The Phytophthora sojae effector PsFYVE1 modulates immunity-related gene expression by targeting host RZ-1A protein

Abstract

Oomycete pathogens secrete numerous effectors to manipulate plant immunity and promote infection. However, relatively few effector types have been well characterized. In this study, members of an FYVE domain-containing protein family that are highly expanded in oomycetes were systematically identified, and one secreted protein, PsFYVE1, was selected for further study. PsFYVE1 enhanced Phytophthora capsici infection in Nicotiana benthamiana and was necessary for Phytophthora sojae virulence. The FYVE domain of PsFYVE1 had PI3P-binding activity that depended on four conserved amino acid residues. Furthermore, PsFYVE1 targeted RNA-binding proteins RZ-1A/1B/1C in N. benthamiana and soybean (Glycine max), and silencing of NbRZ-1A/1B/1C genes attenuated plant immunity. NbRZ-1A was associated with the spliceosome complex that included three important components, glycine-rich RNA-binding protein 7 (NbGRP7), glycine-rich RNA-binding protein 8 (NbGRP8), and a specific component of the U1 small nuclear ribonucleoprotein complex (NbU1–70K). Notably, PsFYVE1 disrupted NbRZ-1A–NbGRP7 interaction. RNA-seq and subsequent experimental analysis demonstrated that PsFYVE1 and NbRZ-1A not only modulated pre-mRNA alternative splicing (AS) of the necrotic spotted lesions 1 (NbNSL1) gene, but also co-regulated transcription of hydroxycinnamoyl-CoA shikimate/quinate hydroxycinnamoyl transferase (NbHCT), ethylene insensitive 2 (NbEIN2), and sucrose synthase 4 (NbSUS4) genes, which participate in plant immunity. Collectively, these findings indicate that the FYVE domain-containing protein family includes potential uncharacterized effector types and also highlight that plant pathogen effectors can regulate plant immunity-related genes at both AS and transcription levels to promote disease.

Introduction

Oomycetes such as Phytophthora spp. include a wide variety of plant pathogens that cause devastating diseases in a wide range of crops (Oh et al., 2010). During infection, Phytophthora pathogens secrete several types of effector proteins into hosts to subvert plant defense responses (Bozkurt et al., 2012). Apoplastic effectors are one type and include elicitins, necrosis- and ethylene-inducing peptide 1 (NEP1)-like proteins (NLPs), glycoside hydrolase family 12 (GH12) proteins, and cellulose-binding elicitor lectin (CBEL), which contain the signal peptides and are secreted into the host extracellular space during interaction (Raaymakers and Van den Ackerveken, 2016; Rocafort et al., 2020). Another type is cytoplasmic effectors, such as RXLR and Crinkler (CRN) that contain an N-terminal signal peptide and are delivered into plant cells (Dou and Zhou, 2012). Recently, some specific effectors are also reported in oomycetes; for example, the unconventionally secreted effectors (Liu et al., 2014), the Albugo laibachii CHXC effectors (Kemen et al., 2011), and the small open reading frame-encoded polypeptides (Wang et al., 2021). However, investigations on virulence strategies of oomycete pathogens tend to be restricted to conventional effector types. Furthermore, the prediction of oomycete-secreted effectors by common sequence features will miss bona fide effectors that do not carry such signatures (Sperschneider and Dodds, 2021).

RXLR effectors have been studied extensively, because they can target plant proteins to interfere with host immunity through sophisticated pathogenicity mechanisms, in turn, some of them may be recognized by plant NLR-mediated immune responses (Wang et al., 2011; Anderson et al., 2015; He et al., 2020). However, how RXLR effectors enter host cells is subject to debate and even controversy. The RXLR domain may bind to phosphatidylinositol 3-phosphate (PI3P) to mediate translocation of effector proteins into plant cells and to increase their stability in planta (Kale et al., 2010). Nevertheless, Gan et al. do not demonstrate that flax rust effectors AvrM and AvrL567 require PI3P binding to internalize into plant cells (Gan et al., 2010), and AVR3a binds to PI3P via the effector domain, not the RXLR domain, although PI3P-binding ability is essential for its accumulation inside host cells (Yaeno et al., 2011).

The phosphoinositide PI3P associates with endosomal membranes to regulate protein trafficking and endocytic pathways, autophagy, and cytokinesis in various eukaryotes (Stenmark et al., 2002). During interactions between plants and pathogens, phosphoinositides are crucial for recognition and membrane trafficking to hijack host cellular processes. For example, during Colletotrichum higginsianum infection, both PI(4,5)P₂ and PI4P are at the extra-invasive hyphal membrane (EIHM), a plant cell-derived membrane that surrounds the hemibiotrophic hyphae (Shimada et al., 2019). In the Arabidopsis-powdery mildew interaction, only PI(4,5)P₂ is recruited to the extrahaustorial membrane (EHM) derived from the host plasma membrane (PM). Deletion of PIP5K genes responsible for PI(4,5)P₂ biosynthesis prevents susceptible responses and colonization of powdery mildew in leaf tissues (Qin et al., 2020; Qin and Wei, 2021). In our previous study, Phytophthora-derived PI3P is exposed at the surface of infection hyphae to aid the actions of RXLR effectors (Lu et al., 2013). Thus, the compound can be used to guide anti-microbial peptides produced by plant cells to accumulate specially at pathogen tissues to achieve high disease resistance (Zhou et al., 2021), further suggesting important roles of PI3P in Phytophthora infection.

The PI3P-binding FYVE domain is named after the first four FYVE proteins, Fab1p, YOTB, Vac1p, and EEA1. The FYVE domain contains conserved coordinating residues: the N-terminal WxxD, the central R(R/K)HHCR, and C-terminal R(V/I)C residues, which are the principal sites of compact binding with PI3P (Stenmark et al., 2002; Kutateladze, 2006). The FYVE domain is widely distributed in eukaryotes. According to previous studies, 27 proteins contain the FYVE domain in humans (Homo sapiens), 5 in yeast (Saccharomyces cerevisiae), and 15 in Caenorhabditis elegans (Stenmark et al., 2002). The FYVE proteins regulate membrane trafficking, receptor signaling, and cytoskeletal dynamics by specific binding to PI3P (Gillooly et al., 2001). The human FYVE protein SARA (Smad anchor for receptor activation) initiates a signaling cascade by recruiting intracellular signaling mediators to receptors (Tsukazaki et al., 1998). Recently, one Phytophthora sojae FYVE protein PsFP1 is found to play an important role in mycelial growth, pathogenicity, and oxidative stress response (Zhang et al., 2021). However, the distribution and biological functions of other FYVE proteins in oomycete pathogens are poorly known.

Transcriptional regulation and alternative pre-mRNA splicing are important in plant adaptations to the environment (Rigo et al., 2019). The processes provide an important RNA-based layer of protein diversity regulation in development and stress response (Reddy et al., 2013). For example, a plant peptide hormone and its receptor complex RALF1-FER modulate dynamic alternative splicing (AS) in cold response and ABA response by interacting with glycine-rich RNA-binding protein 7 (GRP7) (Wang et al., 2020c). Moreover, evidence is emerging that various biotic (viral and oomycete pathogens) and abiotic stresses (heat and cold) can trigger feedback on transcriptions and AS. For example, P. sojae effector PsCRN108 inhibits expression of plant heat shock protein (Hsp)-encoding genes by directly targeting their promoters to suppress plant HSP gene transcription (Song et al., 2015). In addition, an avirulent effector PsAvr3c derived from P. sojae reprograms host pre-mRNA splicing by binding to serine/lysine/arginine-rich proteins SKRPs associated with spliceosome components (Huang et al., 2017). These observations indicate that gene transcription and AS are important components of host transcriptome reprograming in response to pathogen infection.

Sequence-specific glycine-rich RNA-binding proteins (GRPs) are critical for RNA processing in various organisms and have been intensively investigated in the plant kingdoms. Plant GRPs usually contain a canonical RNA recognition motif (RRM) or a cold-shock domain (CSD) at the N-terminus and a glycine-rich region (Gly, 20%–70%) at the C-terminus (Ma et al., 2021). The GRPs harboring characteristic internal CCHC-type zinc finger are designated as RZ-1 proteins and include three members in Arabidopsis (Arabidopsis thaliana) (AtRZ-1A, AtRZ-1B, and AtRZ-1C) (Kim et al., 2010a; Ciuzan et al., 2015; Czolpinska and Rurek, 2018). Evidence is increasing that GRPs are linked to plant growth and development as well as stress responses by regulating transcriptional and post-transcriptional processes (Ma et al., 2021). For example, A. thaliana RZ-1B and RZ-1C interact with a spectrum of splicing-related proteins and directly bind to FLOWERING LOCUS C (FLC) chromatin to regulate splicing of FLC introns (Wu et al., 2016). Of the GRPs, GRP7 is the most intensively studied in A. thaliana, and it controls the transcription level of oscillation-related genes in a clock-regulated negative feedback circuit (Staiger et al., 2003). In addition, A. thaliana GRP8, the putative paralog of GRP7, can regulate AS events via direct interaction with transcripts (Streitner et al., 2012). Although detailed mechanism remains unknown, current findings provide sufficient evidence that GRPs participate in transcriptional and post-transcriptional regulation in cold, salt, and dehydration stress (Kim et al., 2005; Kim et al., 2010b; Xu et al., 2014).

Based on advances in high-throughput transcriptome sequencing, up to 70% of all plant multi-exon genes are estimated to undergo AS and most AS events require the spliceosome (Chaudhary et al., 2019). The spliceosome consists of five snRNAs (U1, U2, U4, U5, and U6) and nearly 300 proteins. U1–70K directly interacts with U1 snRNA to recognize the conserved GU dinucleotide at the 5′ donor site of intron and act as a signal for accurate splicing of pre-mRNAs (Reddy, 2007; Buratti and Baralle, 2010). Combined effect of GRP7, GRP8, U1–70K, and other spliceosomal proteins constitutively regulate pre-mRNA AS to alter the transcriptome (Reddy, 2007). Although plant spliceosomes have not been isolated to date, analysis of the known spliceosomal proteins indicates that core spliceosome machinery is conserved among different species.

In this study, the FYVE domain-containing proteins were systematically identified and analyzed in oomycete, and one secreted FYVE domain-containing protein from P. sojae (PsFYVE1) was studied. PsFYVE1 was essential for P. sojae virulence and suppressed plant immunity. PsFYVE1 bound to PI3P via its conserved FYVE domain in a binding sites-dependent manner. PsFYVE1 targeted RZ-1A in Nicotiana benthamiana and soybean (Glycine max), and silencing of NbRZ-1A increased Phytophthora infection and suppressed plant defense. GmRZ-1A/NbRZ-1A complementation in NbRZ-1A-silenced-plants could recover the immunity phenotype. Notably, NbRZ-1A was associated with three key splicing factors, NbGRP7, NbGRP8, and NbU1–70K. The binding affinity between NbRZ-1A and NbGRP7 would be reduced by increasing protein accumulation of PsFYVE1. Moreover, according to combined RNA-seq and experimental analyses, PsFYVE1 and NbRZ-1A co-regulated pre-mRNA AS and transcription of immunity-related genes. Overall, the results shed light on P. sojae effector PsFYVE1 which can regulate AS and transcription of immunity-related genes by targeting plant RZ-1A proteins to promote Phytophthora infection.

Results

Expanded FYVE domain-containing protein family exhibits distinct characteristics in oomycetes

According to bioinformatics analysis, an average of 109 FYVE proteins was predicted in the genus Phytophthora, compared with 99 in Pythium ultimum, 88 in Hyaloperonospora arabidopsidis, and 97 in Albugo laibachii (Figure 1A, Supplemental Table S1). Among the examined oomycetes, the fish pathogen Saprolegnia parasitica contained the most members with 177. By contrast, none of the five bacterial genomes examined contained FYVE proteins, and much smaller numbers of FYVE proteins were predicted in fungi, plants, and metazoans (Figure 1A). Notably, the number of FYVE proteins was not directly associated with the number of total proteins for a given organism in eukaryotes (Figure 1A). Although two diatoms, Phaeodactylum tricornutum and Thalassiosira pseudonana, were closely related to oomycetes, they contained only four and five FYVE proteins, respectively. Therefore, these results suggest that FYVE proteins expanded widely in oomycetes, and that some may contribute to pathogen lifestyles (Figure 1A).

Figure 1.

FYVE domain-containing protein family exhibits distinct characteristics in oomycetes. A, Comparison of FYVE domain-containing protein numbers and total protein numbers among representative oomycetes, diatoms, fungi, plants, metazoan, and bacteria. B, Distribution of additional domains in FYVE domain-containing proteins among eukaryotes. C, Four expression patterns of FYVE genes based on P. sojae transcriptome data at 10 different developmental and infection stages. D, Flow chart of how to select PsFYVE1 among P. soaje FYVE domain-containing proteins. E, Phylogenetic analysis and domain distribution of PsFYVE1 and its orthologs among oomycetes.

To infer the evolutionary history of oomycete FYVE proteins, P. sojae was selected to represent the oomycetes, and a phylogenetic analysis of FYVE proteins was conducted with representatives of other eukaryotes. In the phylogenetic tree, only a few P. sojae FYVE proteins clustered with homologs from other eukaryotes. Most P. sojae FYVE proteins formed independent clades that did not clearly cluster with those of other eukaryotes (Supplemental Figure S1). The results indicate that some FYVE proteins evolve specifically in oomycetes.

Conservation of FYVE domains in oomycetes was investigated. Notably, 92.5% of oomycete FYVE proteins contained only one FYVE domain, whereas the others contained two to five FYVE domains. According to further sequence analysis, 16.4% of oomycete FYVE domains were of the classic FYVE domain type, but most had variable amino acid composition in the three regions. In addition to the FYVE domain, half of oomycete FYVE proteins had at least one other additional domain. A total of 111 diverse types of domains were identified in oomycete FYVE proteins (Supplemental Table S1), substantially more than in proteins of diatoms (2), fungi (6), plants (21), and metazoans (39). Of these domains, oomycetes shared 16 domains with metazoans, 7 with plants, 4 with fungi, and 2 with diatoms (Figure 1B). Notably, 95 domains were found only in oomycetes, and 15 of those were distributed in at least eight of the nine studied oomycetes. These results collectively indicate that oomycete FYVE proteins evolve to exhibit variable FYVE domains and diverse domain architectures. Digital gene expression profiling analysis of P. sojae at 10 different developmental and infection stages (Ye et al., 2011) showed that P. sojae FYVE genes were classified into four major clusters (C1–C4) (Figure 1C). The largest cluster was C1, and members had high expression levels at each infection stage compared with those in developmental stages. By contrast, C2 members were highly expressed at zoospore (ZO) and cyst germination (CY) stages. These results suggest that P. sojae FYVE genes may have diverse roles during the lifecycle.

To further investigate biological functions of FYVE proteins during interaction between oomycetes and plants, the P. sojae FYVE protein PsFYVE1 (Ps136834) was selected based on several criteria and we performed subsequent experimental analysis (Figure 1D). PsFYVE1 encoded a FYVE protein that contained an N-terminal signal peptide, one classic FYVE domain, and two oomycete-specific FIST domains (Figure 1E; Supplemental Figure S2A). Both RNA-seq and reverse transcription quantitative PCR (RT-qPCR) data verified that PsFYVE1 was up-regulated in the infection process (Figure 1C; Supplemental Figure S2B). Further analysis showed that PsFYVE1 was highly conserved in other oomycetes (Figure 1E), but had little similarity with those in other eukaryotes. Notably, only PsFYVE1 contained the N-terminal signal peptide, and no polymorphism was found among P. sojae isolates, suggesting that PsFYVE1 is unique in P. sojae.

PsFYVE1 carries a functional secretory signal peptide and underpins translocation to plant cells

A yeast signal sequence trap system was used to functionally validate the predicted signal peptide of PsFYVE1. The Avr1b signal peptide and the pSUC2 empty vector were used as positive and negative controls, respectively (Dong et al., 2015; Song et al., 2015). The PsFYVE1 N-terminal signal peptide sequence was cloned into the yeast vector pSUC2, followed by introduction into invertase secretion-defective yeast YTK12. The YTK12 strain carrying the signal peptide of PsFYVE1 induced secretion of yeast invertase and grew on YPRAA medium (with raffinose instead of sucrose), whereas the YTK12 strain and the YTK12 strain carrying pSUC2 did not induce secretion (Supplemental Figure S3). Invertase activity was further confirmed by reduction of the dye TTC to the insoluble red color of triphenylformazan (Supplemental Figure S3). The results suggest that PsFYVE1 carries a functional secretory signal peptide.

To further confirm the entire protein of PsFYVE1 can be secreted and translocated into plant cells, the full-length sequence of PsFYVE1 was cloned into a plant binary vector to investigate PsFYVE1 cell-to-cell movement. The pmCherry-HDEL//GFP-GFP vector were used as negative control (Feng et al., 2016). PsAvr1b, a well-known P. sojae secreted full-length effector, was used as a positive control (Dou et al., 2008). When the Agrobacterium containing the pmCherry-HDEL//PsFYVE1-GFP construct was infiltrated at an optical density final OD₆₀₀ of 0.001, green fluorescence signals from PsFYVE1-GFP and PsAvr1b-GFP moved into multiple cells, while red fluorescence signals from mCherry-HDEL accumulated in the initial cells (Figure 2A). However, the green and red fluorescence signals of pmCherry-HDEL//GFP-GFP vector were only observed in the initial cells (Figure 2A). These results showed that the entire protein of PsFYVE1 could be secreted and further enter into the surrounding cells, suggesting that PsFYVE1 acted as a potential effector.

Figure 2.

PsFYVE1 trafficks in plant cells and is required for virulence. A, Cell-to-cell movement analysis of the binary construct to co-express mCherry-HDEL and PsFYVE1-GFP in N. benthamiana leaves. Agrobacterium containing the construct to co-express mCherry-HDEL and PsFYVE1-GFP/PsAvr1b-GFP/PsGFP-GFP was diluted 500 times and got an OD₆₀₀ of 0.001 for expression in a single epidermal cell. The cell-to-cell movement of PsFYVE1-GFP, GFP-GFP, and PsAvr1b-GFP was analyzed at 24 h after infiltration. The moved green fluorescence signals were indicated with arrows. The numbers in each image are the number of scans displaying protein distribution equivalent to the image shown out of the total number of scans collected from three independent experiments. The ratios represent the number of green fluorescence signals movement to the number of red fluorescence signals collected from three independent experiments. The sizes of scale bars were labeled in each group of pictures. B, Disease symptoms on etiolated hypocotyls infected by PsFYVE1-silenced transformants. Wild-type (WT), PsFYVE1-silenced transformants (T5/8/15), and T9 (a non-silenced transgenic line used as a negative control) were incubated on soybean for 36 hpi. Experiments were repeated three times with similar results. Scale bars represent 1 cm. C, Average lesion lengths of the infected etiolated soybean seedlings. Lesion lengths (cm) were measured at 36 hpi. All data are from three independent biological replicates using at least five seedlings in each replicate. Error bars represent mean ± SD (n > 15). Asterisks indicate statistical significance: **P < 0.01, two-sided unpaired t test. D, Relative biomass quantification of the inoculated soybean seeding. DNA from P. sojae-infected regions was isolated at 36 hpi and primers specific for the soybean and P. sojae actin genes were used for qPCR. Data are shown as mean ± SD of three replicates. Asterisks indicate statistical significance: **P < 0.01, two-sided unpaired t test.

To further investigate which region of PsFYVE1 directed its secretion and uptake by plant cells, deletion mutants of signal peptide, FYVE and FIST domains were fused to the plant binary vector and expressed in N. benthamiana (Supplemental Figure S4A). As shown in Supplemental Figure S4B, both green and red fluorescence signals of PsFYVE1^ΔSP and PsFYVE1^ΔFYVE remained in the primary expressing cell. The fluorescence signals accumulated in the nucleus when signal peptide was deleted, but exported from the cells and accumulated in the nucleus and apoplast when FYVE domain was deleted. By contrast, when PsFYVE1^ΔFIST_C and PsFYVE1^ΔFIST_N/C was fused to the plant binary vector, green fluorescence signals moved into multiple cells, while red fluorescence signals accumulated in the initial cells, which was similar with full-length PsFYVE1. When we further deleted FYVE domain of PsFYVE1^ΔFIST_N/C, both green and red fluorescence signals of PsFYVE1^{ΔFYVE_ΔFIST_N/C} only accumulated in the initial cells, which was the same as PsFYVE1^ΔFYVE (Supplemental Figure S4B). These results indicated that the signal peptide acted to enable PsFYVE1 protein to secrete and the FYVE domain rather than FIST domain of PsFYVE1 was indeed essential for uptake by plant cells.

PsFYVE1 contributes to P. sojae virulence by suppressing plant defense response

To investigate the contribution of PsFYVE1 to P. sojae virulence, we utilized a gene silencing technology and successfully obtained three independent PsFYVE1-silenced transformants. The RT-qPCR analysis showed that transcript levels of the three silenced transformants (T5, T8, and T15) decreased significantly to 31%–42% of those of the WT strain. In the non-silenced transformant T9, transcript levels were not affected (Supplemental Figure S5A). Analysis of mycelial colony growth revealed that filamentous growth in the silenced transformants was similar to the WT strain (Supplemental Figure S5, B and C). The virulence of the transformants was also tested on etiolated soybean seedlings. Virulence of the three PsFYVE1-silenced transformants decreased significantly compared with that of the WT strain (Figure 2, B and C). A relative P. sojae biomass (ratio of P. sojae DNA to soybean DNA) also indicated that the three silenced transformants were less virulent compared with the WT stain (Figure 2D). These results indicate that PsFYVE1 is an essential virulence factor for P. sojae infection.

To explore the virulence functions of secreted PsFYVE1 in planta, N. benthamiana leaves overexpressing signal peptide-removed PsFYVE1-GFP or GFP alone (negative control) were inoculated with Phytophthora capsici. Moreover, we constructed mouse FYVE protein EEA1-GFP used as a control to rule out the possibility of PsFYVE1 overexpression-mediated cellular function disruption. The stability of PsFYVE1-GFP, EEA1-GFP, and GFP fusion proteins was detected by western blotting (Figure 3A). Expression of PsFYVE1-GFP in N. benthamiana significantly promoted P. capsici infection, compared with the GFP control (Figure 3, B and C). Consistently, the relative P. capsici biomass with PsFYVE1-GFP overexpression was higher than that in the GFP control (Figure 3D). However, there was no significant difference between EEA1-GFP and GFP alone in terms of lesion size and relative biomass (Figure 3, B–D). Taken together, these results reveal that PsFYVE1 can impair plant defense and promote P. capsici infection. As the activation of plant immune responses is often accompanied by accumulation of reactive oxygen species (ROS), and successful pathogens may deploy virulence factors that conquer such plant basal resistance (Chinchilla et al., 2006; Klauser et al., 2013; Wang et al., 2020a). Here, flg22- and chitin-triggered ROS were selected as examples, and responses to PsFYVE1 were examined. As shown in Figure 3E, flg22-induced ROS production in PsFYVE1-expressing leaves was reduced by 83% compared with that in leaves with the empty vector. Similarly, chitin treatment generated a 37% reduction in ROS accumulation in PsFYVE1-overexpressing leaves compared with that in leaves with the empty vector, whereas EEA1 could not suppress ROS accumulation induced by neither Flg22 nor chitin (Figure 3, E and F) Thus, these results suggest that PsFYVE1 can suppress plant defense by inhibiting ROS accumulation.

Figure 3.

PsFYVE1 promotes Phytophthora infection and inhibits plant immunity. A, Protein expression of validated by western blotting using an anti-GFP antibody. Protein loading is visualized by Ponceau stain. B, Enhanced Phytophthora infection in N. benthamiana leaves expressing PsFYVE1. GFP alone and the mouse FYVE protein, EEA1, were used as negative controls. The N. benthamiana leaves were infiltrated to express the indicated proteins, and then inoculated with P. capsici 24 h later. Photographs were taken at 36 hpi under UV light. Dashed lines indicate the lesion areas.(C) Average lesion sizes of infected N. benthamiana leaves. Lesion area (cm²) was measured at 36 hpi. Error bars represent the mean ± SD (n = 24). Asterisks indicate statistical significance: **P < 0.01, two-sided unpaired t test. D, Relative biomass quantification of inoculated N. benthamiana leaves. DNA from P. capsica-infected regions was isolated at 36 hpi and primers specific for the N. benthamiana and P. capsici actin genes were used for qPCR. qPCR was performed and normalized to GFP. Data are shown as mean ± SD of three replicates. Asterisks indicate statistical significance: **P < 0.01, two-sided unpaired t test. E-F, Flg22- and chitin-induced reactive oxygen species (ROS) burst in the GFP, PsFYVE1-GFP, and EEA1-GFP overexpressed N. benthamiana leaves. Relative luminescence units (RLUs) indicate relative amounts of H₂O₂ production induced by 1 μM flg22 (E) and 100 μg ml^–1 chitin (F) in leaf discs (left). And the statistic of total RLUs were performed significance analysis (right). Error bars represent the mean ± SD (n > 9). Asterisks indicate statistical significance: **P < 0.01; *P < 0.05, two-sided unpaired t test. Experiments were repeated three times with similar results.

To further study the subcellular localization of PsFYVE1, N. benthamiana leaves overexpressing PsFYVE1-GFP were performed confocal microscopy 48 h after agroinfiltration. The results showed that PsFYVE1-GFP localized in the both nucleoplasm and nuclear speckles (Supplemental Figure S6A). In addition, the statistics for two types of subcellular localization were obtained from a total of 105 green fluorescent observation of PsFYVE1-GFP. The ratio statistical results showed that more than 70% PsFYVE1-GFP predominantly accumulated in the nuclear speckles, and the remaining had nucleoplasm localization (Supplemental Figure S6B).

FYVE domain of PsFYVE1 has PI3P-binding activity

As the FYVE zinc finger domain usually functions in binding to PI3P (Gillooly et al., 2000; Vermeer et al., 2006), whether PsFYVE1 and its FYVE domain could bind to PI3P and other phosphoinositides was tested. The R and H residues in the FYVE domain provide critical hydrogen bonds to PI3P (Stenmark et al., 2002; Kutateladze, 2006). Therefore, a PsFYVE1^AAAA mutant gene was synthesized with the first R and two H residues of the R(R/K)HHCR motif and the R of the R(V/I)C motif substituted into A (Supplemental Figure S7A). In addition, a construct encoding a tandem repeat of the PsFYVE1 FYVE domain (2×FYVE) was generated to test phosphoinositide-binding activities. Six PIs were spotted onto Hybond-C membranes, which were incubated with purified protein of GST fused to the N-terminal full-length PsFYVE1, PsFYVE1^AAAA, or 2×FYVE. The mouse FYVE protein, EEA1, was the positive control; whereas the empty vector was the negative control. As shown in Supplemental Figure S7B, GST-PsFYVE1 and GST-2×FYVE bound to PI3P, whereas GST-PsFYVE1^AAAA did not bind, suggesting that PsFYVE1 specifically bind to PI3P on the basis of the conserved amino acids of the FYVE domain. Furthermore, GST-PsFYVE1, GST-PsFYVE1^AAAA, and GST-2×FYVE could also bind to PI5P, whereas GST-EEA1 could not (Supplemental Figure S7B).

PsFYVE1 interacts with plant Rz-1a, Rz-1b, and Rz-1c through its FIST_C domain

To identify plant targets of PsFYVE1, PsFYVE1-GFP was transiently expressed in N. benthamiana leaves followed by immunoprecipitation (IP) assay using GFP-Trap_A beads with GFP as the control. Immuno-purified proteins were then analyzed by LC-MS/MS, and a total of 52 N. benthamiana proteins in parallel repetitions potentially interacted with PsFYVE1 (Supplemental Table S2). Based on sequence similarity, the largest group of proteins extensively linked with the DNA-/RNA-binding processes. Of the DNA-/RNA-binding proteins, the homolog of NbRZ-1B in A. thaliana has a nuclear speckles-localization and binds to chromatin (Wu et al., 2016), but its role in plant immunity remains unknown. Therefore, NbRZ-1B was selected as the potential target for subsequent study.

NbRZ-1B is in the zinc finger-containing glycine-rich RNA-binding protein family, which has three members in A. thaliana (AtRZ-1A, AtRZ-1B, and AtRZ-1C) (Kim et al., 2010; Wu et al., 2016; Czolpinska and Rurek, 2018). Bioinformatics analysis showed that NbRZ-1A and NbRZ-1C were the paralogs of NbRZ-1B. To confirm the interactions between PsFYVE1 and NbRZ-1A/1B/1C, a Co-IP assay was performed. The results clearly showed that PsFYVE1 was substantially enriched in NbRZ-1A/1B/1C precipitates but not in that of the GFP control (Supplemental Figure S8A). The interactions were further validated via the bimolecular fluorescence complementation (BiFC) experiments. NbRZ-1A/1B/1C (NbRZ-1A-Yn, NbRZ-1B-Yn, and NbRZ-1c-Yn) and PsFYVE1 (PsFYVE1-Yc) were co-expressed in N. benthamiana. YFP fluorescence was observed exclusively in the nuclear speckles, whereas YFP fluorescence was not detected in the negative control (Supplemental Figure S8B, Supplemental Figure S9A). Taken together, these experiments demonstrate that PsFYVE1 interacts with NbRZ-1A/1B/1C and that the proteins co-localize in nuclear speckles.

To test whether PsFYVE1 also interacted with NbRZ-1A/1B/1C homologs in soybean, the natural host of P. sojae, we cloned GmRZ-1A/1B/1C full-length cDNA sequences using Glycine max, cv. William 82 cultivar. The interactions between GmRZ-1A/1B/1C and PsFYVE1 were validated via the Co-IP experiment (Figure 4A) and the BiFC assay (Figure 4B, Supplemental Figure S9B). Moreover, the interactions were tested using a luciferase complementation assay. When PsFYVE1 was fused with nLUC at the C-terminal, GmRZ-1A/1B/1C fused with cLUC at the N-terminal showed clear interactions with cLUC-PsFYVE1 in N. benthamiana leaves (Figure 4C, Supplemental Figure S9C). No luminescent signal could be detected in negative controls (Figure 4C, Supplemental Figure S9C). Therefore, all evidence demonstrates that RZ-1A/1B/1C proteins in N. benthamiana and soybean are the targets of PsFYVE1.

Figure 4.

PsFYVE1 interacts with GmRZ-1A, GmRZ-1B and GmRZ-1C in vivo. A, Co-immunoprecipitation (Co-IP) assay of PsFYVE1 and GmRZ-1A/1B/1C. Co-IP of protein extracts from agroinfiltrated leaves using GFP-Trap confirmed that PsFYVE1-HA was associated with GmRZ-1A/1B/1C, but not the GFP control. The positions of the expected protein bands are indicated by asterisks. Protein sizes are indicated in kilodaltons (kDa), and protein loading is visualized by Ponceau stain. B, Interaction and co-localization of PsFYVE1 and GmRZ-1A/1B/1C in bimolecular fluorescence complementation (BiFC) assay. A. tumefaciens cells harboring the indicated YFPn (Yn)- or YFPc (Yc)-fused constructs were co-infiltrated into N. benthamiana leaves. The fluorescent signals were detected by confocal laser microscopy at 48 h after infiltration and 4′,6-diamidino-2-phenylindole (DAPI) were used to a nucleus staining. Scale bars represent 20 μm. C, Association of PsFYVE1 and GmRZ-1A/1B/1C in split-luciferase complementation assay. Leaves were used to measure the LUC activity and 24leaf discs were used to measure the luminescence 48 h after co-expression of the indicated proteins. Error bars represent the mean ± SD (n = 24). All the experiments were repeated three times with similar results.

To determine the precise subsections of PsFYVE1 that dominated its interactions with RZ-1 proteins, four PsFYVE1 residue-exchange and deletion mutants were constructed (Figure 5A). In the luciferase complementation assay, the luminescence signal was not detected in leaves expressing PsFYVE1^ΔFIST_N-nLUC or PsFYVE1^ΔFIST_C-nLUC with cLUC-GmRZ-1A/cLUC-NbRZ-1A (Figure 5A). The stability and size of each recombinant protein was checked by western blot (Supplemental Figure S10, A and B). LUC activity measurements also confirmed those results (Figure 5B). Thus, the FIST_N and FIST_C domains, rather than the FYVE domain, are essential for interactions between PsFYVE1 and RZ-1A proteins.

Figure 5.

FIST domain is essential for the interaction between PsFYVE1 and RZ-1A as well as the promotion of Phytophthora infection. A, Verifying the interaction between PsFYVE1 mutants and RZ-1A in split-LUC assays. Schematic drawings of PsFYVE1 and mutants are shown on the left. The interaction between PsFYVE1 mutants and RZ-1A is shown on the right. B, Statistics analysis of (A). Twelve leaf discs were used to measure the luminescence 48 h after co-expression of the indicated proteins. Error bars represent the mean ± SD (n = 12). C, P. capsici infection assay on leaves expressing PsFYVE1-GFP, PsFYVE1^ΔFIST_N-GFP or PsFYVE1^ΔFIST_C-GFP. P. capsici mycelia were inoculated on the leaves at 24 h after Agroinfiltration. Photographs were taken at 36 hpi under UV light. Dashed lines indicate the lesion areas. D, Protein expression of GFP, PsFYVE1-GFP PsFYVE1^ΔFIST_N-GFP and PsFYVE1^ΔFIST_C-GFP in western blotting. The positions of the expected protein bands are indicated by asterisks. Molecular weight markers are indicated in kDa, and protein loading is visualized by Ponceau stain. E, Average lesion sizes of infected N. benthamiana leaves. Lesion sizes (cm²) were measured at 36 hpi. Data are shown as mean ± SD (n = 16). Asterisks indicate statistical significance: **P < 0.01, two-sided unpaired t test. F, Relative biomass quantification of inoculated N. benthamiana leaves. DNA from P. capsici-infected regions was isolated at 36 hpi and primers specific for the N. benthamiana and P. capsici actin gene were used for qPCR. qPCR was performed and normalized to GFP. Data are shown as mean ± SD of three replicates (**P < 0.01, *P < 0.05; two-sided unpaired t test). All the experiments were repeated three times with similar results.

We found that ectopic expression of the PsFYVE1^ΔFIST_N mutant in N. benthamiana did not enhance plant susceptibility to P. capsici, but its protein accumulation level was decreased substantially (Figure 5, C and D). Moreover, the FIST_C deletion mutant did not enhance N. benthamiana susceptibility, and protein expression was stable (Figure 5, C and D). The lesion areas and relative biomass assay also confirmed the results (Figure 5, E and F). Therefore, PsFYVE1 promotes Phytophthora infection via its linkage to GmRZ-1A/NbRZ-1A, which depends on its FIST_C domain.

Disease resistance and ROS are attenuated in NbRZ-1a-silenced leaves, which can be complemented by GmRZ-1a and NbRZ-1a

A. thaliana RZ-1 proteins are known to participate in abiotic stress responses (Kim et al., 2010; Kim et al., 2010), ABA signaling (Kim et al., 2007) and plant development (Wu et al., 2016). However, whether RZ-1 proteins are associated with plant defenses is largely unknown. To further characterize the roles of RZ-1 proteins in defense response, NbRZ-1A, NbRZ-1B, and NbRZ-1C genes in N. benthamiana were silenced using Agrobacterium-mediated VIGS technology. Silencing efficiency of NbRZ-1A, NbRZ-1B, and NbRZ-1C ranged from 64% to 51% that was indicated by RT-qPCR data (Figure 6A). However, TRV:NbRZ-1C was not specifically silenced and interfered with the NbRZ-1B gene (Figure 6A). Silencing of NbRZ-1B and NbRZ-1C resulted in stunted plants with smaller leaves, whereas silencing of NbRZ-1A did not alter growth compared with that in TRV:GUS (negative control) plants (Figure 6B). Then P. capsici infection assay was conducted on silenced leaves. NbRZ-1A/1B/1C-silenced-plants showed significantly larger disease lesions than GUS-silenced control (Figure 6, B and C), and relative biomass assay also confirmed those results (Figure 6D). Afterward, we tested effects of silencing NbRZ-1A/1B/1C on ROS accumulation triggered by flg22 and chitin. As shown in Figure 6E, flg22-induced ROS production in NbRZ-1A/1B/1C-silenced leaves decreased by 60%, 50%, and 73%, respectively, compared with leaves in the GUS-silenced control. Similarly, chitin treatment generated 67%, 53%, and 55% reduction in ROS accumulation in NbRZ-1A/1B/1C-silenced leaves compared with leaves in the GUS-silenced control (Figure 6F). Thus, the results suggest that NbRZ-1A/1B/1C are positive regulators of plant immunity.

Figure 6.

Silencing of NbRZ-1 in N. benthamiana attenuates plant immunity. A, Silencing efficiency of NbRZ-1A/1B/1C genes. The RT-qPCR results show that each of three genes is significantly silenced by virus-induced gene silencing (VIGS) technology. TRV:GUS was used as a control. Data are shown as means ± SD. B, Lesions caused by P. capsici on the NbRZ-1A/1B/1C-silenced N. benthamiana leaves. The silenced plants were photographed at 30 days after infiltration (upper panel). P. capsici infection photographed on leaves (down panel) were taken at 36 hpi under UV light. The dashed lines indicate the lesion areas. Scale bars represent 1 cm. C, The average lesion areas caused by P. capsici at 36 hpi. The data are shown as mean ± SD (n = 18). Asterisks indicate statistical significance: **P < 0.01; *P < 0.05, two-sided unpaired t test. D, Relative P. capsici biomass quantified by qPCR. DNA from infected regions was isolated at 36 hpi for qPCR analyses. Means and standard deviations from three replicates (**P < 0.01; *P < 0.05, two-sided unpaired t test). E–F, Suppression of flg22- and chitin-induced ROS production in the NbRZ-1A/1B/1C-silenced leaves. Relative luminescence units (RLUs) indicate the relative amounts of H₂O₂ production induced by 1 μM flg22 (E) and 100 μg ml^–1 chitin (F). Also, the significance analysis of total RLUs was shown in the lower panel. Error bars represent the mean ± SD (n > 9). Asterisks indicate statistical significance: **P < 0.01, two-sided unpaired t test. All experiments were repeated three times with similar results.

Based on public RNA-seq data (Jing et al., 2016; Yu et al., 2019), transcription level of GmRZ-1A and NbRZ-1A are higher than those of GmRZ-1B/1C and NbRZ-1B/1C (Supplemental Figure S11A). Therefore, GmRZ-1A and NbRZ-1A were investigated further in this study. According to multiple sequence alignments of full-length amino acid sequences of GmRZ-1A and NbRZ-1A, GmRZ-1A shared 64% identity with NbRZ-1A, and both contained a glycine, arginine, and aspartic acids-rich region in the C-terminus (Supplemental Figure S11B) consistent with A. thaliana RZ-1 proteins (Lorkovic and Barta, 2002; Wu et al., 2016).

To further confirm silencing NbRZ-1A caused the susceptible phenotype, the synthetic GmRZ-1A^syn or NbRZ-1A^syn, synonymous mutation of GmRZ-1A or NbRZ-1A that was transiently expressed to complement the NbRZ-1A-silenced N. benthamiana leaves. After P. capsici challenge, overexpression of GmRZ-1A^syn-GFP and NbRZ-1A^syn-GFP suppressed NbRZ-1A silencing-mediated infection promotion and ROS inhibition (Supplemental Figure S12, A and B). Western blotting was used to analyze the protein stability (Supplemental Figure S12C). Leaves with GmRZ-1A^syn/NbRZ-1A^syn developed significantly smaller lesion than those in leaves with the GFP control in NbRZ-1A-silenced N. benthamiana (Supplemental Figure S12, A and D). The ROS accumulation could be restored upon GmRZ-1A^syn/NbRZ-1A^syn overexpression (Supplemental Figure S12B). The results suggest that NbRZ-1A and GmRZ-1A act as orthologous proteins and maintain an identical positive regulatory role in plant immunity.

RZ-1A c-terminus is essential for interaction with PsFYVE1 and immune function

To test the relation between immune-regulated function and different domains of RZ-1A, a series of truncated versions of RZ-1A were generated, including GmRZ-1A^ΔCT (amino acids 1–110), GmRZ-1A^ΔRRM (amino acids 80–208), NbRZ-1A^ΔCT (amino acids 1–110), and NbRZ-1A^ΔRRM (amino acids 78–209) (Supplemental Figure S13A). To determine the precise subsections of RZ-1A that dominated the interaction of PsFYVE1 and GmRZ-1A/NbRZ-1A, luciferase complementation assays were performed. The results showed that GmRZ-1A^ΔRRM and NbRZ-1A^ΔRRM interacted with PsFYVE1 but GmRZ-1A^ΔCT and NbRZ-1A^ΔCT did not (Supplemental Figures S13, A and B and S14, A and B), suggesting the interaction of RZ-1A and PsFYVE1 depends on the C-terminus of RZ-1A proteins. Then, P. capsici was inoculated on the N. benthamiana leaves transiently expressing the above RZ-1A mutants. GmRZ-1A^ΔCT and NbRZ-1A^ΔCT did not positively regulate plant defense against Phytophthora infection compared with GmRZ-1A/NbRZ-1A, whereas the expression of GmRZ-1A^ΔRRM and NbRZ-1A^ΔRRM could reduce P. capsici infection (Supplemental Figure S13C). Lesion areas with GmRZ-1A^ΔCT and NbRZ-1A^ΔCT overexpression were not statistically different from those of the GFP control, but GmRZ-1A^ΔRRM and NbRZ-1A^ΔRRM could significantly suppress expanding lesion (Supplemental Figure S13D). Stability and size of each fusion protein were checked by western blotting (Supplemental Figure S13E). In addition, GmRZ-1A^ΔCT and NbRZ-1A^ΔCT mutants altered nuclear speckles-localization of GmRZ-1A-GFP and NbRZ-1A-GFP, whereas GmRZ-1A^ΔRRM and NbRZ-1A^ΔRRM also localized in the nuclear speckles (Supplemental Figure S13F). The results confirm that the RZ-1A C-terminus is essential not only for its interaction with PsFYVE1 but also for its function in immune regulation. Nuclear speckle localization of RZ-1As is in a C-terminus-dependent manner.

PsFYVE1 disrupts association of NbRZ-1a and splicing factors

Despite potential roles of A. thaliana RZ-1B and RZ-1C in pre-mRNA splicing and general gene expression, reports demonstrating transcription and splicing functional roles of RZ-1A proteins are very limited (Hanano et al., 1996; Wu et al., 2016). Therefore, we expressed NbRZ-1A in N. benthamiana and its binding proteins were harvested by co-IP. IP samples were in-gel separated and subjected to LC-MS/MS analyses. A total of 51 N. benthamiana proteins in parallel repetitions were identified to potentially interact with NbRZ-1A (Supplemental Table S3). Among them, peptides matching glycine-rich RNA-binding protein 7 (NbGRP7) was the ortholog of A. thaliana GRP7, which is a well-studied plant spliceosomal protein.

In previous studies, A. thaliana GRP7 associates with its homolog GRP8 or core spliceosome component U1–70K to modulate dynamic AS (Schoning et al., 2007; Filichkin and Mockler, 2012; Kruse et al., 2020; Wang et al., 2020c). Therefore, full-length cDNAs of N. benthamiana NbGRP7 (Niben101Scf03953g01007.1), NbGRP8 (Niben101Scf09268g00007.1), and NbU1–70K (Niben101Scf05610g01007.1) were cloned, and their interactions with NbRZ-1A were investigated by luciferase complementation (Figure 7A, Supplemental Figure S15A), Co-IP (Figure 7B), and BiFC (Figure 7C, Supplemental Figure S15B) assays. The results showed that NbRZ-1A interacted with components of the spliceosome complex and likely participated in gene transcriptional and post-transcriptional regulations.

Figure 7.

PsFYVE1 disrupts association of NbRZ-1A and plant splicing factor. A, Association of NbRZ-1A with N. benthamiana GRP7, GRP8, and U1–70K in split-LUC assay. Leaves and fifteen leaf discs were used to measure the LUC activity and 48 h after co-expression of the indicated proteins. Error bars represent the mean ± SD (n = 15). B, Co-IP assay of PsFYVE1 and NbGRP7/NbGRP8/NbU1–70K. Co-IP of protein extracts from agroinfiltrated leaves using GFP-Trap confirmed that HA-tagged NbGRP7, NbGRP8, and NbU1–70K were associated with NbRZ-1A-GFP, but not the GFP-negative control. The positions of the expected protein bands are indicated by asterisks. Protein sizes are indicated in kDa, and protein loading was visualized by Ponceau stain. C, Interaction and co-localization of NbRZ-1A and NbGRP7/NbGRP8/NbU1–70K in BiFC assay. A. tumefaciens cells harboring the indicated YFPn (Yn)- or YFPc (Yc)-fused constructs were co-infiltrated into N. benthamiana leaves. The fluorescent signals were detected by confocal laser microscopy and DAPI were used to a nucleus staining. Photographs were taken 48 h after infiltration. Scale bars represent 20 μm. D, Interference of interaction between NbRZ-1A and plant splicing factors by PsFYVE1. cLUC-NbRZ-1A and NbGRP7/NbGRP8/NbU1–70K-nLUC were co-expressed in N. benthamiana leaves in the presence of PsFYVE1-GFP or PsFYVE1^ΔFIST_C-GFP. The concentrations (OD₆₀₀) of Agrobacteria carrying PsFYVE1-GFP or PsFYVE1^ΔFIST_C-GFP were 0, 0.01, 0.05, and 0.2 respectively. PsFYVE1^ΔFIST_C-GFP indicates a mutant described in Figure 5A and is used as a control. Nine leaf discs were used to measure the luminescence 48 h after co-expression of the indicated proteins. Error bars represent the mean ± SD (n > 9). All experiments were repeated three times with similar results.

To test PsFYVE1 effects on the spliceosome complex of NbRZ-1A–NbGRP7, NbGRP8, and NbU1–70K complex, luciferase complementation assays of NbRZ-1A and NbGRP7/NbGRP8/NbU1–70K were performed in the presence of different concentrations of PsFYVE1-GFP, with non-interacting PsFYVE1^ΔFIST_C-GFP as the negative control. We observed that the luminescence signal of cLUC-NbRZ-1A and NbGRP7-nLUC binding gradually decreased with the increase in concentrations of PsFYVE1-GFP (Figure 7D, Supplemental Figure S15C). The luminescence signal of cLUC-NbRZ-1A and NbGRP8/NbU1–70K-nLUC binding also decreased as the concentration of PsFYVE1-GFP/PsFYVE1^ΔFIST_C-GFP increased, but the signal was not significantly different at the same concentration of PsFYVE1-GFP/PsFYVE1^ΔFIST_C-GFP (Figure 7D, Supplemental Figure S15C). To further validate PsFYVE1 mechanism of action, Co-IP and BiFC assays were used to confirm that PsFYVE1 could disrupt the association of NbRZ-1A and NBGRP7. With the increasing concentration of PsFYVE1 added to the system, the enrichment of NbGRP7 in the NbRZ-1A-bound resins gradually decreased in Co-IP assay (Supplemental Figure S16A). The split-LUC assay showed that the high concentration of PsFYVE1 reduced the number of YFP-positive cells which induced by interaction of NbRZ-1A-Yn and NbGRP7-Yc (Supplemental Figure S16B). Therefore, we conclude that PsFYVE1 can disrupt the association of NbRZ-1A and NbGRP7.

PsFYVE1 and NbRZ-1a modulate pre-mRNA AS of defense-related genes

On the basis of above results, RNA-seq technology was used to further explore whether PsFYVE1 and NbRZ-1A regulate alternative pre-mRNA splicing and transcription in planta. N. benthamiana leaves overexpressing PsFYVE1 and overexpressing GFP (control) and NbRZ-1A-silenced and GUS-silenced (control) leaves were inoculated with P. capsici for 36 hpi, followed by subsequent RNA sequencing. Notably, we found 120 AS events were differentially spliced in the above RNA-seq data, suggesting that PsFYVE1 and NbRZ-1A co-regulate pre-mRNA AS (Supplemental Table S4). Among the 120 overlapped differential AS events, skipped exon (SE, 40 events) was the most abundant type, followed by retained intron (RI, 31 events), alternative 3′ splice site (A3SS, 31 events), and alternative 5′ splice site (A5SS, 18 events).

To validate differential AS events identified in the RNA-seq analysis, four AS events including one SE event and three RI events, were sure tested by performing RT-qPCR and measuring transcript levels of different isoforms (Figure 8A ; Supplemental Figure S17, A to C). For RI events, AS ratios were calculated using the following formula: the relative expression level of the spliced isoform/the unspliced isoform, while splicing ratios of SE event were calculated by the relative expression level of the unspliced isoform/the spliced isoform. For Niben101Scf01240g06009 and Niben101Scf01702g02006, splicing ratios (transcript level of the spliced isoform divided by the transcript level of the unspliced isoform, isoform1/2) increased during PsFYVE1 overexpression or NbRZ-1A silencing (Supplemental Figure S17, A and B), whereas that of Niben101Scf03790g01009 (isoform1/2) decreased (Supplemental Figure S17C). Based on the biological function of homologous genes encoding transcripts with AS events in A. thaliana, Niben101Scf01008g02005, a member of membrane-attack complex/perforin (macpf) family (NbNSL1), is a defense-related gene in plant immunity (Noutoshi et al., 2006). NbNSL1.1 transcript produces a functional protein of macpf family, whereas transcript of NbNSL1.2 undergoes SE, which results in the production of truncated proteins that lacks the macpf domain (Figure 8A). Splicing ratios (transcript level of the unspliced isoform divided by the transcript level of the spliced isoform) of NbNSL1 (NbNSL1.1/NbNSL1.2) increased in leaves with PsFYVE1-overexpressed or NbRZ-1A-silenced (Figure 8A).

Figure 8.

PsFYVE1 involves in pre-mRNA alternative splicing of the defense-related gene to suppresses plant immunity. A, Splicing ratios of the defense-related gene in RT-qPCR analysis. The schematic shows models of two isoforms of NbNSL1 gene (upper panel), in which the arrows indicate the locations of primers used to measure the splicing efficiency of introns and the asterisks indicate a premature termination codon. RT-qPCR was performed using isoform-specific primers to measure the splicing ratios of NbNSL1 gene (lower panel). Total RNA was extracted from the indicated N. benthamiana leaf tissues infected with P. capsici at 36 hpi, including overexpression of GFP (a control), PsFYVE1-GFP, TRV:GUS (a control), and NbRZ-1A-silenced (TRV:NbRZ-1A). Mean values and standard deviations are shown from three replicates. Asterisks indicate statistical significance: **P < 0.01, two-sided unpaired t test. B, P. capsici infection assay on the silenced N. benthamiana leaves with silencing of NbNSL1 gene. P. capsici mycelia were inoculated on the leaves at 30 days after agroinfiltration. The photographs and lesion sizes (cm²) were taken and measured at 36 hpi. Data are shown as mean ± SD (n = 18). Asterisks indicate statistical significance: *P < 0.05, two-sided unpaired t test. All experiments were repeated three times with similar results. C, Mechanism of PsFYVE1 suppressing plant immunity. A schematic diagram illustrating that PsFYVE1 regulates alternative splicing and transcription through targeting RZ-1A protein and disrupting RZ-1A-spliceosome interaction. The secreted effector PsFYVE1 enters plant nucleus, resulting in weak resistance and being more susceptible during infection. One-way solid arrows represent promotion, including increasing RZ-1A interaction with PsFYVE1 in the presence of PsFYVE1 that replacing association with spliceosome, as well as inducing AS event (exon skipping of NbNSL1) and transcription regulation (activation of NbHCT and suppression of NbEIN2/NbSUS4) to increase plant susceptibility.

NSL1 has been reported to negatively regulate salicylic acid (SA)-mediated pathway of cell death programs and defense responses in A. thaliana (Noutoshi et al., 2006). To further proceed functional studies of the protein isoforms produced by AS, the putative defense-related genes NbNSL1 was silenced. NbNSL1-silenced leaves were more resistant to P. capsici (Figure 8B). The results indicated that the NbNSL1 functional isoform was a negative regulator of plant immunity to Phytophthora attack. These results all demonstrate that PsFYVE1 interferes with RZ-1A functions to regulate plant immunity by promoting AS of susceptibility factors of plant immunity.

PsFYVE1 and NbRZ-1a co-regulate transcription of defense-related genes

A total of 79 differentially expressed genes (DEGs) were identified in both PsFYVE1-overexpressing and NbRZ-1A-silenced samples (Supplemental Figure S18A ; Supplemental Table S5). Among them, 60 genes were up-regulated and 19 genes were down-regulated. GO enrichment analysis indicated that the 79 DEGs primarily participated in immune system process, response to stimulus, metabolic process, and regulation of biological process (Supplemental Figure S18B).

Based on the biological function of homologous genes encoding transcripts with 79 DEGs in A. thaliana, three defense-related DEGs were selected: a transferase of plant secondary metabolites (NbHCT), a central factor in ET pathways (NbEIN2), and a sucrose-cleaving enzyme (NbSUS4) for validation of transcriptional changes by RT-qPCR analysis. The RT-qPCR results showed that PsFYVE1 overexpression or NbRZ-1A silencing promoted NbHCT transcription but repressed NbEIN2 and NbSUS4 transcription (Supplemental Figure S18, C–E), which were results consistent with RNA-seq data. In previous studies, it has been proven that HCT can subvert the expression of pathogenesis-related (PR) genes by modifying lignin content and composition (Gallego-Giraldo et al., 2020), and therefore may be a negative regulator of plant immunity. By contrast, EIN2 and SUS4 are positive regulators of plant immunity. A. thaliana EIN2 is involves in trafficking signal inducing-ethylene response, and its deficient mutant enhances susceptibility to Macrophomina phaseolina (Schroeder et al., 2019). Another SUS4 gene can activate plant immune responses against pathogens by modifying sugar metabolism and content (Goren et al., 2011; Tauzin and Giardina, 2014). To analyze functions of the three putative defense-related genes, we inoculated P. capsici to observe infection in NbHCT-, NbEIN2-, and NbSUS4-silenced N. benthamiana leaves. NbHCT-silenced leaves were more resistant to P. capsici; whereas NbEIN2- and NbSUS4-silenced leaves were more susceptible (Supplemental Figure S18F). The results indicate that NbHCT negatively regulates immunity to Phytophthora attack, whereas NbEIN2 and NbSUS4 are positive regulators of plant immunity to Phytophthora attack. Overall, the results suggest that PsFYVE1 interferes with RZ-1A functions to regulate plant immunity by promoting transcription of susceptibility factors and repressing transcription of positive regulators of plant immunity.

Discussion

Phytophthora pathogens secrete diverse groups of effectors into plants to interfere with plant immune responses and promote infection. Although some types of effectors, such as RXLRs, CRNs, elicitins, and NLPs, have been identified and widely studied, there are still many unknown effectors to be excavated. In this study, an FYVE domain-containing protein family that is highly expanded and with distinct characteristics in oomycetes provided a potential library of uncharacterized effectors. The FYVE domain-containing protein PsFYVE1 was selected and functional analysis was performed. We found that PsFYVE1 contained a functional signal peptide and exhibited cell-to-cell movement inside plant cells. PsFYVE1-silenced transformants significantly decreased pathogen virulence, and overexpression of PsFYVE1 in N. benthamiana could suppress plant immune response. The conserved FYVE domain of PsFYVE1 is confirmed to be essential for its PI3P-binding activity, suggesting that this class of secreted FYVE proteins could act as potential effectors for Phytophthora pathogens. PsFYVE1 interacted with the RNA-binding proteins NbRZ-1A of N. benthamiana and GmRZ-1A of host soybean. Silencing of NbRZ-1A in plants suppressed Phytophthora infection and ROS accumulation, indicating the NbRZ-1A protein contributes positively to plant resistance. In addition, NbRZ-1A formed a splicing-related complex with GRP7, GRP8, and U1–70K in N. benthamiana. PsFYVE1 disrupted association of NbRZ-1A and splicing factors GRP7, suggesting that PsFYVE1–NbRZ-1A interaction interferes with plant immunity at the level of RNA regulation. Both RNA-seq and RT-qPCR analyses indicated that PsFYVE1 and NbRZ-1A co-affected transcription and AS of immune-related genes. Overall, our results reveal that PsFYVE1 is a P. sojae effector that regulates plant immunity by interfering with the functions of RZ-1A to reprogram gene transcriptional and post-transcriptional process (Figure 8C).

Although the FYVE domain-containing protein family has been widely investigated in humans, yeasts and plants, its distribution, evolution, and functions are poorly known in oomycetes. In this study, FYVE domain-containing proteins were systematically identified and analyzed in oomycetes, especially P. sojae. We observed that all the oomycete genomes contained much larger numbers of FYVE proteins compared with other eukaryotes. In addition to conserved FYVE domain, oomycete FYVE proteins also contained a relatively large number of specific additional domains, which might help oomycetes adapt to their distinct environments during evolution. According to published RNA-seq data of P. sojae, many FYVE genes were up-regulated at different infection stages, whereas some were highly expressed in developmental stages, suggesting that P. sojae FYVE genes may be involved in vegetative growth, stress response, and virulence. In this study, PsFYVE1 contained a signal peptide, and was confirmed to be a potential effector. When the large number of proteins, diverse domain compositions, and variable transcriptional profiles are considered, it is likely that other signal peptide-containing FYVE proteins derived from filamentous pathogens, including oomycetes and fungi, are also potential effectors, which need to be investigated further in the near future.

Our results showed that PsFYVE1 carried a functional secretory signal peptide and a conserved FYVE domain, which was required for binding to PI3P. The full-length PsFYVE1 and its FYVE domain could bind to PI3P, and this binding depends on four conservative amino acid residues. Whether transmembrane transfer of PsFYVE1 depends on its PI3P-binding ability remains to be determined. It is difficult to study the host cell translocation of RXLR effector, as these processes occur only at the interface of the pathogen with the host and only during infection. Although it is widely accepted that the sequence of signal peptide and RXLR-motif mediate RXLR effectors translocation into host cells and/or increase their stability in planta, how this occurs is still under debate. It has been shown that RXLR effectors enter plant cells via their RXLR domain-mediated association with PI3P (Kale et al., 2010). By contrast, a P. infestans RXLR effector, AVR3a, binds to PI3P via its effector domain to enhance its stability during infection (Yaeno et al., 2011). We have shown that Phytophthora could produce PI3P to aid RXLR effectors action (Lu et al., 2013). Interestingly, RXLR motif of AVR3a is cleaved before secretion and the potential cleavages of other RXLR effectors remain largely unknown (Wawra et al., 2012; Wawra et al., 2017). Whether the FYVE domain also has similar functions as RXLR motif in effector translocation is a question that needs to be addressed in the future. Here, we performed an alternative method, which analyzed the cell-to-cell movement of full-length PsFYVE1 using a plant binary vector. In several literatures, host cell-to-cell movement has been used to prove the effector could be secreted and trafficked inside the host cells (Khang et al., 2010; Cao et al., 2018; Li et al., 2021). We performed a host cell-to-cell movement experiment, which was generated to test Tomato spotted wilt tospovirus (TSWV) NSm (Feng et al., 2016). We found that PsFYVE1 has similar intercellular movement as TSWV NSm, suggesting that PsFYVE1 could be secreted from plant cell and trafficked in the surrounding cells. In addition, PsFYVE1 contained FIST_N and FIST_C domains, which are found in signal transduction proteins derived from bacteria, archaea, and eukaryotes (Borziak and Zhulin, 2007). FIST_C domain of PsFYVE1 was essential for its interaction with target proteins, which is consistent with a previous study that found FIST domains bind amino acids ligands (Borziak and Zhulin, 2007). Moreover, FIST_C domain of PsFYVE1 is required for PsFYVE1 enhanced susceptibility suggesting that FIST_C specially contributes to virulence function. Overall, PsFYVE1 is a virulence protein secreted by P. sojae, and its two functional domains cooperate to perform different functions for successful infection.

RZ-1 proteins, including RZ-1A, RZ-1B, and RZ-1C, constitute a small group of plant GRPs that are widespread in land plants, acting as important regulators of seed maturation, flowering time and freezing tolerance (Kim et al., 2007; Kim et al., 2010a; Wu et al., 2016). However, biological roles of RZ-1 proteins in plant immune responses remain unclear. In this study, PsFYVE1 targeted RZ-1A/1B/1C proteins in N. benthamiana and host soybean, and silencing NbRZ-1A/1B/1C proteins, suppressed plant immunity, suggesting that RZ-1 proteins positively regulated plant resistance. In addition, complementation of GmRZ-1A or NbRZ-1A in NbRZ-1A-silenced leaves rescued susceptibility phenotypes, indicating that, at least, RZ-1A proteins play a conserved role in immunity in diverse plants. Thus, these results extend the functions of RZ-1A in plant immunity.

In this study, PsFYVE1–NbRZ-1A interaction affected transcription and pre-mRNA AS of immune-related genes, which provides further characterization of how Phytophthora pathogens reprogram plant immunity by changing transcriptional and post-transcriptional processes. To successfully colonize plants, pathogens produce intracellular effectors that inhibit defense responses by different molecular mechanisms, including modulating gene expression. Two nuclear effectors of Magnaporthe oryzae, MoHTR1 and MoHTR2, modulate disease susceptibility of rice (Oryza sativa) via reprograming the transcription of many immunity-associated genes (Kim et al., 2020). In addition to transcriptional control of gene expression, post-transcriptional processes, notably AS, emerge as key mechanisms for gene regulation during infection. For example, P. infestans RXLR effector SRE3 interacts directly with splicing factors U1–70K, SR30, and SR45 to manipulate AS of plant immune processes (Huang et al., 2020). In this study, NbRZ-1A associated with the components of spliceosome complex, including GRP7, GRP8, and U1–70K, and PsFYVE1 attacked the NbRZ-1A–GRP7 interaction. The results suggest that PsFYVE1 can manipulate AS through interaction with NbRZ-1A to indirectly interfere with the spliceosome. Furthermore, RNA-seq analysis of PsFYVE1-expressing and NbRZ-1A-silenced leaves shed light on that PsFYVE1 and NbRZ-1A co-regulated transcription and pre-mRNA AS of many plant genes. Biochemical experiments combined with molecular-genetic analysis revealed that overexpression of PsFYVE1 and silencing of NbRZ-1A subverted plant defense responses by promoting transcription or AS of susceptibility factors and repressing transcription of positive regulators of plant immunity. These results provide evidence that transcriptional changes and AS regulation can co-exist simultaneously in immunomodulation, in connection with regulation properties of RNA-binding proteins.

In addition, the significant changes in different splicing variants of three genes were verified. PsFYVE1 overexpression or NbRZ-1A silencing induced intron retention of Niben101Scf01240g06009, Niben101Scf01702g02006, and Niben101Scf03790g01009 (Supplemental Figure S17, A to C). Sequence analysis of Niben101Scf01240g06009 and Niben101Scf03790g01009 showed that they share similarity with genes encoding eukaryotic initiation factors in A. thaliana, which help stabilize the formation of the translation initiation complex to provide regulatory mechanisms in translation initiation (Li et al., 2017; Gallie, 2018). The sequence of Niben101Scf01702g02006 is similar to that of a gene encoding a vacuole-sorting protein in A. thaliana, which are of great importance in autophagy and auxin transport in plants (Jaillais et al., 2007; Jha et al., 2018). These imply that PsFYVE1 and NbRZ-1A may have a fine-tuning role in developmental pathways.

In conclusion, we systematically analyzed the FYVE domain-containing protein family in oomycetes by bioinformatics, and identified functional characteristics of PsFYVE1, a FYVE protein secreted from the plant pathogen P. sojae. PsFYVE1 is a potential effector that interacts with plant RZ-1A proteins to manipulate the transcription and splicing process by disrupting the association of RZ-1A with splicing factor. These findings highlight a previously uncharacterized mechanism in which a virulence factor of Phytophthora hijacks plant RZ-1A proteins to overcome plant defenses.

Materials and methods

Bioinformatics analyses

The genomes of organisms used in this study were retrieved from NCBI and JGI databases. To predict FYVE domain-containing proteins in each genome, the Hidden Markov Model (HMM) profile of the FYVE domain (PF01363) was obtained from the Pfam 34.0 database (http://pfam.xfam.org/), followed by searching against each genome using the HMMER tool (E value cut-off <1e-5) (Johnson et al., 2010). To investigate phylogenetic relations of FYVE proteins between oomycetes and other organisms, the selected FYVE proteins were used to construct a phylogenetic tree following a neighbor-joining algorithm with 1,000 bootstrap replicates in MEGA 7 software. SignaIP 3.0 online tool was used to predict the signal peptide. Potential additional domains in each FYVE protein were predicted by searching against the Pfam database.

Plants and microbe cultivation

N. benthamiana plants were grown in a greenhouse under a 16 h day at 25°C and an 8 h night at 22°C. Four-leaf-stage N. benthamiana plants were used for VIGS and 6–8 weeks old N. benthamiana leaves were used for inoculations of Phytophthora. Etiolated soybean (Glycine max) seedlings were grown without light for 5 days and then under a 16 h day at 25°C and an 8 h night at 20°C before inoculation. Phytophthora sojae (P6497) and P. capsici (LT263) strains were routinely maintained on 10% (v/v) vegetable (V8) juice medium at 25°C in the dark.

Yeast signal sequence trap system

The yeast signal trap system was performed as following described (Oh et al., 2009; Xu et al., 2020). The signal peptide of PsFYVE1 or Avr1b (as positive control) was ligated into vector pSUC2T7M13ORI (pSUC2), which carries a truncated invertase gene (SUC2) lacking both the initiation Met and signal peptide. pSUC2 recombinant plasmids were transformed into yeast strain YTK12 using the lithium acetate method. Transformed colonies could grow on CMD-W (minus Trp) plates after 48 h of incubation at 30°C. To assay for invertase secretion, colonies were transferred to YPRAA plates. Invertase enzymatic activity was detected by the reduction of 2, 3, 5-triphenyltetrazolium chloride (TTC) to insoluble red-colored triphenylformazan. The pSUC2 empty vector was used as the negative control.

Plasmid construction

PsFYVE1 without a signal peptide and the FYVE domain fragment was cloned from the complementary DNA (cDNA) of P. sojae mycelia. In the P. sojae transformation experiment, the FYVE domain fragment was ligated into pTOR using a one-step cloning kit (Vazyme Biotech, Nanjing, China). In the inoculation assays, the amplified PsFYVE1 fragment was ligated into pBinGFP4 (C-terminal tag green fluorescent protein [GFP] fusion) (Ma et al., 2017). Full-length RZ-1 genes were cloned from N. benthamiana and soybean cDNAs. Full-length NbGRP7, NbGRP8, and NbU1–70K genes were coned from N. benthamiana using gene-specific primers. GmRZ-1A^syn and NbRZ-1A^syn were synthesized by Nanjing Genscript (Nanjing, China) and ligated into pCAM1300. Individual colonies of each construct were tested by PCR and verified by sequencing. The cloning primers are shown in Supplemental Table S6. Primers were synthesized by Sangon Biotech (Shanghai, China).

P. sojae transformation and silencing efficiency detection

The P. sojae transformation was conducted as previously described (Fang and Tyler, 2016; Wang et al., 2020b). The antisense silencing sequence was designed in the FYVE region of PsFYVE1. P. sojae strain P6497 was maintained on V8 juice agar. Before liquid culture, P. sojae discs were inoculated onto nutrient pea agar plates. Mycelia were harvested and then cultured in nutrient pea broth liquid medium for 2 days. Mycelia were washed in 0.8 M mannitol and then placed in a mixture of lysing enzyme and cellulysin solution to incubate for 40–50 min at 25°C with gentle shaking. Protoplasts were harvested by centrifugation at 282 g for 3 min and particles were removed in W5 solution (5 mM KCl, 125 mM CaCl₂·2H₂O, 154 mM NaCl, and 177 mM glucose). After 30 min, protoplasts were collected by centrifugation at 282 g for 4 min, and adjusted concentration in MMg solution to maintain osmotic pressure of protoplasts. One milliliter of protoplasts was added to 35 µg of transforming DNA and incubated for 10 min on ice. Then, three successive aliquots of 580 µl of fresh 40% polyethylene glycol (PEG4000) solution were pipetted into the protoplast suspension and gently mixed. After a 20-min incubation on ice, 2, 8, 10 ml of pea broth liquid medium containing 0.5 M mannitol was added to the mixed solution every 2 min, and protoplasts were incubated for 14–16 h to regenerate in the dark. Regenerated protoplasts were harvested by centrifugation at 501 g for 5 min, mixed thoroughly in pea broth solid medium containing 50 µg ml^–1 G418 and 50 µg ml^–1 ampicillin, and then poured in plates. Transformed colonies were observed after 2 days of incubation at 25°C and again covered with V8 juice medium. After 2–3 days, transformants were selected based on the transcription level of PsFYVE1. Total RNA was extracted from mycelia of transformants cultured in V8 juice medium by using an RNA-simple Total RNA Kit (Tiangen Biotech, Beijing, China). Primers specific for PsFYVE1 gene and P. sojae actin gene (Supplemental Table S6) were used to quantify the relative expression level and evaluate silencing efficiency.

Soybean and N. benthamiana inoculation assays

In soybean inoculation assays, P. sojae transformants were cultured in V8 juice medium without G418 selection for 5 days. Hypocotyls of etiolated soybean seedlings (Chinese susceptible cv. HF47) were inoculated with P. sojae transformant mycelia. Lesion lengths were measured at 48 h post inoculation (hpi), and infection photographs were also taken at 48 hpi. Infected soybean seedlings of equal quality were collected to quantify the relative biomass of P. sojae by qPCR. Primers of the P. sojae actin gene and soybean actin gene are listed in Supplemental Table S6. qPCR reactions were performed on an ABI Prism 7500 Fast real-time PCR System (Applied Biosystems, Foster City, CA, USA).

For P. capsici infection on N. benthamiana leaves, overexpression of all proteins was confirmed by western blotting, and pBinGFP4 empty vector was used as the negative control. Leaves were detached 24 h after agroinfiltration and then inoculated with 6-mm disks of 4-day growth P. capsici medium, and the hyphal side kept close to the leaves. The P. capsici-inoculated leaves were put into a growth chamber at 25°C, and lesion areas (cm²) were scored at 36 hpi under UV light. Primers specific for P. capsici and N. benthamiana actin genes (Supplemental Table S6) were used to quantify the relative biomass of P. capsici.

Protein expression and lipid filter-binding assays

Escherichia coli cells containing plasmids encoding each fusion protein were grown in LB (Luria-Bertani) liquid medium at 37°C to an OD₆₀₀ of 0.4–0.6. Protein expression was induced with 0.1–0.5 mM IPTG at 16°C for 16–20 h. The GST-fusion proteins were collected, sonicated, and purified using glutathione-sepharose 4B beads (GE Healthcare, China).

Lipid filter-binding assays were conducted as previously described (Kale et al., 2010; Zhou et al., 2021). Lipids were dissolved in a mixture of methanol:chloroform:water (20:10:8) and then spotted at 100 pmol onto Hybond-C extra membranes. The membranes were dried at room temperature for 1 h and then incubated in blocking buffer (50 mM Tris-HCl, 150 mM NaCl, 0.1% [v/v] Tween 20, and 2 mg ml^–1 fatty acid-free BSA, pH 7.5) at room temperature for 1 h. Purified GST-fusion proteins, 20–30 mg, were added to the blocking buffer, and the membranes were incubated overnight at 4°C. The membranes were washed in TBST (50 mM Tris-HCl, 150 mM NaCl, and 0.1% [v/v] Tween 20) 10 times over 50 min. Bound proteins were detected with mouse anti-GST antibody at a ratio of 1:2,000. After inoculation for 1 h, the membranes were washed using TBST 12 times over 1 h and incubated with a 1:5,000 dilution of the HRP-conjugated anti-mouse secondary antibody for 1 h. The membranes were washed in TBST 12 times over 1 h, and then visualized by an Enhanced Chemiluminescence (ECL) reagents (Thermo Fisher Scientific, Waltham, USA) according to the manufacturer's instructions.

Oxidative burst assay

The oxidative burst assay was carried out as previously described (Li et al., 2019; Liang et al., 2021), and used to determine the role of indicated construct proteins in the ROS burst in response to flg22 and chitin. A luminol-based assay was performed as following described. After protein overexpression for 48 h, leaf discs were incubated with 200 µl of water in a 96-well plate overnight in the dark at 25°C. The oxidative burst of leaf discs was detected using 200 µl of test buffer (1 µM flg22, 100 µM luminol, and 20 µg ml^–1 horseradish peroxidase) (100 µg ml^–1 chitin, 100 µM luminol, and 20 µg ml^–1 L-012) and recorded by a Promega GloMax Navigator microplate luminometer (Promega Biotech, Beijing, China). GraphPad Prism software was used to analyze data.

Western blotting

After agroinfiltration for 48 h, proteins were extracted and fractionated by SDS-PAGE. Separated proteins were transferred from the gel to a PVDF membrane and then blocked using PBST (PBS with 0.1% [v/v] Tween 20) containing 5% (w/v) non-fat milk at room temperature for 30–40 min. Anti-GFP (1:5,000; #M20004; Abmart Biotech, Shanghai, China) and anti-HA (1:5,000; #M20003; Abmart Biotech, Shanghai, China) antibodies were added to PBSTM (PBST with 5% non-fat dry milk) and incubated at room temperature for 180 min, followed by washes with PBST three times (5 min each). The membranes were incubated with a goat anti-mouse IRDye 800CW antibody (Odyssey, no. 926-32210; Li-Cor) at a ratio of 1:10,000 in PBSTM for 30 min at room temperature and then washed three times with PBST. Proteins were visualized by excitation at 800 nm.

Co-IP assays

N. benthamiana leaves were collected 48 h after initial agroinfiltration and then frozen in liquid nitrogen and ground to powder using mortar and pestle. The total proteins were extracted using RIPA lysis buffer (50 mM Tris pH 7.4, 150 mM NaCl, 1% [v/v] Triton X-100, 1% [w/v] sodium deoxycholate and 0.1% [w/v] SDS) and centrifuged at 4°C for 10 min at 12,000 × g. The supernatant was transferred to a new tube, and appropriate SDS-PAGE sample loading buffer was added into tubes to detect proteins by western blotting. For GFP-IP, 1–4 ml of supernatant was incubated at 4°C with GFP-Trap_A beads (Chromotek, Planegg-Martinsried, Germany) overnight. Beads were collected by centrifugation at 2,500 × g and then washed three times in 1–4 ml of washing buffer (10 mM Tris-Cl, 150 mM NaCl, and 0.5 mM EDTA, pH 7.5). Appropriate SDS-PAGE sample loading buffer was added to bound proteins and boiled for 8–10 min. The results were detected by western blotting using anti-GFP and anti-HA antibodies. The images were caught by Odyssey CLx System.

VIGS of NbRZ-1 genes in N. benthamiana

Tobacco Rattle Virus (TRV)-based VIGS system was used to silence NbRZ-1A, NbRZ-1B, and NbRZ-1C gene in N. benthamiana. Fragments from each gene were cloned into the pTRV2 vector. The TRV:GUS vector was used as the control. Four-leaf stage N. benthamiana plants were used in agroinfiltration containing a mixture of pTRV1 and pTRV2 vectors at OD₆₀₀ of 0.5 each (Velasquez et al., 2009). Silenced leaves were inoculated, and primers specific for silencing genes and N. benthamiana actin gene were used to detect silencing efficiency at 30 days post-infiltration (dpi) of Agrobacterium containing pTRV vectors.

Confocal microscopy and bimolecular fluorescence complementation

The N. benthamiana leaves were cut and mounted in water and analyzed using an LSM 710 laser scanning microscope with a × 63 objective lens (Carl Zeiss, Jena, Germany). GFP, mCherry, YFP or DAPI fluorescence was observed at an excitation of 488, 587, 514, or 358 nm, respectively.

RZ-1 genes and PsFYVE1 (without signal peptide) sequences were cloned into YFPn and YFPc vectors separately. The A. tumefaciens strains GV3101 harboring indicated plasmids were mixed at final OD₆₀₀ of 0.5 each, and 6-week-old N. benthamiana plants were used for infiltration. Co-locational fluorescence was observed 36–48 hpi using confocal microscopy.

Luciferase complementation assays

Lluciferase complementation assays were carried out as previously described (Chen et al., 2008; Huang et al., 2020). PsFYVE1, PsFYVE1 mutant genes (without a signal peptide), and N. benthamiana GRP7, GRP8, and U1–70K genes were inserted into pCAMBIA1300-nLUC (nLUC), whereas soybean RZ-1A/1B/1C genes and N. benthamiana RZ-1A/1B/1C genes were inserted into pCAMBIA1300-cLUC (cLUC).

A. tumefaciens strains GV3101 harboring indicated plasmids were mixed at final OD₆₀₀ of 0.5 each and then infiltrated into N. benthamiana leaves. The N. benthamiana leaves were collected at 48 h after infiltration and then incubated with 1 mM luciferin substrate (catalog no. N1110; Promega Biotech, Beijing, China). LUC images were captured with a Tanon-5200 Multi Chemiluminescent Imaging System (Tanon, China). The N. benthamiana leaf discs were incubated with 1 mM luciferin substrate in a 96-well plate, and activity of luciferase reporter gene (LUC) was detected using Promega GloMax® Navigator microplate luminometer.

RNA-Seq data analysis

1. N. benthamiana leaves overexpressing PsFYVE1 and overexpressing GFP (control) and NbRZ-1A-silenced and GUS-silenced (control) leaves were inoculated with P. capsici for 36 hpi. Two biological repeats were performed. The total RNA was extracted using an RNA-simple Total RNA Kit (Tiangen Biotech, Beijing, China), followed by constructing RNA-Seq libraries with the Illumina TruSeq RNA Sample Preparation Kit v2 (Illumina, San Diego, USA) following the manufacturer's recommendations. These libraries were sequenced on the Illumina HiSeq X Ten platform, and 150 bp pair-end reads were produced. The reads containing poly-N and the low-quality reads were removed to obtain clean reads. The clean reads were mapped to N. benthamiana v1.0.1 reference genome using HISAT2 (Kim et al., 2015), and only the uniquely mapped reads were used for subsequent processing. The expression level of each gene was calculated by normalizing to the fragment per kilobase of exon per million mapped reads (FPKM) value. DEGs were identified by DESeq R package (Anders and Huber, 2010), and genes with at least two fold change and P value < 0.05 were considered differentially expressed. AS events were identified using rMATS v4.1.0 (Shen et al., 2014), and events with P value < 0.05 were regarded as the differential AS events.

Accession numbers

Genes referenced in this article can be found in the GenBank database under the following accession numbers: PsFYVE1 (OL848976), NbRZ-1A (OL848977), NbRZ-1B (OL848977), NbRZ-1C (OL848977), GmRZ-1A (OL848977), GmRZ-1B (OL848981), and GmRZ-1C(OL848982). The raw data of RNA-Seq has been submitted to NCBI SRA database (accession number PRJNA778182).

Supplemental data

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

Supplemental Figure S1 . Phylogenetic analysis of FYVE domain-containing proteins.

Supplemental Figure S2 . PsFYVE1 contains classic FYVE domain and shows high expression during infection.

Supplemental Figure S3 . PsFYVE1 carries a functional secretory signal peptide.

Supplemental Figure S4 . Signal peptide and FYVE domain are necessary to deliver PsFYVE1 into plant cells.

Supplemental Figure S5 . Silencing of PsFYVE1 does not affect mycelial growth.

Supplemental Figure S6 . Phosphoinositide-binding activities of PsFYVE1 and its FYVE domain.

Supplemental Figure S7 . PsFYVE1 interacts with NbRZ-1A, NbRZ-1B, and NbRZ-1C.

Supplemental Figure S8 . Transcription profiles and sequence analysis of RZ-1A proteins.

Supplemental Figure S9 . Complementation of GmRZ-1A^syn and NbRZ-1A^syn recovers the TRV:NbRZ-1A phenotype.

Supplemental Figure S10 . RZ-1A C-terminus is essential for interaction with PsFYVE1 and immune-function.

Supplemental Figure S11 . Both PsFYVE1 and NbRZ-1A regulate transcription of plant immunity-related genes.

Supplemental Figure S12 . Both PsFYVE1 and NbRZ-1A regulate pre-mRNA alternative splicing.

Supplemental Figure S13 . RZ-1A C-terminus is essential for interaction with PsFYVE1 and immune function.

Supplemental Figure S14 . The protein expressions of RZ-1A mutants and PsFYVE1 in split-LUC assay were detected by western blotting.

Supplemental Figure S15 . The protein co-expression of PsFYVE1, NbRZ-1A and plant splicing factors in split-LUC and BiFC assays were detected by western blotting.

Supplemental Figure S16 . PsFYVE1 disrupts association of NbRZ-1A with NbGRP7.

Supplemental Figure S17 . Both PsFYVE1 and NbRZ-1A regulate pre-mRNA alternative splicing.

Supplemental Figure S18 . Both PsFYVE1 and NbRZ-1A regulate transcription of plant immunity-related genes.

Supplemental Table S1 List of predicted FYVE domain-containing proteins in oomycetes.

Supplemental Table S2 Putative targets of PsFYVE1 by immunoprecipitation assay.

Supplemental Table S3 Putative targets of NbRZ-1A by immunoprecipitation assay.

Supplemental Table S4 List of changed alternative splicing events.

Supplemental Table S5 Differentially expressed genes identified in both PsFYVE1-overexpressing and NbRZ-1A-silenced samples at 36 hpi.

Supplemental Table S6 Primers used in this study.

Funding

This work was supported by the National Natural Science Foundation of China (32070139, 31625023, and 32072507) and the Fundamental Research Funds for the Central Universities (KYT202001 and JCQY201904).