Introduction

Plants fight pathogen attack via a complex immune system that includes basal and highly specific defense responses1,2,3. The first line of defense against the vast majority of potential pathogens is termed pathogen-associated molecular pattern (PAMP)-triggered immunity (PTI)1,2. To circumvent PTI, many host-adapted pathogens produce different classes of effector molecules1,2. In turn, plants have evolved a second surveillance mode to recognize pathogen-derived effector proteins. This second mode or line of defense is triggered by specialized molecules such as resistance (R) proteins, which recognize the effectors and activate effector-triggered immunity (ETI)1,2. When R genes recognize effector proteins, the typical resistance responses include the hypersensitive response (HR) and extreme resistance (ER)4,5. HR results in localization of the pathogen at the inoculation site as a consequence of the rapid death of infected and neighboring cells5. In contrast, ER is characterized by the arrest of pathogens at the inoculation site without any macroscopic or microscopic cell death lesions4,6.

Soybean mosaic virus (SMV), a member of the genus Potyvirus, is one of the most common viruses of soybean (Glycine max (L) Merr.)7. SMV has a single-stranded positive-sense RNA genome of approximately 9.6 kb. This genome encodes one large polyprotein, which is proteolytically cleaved to generate at least 11 mature proteins: P1, helper-component proteinase (HC-Pro), P3, PIPO, 6K1, cylindrical inclusion (CI), 6K2, VPg, NIa-Pro, NIb and coat protein (CP)8. SMV has been classified into several distinct strains based on disease reactions on various soybean genotypes9,10. Genetic mapping studies have identified three distinct R genes (Rsv1, Rsv3 and Rsv4) that confer strain-specific resistance to SMV11,12,13,14. Extensive studies have identified the SMV effector proteins that interact with each Rsv gene: SMV strain-specific P3 protein is an effector of both Rsv1- and Rsv4-mediated disease resistance15,16 while SMV CI protein is an effector of Rsv3-mediated ER17,18.

The plant hormone abscisic acid (ABA) not only regulates various plant developmental processes but also has a primary function in plant response to biotic and abiotic stress19,20. Accumulating evidence suggests that ABA has many effects on defense responses, acting as both a positive and negative modulator of associated signaling pathways; its effects depend on the specific plant–pathogen interaction19. The key regulators of the ABA signaling network are a subclass of serine/threonine phosphatases, represented by type 2C protein phosphatases (PP2Cs)21,22. The large number of PP2C genes in plant genomes suggests that the integrative network of ABA signaling is finely orchestrated by the diversity and specificity of PP2Cs22.

Although Rsv3-mediated ER induced in response to SMV attack was phenotypically characterized in the previous studies17,18, the involved molecular mechanism remains unclear. In the current study, we gained insight into the molecular mechanism involved in Rsv3-mediated ER by using RNA sequencing (RNA-Seq) to analyze the genome-wide transcriptional reprograming that occurs in soybean expressing ER to SMV. In Rsv3-mediated ER, we observed a significant up-regulation of a subset of the PP2C genes that are induced by ABA. we focused on the PP2C gene that was most up-regulated, which was GmPP2C3a. Using infectious clones of avirulent and virulent strains of SMV (SMV-G5H and -G7H, respectively) tagged with a β-glucuronidase (GUS) reporter gene, we show that Rsv3-mediated ER restricts virus accumulation in the initially inoculated cells. Synchronized over-expression of GmPP2C3a in the cell inoculated with SMV-G7H results in ER, suggesting that GmPP2C3a is a key regulatory gene conferring ER. Further evidence suggests that inhibition of viral cell-to-cell movement by Rsv3-mediated ER is primarily caused by callose deposition in an ABA signaling-dependent manner.

Results

The strategy used here for comparative analysis of genome-wide transcriptional regulation of Rsv3-mediated ER is outlined in Figure 1. Infectious sap of SMV-G5H (an avirulent strain) was mechanically inoculated on the fully expanded first trifoliolate leaves of the soybean line L29 carrying Rsv3. For controls, the soybean leaves were inoculated with mock and SMV-G7H (a virulent strain) in the same way. For time-course analysis, mRNAs, which were extracted from the inoculated leaves and untreated healthy leaves at 8, 24 and 54 hours post inoculation (hpi), were subjected to RNA-Seq. To identify genes specifically involved in Rsv3-mediated ER, we evaluated differentially expressed genes among the samples, i.e., the expression levels of the genes were calculated relative to those of the corresponding genes in the untreated, healthy control (see Methods).

Figure 1
figure 1

The strategy for comparative analysis of genome-wide transcriptional regulation of Rsv3-mediated extreme resistance.

The fully expanded first trifoliolate leaves of the soybean line L29 carrying Rsv3 were mechanically inoculated with either mock, SMV-G5H (avirulent stran), or SMV-G7H (virulent strain). For time-course analysis, mRNAs were extracted from the inoculated leaves and untreated healthy leaves at 8, 24 and 54 hours post inoculation (hpi) and subjected to RNA-Seq. To identify genes specifically involved in Rsv3-mediated ER, comparative analyses of the transcriptome profiles were performed. The image of the leaf in Fig. 1 is a photograph taken by one of our co-authors.

Full size image

Specific up-regulation of a subset of the PP2C gene family in Rsv3-mediated ER

We performed extensive analysis of the RNA-Seq data for gene families categorized by biological function (according to the SoyBase database). Using this approach, we found a significant up-regulation of a subset of the PP2C gene family in Rsv3-mediated ER at an early time point (8 hpi) (Figs. 2 and S1). Phylogenetic analysis of the soybean PP2C gene family revealed that the up-regulated PP2C genes belonged to the same cluster that contains the soybean orthologs of ArabidopsisABI1 and ABI2, which are negative regulators of ABA responses23,24 (Fig. S1B). However, the expression levels of the soybean orthologs of ABI1 and ABI2 were not significantly different from those of the controls, indicating that the regulatory mechanism of Rsv3-mediated ER differs, at least in the identity of the ABA-inducible genes, soybean vs. Arabidopsis (Fig. S1). The expression levels of the significantly up-regulated PP2C genes were quantified as reads per kilobase of the transcript per million mapped reads to the transcriptome (RPKM). Among them, Glyma14g32410, which is referred to as GmPP2C3a in this study, was the most highly expressed in Rsv3-mediated ER at an early time point (8 hpi) (Fig. 2C). The high GmPP2C3a expression level indicated by RNA-Seq was confirmed by real-time PCR (Fig. 2D). Thus, we subjected GmPP2C3a to detailed analysis with the goal of determining its functional involvement in Rsv3-mediated ER.

Figure 2
figure 2

Expression levels of the PP2C genes inducible by ABA in Rsv3-mediated ER.

(A) The heat map showing the hierarchically clustered expression patterns of 16 soybean PP2C genes inducible by ABA. The heat map was produced using a log scale of the values of RPKM (reads per kilobase of the transcript per million mapped reads to the transcriptome) obtained by RNA-Seq. Red and green indicate up-regulation and down-regulation, respectively. (B) Phylogenetic analysis of 16 soybean PP2C genes inducible by ABA. The analysis was performed using the deduced amino acid sequences by the maximum likelihood method. (C) The RPKM values of five significantly up-regulated PP2C genes. (D) Real-time PCR validation of the expression level of Glyma14g32410, which is referred to as GmPP2C3a in this study. Symbols indicate the significantly up-regulated PP2C genes in Rsv3-mediated ER at 8 hpi.

Full size image

ABA accumulation accompanies Rsv3-mediated ER

Because some ArabidopsisPP2C genes including ABI1 and ABI2 are transcriptionally up-regulated by ABA25,26, the endogenous levels of ABA were analyzed in soybean leaves inoculated with either SMV-G5H or SMV-G7H. As expected, ABA accumulation was significantly higher in the leaves inoculated with SMV-G5H than in those inoculated with mock or SMV-G7H (Fig 3A). Because ABA has an antagonistic effect on SA signaling that plays an important role in activation of defense responses, the accumulation levels of SA were also analyzed in the same set of the soybean leaf samples. Interestingly, SA accumulation was dramatically increased in the SMV-G7H-inoculated leaves while no significant difference in SA accumulation was observed between SMV-G5H- and mock-inoculated leaves (Fig. 3B). Taken together, the results suggested that the ABA signaling pathway might be involved in conferring Rsv3-mediated ER and that GmPP2C3a could function as a key regulator of this signaling pathway. In addition, it seemed likely that although the SA signaling pathway could be activated by SMV infection, the activation was insufficient to confer resistance to virulent strains of SMV.

Figure 3
figure 3

Time-course analysis of ABA and SA levels in Rsv3-mediated ER.

Endogenous ABA (A) and SA (B) levels in soybean leaves inoculated with mock, SMV-G5H, or SMV-G7H were quantified at 8, 24 and 54 hpi. Values are the means ± SD from three independent experiments.

Full size image

Synchronized overexpression of GmPP2C3a inhibits virus cell-to-cell movement

We previously developed a useful viral vector system by engineering SMV-G7H to express foreign genes in soybean and by tagging the SMV vector with a GUS gene; using this system, we were able to visualize virus infection and movement in plant cells27,28. The same strategy was used to tag SMV-G5H with GUS (Fig. 4A). We hypothesized that if GmPP2C3a is a positive regulator of Rsv3-mediated ER, its ectopic overexpression would confer ER even to a virulent strain of SMV. Testing this hypothesis required that GmPP2C3a be overexpressed only in the cells infected with SMV. Thus, we used the SMV-G7H vector for the synchronized overexpression of GmPP2C3a and we named the resulting construct that was engineered to overexpress GmPP2C3a as pSMV-G7H-GmPP2C3a (Fig. 4).

Figure 4
figure 4

Schematic representation of the SMV infectious constructs.

The gene of interest was inserted between either P1 and HC-Pro or between NIb and CP to be generated by proteolysis when the virus replicates. In vivo transcription of the SMV infectious constructs is under the control of a double 35S promoter (P35SX2), a cis-cleaving ribozyme sequence (Rz) and a NOS terminator (NOSt).

Full size image

We first examined the infectivity of pSMV-G7H-GmPP2C3a. Plasmid DNAs of pSMV-G7H-GmPP2C3a and other controls were rub-inoculated onto the first trifoliolate leaves of soybean line L29 (Rsv3). While all of the soybean plants inoculated with pSMV-G7H and pSMV-G7H-GUS were systemically infected (Table 1), no infection of the soybean plants inoculated with pSMV-G5H or pSMV-G7H-GmPP2C3a was evident even at 25 days post inoculation (dpi), as determined by subjecting the total RNA extracted from upper non-inoculated leaves to RT-PCR specific for SMV (Table 1). The results suggested that the synchronized overexpression of GmPP2C3 might induce resistance to SMV-G7H even in the absence of Rsv3-dependent activation. However, we could not totally exclude the possibility that the presence of GmPP2C3a DNA sequence in the viral vector genome is responsible for the impaired infectivity, although foreign gene expression by proteolytic processing from the SMV polyprotein precursor is a promising strategy. Therefore, we performed additional experiments as described below.

Table 1 Effect of the synchronized expression of GmPP2C3a on SMV infectivity in three independent experiments
Full size table

The most striking feature of ER is the rapid arrest in an HR-independent manner of virus accumulation in the initially infected cell4. Thus, we next examined the phenotypic characteristics of Rsv3-mediated ER on virus accumulation at the microscopic level. To this end, we inoculated L29 (Rsv3) with pSMV-G5H-GUS and pSMV-G7H-GUS and performed a histochemical GUS assay at 5 dpi to visualize virus accumulation and cell-to-cell movement because only infected cells express GUS activity as a reporter of virus accumulation28. SMV-G7H-GUS replicated actively and spread to form infection foci > 300 μm in diameter (Fig. 5A and Table S1). In contrast, accumulation of SMV-G5H-GUS was restricted to the initially inoculated single cells (Fig. 5A and Table S1). However, Rsv3-mediated ER did not seem to inhibit virus replication in the initially infected cells because high levels of GUS activity produced by replication of SMV-G5H-GUS were detectable (Fig. 5A).

Figure 5
figure 5

GmPP2C3a restricts virus accumulation and enhances callose deposition in Rsv3-mediated ER.

In situ detection of GUS activity (A) and callose deposition (B) in soybean leaves. The inocula are indicated above each image. Red arrows indicate accumulation of callose stained with aniline blue. The GUS foci and callose deposition were photographed at 5 dpi. (C) The appearance of HR-like necrotic lesions on the leaves inoculated with SMV-G5H and SMV-G7H-GmPP2C3a when treated with DDG. The enlargements of the boxed areas are shown below each image. Images were captured at 8 dpi.

Full size image

We next determined whether the synchronized overexpression of GmPP2Ca results in restriction of virus accumulation in the initially infected cells as observed above. Thus, we engineered the GmPP2C3a coding sequence into the polyprotein ORF between the NIb and CP cistrons of pSMV-G7H-GUS to produce mature GmPP2C3a by proteolysis (Fig. 4). The resulting construct, named pSMV-G7H-GUS-GmPP2C3a, could produce both GUS and GmPP2C3a upon virus replication. Leaves of L29 (Rsv3) were inoculated with pSMV-G7H-GUS-GmPP2C3a and subjected to a GUS assay. As expected, virus accumulation was restricted to the initially inoculated single cell and obvious detection of the GUS expression confirmed that the insertion of GmPP2C3a in the viral genome did not impair virus replicative ability (Fig. 5A and Table S1). This result suggested that ectopic overexpression of GmPP2C3a could activate ER in the absence of effector recognition by Rsv3.

Callose deposition is critical for Rsv3-mediated ER

Cell-to-cell movement of a plant virus through plasmodesmata (PD) is a prerequisite for systemic infection29,30. Callose deposition is known to decrease the permeability of PD and therefore virus movement through PD31,32,33. Thus, we determined whether Rsv3-mediated ER is accompanied by callose deposition at PD. Leaves of L29 (Rsv3) that had been inoculated with pSMV-G7H-GUS, pSMV-G5H-GUS, or pSMV-G7H-GUS-GmPP2C3a were stained with aniline blue and then examined microscopically at 5 dpi. Interestingly, significant quantities of callose accumulated along the plasma membrane of the initially inoculated single cells in the leaves inoculated with pSMV-G5H-GUS and pSMV-G7H-GUS-GmPP2C3a but not in leaves inoculated with mock or pSMV-G7H-GUS (Fig. 5B).

To further confirm whether the callose deposition detected by microscopic observation was relevant for Rsv3-mediated ER, we performed inoculation experiments in L29 (Rsv3) and included treatments of 500 μM 2-deoxy-D-Glc (DDG), which is a callose synthesis inhibitor34. DDG was applied two-times: first at 24 hours before inoculation and then at 24 hours after inoculation. Interestingly, inoculation of pSMV-G5H and pSMV-G7H-GMPP2C3a resulted in HR-like necrotic lesions on the leaves treated with DDG but not on those treated with H2O, while no visible lesions were detected on the leaves inoculated with pSMV-G7H and other controls (Fig. 5C). In addition, a significant accumulation of viral RNA was demonstrated by RT-PCR of leaf tissues with HR-like necrotic lesions (data not shown). The results indicated that inhibition of callose deposition resulted in the failure to arrest viruses at the initially inoculated sites even though Rsv3-mediated defense signaling was activated. Furthermore, our results suggest that Rsv3-mediated signaling has the potential to induce a cell death response although it seems that ER is epistatic to the cell death response in Rsv3-mediated resistance. This further suggested that GmPP2C3a induction could be associated with both ER and HR as a general feature of ETI in soybean.

Discussion

ETI is a rapid and strong immune response activated by interactions between host R proteins and their cognate effectors1,2. Although HR, a form of programmed cell death, is a common manifestation of ETI, various R genes mediate HR-independent processes that confer resistance against viruses; these HR-independent responses are referred to as extreme resistance or ER4,17,35. While the HR signaling pathway has been extensively studied in various host-pathogen systems1,2, very limited information is available on the signaling pathway involved in ER.

The soybean-SMV pathosystem is a useful model for studying molecular interactions between host and viruses, because the genotype-specific disease responses are well characterized for various SMV strains17,35,36,37. In this study, we used the soybean (Rsv3)-SMV (CI) pathosystem to gain insight into the molecular signaling pathway involved in ER. Genome-wide transcriptome analysis enabled the identification of a subset of PP2C genes that are significantly induced in Rsv3-mediated ER. The most significantly expressed soybean PP2C gene, GmPP2C3a, was selected for in depth study to further characterize its functional involvement in Rsv3-mediated ER.

A major mechanism of signal transduction is reversible phosphorylation of protein components in the particular pathway21. PP2Cs form the largest protein phosphatase family in plants, a family that is also highly complex. In eukaryotes, PP2Cs act as negative regulators of stress-activated protein kinase cascades21,22. In plants, substantial research has shown that PP2Cs function as negative regulators of ABA signaling21,22. ABI1 and ABI2 are the best-characterized plant PP2Cs and are partially redundant in their repression of ABA responses24,25. ABI1 and ABI2 function as key regulators of ABA signaling by being responsible for nearly 50% of the ABA-induced PP2C activity24. However, this also indicates that other PP2Cs are involved in the regulation of the integrative network of ABA signaling. Indeed, accumulating evidence shows that PP2Cs are involved in regulation of plant immunity as downstream factors of ABA signaling38,39,40. In this study, we identified a subset of the PP2C gene family that is significantly up-regulated soon after inoculation with the SMV avirulent strain G5H (Fig. 2 and S1). However, this PP2C gene subset did not include the soybean orthologs of ABI1 and ABI2 (PP2Cs that are up-regulated by ABA in Arabidopsis), although all of these PP2Cs are closely related according to phylogenetic analysis (Fig. S1). Our results suggest that the up-regulated PP2C gene subset may be specifically involved in Rsv3-mediated ER in an ABA-dependent manner. This further suggests that the PP2C activity is regulated by protein expression level and that PP2Cs finely modulate certain signal pathways of ABA signaling through their diversity and specificity to substrates and other signaling components.

Rsv3-mediated ER was accompanied by ABA accumulation (Fig. 3A). In Arabidopsis, some of the PP2Cs, including ABI1, ABI2 and AtPP2CHA, are transcriptionally up-regulated by ABA21. In Rsv3-mediated ER, however and as noted earlier, elevated ABA triggered up-regulation of a certain subset of soybean PP2Cs including GmPP2C3a but not the orthologs of ABI1 and ABI2 (Fig. 2 and S1). Because the effect of ABA on plant defense was shown to be ambivalent, researchers have proposed various modes of action of ABA on disease resistance, including antagonistic suppression of SA- and ET/JA-mediated basal defense, synergistic stimulation of JA signaling, induction of stomatal closure and priming for callose deposition19,20,41. Thus, it is likely that the complex role of ABA in defense responses is modulated by divergent signaling pathways and that phase-specific activation of certain PP2Cs determines the modes of action of ABA-dependent disease responses.

As mentioned above, one of the positive effects of ABA on plant immunity is its ability to trigger callose deposition32. Callose can serve as a matrix in which defense molecules are deposited and its deposition and degradation at the neck region of plasmodesmata (Pd) is one of the regulatory mechanisms controlling Pd permeability32,33. Successful infection of host plants by viruses requires cell-to-cell movement through Pd and long-distance movement through the vascular tissue42. In this regard, callose deposition at PD can serve as an antiviral defense response by forming a physical barrier, thereby preventing virus cell-to-cell movement and restricting virus accumulation at the initially infected cells33. Our results showed that Rsv3-mediated ER is associated with callose deposition and that ectopic overexpression of GmPP2C3a could activate callose deposition (Fig. 5), indicating that GmPP2C3a is a positive regulator of callose deposition. Accumulation of callose is primarily controlled by the balance of activities of callose synthases and the β-1,3-glucanases that degrade callose33. Thus, GmPP2C3a may function upstream of callose synthases and/or β-1,3-glucanases in regulation of callose accumulation.

Based on our findings, we suggest a hypothetical model for Rsv3-mediated ER against SMV (Fig. 6). In the initially infected cells, SMV replicates and produces the effector CI. The recognition of CI by the R protein Rsv3 results in accumulation of ABA and thus activation of the ABA signaling pathway. In the ABA signaling pathway linked to Rsv3-mediated ER, the transcriptionally up-regulated PP2C genes, including GmPP2C3a, function as positive regulators of the signaling to stimulate callose deposition. Eventually, callose deposition at PD inhibits virus cell-to-cell movement and restricts virus accumulation to the initially infected cells.

Figure 6
figure 6

A model for Rsv3-mediated ER against SMV.

Full size image

ETI is the outcome of dynamic interactions between plants and pathogens1,2,3. Various factors including interaction affinity between an R protein and its effector, levels of interacting components and timing may affect the fine tuning of the defense signaling network1,2,3. Thus, it is important to find and characterize key regulators that decide the fate of resistance interactions. Based on genome-wide transcriptome analysis, we have identified a soybean PP2C gene, GmPP2C3a, that is a key regulator of Rsv3-mediated ER. To further dissect Rsv3-mediated ER, researchers will need to isolate GmPP2C3a-interacting proteins and signaling components that link GmPP2C3a to callose deposition.

Methods

Plants, virus sources and inoculation

Soybean line L29 (Rsv3) was grown in a growth chamber at 25°C under a 16/8-h photoperiod. Soybean seedlings were selected for inoculation when the first trifoliolate leaves were fully expanded. Infectious cDNA clones of SMV-G7H (pSMV-G7H)28 and SMV-G5H (pSMV-G5H)43 were used as the virus sources. SMV G7H and G5H strains were propagated in the soybean cultivar Lee 74 (rsv) and were used for sap inoculation. The saps and cDNA plasmids of infectious SMV constructs were mechanically inoculated as described previously28. Where indicated, soybean leaves were treated with either 500 μM DDG or H2O two times: at 24 hours before inoculation and at 24 hours after inoculation.

Transcriptome analysis

Inoculation experiments described in Figure 1 were carried out independently three times, with each treatment being represented by at least 10 plants in total. In this study, one of the primary goals of transcriptome analysis was to identify candidate genes involved in the plant immune responses against viral infection. Although it is important to use biological replication to obtain robust conclusions about biological differences among samples, the cost of RNA-Seq is high and our analysis was therefore limited to a single biological sample of pooled tissues. Total RNA was extracted from the inoculated leaves and pooled in equal quantities for each experimental set. RNA libraries were generated using the Illumina TruSeq RNA sample prep kit (Illumina, Inc., USA) with no modifications to the standard protocol. Sequencing on an Illumina HiSeq2000 sequencer (Illumina, Inc., USA) was performed by Macrogen Inc. (Seoul, South Korea). Raw sequence reads were filtered by the Illumina pipeline and mapped to reference sequences (Phytozome v9.1, http://phytozome.net/) of the soybean genome (Glycine max). To identify differentially expressed genes and obtain expression values, statistical comparison between the libraries was performed by calculating RPKM after normalization using the NGS Cell V 4.06 program with default parameters (CLC Bio, Denmark). Heat maps were generated and visualized using Genesis44. To validate RNA-seq data, real-time PCR was performed using ACT11 as a reference gene as described previously45.

Measurement of ABA and SA Levels

Endogenous ABA and SA were extracted from soybean leaves and then quantified with the Phytodetek ABA Test Kit (Agdia, USA) for ABA and with reversed-phase HPLC (Waters, USA) for SA as described previously40.

Engineering of the SMV viral constructs

The coding region of the GmPP2C3a gene was amplified by PCR using a primer pair harboring XbaI sites (5′- TGCTCTAGAATGTTCAGTTCGTTGCATGCAAAA-3′ and 5′-TGCTCTAGAATCAGTTGTTACTGATCCTCTTAG-3′). The amplified fragments were digested with XbaI and cloned into pSMV-MCS46, which was opened with XbaI. The resulting construct with the GmPP2C3a insert in the correct orientation was named pSMV-G7H-GmPP2C3a. The same strategy was used to construct SMV-G5H-based viral vector carrying the cloning sites (XbaI and MluI) between P1 and HC-Pro as described previously28 and the amplified GUS gene was inserted into the XbaI site to generate pSMV-G5H-GUS as described previously27. For simultaneous expression of GUS and GmPP2C3a, an additional NIa-Pro cleavage site and the cloning sites (SalI and SnaBI) were engineered into the polyprotein ORF between the Nib and CP cistrons of pSMV-G7H-GUS as described previously47; the resulting construct was named pSMV-G7H-GUS-MCS. The amplified GmPP2C fragment was inframe inserted into the SalI site of pSMV-G7H-GUS-MCS and the resulting construct, which is capable of simultaneously expressing GUS and GmPP2C3a, was named pSMV-G7H-GUS-GmPP2C3a.

Detection of GUS activity and callose deposition

GUS expression driven by virus infection was examined at 5 dpi by histochemical GUS assays as described previously27. Detection of callose deposition was performed at 5 dpi by the aniline blue staining method as described previously48.