Transcriptome-based discovery of genes and networks related to R_SC3Q-mediated resistance to Soybean mosaic virus in soybean

Keywords: CML, Glycine max, NILs, plant-hormone signal-transduction pathway, Soybean mosaic virus.

Soybean mosaic virus (SMV) is a member of the largest genus of known plant viruses, Potyvirus (Adams et al. 2005). SMV is a worldwide disease of soybean (Glycine max (L.) Merr.) and can cause serious reduction in yield and seed quality. Based on their differential responses on resistant and susceptible soybean lines, SMV isolates have been grouped into seven strains (G1–G7) in the United States (Cho and Goodman 1979; Cho and Goodman1982), five strains (A–E) in Japan (Takahashi et al. 1980; Nakano 1982), and 22 strains in China (Li et al. 2010; Wang et al. 2018). Soybean has a two-layered innate immune system for combatting SMV: pathogen-associated molecular pattern-triggered immunity (PTI), and effector-triggered immunity (ETI) (Jones and Dangl 2006). ETI is induced when a strain-specific avirulent (Avr) protein from the pathogen associates directly or indirectly with a cognate plant resistance (R) protein (Jones and Dangl 2006). Currently, soybean resistance genes to SMV, Rsv1, Rsv3, Rsv4 and Rsv5, are identified as single dominant SMV resistance gene (R) loci in the USA (Yu et al. 1994; Hayes et al. 2000; Jeong et al. 2002; Klepadlo et al. 2017). Resistance to the strains from China is derived from the single dominant SMV Rsc loci, and these have been mapped to chromosomes 2, 13, 14, and 6 in the respective cultivars Kefeng-1, Qihuang-1, Dabaima and RN-9 (Ma et al. 2011; Wang et al. 2011; Zheng et al. 2014; Rui et al. 2017). To date, only the Rsv4 gene has been cloned (Ishibashi et al. 2019). Some downstream signalling components and pathways of Rsv-mediated resistance have been reported; however, there are no reports about Rsc-mediated resistance (Fu et al. 2009; Zhang et al. 2012; Wang et al. 2014).

Plant-hormone signalling pathways also play a crucial role in the process of the plant defence response, involving hormones such as salicylic acid (SA), jasmonic acid (JA) and abscisic acid (ABA). SA, a phenolic compound, is usually required for triggering innate immune responses (i.e. PTI and ETI). Localised resistance responses of plants, activation of programmed cell death, systemic acquired resistance (SAR) and small interfering RNA (siRNA) antiviral machinery are associated with SA and/or SA derivatives (Alamillo et al. 2006; Hunter et al. 2013; Baebler et al. 2014; Shigenaga and Cristiana 2016). JA is a lipid-derived signalling molecule. In the JA signal-transduction process, JASMONATE ZIM-DOMAIN (JAZ) proteins, as a subfamily of TIFY, are key regulators in JA hormonal response (Vanholme et al. 2007). Upon stress perception, plants accumulate the bioactive JA-Ile molecule ((+)-7-iso-JA-Ile), which induces the interaction between F-box CORONATINE INSENSITIVE 1 (COI1) and JAZ (Fonseca et al. 2009; Sheard et al. 2010). The JA-Ile-mediated COI1–JAZ interaction leads to the ubiquitination and subsequent degradation of the JAZ proteins. Several transcription factors including the JA master regulator MYC2 are liberated, which in turn induce JA-specific cellular reprogramming, such as defence responses (Chini et al. 2007; Kazan and Manners 2013). JA reportedly has positive and negative effects on defence against viruses (Alazem et al. 2018); at the early stage of virus infection, JA seems to support plant defence, but it decreases plant resistance at a later stage when it is induced or applied (Pacheco et al. 2012; Garcia-Marcos et al. 2013). ABA is a sesquiterpene compound produced by the cleavage of γ-carotene. When ABA receptors PYR/PYL/RCAR bind accumulated ABA, dephosphorylation of SnRK2s (SNF1-related protein kinase 2 family) mediated by protein phosphatase 2C (PP2C) is prevented. The active SnRK2 kinases phosphorylate and activate downstream transcription factors, and then induce the transcription of ABA‐responsive genes for developmental process and adaptive stress responses (Hauser et al. 2011; Alazem et al. 2018). During plant defence, the role of ABA depends on the stage of virus infection. ABA can resist viruses by mediating stomata closure or increasing callose deposition on plasmodesmata to restrict movement of viruses at early stages of infection. However, ABA can suppress hypersensitive response to decrease the production of reactive oxygen species (ROS) and SA, and weak distal SAR and siRNA systems (Alazem et al. 2018).

Qihuang-1 is an important resistance source for SMV in China. It carries a resistance gene (R_SC3Q) to SMV strain SC3, a prevalent SMV strain in Huang, Huai and Chang Jiang Valleys in China (Wang et al. 2003). However, the mechanism underlying SMV resistance in Qihuang-1 is not clear. Next-generation sequencing has become the first choice for researching SMV resistance mechanisms, with a reduction in expense and time needed in sequencing. In this study, we applied transcriptome analysis of a pair of near-isogenic lines (NILs) from the cross Qihuang-1 (resistant to SC3) × Nannong 1138-2 (susceptible to SC3) in order to explore the transcript-accumulation patterns after SMV infection. This enabled us to discover genes and networks in Qihuang-1 that are regulated by R_SC3Q-mediated resistance.

Near-isogenic lines (R and S lines) that have similar genetic background were derived from the cross Qihuang-1 × Nannong 1138-2 in a previous study (Zheng et al. 2014). The R line carries a resistance gene to SC3 (R_SC3Q), like Qihuang-1, whereas S line carries a susceptible gene (r_SC3Q), like Nannong 1138–2.

Qihuang-1, Nannong 1138-2, the R line and the S line were planted in a mixture of soil, perlite and vermiculite (volume ratio 3 : 1 : 1). Plants were grown in a greenhouse with temperatures of 25°C (day) and 20°C (night). Plants were inoculated with the SMV strain SC3 or phosphatic buffer solution (control) when the unifoliolate leaves were fully unfolded (V1 stage). The inoculation method was according to the previous study of Li et al. (2017).

The RNA-Seq data from the NILs (R line and S line) were obtained previously (Li et al. 2017; Yuan et al. 2020). In brief, leaves were collected as samples from the two lines independently at 0, 6, 20 and 48 h post-inoculation (hpi) with SC3. After extracting total RNA from the samples, libraries were generated from eight samples with Ultra (New England Biolabs, Ipswich, MA, USA), following the manufacturer’s instructions, and samples were sequenced on a HiSeq 2000 (Illumina, San Diego, CA, USA). After removing impure data from the raw sequence (using FASTX-Toolkit version 0.0.13; Gordon and Hannon 2010), clean reads were obtained. The clean reads were aligned to the Williams 82 soybean mRNA reference and genome Glyma.Wm82.a1.v1 (Schmutz et al. 2010) by using SOAP2 version 2.21 with default settings (Li et al. 2009). Differentially expressed genes (DEGs) (|log₂(fold-change)| ≥2 and P ≤ 0.01) were obtained by comparison of SMV-infected samples (at 6, 20 and 48 hpi) with 0 hpi in the R line and the S line, using Cufflinks v1.1.0 software (Trapnell et al. 2010). All of the original RNA-sequencing data have been submitted to the Sequence Read Archive database (SRA accession no. PRJNA668549, www.ncbi.nlm.nih.gov/sra; BioSample accession no. SAMN16414956, www.ncbi.nlm.nih.gov/biosample).

Soybean pathway enrichment analysis of the DEGs was performed by using EXPath 2.0 (Zheng et al. 2017). A heatmap graph was drawn by Origin 2019b (www.originlab.com/2019b) based on log₂(fold-change) values.

Extraction of RNA from leaf tissues of soybean plants and quantitative real-time polymerase chain reaction (qPCR) were performed according to the previous study (Zheng et al. 2014; Luan et al. 2016). Each gene was tested with at least three biological replicates and the experiment was repeated at least twice. Gene IDs and primer sequences are provided in Supplementary Material Table S1 (available at the journal’s website).

In terms of SMV strain SC3 content in R and S lines after SC3 inoculation (Fig. 1a), there was a continuous decrease in the R line after 48 hpi, and virus could hardly be detected at 5 days post-infection (dpi). The SMV level in S line showed a trend of fluctuation. A small amount of virus was detected in inoculated leaves in the S line, but an abundance of virus in uninoculated distal leaves (the first trifoliate leaves at 7 dpi) (Fig. 1a). This suggested that the S line without R_SC3Q also had some resistance to SC3; however, this resistance did not stop the movement of the virus. Some virus moved along the veins and stems to the distal leaves and replicated, which was detected at 7 dpi in the S line (Fig. 1a).

Fig. 1.  (a) Content of virus in soybean after inoculation with Soybean mosaic virus (SMV) strain SC3 of seedlings of R and S lines; and (b–d) Venn diagrams illustrating the clustering of differentially expressed genes (DEGs). Samples were collected at 0, 6, 24, 48 h post inoculation (hpi) and at 3 and 5 days post-inoculation (dpi) from inoculated leaves (local leaves), and at 7 dpi from the first trifoliate leaves (distal leaves). SMV content was determined by quantitative real-time PCR analysis using SMV-specific primers, with soybean tubulin transcript level as an internal control; **P < 0.01 (t-test). DEGs (fold-change ≥2 or ≤0.5, and P ≤ 0.01) were obtained by comparison of SMV-infected sample at 6, 20 and 48 hpi with 0 hpi in R and S lines. Diagrams show number of DEGs with upregulated (↑) and downregulated (↓) expression from both lines. Boxed 1 indicates that one DEG had the opposite expression pattern in R and S lines.

The eight samples from the R and S lines at four time-points subjected to Illumina sequencing generated 43–67 million clean reads (Table S2), 79.72–92.57% of which were mapped to the soybean reference genome. Around 40 000 genes were found to be expressed in each sample (Table S2).

The reliability of transcriptome analysis was tested via qPCR analysis on eight randomly selected genes. The results showed that seven of the genes had similar expression trends in the transcriptome and qPCR analysis, which indicated that transcriptome data were reliable (Fig. S1).

In total, 779, 189 and 84 DEGs were obtained at three time-points (6, 20, and 48 hpi, respectively) in the R line, and 390, 162 and 65 DEGs were discovered at the same three time-points in the S line (Fig. 1b–d). The number of DEGs was greater in the R line than the S line at each time-point. Comparable numbers of DEGs with upregulated and downregulated expression were apparent; in the R line, more were upregulated than downregulated at three time-points, but in the S line, more were downregulated (Fig. 1b–d). Most of the DEGs were mobilised in response to SC3 infection before 20 hpi in both lines, which was likely an important aspect of soybean defence.

Pathway enrichment analyses by EXPath 2.0 showed that, whether in the R or S line, DEGs involved in metabolic-related pathways represented a large proportion at the three time-points (6, 20 and 48 hpi; Fig. S2). Among the different pathways, plant–pathogen interaction and plant-hormone signal-transduction pathways play important roles in plant defence.

Three CML genes (Glyma03g28650, Glyma19g31395 and Glyma11g33790) with downregulated expression were found in the plant–pathogen interaction pathway in the S line only. CMLs are the members of the Ca²⁺ sensors, which interact with Ca²⁺ and regulate the function of diverse target proteins by direct binding or through phosphorylation (Aldon et al. 2018). CMLs are reported to be important regulators of plant defence against pathogens (Xu et al. 2017; Zhu et al. 2017; Lu et al. 2018). Therefore, we speculated that the three identified CML genes in the S line at 6 hpi probably positively regulated the early resistance response of soybean to SC3. Further, qPCR analysis was performed on the parent plants of R and S lines (Qihuang-1 and Nannong 1138-2). As shown in Fig. 2, expression of the three CML genes increased in Qihuang-1 but decreased significantly in Nannong 1138-2 at 6 hpi, indicating that inhibition of expression of the three CML genes was very likely to promote SMV infection.

Fig. 2.  Detected expression of genes for calmodulin-like protein (CML) by quantitative real-time PCR, at 0, 6, 24, and 48 h after inoculation with Soybean mosaic virus strain SC3 of Qihuang-1 (QH1, resistant to SC3) and Nannong 1138-2 (NN1138-2, susceptible to SC3). *P < 0.05, **P < 0.01 (t-test).

According to the pathway enrichment analyses, 14 DEGs were found in the plant-hormone signal-transduction pathway (Fig. 3). Glyma01g41290 and Glyma11g04130 encoding TIFY/JAZ proteins had downregulated expression in the S line, and upregulated in R line (Fig. 3), indicating that induction of Glyma01g41290 and Glyma11g04130 expression is advantageous for soybean resistance during early stages of SC3 infection. TIFY/JAZ proteins are a key negative regulator in JA hormonal response (Fonseca et al. 2009). This means that JA probably plays a role in response to SC3 infection, and repression of JA probably enhances soybean resistance to SC3 during early stages of SC3 infection.

Fig. 3.  Heatmap of differentially expressed genes (DEGs) in plant hormone, MAPK and autophagy signal pathways at different times after inoculation. DEGs (|log₂(fold-change)| ≥2, and P ≤ 0.01) were obtained by comparison of SMV-infected sample at 6, 20, and 48 h post inoculation (hpi) with 0 hpi in R and S lines. Hierarchical cluster heatmap was drawn by Origin 2019b software, and log₂(fold-change) values for these DEGs are indicated in different colours.

Two DEGs (Glyma14g06100 and Glyma14g32430) were related to the ABA-signalling pathway. Glyma14g06100 encodes a PYR1-like protein (PYL), which is an ABA receptor and positively responds to ABA regulation (Hauser et al. 2011). Its expression was significantly downregulated in the S line. Glyma14g32430, as an ABA-induced PP2C3a (encoding a subset of the Type 2C protein phosphatase), is reported to be a key positive regulator of Rsv3-mediated extreme resistance against SMV strain G5H in soybean (Seo et al. 2014). In our study, PP2C3a had significantly upregulated expression at 6 hpi in the R line, but it had almost had no expression in the S line (Fig. 3). These results suggest that ABA was involved in plant response after SC3 infection and probably promoted R_SC3-mediated resistance.

The auxin (IAA)-related DEGs with different expression patterns were mostly in the plant-hormone signal-transduction pathway (Fig. 3). This shows that the expression of DEGs was greatly affected by SMV infection at early time-points in the R line, but not in the S line. We speculate that growth and development were greatly affected through R_SC3Q-mediated resistance responses during the early stages of SC3 infection. There was one DEG in each of the ethylene (ET) and gibberellin (GA) signalling pathways, encoding ET receptor EIN4 (ETHYLENE INSENSITIVE4) and GA receptor GID1B (GIBBERELLIN INSENSTIVE DWARF1), respectively (Fig. 3). This indicated that ET and GA participated in response to SC3 inoculation.

Mitogen-activated protein kinase (MAPK) signalling pathway plays a critical role in plant immunity. Five DEGs were identified in the MAPK signalling pathway at 6 hpi in the R and S lines (Fig. 3). We speculated that these DEGs were probably regulated by R_SC3Q-mediated resistance. Three of them (Glyma05g05540, Glyma14g32430 (PP2C3a) and Glyma19g42220) had significantly upregulated expression at 6 hpi in the R line, but they had almost no expression in the S line (Fig. 3). We speculate that these three DEGs maybe more important in R_SC3Q-mediated resistance. Autophagy responses have also been reported in plant defences (Alazem et al. 2018). Only one DEG was involved in the autophagy pathway (Glyma01g32400), which had downregulated expression in both R and S lines.

Qihuang-1 (R_SC3Q) is an important resistance source for SMV strain SC3 in China, but the published literature includes little information on R_SC3Q-regulated resistance mechanisms or pathways. Our findings provide some R_SCQ-mediated, resistance-related genes and a general understanding of transcriptional activation of defence-related genes. Immediately downstream of the initial elicitor–receptor recognition, the activation of ion fluxes (including Ca²⁺) and the production of H₂O₂ are the initial responses detected in plant cells, which occur before transcriptional activation of defence-related genes (Ramos et al. 2008). In response to the stimuli, cytosolic free Ca²⁺ rises and binds to a plethora of sensors, including CML, which in turn activate subsequent reactions of plant immunity (Aldon et al. 2018). In our study, three CML genes (Glyma03g28650, Glyma19g31395 and Glyma11g33790) were enriched in the plant–pathogen interaction pathway in the S line only. Their expression was downregulated in the S line and the susceptible parent Nannong 1138-2 at 6 and 24 hpi, but was upregulated in resistant parent Qihuang-1 (Fig. 2). We consider that these three CML genes are beneficial for R_SC3Q-mediated resistance. The functions of CML genes are diverse. CML8 and CML9 can promote Arabidopsis resistance against the phytopathogenic bacteria Pseudomonas syringae pv. tomato (strain DC3000) (Zhu et al. 2017). Disease-resistance pathways involving CML8 and CML9 are different. CML8 plays a role in SA-dependent processes, probably by modulating the effect of bacterial effectors. CML9 contributes to PTI (via a flagellin-dependent pathway) but also to SA-dependent processes (Zhu et al. 2017). CML9 as a negative regulator is involved in drought and salt stress. Arabidopsis CML42 and CML37 are also involved in plant defence and drought response (Vadassery et al. 2012a, 2012b; Scholz et al. 2014). Therefore, the functions of the three CML genes in this study can be investigated jointly from the aspects of plant defence and abiotic stress in subsequent experiments. These three genes might have different functions.

Some DEGs identified had involvement in plant-hormone signal transductions (Fig. 3). Plants defend against pathogen attack by modulating plant-hormone signalling pathways, such as those involving SA, JA, ET and ABA (Robert-Seilaniantz et al. 2011; Pieterse et al. 2012; Vos et al. 2013).

The role of JA in plant defence against viruses is controversial (Alazem and Lin 2015). JA has been known for its positive roles in a few compatible interactions; for example, the silencing of JA receptor gene COI1 can enhance virulence of Potato virus X and Potato virus Y in Nicotiana benthamiana (Garcia-Marcos et al. 2013). However, for incompatible interactions between Tobacco mosaic virus (TMV) and N. benthamiana, JA had negative roles. Oka et al. (2013) found that N-mediated resistance to TMV was enhanced in NtCOI1-RNAi line. There was an incompatible interaction between the R line and SC3, and a compatible interaction between the S line and SC3. Two DEGs (Glyma01g41290 and Glyma11g04130) encoding TIFY/JAZ, a negative regulator in JA pathway, had significantly downregulated expression in the R line and upregulated in S line following SC3 infection (Fig. 3). Therefore, we speculate that JA probably had a positive role in R_SC3Q-mediated resistance during the early stages of infection, which was similar to previously reported findings as mentioned above (Oka et al. 2013; Alazem et al. 2018).

Xun et al. (2019) found that ABA might be related to R-mediated resistance to SMV in soybean. Overexpression of a candidate R gene against SMV in soybean increased ABA accumulation. In the present study, we found that ABA-related genes PYL (Glyma14g06100) and PP2C3a (Glyma14g32430) were differentially expressed. As shown in Fig. 3, PP2C3a was rapidly induced in the R line at 6 hpi, but not in the S line. PP2C3a, encoding a Type-2C protein phosphatase, was predicted to participate not only in the ABA signalling pathway, but also in the MAPK signalling pathway (Fig. 3). When plants are invaded by pathogens, callose deposition increases on plasmodesmata and restricts cell-to-cell movement of viruses (Nakashima et al. 2003). In this process, the MAPK signalling pathway plays a positive regulatory role. Inhibition of the MAPK signalling pathway will decrease callose deposition and reduce resistance of plants (Xu et al. 2019). It has been verified that PP2C3a is specifically involved in Rsv3-mediated extreme resistance in an ABA-dependent manner against SMV (strain G5H) in soybean. It controls the rapid accumulation of callose at the points of G5H infection to stop virus spread (Seo et al. 2014; Alazem et al. 2019). All of the above evidence indicates that ABA plays an important role in R-mediated resistance against SMV and that PP2C3a might be a key gene connecting the MAPK and ABA signalling pathways. The role of PP2C3a in two pathways needs further verification.

In plant defence against viruses, SA signalling constitutes the major defensive pathway, and is tightly connected to the majority of R genes (Alazem and Lin 2015). Although SMV multiplication was reported to be inhibited by SA treatment (Zhao et al. 2018), no SA-related DEGs were detected in our study. Perhaps SA did not act as a main regulator of R_SC3Q-mediated resistance during the early stages of SC3 infection.

In addition to PP2C3a, Glyma19g42220 (encoding respiratory burst oxidase homologue B, RBOHB) was significantly upregulated at 6 and 20 hpi in the MAPK signalling pathway in the R line. RBOH is also called NADPH oxidase (NOX), a key enzyme of ROS generation, and playing vital roles in various biological processes including plant immunity. The important role of NOX/RBOH in plant immunity has been well reported, especially in Arabidopsis, tobacco and rice (Hu et al. 2020). However, there is little research into the NOX/RBOH–MAPK pathway in plant immunity. In N. benthamiana, NbRBOHB is an important player in both the PTI ROS burst and the ETI ROS burst. It is noteworthy that MAPK is responsible for the ETI ROS burst by transactivation of NbRbohB, but not for the PTI ROS burst (Yoshioka et al.2016). We assume that Glyma19 g42220 (RBOHB) has a similar function to NbRBOHB, which participates in the ETI ROS burst (the R-mediated resistance) through the MAPK cascade-signalling pathway.

In summary, CML, JA, ABA and MAPK, but not SA, are involved in the plant defence response during the early stages of SC3 infection. Suppression of JA or induction of ABA probably increased R_SC3Q-mediated SMV resistance. Soybean resistance against SMV was mediated by complex gene families at different loci, and the resistance was specified via both SMV strain and soybean variety. Our understanding of R_SC3Q-mediated resistance against SC3 would be enhanced by identification of factors associated with R_SC3Q that help to trigger the resistance response.

We declare no conflict of interest.

HZ, Y Yuan and Y Yang conceived and designed the experiments; Y Yuan wrote the article and performed major parts of the experiments; JL modified the language and provided technical support for RNA-Seq; LW performed parts of RNA-Seq experiment; YS and BL performed parts of plants cultivation and qRT experiment; HZ re-edited the article.