0.15% [v/v] Triton X-100, 2 mM MgCl₂, 5 mM β-mercaptoethanol, 0.1 mM PMSF, and complete protease inhibitor cocktail). After centrifuging at 12,000 g for 1 h at 4°C, the supernatant was removed and the nuclear pellet was resuspended in sonication buffer (50 mM Tris–HCl pH 8.0, 10 mM EDTA, 1% [w/v] SDS, 0.1 mM PMSF, and complete protease inhibitor cocktail). The chromatin was sheared into ∼300-bp fragments with 20 pulses of 10 s each, followed by resting for 20 s at 35% amplitude (Qsonica * sonicator, Q125, Branson, USA). After centrifugation, the supernatant was diluted with 10× ChIP dilution buffer (1.1% [v/v] Triton X-100, 1.2 mM EDTA, 16.7 mM Tris–HCl pH 8.0, 167 mM NaCl, 0.1 mM PMSF, and complete protease inhibitor cocktail) and 10% was set aside as input DNA. Immunoprecipitation was performed using anti-Flag magnetic beads (Thermo Scientific, No. 2162708). The beads were successively washed in low salt wash buffer (one time; 150 mM NaCl, 20 mM Tris–HCl pH 8.0, 0.1% [w/v] SDS, 0.5% [v/v] Triton X-100, and 2 mM EDTA), high salt wash buffer (one time; 500 mM NaCl, 20 mM Tris–HCl pH 8.0, 0.1% [w/v] SDS, 0.5% [v/v] Triton X-100, and 2 mM EDTA), LiCl wash buffer, and TE buffer (two times; 100 mM Tris–HCl pH 8.0 and 10 mM EDTA). DNA was then eluted from the beads with elution buffer (1% [w/v] SDS and 0.1 M NaHCO₃) and digested with proteinase K. DNA was then precipitated by the addition of 2.5 volumes 100% ethanol, 1/10 volume 3 M NaOAc at pH 5.4, and 1 µL glycogen overnight at –20°C. The purified DNA was resuspended in 30 µL Milli-Q water.

ChIP-qPCR was performed with three biological replicates and results were normalized to input DNA. Independent ChIP-qPCR experiments were performed twice with similar results. Primers used for ChIP-qPCR are listed in Supplemental Data Set S2.

Protein production and P. capsici infection in N. benthamiana

For protein production and disease infection in N. benthamiana, the sequences encoding GFP, PsAvh110, PsAvh110 derivatives, GmLHP1-2 derivatives, and GmPHD6 were cloned into pBIN-p35S:GFP or pBIN-p35S:Flag vectors. The resulting vectors were transformed into Agrobacterium strain GV3101. Positive Agrobacterium colonies were cultured in LB medium containing 50 μg/mL rifampicin and kanamycin at 28°C overnight; the cells were harvested by centrifugation at 2,500 g at room temperature for 3–5 min and resuspended in infiltration buffer (10 mM MgCl₂, 10 mM MES pH 5.6, and 100–200 µM acetosyringone). The cell density was adjusted to OD₆₀₀ = 0.5 before the cell suspensions were infiltrated into 6-week-old N. benthamiana leaves. The infiltrated leaves were inoculated with P. capsici mycelium at 12 hpi, and leaf lesions were photographed under UV light at 36 or 48 hpi. The accumulation of the protein encoded by the infiltrated constructs was confirmed by immunoblotting.

Virus-induced gene silencing of NbLHP1 in N. benthamiana

The virus-induced gene silencing (VIGS) assay was performed as described previously (Li et al., 2014). Briefly, ∼200-bp NbLHP1 DNA fragment was designed via the Solanaceae Genomics Network (https://solgenomics.net) and was amplified from N. benthamiana cDNA with the primers listed in Supplemental Data Set S2, and was ligated into the pTRV2 vector to generate the pTRV2-NbLHP1 construct. The pTRV2-GFP vector (EV) was used as control. The pTRV2-NbLHP1 and pTRV2-GFP constructs were transformed into Agrobacterium strain GV3101 by electroporation. Bacterial cells were grown in LB medium containing 50 μg/mL kanamycin and rifampicin at 28°C overnight. The bacterial cells were then harvested by centrifugation at 4,000–5,000 rpm 2,500 g at room temperature for 3–5 min and resuspended in infiltration buffer (10 mM MgCl₂, 10 mM MES pH 5.7, and 100–200 µM acetosyringone). The bacterial suspension (adjusted to OD₆₀₀ = 1.0) containing pTRV1, or pTRV2-NbLHP1 were mixed in a 1:1 ratio and infiltrated into the first pair of true leaves of 2-week-old soil-grown N. benthamiana seedlings using a syringe. The silencing efficiency of NbLHP1 homologs was tested by RT-qPCR with specific primers listed in Supplemental Data Set S2.

BiFC assays

The sequence encoding PsAvh110 without the signal peptide was cloned into the pSY-736-35S-nYFP vector. The full-length sequence of NbLHP1-1 was cloned into the pSY-735-35S-cYFP vector. The resulting constructs were transiently infiltrated into N. benthamiana leaves by Agrobacterium-mediated infiltration. At 36 hpi, the YFP signal was examined at an excitation wavelength of 514 nm and an emission wavelength of 530 nm, and images were captured using an LSM 710 confocal laser scanning microscope (Carl Zeiss, Germany).

Confocal microscopy

The fluorescence of GFP and RFP was observed using a LSM 710 confocal laser scanning microscope (Carl Zeiss, Germany) with a 20×, 40×, or 60× objective lens. The excitation wavelength of GFP and RFP was 488 and 588 nm, respectively. The emission wavelength for GFP and RFP was 490–530 and 590–620 nm, respectively. The autofluorescence of chloroplasts was excited at 630 nm and the emission wavelength for detecting chloroplast signal is 690–700 nm. Images were acquired and analyzed using Zen 2010 software.

P. sojae transformation and inoculation assays

Phytophthora sojae Psavh110 mutants and PsAvh110^mNLS replacement strains were generated using the CRISPR/Cas9 system as previously described (Fang and Tyler, 2016; Qiu et al., 2021). Briefly, three single-guide RNAs (sgRNA-1: TCGATCCAGAACGCTGCACG, sgRNA-2: GCTTACCAAACCTCGCGAAG, and sgRNA-3: GGCGTAGCGAGTGTCCGCGT) targeting PsAvh110 were selected based on high confidence predictions (http://www.broadinstitute.org/rnai/public/analysis-tools/sgrna-design), and were individually ligated into the pYF515 vector. The replacement sequences together with 1 kb of homologous 5′ and 3′ flanking arms outside the target region were amplified and cloned into the pBluescript II KS+ vector by In-Fusion HD Cloning Kit (Clontech). These vectors were transformed into WT P. sojae strain P6497 using the PEG-mediated P. sojae protoplast transformation method as previously described (Huang et al., 2017). The transformants were screened by genomic PCR using the primers listed in Supplemental Data Set S2, and the PCR products were sequenced using Sanger sequencing to identify Psavh110 mutants. The virulence of P. sojae Psavh110 mutants was tested by inoculating the hypocotyls of etiolated soybean seedlings. Etiolated soybean seedlings were grown in darkness for 5 days at 25°C in the growth chamber. The Psavh110 mutants were grown on V8 medium plates for 4 days at 25°C; the plates were washed with sterile water in 30-min intervals three to four times, before being placed in a 25°C incubator in the dark for 6–8 h until zoospore release. One hundred zoospores of WT or the Psavh110 mutants were inoculated onto soybean hypocotyls; the inoculated seedlings were then incubated at 25°C in darkness for 2–3 days. More than 12 etiolated soybean hypocotyls were tested for each biological replicate. The infected seedlings were photographed at 2–3 days post-inoculation (dpi) and the samples were harvested and stored at −20°C for P. sojae biomass detection.

Statistical analysis

Data are shown as mean ± standard deviation (SD) or standard error of the mean (SEM). Statistical significance was determined by Student’s t test or one-way ANOVA test (*P < 0.05; **P < 0.01; ***P < 0.001; and ****P < 0.0001) using GraphPad Prism 8.0 software.

Accession numbers

The RNA-seq data have been deposited at the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) under the BioProject ID PRJNA869950. Gene sequences, used in this study, were obtained from Phytozome (https://phytozome.jgi.doe.gov/) and Sol Genomics Network (http://solgenomics.net/) database. The accession numbers of genes are as follows: GmLHP1-2 (Glyma.03G094200), GmPHD6 (Glyma.09G068500), NbLHP1-1 (Niben101Scf00215g06004), NbLHP1-2 (Niben101Scf01538g00023), and NbLHP1-3 (Niben101Scf11379g00011). Alignments and machine-readable tree files for phylogenetic analysis are provided in Supplemental Files S1–S4.

Supplemental data

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

Supplemental Figure S1. PsAvh110 knockout in P. sojae using CRISPR/Cas9.

Supplemental Figure S2. Generation of mutated NLS and complementation strains with PsAvh110 in P. sojae using a CRISPR/Cas9-mediated in situ complementation method.

Supplemental Figure S3. PsAvh110 regulates the expression of immune-associated genes.

Supplemental Figure S4. PsAvh110 interacts with plant LHP1-2 homologs.

Supplemental Figure S5. GmLHP1-2 and GmPHD6 enhance plant immunity against P. capsici infection.

Supplemental Figure S6. The immune-associated genes regulated by PsAvh110 contain G-rich elements in their promoters.

Supplemental Figure S7. GmPHD6 binds to the G-rich elements in the promoters of defense genes to activate their transcription in a GmLHP1-2-dependent manner.

Supplemental Data Set S1 . List of DEGs in soybean hairy roots infected by P. sojae WT P6497 or PsAvh110-knockout mutant T52 at 6 hpi.

Supplemental Data Set S2. Primers used in this study.

Supplemental Data Set S3. Summary of statistical analyses.

Supplemental File S1. Protein sequence alignment used to generate the phylogenetic tree shown in Supplemental Figure S3B.

Supplemental File S2. Newick file of the phylogenetic tree shown in Supplemental Figure S3B.

Supplemental File S3. Protein sequence alignment used to generate the phylogenetic tree shown in Supplemental Figure S4A.

Supplemental File S4. Newick file of the phylogenetic tree shown in Supplemental Figure S4A.

 

X.Q., L.K., and Y.W. conceived, designed the experiments, analyzed the data, and wrote the manuscript with input from all the authors. X.Q. performed most of the molecular, biochemical, confocal microscope, and pathogen infection experiments presented in the manuscript. L.K. and X.Q. initiated the project by identifying GmLHP1-2 as PsAvh110-interacting protein through a Y2H screen, and performing phenotypical analysis. X.Q., L.K., H.C., Y.L., S.T., L.W., J.X., Z.C., M.Z., K.D., P.Y., M.Q., W.Y., Y.W., and S.D. helped to analyze data, provided critical feedback, and helped shape the research.

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: Yuanchao Wang (wangyc@njau.edu.cn).