Several genera of plants can benefit from atmospheric nitrogen fixation through symbiosis with particular microorganisms. The most widespread terrestrial mutualistic symbiosis is that between plants and arbuscular mycorrhizal fungi. The uptake of water and mineral nutrients (particularly phosphate) from soil by the fungal partner promotes the growth and disease resistance of the host plant (Parniske, 2008; Bonfante and Genre, 2010). A more typical symbiosis is that between most species of legume plants and N₂-fixing bacteria, such as Aminobacter, Azorhizobium, Bradyrhizobium, Devosia, Mesorhizobium, Methylobacterium, Microvirga, Ochrobactrum, Phyllobacterium, Rhizobium, Sinorhizobium, Burkholderia, Cupriavidus, and Herbaspirillum (collectively called rhizobia) (Laranjo et al., 2014). The legume-rhizobium symbiosis results in the formation of a completely new organ, the root nodule, where the rhizobia are intracellularly hosted and fix atmospheric nitrogen into ammonia that can be assimilated by the plant (Oldroyd and Downie, 2008). There is a high specificity between a host legume species and its partner rhizobium in the establishment of a nitrogen-fixing symbiosis. For instance, the bacterium Sinorhizobium meliloti induces the formation of root nodules on Medicago, Melilotus, and Trigonella. Mesorhizobium loti is compatible with the Lotus species, whereas Rhizobium etli is compatible with species of the same genus (Gibson et al., 2008). NBS-LRR type disease resistance (R) genes are involved in determination of host specificity in the legume-rhizobia symbiosis (Yang et al., 2010). A successful interaction requires the strict coordination of two processes: rhizobium infection and nodule organogenesis (Oldroyd et al., 2011). Bacterial infection can occur either through root hair curls or cracks in the epidermis (Oldroyd and Downie, 2008). In most legumes, infection threads initiate from these infection sites, and allow the invasion of bacteria into the cortex. Finally, the cortical cells divide to form a nodule meristem, which further develops into a nodule (Kouchi et al., 2010). Symbiotic nitrogen fixation makes a major contribution to soil fertility and plays a critical role in sustainable agriculture. It is significant, as well as challenging, to transfer the establishment of such symbiosis to non-legumes for future agriculture development (Gewin, 2010). In rhizobial symbiosis, the establishment of a successful interaction requires the presence of signaling molecules called Nod factors (NFs) (Denarie and Cullimore, 1993). NFs are required for early steps of legume infection and root nodule organogenesis (Denarie et al., 1996). In addition, they are recognized by the LysM domain receptor kinases NFR1 and NFR5 in Lotus japonicus and LYK3 and NFP in Medicago truncatula (Amor et al., 2003; Limpens et al., 2003; Madsen et al., 2003; Radutoiu et al., 2003) to elicit calcium spiking and finally to reprogram the expression of specific symbiosis genes (Oldroyd, 2013). It has been reported that NFs also play a regulatory role in the production of reactive oxygen species (ROS) (Lohar et al., 2007; Cardenas et al., 2008). The mitogen-activated protein kinase (MAPK) cascade is associated with the NF signal transduction (Chen et al., 2012). M. truncatula NFP, initially described as a putative NF receptor, can also play a role in defense against pathogens, resistance to both the oomycete Aphanomyces euteiches and the fungus Colletotrichum trifolii (Rey et al., 2013).

Higher plants are continually exposed to various attacks, including bacteria, fungi, viruses, insects, and nematodes. To prevent invasion by these hostile microbes, plants appear to contain at least two branches of innate immune signaling: pathogen-associated molecular pattern-triggered immunity (PTI) and effector-triggered immunity (ETI) (Jones and Dangl, 2006). PTI is based on pattern recognition receptors (PRRs) that mediate recognition of microbe-associated molecular patterns (MAMPs). One of the best-characterized MAMP/PRR pair is flagellin/FLS2 in Arabidopsis thaliana (Zipfel et al., 2004; Boller and Felix, 2009). Flg22, a conserved 22-amino acid peptide from the bacterial flagellin protein, activates signaling cascade that includes calcium influx (Ranf et al., 2011), ROS production in surrounding infection sites (Torres et al., 2005), and activation of a MAPK cascade (Asai et al., 2002). Subsequently, defense responses, such as callose deposition, are induced (Clay et al., 2009). Flg22 signaling triggers the expression of many genes, such as WRKY transcription factors (Denoux et al., 2008). To avoid PTI, pathogens have evolved effector proteins that are directly delivered into the plant cell through translocation systems such as the bacterial types III and IV protein secretion systems (Feng and Zhou, 2012). R genes recognize these pathogen effector proteins according to the gene-for-gene theory (Block and Alfano, 2011). This form of resistance (termed as ETI) is often accompanied by a hypersensitive response at the infection site (Jones and Dangl, 2006). ETI or PTI receptors activate MAPK cascades that phosphorylate the downstream transcription regulators controlling the expression of early defense genes (van den Burg and Takken, 2010). Interestingly, some of the defense responses have also been detected in epidermal cells of legume roots in response to NF application (Ramu et al., 2002; Nakagawa et al., 2011), and the detected defense responses are transient and local (Cardenas et al., 2008; Jardinaud et al., 2016). Bradyrhizobium japonicum NFs have been found to suppress immune responses in both legumes and non-legumes (Liang et al., 2013). The overlap of defense and symbiotic signaling pathways was highlighted in a study of L. japonicus roots inoculated with its symbiotic partner M. loti and treated with the elicitor flg22 (Lopez-Gomez et al., 2012). There is increasing evidence that the perception of NFs might have evolved from chitin perception, which is related to PTI (Liang et al., 2014). Lipopolysaccharide (LPS) produced by some bacteria often acts as a virulence factor (Petrocelli et al., 2012). Purified LPS from S. meliloti can suppress an oxidative burst in M. truncatula a suspension of cells (Tellstrom et al., 2007). Almost all rhizobia can produce at least one type of exopolysaccharide (EPS) (Skorupska et al., 2006), and S. meliloti produces two types of EPS, namely succinoglycan and galactoglucan (Gibson et al., 2008). Transcriptomic data showed that a S. meliloti EPS-deficient mutant induced the expression of many defense-related genes in the M. truncatula host (Jones et al., 2008). Rhizobial surface polysaccharides (e.g., EPS, LPS, and cyclic glucans) have been implicated in facilitating infection thread formation and nodule development (Jones et al., 2008). It has been proposed that surface polysaccharides play a role in the evasion or suppression of host defense responses, a feature that is shared by pathogenic and symbiotic bacteria (D’Haeze and Holsters, 2004; Soto et al., 2009). Other strains, such as Sinorhizobium sp. NGR234, secrete a few translocate type III effectors including NopM, NopL, NopP, and NopT into host plant cells, where most of the effectors interfere with immune signaling pathways to suppress host defense responses (Skorpil et al., 2005; Dai et al., 2008; Xin et al., 2012; Ge et al., 2016).

Pseudomonas syringae pv. tomato DC3000 (Pst DC3000), a tomato pathogen of tomato (Cuppels, 1986), is also highly pathogenic to A. thaliana (Whalen et al., 1991). Pst DC3000 has been widely used to elucidate the general principles underlying plant immune response and bacterial pathogenesis (Xin and He, 2013; Yao et al., 2013). Plant roots are in contact with various microbes, including symbiotic and pathogenic microbes. However, how plants distinguish symbionts from pathogens is poorly understood. In this study, we successfully established a Pst DC3000-M. truncatula system, and the defense and symbiotic responses of M. truncatula plants were investigated. The interplay was studied by determining the expression of symbiosis marker genes and defense-related genes in plants co-inoculated with S. meliloti Sm2011 and Pst DC3000. These analyses were performed during the rhizobial infection and nodule formation on the roots and the initial hemibiotrophic colonization of Pst DC3000 on the leaves. It was found that the establishment of legume-rhizobium symbiosis suppresses the defense signaling pathways and that Pst DC3000 significantly reduces nodule organogenesis.

Legumes can establish an efficient symbiotic interaction with rhizobia, which results in the production of nodules. However, defense responses are inevitably activated when plants are attacked by pathogens. Thus, a mechanism to successfully avoid or resist plant immune responses is a prerequisite for the establishment of symbiosis between the host and symbiotic rhizobia. In this study, we established a convenient system, to study the interplay between immunity and symbiosis, which consists of M. truncatula A17 plants, pathogenic bacterium Pst DC3000 and rhizobia.

The phosphorylation levels of MAPK3 and MAPK6 were elevated by Pst DC3000 as well as by S. meliloti Sm2011 (Figure 1A), suggesting that the phosphorylation of MAPK3 and MAPK6 is associated with defense and symbiotic responses. However, the phosphorylation level of MAPK3 and MAPK6 induced by rhizobia was lower than that induced by Pst DC3000. Importantly, inoculation with compatible rhizobia could suppress the pathogen-triggered MAPK phosphorylation as shown by three time-course experiments (Figure 1 and Supplementary Figure S2). In addition, treatment of M. truncatula leaves with S. meliloti Sm2011 and Pst DC3000 significantly reduced Pst DC3000-induced ROS production (Figure 2). These results suggest that rhizobia produce elicitors (unknown MAMPs) that are detected by the host plant resulting in suppressed Pst DC3000-triggered immunity in M. truncatula. Bacterial cell surface components, such as oligo- and polysaccharides, are considered as crucial signals for the interactions of bacteria with host plants (Mithofer, 2002). In addition to effectors, several signals from rhizobia have been identified to be associated with the suppression of the host defense in legume roots, including NFs, LPS, and EPS (Luo and Lu, 2014; Gourion et al., 2015; Tóth and Stacey, 2015; Yasuda et al., 2016). It has been proposed that the symbiotic suppression of legume innate immunity is a result of co-evolution (Gourion et al., 2015) and that the identified signals for symbiosis may interact with immune signaling pathways to suppress host defense responses.

To gain a better understanding of the possible overlap of molecular mechanisms underlying symbiosis and disease development, we investigated the expression patterns of symbiosis- and defense-related genes during the early symbiotic stage and disease development caused by Pst DC3000 (Figures 4, 5 and Supplementary Figure S4). The expression of NIN and NPL was lower in co-inoculated roots than in those inoculated with S. meliloti Sm2011 alone (Figure 4A). These results suggest that the symbiotic pathway is suppressed when defense signaling pathways are activated. Similarly, the expression of FLS2, PR10, and WRKY33 was lower in co-inoculated plants than in those inoculated with Pst DC3000 alone (Figure 5), suggesting that defense signaling pathways are suppressed during the establishment of symbiosis. In addition, the transcript levels of defense-related genes MAPKs, RbohC, and WRKY33 exhibited a transient increase in response to rhizobial inoculation, whereas low levels were measured during the formation of nodule primordia (Figures 4B, 5C and Supplementary Figure S4). Hence, a rapid and defense-like response occurred in M. truncatula upon S. meliloti inoculation. These findings are consistent with previous speculations (Tóth and Stacey, 2015).

Plant immune responses are usually accompanied by the production of ROS, one of the earliest responses following pathogen infection in plants. Genetic and biochemical evidence indicates that NADPH oxidases (RBOHs) are key compounds for ROS production involved in plant innate immunity (Arthikala et al., 2014). RBOHs have been described as a major source of ROS during the establishment of root nodules (Puppo et al., 2013). In this study, no significant change in the transcripts level of RbohD was observed between Pst DC3000 and S. meliloti Sm2011 infection, while the expression of RbohC increased significantly at 3 dpi with both strains (Supplementary Figure S3). It is possible that another RBOH gene should be used for detection.

Defense and symbiotic responses in M. truncatula appear to influence each other. Pst DC3000 likely caused a negative effect on the symbiotic interaction between M. truncatula and S. meliloti. This negative effect resulted in a decrease in infection threads and reduced nodulation (Figure 7). Pst DC3000 affected nodule organogenesis. Pst DC3000 caused diffuse chlorosis on leaves, and thus, likely had reduced photosynthesis and uptake of nutrients. In this view, poor nodulation would be the result of weak plant growth. Legumes possess a systemic negative feedback regulatory system called autoregulation of nodulation (AON), which controls the nodule number and the nodulation zone through long-distance root-shoot signaling (Oka-Kira and Kawaguchi, 2006; Ferguson et al., 2010). AON is initiated during nodule development by the synthesis of a root-derived signal named ‘Q’ (Okamoto et al., 2009). Pst DC3000 infected plant leaves and S. meliloti Sm2011 infected plant roots. Maybe there is Q signal by the AON in the leaf results in the production of a novel shoot-derived inhibitor, which appears to enter the phloem and travels down to the root where it acts to inhibit further nodulation events (Lin et al., 2010). The involvement of hormone in AON (Kinkema and Gresshoff, 2008). Plant hormones are likely to be essential throughout nodule organogenesis for integration of developmental and environmental signaling cues into nodule development (Ryu et al., 2012; Ferguson and Mathesius, 2014). Under the condition of pathogen attack, plants generated salicylic acid (SA) and jasmonic acid (JA) acting as secondary signaling molecules to modulate plant defense response. Indeed, the expression of defense-related genes PR5 and PAL were significantly induced when plants were inoculated with Pst DC3000, while co-inoculation with Pst DC3000 and S. meliloti Sm2011decreased the expression levels of PR5 and PAL (Figure 6). These data confirmed that rhizobia repress immune response triggered by bacterial pathogens.

Tripartite legume/rhizobium/pathogen interactions are complex systems and their analysis will require future attention. In our test system, it is hard to monitor the number of mature nodules (2 or 3 weeks post inoculation) since Pst DC3000 could cause the death of M. truncatula leaves and seriously affected the growth of plants, which impaired nodule development. Instead, we took advantage of quantitative PCR to compare the growth of Pst DC3000 in plants in response to rhizobia treatment. The quantification of Pst DC3000 by DNA-based technique is highly reliable and comparable (Brouwer et al., 2003; Ross and Somssich, 2016). S. meliloti Sm2011 significantly inhibits bacterial growth in the leaves of M. truncatula (Figure 8). Besides, some of leaves died (because of the pathogen infection) and fell from the plants may result in lower bacterial growth at 8 dpi than at 5 dpi.

Rhizobia treatment inhibits both growth of Pst DC3000 and defense-related genes expression. The possible mechanism involved in this is that rhizobia may inhibit the growth of Pst DC3000 through an unknown mechanism or compete with Pst DC3000 to infect plants leading to lower number of Pst DC3000 grown in plants. The reduced number of Pst DC3000 might be the direct reason to induce less immune response in plants leading to decreased defense-related gene expression.

Various M. truncatula mutants have been characterized, especially symbiosis mutants. Future gene expression analysis of these mutants inoculated with Pst DC3000 will further enhance our understanding on regarding the overlap of molecular mechanisms underlying symbiosis and disease development.

TC and ZZ designed the research and wrote the paper. TC, BZ, and LD executed the experiments. TC, BZ, LD, HY, HZ, YC, and ZZ performed the data and analyses. All authors read and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.