Report

Nonlegumes Respond to Rhizobial Nod Factors by Suppressing the Innate Immune Response

See allHide authors and affiliations

Science  20 Sep 2013:
Vol. 341, Issue 6152, pp. 1384-1387
DOI: 10.1126/science.1242736

Stealth Nod Factor Recognition

Legumes' symbiotic interaction with nitrogen fixing bacteria supplies the plant with nitrogen. Many important crop plants, however, cannot establish these symbioses and, thus, agriculture depends on externally applied fertilizers. Surprisingly, Liang et al. (p. 1384, published online 5 September) found that several nonleguminous plants, including Arabidopsis, tomato, and corn, were able to respond to the same Nod factors that initiate the microbial symbiosis in soybean.

Abstract

Virtually since the discovery of nitrogen-fixing Rhizobium-legume symbioses, researchers have dreamed of transferring this capability into nonlegume crop species (for example, corn). In general, nonlegumes were assumed to lack the ability to respond to the rhizobial lipo-chitin Nod factors, which are the essential signal molecules that trigger legume nodulation. However, our data indicate that Arabidopsis thaliana plants, as well as other nonlegumes, recognize the rhizobial Nod factor via a mechanism that results in strong suppression of microbe-associated molecular pattern (MAMP)–triggered immunity. The mechanism of action leads to reduced levels of pattern-recognition receptors on the plasma membrane involved in MAMP recognition.

A general theory for the development of commensal and mutualistic symbioses is that they evolved from pathogenic relationships (1). Although the Rhizobium-legume symbiosis is beneficial and benign, the plant initially responds with a pathogen defense response that is quickly suppressed (2, 3). The lipo-chitin nodulation factors (known as Nod factors) are substituted acylate chitin oligomers of three to five N-acetylglucosamine residues and are the key signaling molecules for the establishment of Rhizobium-legume symbioses. These factors induce plant responses that lead to the development of the nodule, which becomes the intracellular home of the invading symbiont (4, 5). Nod factors function at nanomolar levels (6, 7), but on only specific, compatible legume hosts. Extension of nitrogen-fixing symbioses beyond these already compatible legume hosts could help reduce dependence on applied fertilizers (8).

Plant innate immunity can be triggered by the recognition of microbe-associated molecular patterns (MAMPs) (9). The best-studied MAMP is bacterial flagellin (flg22, a conserved 22–amino acid peptide from flagellin), which is recognized by the FLAGELLIN-SENSING 2 (FLS2) receptor located on the plasma membrane (10). Perception of flg22 activates a signaling cascade of defense responses, including calcium influx (11), reactive oxygen species (ROS) production (12), activation of mitogen-activated protein (MAP) kinases (13), gene expression (14), callose deposition (15), and bacterial growth restriction (10). Together, these responses are called MAMP-triggered immunity.

To study the intersection between symbiosis and pathogenesis, we added both flg22 and purified Nod factor from Bradyrhizobium japonicum, the soybean symbiont, to soybean leaves (see supplementary materials and methods). The flg22-triggered ROS production was reduced 25% in the presence of Nod factor (Fig. 1A). Chitin is also a strong MAMP but only when the oligomers have more than six N-acetylglucosamine residues. Nod factor is composed of shorter oligomers (three to five residues). We tested the chitin oligomers containing one to eight residues for their ability to suppress flg22-triggered ROS production. Chitotetraose (C4) and chitopentaose (C5) reduced flg22-triggered ROS production, but not as efficiently as Nod factor (fig. S1). Chitin oligomers C6 through C8 did not affect flg22-triggered ROS production (fig. S1). These results are similar to some symbiotic responses in which simple chitooligomers can suffice but only at higher concentration than the Nod factor (16, 17). Treatment with Nod factor or chitotetraose alone did not induce ROS production (Fig. 1A). Pretreatment of plants with Nod factor or chitotetraose enhanced the suppressive effect of flg22-triggered ROS production (Fig. 1B). For example, flg22-induced ROS production was reduced 60% with 30 min of chitotetraose pretreatment (Fig. 1B).

Fig. 1 Nod factor (NF) and chitotetraose (C4) treatment reduces flg22-triggered immunity in soybean leaves.

(A) Nod factor and chitotetraose reduced flg22-triggered ROS production. Leaf discs were treated with H2O, Nod factor (100 nM), chitotetraose (10 μM), and flg22 (100 nM) with or without the addition of Nod factor or chitotetraose. (B) Pretreatment enhanced the suppressive effect of C4 on flg22-triggered ROS production. (C) Nod factor (1 nM) reduced flg22-triggered ROS production. Leaf discs were pretreated with different concentrations of Nod factor for 30 min before flg22 (100 nM) treatment. (A to C) The data are shown as means ± SD (error bars; n = 6 leaf discs). Asterisks indicate significant differences from flg22 treatment (**P ≤ 0.01, *P ≤ 0.05). All experiments were repeated twice with similar results. (D) Nod factor and chitotetraose reduced flg22-induced MAP kinase (MAPK) phosphorylation. Lower panels shows representative Western blot; upper panel shows quantification of normalized protein levels to flg22 treatment. Data are means ± SE (error bars) from three independent biological repeats.

The suppressive effects of Nod factor or chitotetraose addition were also seen when chitooctaose, a potent MAMP, was used to induce ROS production (fig. S2). During nodulation, application of Nod factors at picomolar to nanomolar concentrations elicits a variety of symbiotic responses (for instance, root hair deformation). The lowest concentration of Nod factor required to reduce flg22-triggered ROS production was 1 nM (Fig. 1C). One of the downstream signals after flg22 treatment in Arabidopsis is MAP kinase phosphorylation, which can be detected immunologically with the use of an anti-phospho-p44/p42 MAP kinase antibody. Similarly, soybean phosphorylates MAP kinase after flg22 elicitation. We observed that Nod factor or chitotetraose pretreatment of soybean leaves reduced flg22-induced MAP kinase activation (Fig. 1D). Nod factor or chitotetraose alone did not induce MAP kinase activation (Fig. 1D).

We also tested the ability of Nod factor and chitotetraose to suppress flg22-triggered immunity in Arabidopsis, which is not a legume and has been the subject of much research on plant immunity, including the mechanisms of MAMP-triggered immunity. A 1-hour pretreatment of Arabidopsis leaves with Nod factor or chitotetraose significantly reduced flg22-induced ROS production (Fig. 2A and fig. S3) and MAP kinase phosphorylation (fig. S4). The suppressive effect of Nod factor is weaker and slower in Arabidopsis than in soybean.

Fig. 2 Nod factor and chitotetraose treatment reduces flg22-triggered immunity in Arabidopsis.

(A) Nod factor and chitotetraose reduced flg22-triggered ROS production in Arabidopsis leaves. (B) Nod factor and chitotetraose reduced flg22-triggered callose deposition. Representative images are shown here, and the quantification data are shown in fig. S5. Scale bar, 1 mm. DMSO, dimethyl sulfoxide. (C) Nod factor and chitotetraose reduced flg22-triggered bacterial growth restriction. Five-week-old Arabidopsis plants were infiltrated with the indicated treatment before Pseudomonas syringae pv. tomato DC3000 inoculation at a concentration of 105 colony-forming units (CFU)/ml. Bacterial growth was determined 3 days after inoculation. The data are shown as means ± SD (error bars) from three replicates (*P ≤ 0.05). (D) Nod factor and chitotetraose reduced flg22-triggered calcium influx in Arabidopsis seedlings. Leaf disc (A) or 6-day-old seedlings (D) were pretreated with H2O, Nod factor (100 nM), and chitotetraose (10 μM) for 1 hour before flg22 (10 nM) treatment. Data are means ± SD (error bars; n = 6 leaf discs; **P ≤ 0.01, *P ≤ 0.05). All experiments were repeated twice with similar results.

Application of Nod factor and chitotetraose to Arabidopsis also interfered with other responses to flg22, such as callose deposition (Fig. 2B and fig. S5) and changes in gene expression (fig. S6). In Arabidopsis, flg22-triggered immunity inhibits growth of the bacterial pathogen Pseudomonas syringae. However, treatment with Nod factor or chitotetraose suppressed this inhibition, resulting in more extensive pathogen colonization (Fig. 2C). One of the earliest flg22 responses upstream of ROS production is the transient increase of cytosolic Ca2+ concentration due to calcium influx. In the presence of either Nod factor or chitotetraose, flg22-triggered calcium influx was significantly reduced in both Arabidopsis seedlings (Fig. 2D) and roots (fig. S7). Thus, Arabidopsis, although not a legume, can recognize Nod factor in a way that results in suppression of flg22-triggered immunity, similar to what is seen in soybean treated with Nod factor.

Calcium influx is induced within minutes after flg22 treatment and acts immediately downstream of flg22 perception. The inhibitory effect of Nod factor and chitotetraose on flg22-triggered calcium influx suggested a mechanism that may involve FLS2, the receptor of flagellin. The levels of FLS2 at the plasma membrane are regulated, in part, by ubiquitination and 26S proteasome degradation induced by flg22 recognition (18). We used Western blot analysis with anti-FLS2 antibody to examine whether Nod factor could directly affect FLS2 protein levels. FLS2 protein levels were significantly reduced 1 hour after Nod factor treatment (Fig. 3A), whereas the transcript level was unaltered (fig. S8). Addition of MG132, an inhibitor of proteasome-mediated degradation, blocked the effect of Nod factor on FLS2 protein levels (Fig. 3A).

Fig. 3 Nod factor induces FLS2 protein degradation in Arabidopsis.

(A) Nod factor induces FLS2 protein degradation, and the protein degradation was inhibited by the addition of MG132. Ten-day-old seedlings were treated with or without MG132 (50 μM) for 1 hour before Nod factor (100 nM) treatment. Total protein was extracted 1 hour after Nod factor treatment. (B) Nod factor–induced FLS2 protein degradation was absent in bak1-4 mutants. Immunoblot analysis was performed with the anti-FLS2 antibody. Ponceau S staining was used for protein loading control. (A and B) Lower panels show representative Western blot; upper panels show quantification of normalized FLS2 levels to the control. The data are shown as means ± SE (error bars) from three independent biological repeats. (C) Nod factor addition results in the reduced level of FLS2:GFP from the plasma membrane. Five-day-old FLS2::FLS2:GFP seedlings were treated with or without Nod factor (100 nM). Images were taken from cotyledon 1 hour after Nod factor treatment. Costaining with FM4-64 (10 μM) highlights the plasma membrane. Scale bars, 10 μm.

The leucine-rich repeat receptor kinase BAK1 is required to recruit specific E3 ligases to FLS2 for the attenuation of flg22 responses in Arabidopsis (18). Nod factor–induced FLS2 degradation was absent in the bak1-4 Arabidopsis mutant (Fig. 3B), indicating that Nod factor action on FLS2 is dependent on BAK1. The addition of Nod factor did not affect BAK1 protein levels (fig. S9). We observed that fluorescence on the plasma membrane derived from a green fluorescent protein (GFP)–tagged FLS2 construct in wild-type Arabidopsis was also reduced within 1 hour of Nod factor treatment (Fig. 3C). Thus, the response in Arabidopsis to Nod factor treatment includes the reduced level of FLS2 from the plasma membrane, a process that requires ubiquitin-dependent protein degradation.

Arabidopsis also responds to MAMPs other than flg22, including the peptide elf26 (fig. S10), chitin, and pep1 (an endogenous peptide from Arabidopsis that induces defense responses), as well as the crude extract of rhizobia. Similar to the results obtained with flg22, ROS production elicited by elf26, pep1, chitin, and crude rhizobial extract was significantly reduced in the presence of Nod factor or chitotetraose (figs. S11 and S12). The B. japonicum Nod factor used in our studies was purified from bacterial cultures. However, chemically synthesized Nod factors also reduced flg22-triggered calcium influx in Arabidopsis (fig. S13). EFR (the receptor of elf26) protein levels were significantly reduced after Nod factor treatment, and the protein degradation was inhibited by the addition of MG132 (fig. S14).

We tested two other nonlegume plant species (tomato and corn) for their ability to recognize and respond to Nod factor. Both Nod factor and chitotetroase could suppress flg22-induced ROS production in tomato (fig. S15), and chitooctaose-triggered ROS production in corn (fig. S16). Thus, the mechanism of Nod factor recognition and action that legumes have so effectively used is conserved in the nonlegumes Arabidopsis, tomato (a dicot), and corn (a monocot).

Mutations of Nod factor receptors have been identified in various legumes (1921) and include soybean nod49 (nfr1) and nod139 (nfr5) mutants (22, 23). We found that, although soybean mutants in Nod factor receptor 1 (NFR1) or NFR5 were unable to induce symbiotic functions (e.g., gene expression) in response to Nod factor addition, these mutants retained the ability to suppress MAMP-triggered immunity upon addition of Nod factor (fig. S17). The soybean NFRs are closely related to the chitin receptor LYK1/CERK1 in Arabidopsis and rice (24, 25). These receptors are all characterized by the presence of one or more lysin motifs (LysMs) in their extracellular domain. In Arabidopsis, five LysM receptor kinases (LYKs), including LYK1/CERK1, have been identified. Therefore, we tested mutants in each of these LYK genes for their ability to recognize Nod factor and suppress flg22-triggered ROS production. Plants defective in the LYK3 protein failed to suppress flg22-triggered ROS production upon Nod factor addition (Fig. 4A), whereas ectopic overexpression of LYK3 enhanced Nod factor–induced suppression of ROS production (Fig. 4B). Similarly, mutation or overexpression of LYK3 also altered the suppressive effect of Nod factor on flg22-triggered MAP kinase phosphorylation (Fig. 4C). The other four lyk mutants did not show altered responses to Nod factor treatment. Therefore, our results suggest that LYK3 is required in Arabidopsis for Nod factor–induced suppression of MAMP-triggered immunity.

Fig. 4 AtLYK3 is required for Nod factor suppression of flg22-triggered immunity.

(A) lyk3 mutants are defective in Nod factor suppression of flg22-triggered ROS production. (B) Transgenic plants ectopically overexpressing LYK3 (LYK3-OX) are hypersensitive to Nod factor suppression of flg22-triggered ROS production. Leaf discs in (A) and (B) were pretreated with H2O or Nod factor (100 nM) for 1 hour before flg22 (10 nM) treatment. The relative ROS production was calculated as a percentage of the value of flg22 plus Nod factor divided by the value of flg22. The data are shown as means ± SD (error bars; n = 6 leaf discs). Asterisks indicate significant difference from the wild type (**P ≤ 0.01, *P ≤ 0.05). All experiments were repeated twice with similar results. (C) lyk3 mutants and LYK3-OX transgenic plants exhibited altered response to Nod factor suppression of flg22-triggered MAP kinase phosphorylation. Lower panels show representative Western blot; upper panel shows quantification of normalized protein levels to flg22 treatment. The data are shown as means ± SE (error bars) from four independent biological repeats.

The lysin motif was first characterized in bacterial proteins involved in remodeling of bacterial peptidoglycan (26). Similar proteins were also shown to recognize lipo-chitin signals produced by mycorrhizal fungal symbionts (27). Collectively, these findings led to the suggestion that chitin recognition is an ancient plant trait that probably evolved from mycorrhizal recognition, whose symbiosis has been documented more than 400 million years ago (28), or perhaps from fungal pathogen-plant interaction, whose origin has not been dated. However, Arabidopsis is not infected by either rhizobia or mycorrhizae (29). Hence, the mechanism of Nod factor recognition in this plant must be independent of these symbiotic pathways. If Nod factor suppression of MAMP-triggered immunity is universal in plants, then it may predate the mycorrhizal symbiosis and could represent the beginning of recognition mechanisms for this interesting molecule.

Supplementary Materials

www.sciencemag.org/content/341/6152/1384/suppl/DC1

Materials and Methods

Figs. S1 to S17

References (3036)

References and Notes

  1. Acknowledgments: We thank A. Bent for providing the anti-FLS2 antibody and EFR::EFR:HA transgenic seeds, M. R. Knight for aequorin transgenic seeds, S. Robatzek for FLS2::FLS2:GFP transgenic seeds, Novozymes (Bagsvaerd, Denmark) for the purified Nod factor, and S. Peck for critical comments to improve the manuscript. This work was supported by the Division of Chemical Sciences, Geosciences, and Biosciences, Office of Basic Energy Sciences of the U.S. Department of Energy through grant DE-FG02-08ER15309 and by the Next-Generation BioGreen 21 Program Systems and Synthetic Agrobiotech Center, Rural Development Administration, Republic of Korea (grant PJ009068 to G.S). Additional support was provided by Novozymes.

Stay Connected to Science

Navigate This Article