Functional Analysis of Chimeric Lysin Motif Domain Receptors Mediating Nod Factor-Induced Defense Signaling in Arabidopsis Thali

The Plant Journal (2014) 78, 56–69 doi: 10.1111/tpj.12450 Functional analysis of chimeric lysin motif domain receptors mediating Nod factor-induced defense signaling in Arabidopsis thaliana and chitin-induced nodulation signaling in Lotus japonicus

Wei Wang1,2, Zhi-Ping Xie1,* and Christian Staehelin1,* 1State Key Laboratory of Biocontrol and Guangdong Key Laboratory of Plant Resources, School of Life Sciences, Sun Yat-sen University, East Campus, Guangzhou 510006, China, and 2Anhui Key Laboratory of Plant Genetics & Breeding, School of Life Sciences, Anhui Agricultural University, 130 Changjiang West Road, Hefei, Anhui 230036, China

Received 12 October 2013; revised 11 January 2014; accepted 16 January 2014; published online 8 February 2014. *For correspondence (e-mails [email protected] or [email protected]).

SUMMARY The expression of chimeric receptors in plants is a way to activate specific signaling pathways by corresponding signal molecules. Defense signaling induced by chitin from pathogens and nodulation signaling of legumes induced by rhizobial Nod factors (NFs) depend on receptors with extracellular lysin motif (LysM) domains. Here, we constructed chimeras by replacing the ectodomain of chitin elicitor receptor kinase 1 (AtCERK1) of Arabidopsis thaliana with ectodomains of NF receptors of Lotus japonicus (LjNFR1 and LjNFR5). The hybrid constructs, named LjNFR1–AtCERK1 and LjNFR5–AtCERK1, were expressed in cerk1-2, an A. thaliana CERK1 mutant lacking chitin-induced defense signaling. When treated with NFs from Rhizobi- um sp. NGR234, cerk1-2 expressing both chimeras accumulated reactive oxygen species, expressed chitin- responsive defense genes and showed increased resistance to Fusarium oxysporum. In contrast, expression of a single chimera showed no effects. Likewise, the ectodomains of LjNFR1 and LjNFR5 were replaced by those of OsCERK1 (Oryza sativa chitin elicitor receptor kinase 1) and OsCEBiP (O. sativa chitin elicitor-binding protein), respectively. The chimeras, named OsCERK1–LjNFR1 and OsCEBiP–LjNFR5, were expressed in L. japonicus NF receptor mutants (nfr1-1; nfr5-2) carrying a GUS (b-glucuronidase) gene under the control of the NIN (nodule inception) promoter. Upon chitin treatment, GUS activation reflecting nodulation signaling was observed in the roots of NF receptor mutants expressing both chimeras, whereas a single construct was not sufficient for activation. Hence, replacement of ectodomains in LysM domain receptors provides a way to specifically trigger NF-induced defense signaling in non-legumes and chitin-induced nodulation signaling in legumes.

Keywords: defense signaling, chitin, nodulation signaling, Nod factors, LysM domain receptor, chimeric receptor, Arabidopsis thaliana, Lotus japonicus.

INTRODUCTION

The activation of plant defense responses against the important structural component in fungi. Chitin fragments unwanted invasion of bacteria or fungi (innate immunity) (chitooligosaccharides) released from fungal cell walls by includes recognition of microbe-associated molecular plant chitinase function as MAMPs in a similar way (Felix patterns (MAMPs) by pattern recognition receptors (PRRs). et al., 1993; Boller and Felix, 2009). An example of a bacte- Microbe-associated molecular patterns are conserved rial MAMP is peptidoglycan, which is a conserved structural molecular components from a speciﬁc class of microbes part of the bacterial cell wall and contains alternating that function as elicitors of defense signaling, resulting in residues of b-1,4-linked N-acetylglucosamine and N-acetyl- immunity to fungi and bacteria (Boller and Felix, 2009). A muramic acid residues (Gust et al., 2007). Chitin, chitooligo- typical MAMP is chitin (a polymer with b-1,4-linked N-ace- saccharides and peptidoglycan are recognized by PRRs tyl-D-glucosamine residues; poly-GlcNAc), which is an containing extracellular lysin motif (LysM) domains, which

56 © 2014 The Authors The Plant Journal © 2014 John Wiley & Sons Ltd Chimeric LysM domain receptors 57 are known to bind to GlcNAc-containing molecules. Ligand- molecular flux through plasmodesmata, a chitin response mediated receptor activation triggers defense signaling, that is independent of AtCERK1-regulated defense gene which culminates in accumulation of reactive oxygen spe- activation (Faulkner et al., 2013). The related AtLYM1 and cies (ROS), mitogen-activated protein (MAP) kinase signal- AtLYM3 proteins are involved in peptidoglycan perception ing, activation of defense genes and ultimately in immunity and presumably form a receptor complex with AtCERK1 against invading microbes (Hamel and Beaudoin, 2010; (Willmann et al., 2011). Similar to rice, activated AtCERK1 Gust et al., 2012). initiates MAP kinase signaling (via AtMPK3/AtMPK6), In rice (Oryza sativa L.), OsCEBiP (chitin elicitor-binding which triggers the expression of chitin-responsive defense protein) and OsCERK1 (chitin elicitor receptor kinase 1) are genes (Wan et al., 2004; Miya et al., 2007). LysM domain PRRs required for the induction of chitin- Nod factors (nodulation factors; NFs) are chitin-related induced defense signaling. The chitin-binding OsCEBiP is bacterial signals that are secreted by nitrogen-fixing rhizo- a glycosylphosphatidylinositol-anchored protein, whereas bia to initiate nodule symbiosis with host legumes. Nod OsCERK1 possesses an intracellular kinase domain. Bio- factors are lipo-chitooligosaccharides, i.e. they consist of a chemical studies suggest that the two proteins form hete- carbohydrate moiety usually containing four or five rodimers in a ligand-dependent manner to initiate defense b-1,4-linked GlcNAc oligomers and an acyl residue (fatty signaling (Kaku et al., 2006; Shimizu et al., 2010). More- acid) at the non-reducing end. Furthermore, the terminal over, rice possesses additional LysM domain PRRs, namely sugars of NFs carry strain-specific modifications, which OsLYP4 and OsLYP6 (lysin motif-containing proteins 4 and may be required for nodulation of certain host plants 6), which seem to play a dual role in perception of chitin (Perret et al., 2000). Arbuscular mycorrhizal fungi pro- and peptidoglycan (Liu et al., 2012a). The OsCERK1 and duce structurally related lipo-chitooligosaccharides, the OsCEBiP receptors appear to interact with the Rho-like so-called Myc factors, but their role in establishing symbio- small GTPase OsRac1 and the guanine nucleotide exchange sis with host roots remains elusive (Maillet et al., 2011; factor OsRacGEF1, forming a so-called ‘defensome’. Phos- Genre et al., 2013). Similar to the PRRs perceiving chitin, phorylation of OsRacGEF1 by the kinase domain of chitooligosaccharides and peptidoglycan, specific LysM OsCERK1 was found to be required for downstream domain receptors of host plants recognize NFs and Myc fac- defense signaling (Akamatsu et al., 2013). Furthermore, tors to initiate symbiotic signaling (Ferguson et al., 2010; OsCERK1 phosphorylates the receptor-like cytoplasmic Oldroyd, 2013). LjNFR1 and LjNFR5 (Nod factor receptors 1 kinase OsRLCK185, which is a downstream signaling com- and 5) of Lotus japonicus are prime examples of NF recep- ponent required for activation of the chitin-induced MAP tor genes, which have been also characterized in other kinase cascade OsMKK4–OsMPK3/OsMPK6 (Kishi-Kaboshi legumes such as soybean (Glycine max) and Medicago et al., 2010; Yamaguchi et al., 2013). truncatula (Limpens et al., 2003; Madsen et al., 2003; Radu- In Arabidopsis thaliana, AtCERK1 (also called LYK1; toiu et al., 2003; Arrighi et al., 2006; Indrasumunar et al., LysM-containing receptor-like kinase 1) is a similar LysM 2010, 2011). Nod factors of Mesorhizobium loti bind to domain PRR required for chitin perception and activation LjNFR5 and LjNFR1 at nanomolar concentrations and the of chitin-responsive defense genes (Miya et al., 2007; Wan two receptors seem to form a ligand-dependent heterodi- et al., 2008). AtCERK1 possesses a functional intracellular merization complex (Madsen et al., 2011; Broghammer kinase domain and its activation does not require interac- et al., 2012). LjNFR5 interacts with the Rho-like small tion with an OsCEBiP-related protein (Shinya et al., 2012; GTPase LjROP6, which positively regulates the develop- Wan et al., 2012). Biochemical data and crystal structure ment of infection threads (Ke et al., 2012). In contrast to analysis showed that chitin and chitooligosaccharides LjNFR5, which lacks protein kinase activity, the kinase activ- [degree of polymerization (DP) = 4–8] bind directly to the ity of LjNFR1 was found to be essential for autophosphory- ectodomain of AtCERK1. Long-chain chitooligosaccharides lation and downstream signaling (Madsen et al., 2011). A

[namely chitooctaose, (GlcNAc)8] and chitin act as bivalent specific component of nodulation signaling is the transcrip- ligands, having the capacity to trigger dimerization of two tion factor NIN (nodule inception), which is required for AtCERK1 ectodomains and subsequent phosphorylation of expression of early nodulin genes and subsequent initiation the kinase domain (Iizasa et al., 2010; Petutschnig et al., of nodule organogenesis (Schauser et al., 1999). The NIN 2010; Liu et al., 2012b). AtLYK4, an AtCERK1-related recep- protein of L. japonicus binds to the promoters of two NF-Y tor with a kinase domain, may assist AtCERK1-dependent (nuclear factor-Y) subunit genes, which are thought to regu- chitin signaling (Wan et al., 2012). The OsCEBiP-related late cell division in nodule primordia (Soyano et al., 2013). protein AtLYM2 (lysin motif domain-containing glycosyl- Chimeric receptors can be designed by swapping spe- phosphatidylinositol-anchored protein 2) probably repre- cific domains from two different receptors. Plants express- sents a component of an additional chitin perception ing such chimeras are useful tools for characterization or system in A. thaliana. Mutant analysis revealed that identification of receptor–ligand interactions (e.g. Albert AtLYM2 is required for a chitin-induced reduction in et al., 2010; Brutus et al., 2010; Nakagawa et al., 2011;

Mueller et al., 2012). In L. japonicus, for example, chimeras Chimeras were constructed in which the ectodomain of containing the ectodomain of LjNFR1 and various modifi- AtCERK1 from A. thaliana was replaced by the ectodomain cations of the protein kinase domain of AtCERK1 were of LjNFR1 or LjNFR5 from L. japonicus. Mutant plants of expressed in a non-nodulating nfr1 mutant to analyze func- A. thaliana (cerk1-2) expressing the chimeras were treated tionality, i.e. establishment of symbiosis with M. loti. The with NFs and characterized with respect to known complementation tests showed that the kinase domain of AtCERK1-mediated responses, namely the formation of AtCERK1 gains specificity for nodulation signaling if short reactive oxygen species (ROS), expression of defense amino acid sequences in the region around the activation genes and resistance to the fungus F. oxysporum. Likewise, loop are replaced by those of LjNFR1 (Nakagawa et al., the ectodomains of the NF receptors LjNFR1 and LjNFR5 2011). Plants can also be engineered that express chimeric were substituted by ectodomains of the chitin perception receptors with completely different ectodomains and there- system of rice (OsCERK1 and OsCEBiP). The chimeras were fore activate signaling pathways in response to different expressed in hairy roots of two NF receptor mutants of ligands. Such altered signal output may improve the L. japonicus that express a b-glucuronidase (GUS) reporter plant’s responsiveness to environmental stimuli. In rice, gene under the control of the NIN promoter. Activation of for example, a chimeric receptor, consisting of the ectodo- symbiotic signaling in response to chitin or chitooligosac- main of BRI1 (recognizing the brassinosteroid hormone) charides was analyzed by GUS staining and by microscopic and the intracellular domain of the disease resistance pro- observations of root-hair deformation responses. tein XA21 (a leucine-rich repeat receptor kinase), can mediate induction of a hypersensitive response (rapid cell RESULTS death) in response to brassinosteroids (He et al., 2000). Chimeric LysM domain receptor constructs More recently, the ectodomain of the chitin-binding protein OsCEBiP was fused to the intracellular kinase domain of A first set of chimeric LysM domain receptors was the resistance proteins XA21 and Pi-d2. Rice plants designed to test whether defense signaling in A. thaliana expressing these constructs and treated with chitohepta- can be specifically activated by rhizobial NFs. The N-termi- ose, (GlcNAc)7, showed a similar hypersensitive response. nal ectodomains of the NF receptor genes LjNFR1 and Consequently, increased resistance was observed when LjNFR5 of L. japonicus were fused to the transmembrane the plants were inoculated with the rice blast fungus and cytoplasmic domains of the chitin receptor gene Magnaporthe oryzae (Kishimoto et al., 2010; Kouzai et al., AtCERK1 of A. thaliana, forming the chimeras LjNFR1– 2013). AtCERK1 and LjNFR5–AtCERK1 (Figure 1). The domains in In the present study, we examined whether hybrid LysM these hybrid receptor constructs were not further modified, domain receptors with exchanged ectodomains can medi- in order to maintain correct ligand-induced conformational ate NF-induced defense signaling in a non-leguminous changes and an unaltered interaction with cytoplasmic plant and chitin-induced nodulation signaling in a legume. partners. The chimeras were placed under the control of

Figure 1. Schematic representation of constructed chimeric lysin motif (LysM) domain receptors used for transformation of Arabidopsis thaliana (chitin receptor mutant cerk1-2) and Lotus japonicus Nod factor (NF) receptor mutants [nfr1-1 (pNIN::GUS) and nfr5-2 (pNIN::GUS)]. The LysM domains and transmembrane regions (TM) are marked in gray and black, respectively. Regions corresponding to the signal peptide (SP) are shownat the N-termini. All constructs were expressed in plants under the control of the CaMV 35S promoter. Letters followed by numbers indicate positions of annotated amino acids forming the fusion points (protein accession numbers: CAE02589 for LjNFR1, CAE02597 for LjNFR5, NP566689 for AtCERK1, BAG99408 for OsCERK1 and BAE95828 for OsCEBiP).

© 2014 The Authors The Plant Journal © 2014 John Wiley & Sons Ltd, The Plant Journal, (2014), 78, 56–69 Chimeric LysM domain receptors 59 the cauliflower mosaic virus (CaMV) 35S promoter and defense genes encoding MAP kinase 3 (MPK3), WRKY then used for Agrobacterium tumefaciens-mediated trans- transcription factors (WRKY33, WRKY53) and a nitrilase formation of A. thaliana cerk1-2 (mutated in the AtCERK1 (Nit4) (Wan et al., 2004). RNA was extracted from leaves gene). Reverse transcription polymerase chain reaction 2-h after F. oxysporum inoculation. Transcript levels of the (RT-PCR) indicated expression of the LjNFR1–AtCERK1 and/ four genes were remarkably upregulated in inoculated or LjNFR5–AtCERK1 chimeras in stably transformed plants wild-type (Col-0) and AtCERK1 complemented plants. On (Figure S1). the other hand, only low transcript levels of these genes A second set of chimeric LysM domain receptors was were found in cerk1-2 mutant derivatives expressing either designed to examine whether nodulation signaling in one or both chimeras (Figure 2c). These data indicate L. japonicus roots can be specifically activated by applica- that the LjNFR1/LjNFR5 ectodomains of the constructed tion of chitin. Chimeras, named OsCERK1–LjNFR1 and chimeras were obviously unable to perceive fungal chitin OsCEBiP–LjNFR5, were constructed by fusing the ectodo- elicitors. mains of the chitin receptor genes OsCERK1 and OsCEBiP The cerk1-2 mutant line co-expressing LjNFR1-AtCERK1 from rice to the transmembrane and intracellular domains and LjNFR5-AtCERK1 perceives NFs of LjNFR1 and LjNFR5, respectively. The CaMV 35S promoter sequence was inserted upstream of these hybrid To test whether the constructed cerk1-2 derivatives receptor constructs (Figure 1). Agrobacterium rhizogenes show responsiveness to rhizobial NFs, preparations of was used to obtain transgenic hairy roots of L. japonicus purified NFs were infiltrated into test leaves with a syringe. NF receptor mutants (nfr1-1 and nfr5-2 derivatives). Trans- The NFs were obtained from cultures of Rhizobium sp. formed hairy roots were subjected to RT-PCR analysis with NGR234, a strain with a broad host-range that nodulates primers specific for OsCERK1 and OsCEBiP. As shown in L. japonicus (Hussain et al., 1999). Upon stimulation with Figure S1, corresponding amplicons were detected, NFs, cerk1-2 lines co-expressing LjNFR1–AtCERK1 and indicating expression of the OsCERK1–LjNFR1 and LjNFR5–AtCERK1 showed rapid generation of ROS as OsCEBiP–LjNFR5 chimeras. determined by DAB staining and a chemiluminescence assay with luminol (Figure 3a,b). No cell death was seen Co-expression of LjNFR1–AtCERK1 and LjNFR5–AtCERK1 when the NF-treated leaves were stained with trypan blue. in the A. thaliana cerk1-2 mutant does not result in Leaves of cerk1-2 lines expressing only one or no chimeras resistance to F. oxysporum that have been treated with NFs did not exhibit increased Arabidopsis thaliana cerk1 mutants lack a functional chitin ROS generation. Control material from NGRDnodABC, perception system and consequently do not accumulate a mutant deficient in NF synthesis (Price et al., 1992), ROS in response to chitin or chitooligosaccharides. Such showed no effects on any tested plants. In contrast to NFs, mutants are more susceptible to infection by biotrophic chitin and chitooligosaccharides were inactive in inducing fungal pathogens (Miya et al., 2007; Wan et al., 2008). To a ROS response in cerk1-2 lines co-expressing LjNFR1–At- examine whether constitutive expression of LjNFR1–At- CERK1 and LjNFR5–AtCERK1. Interestingly, NF also CERK1 and/or LjNFR5–AtCERK1 in the cerk1-2 mutant enhanced ROS levels in wild-type (Col-0) plants as well as affects pathogen-induced ROS formation, 10 ll (about in the cerk1-2 mutant complemented with AtCERK1 (Figure 1 9 103 spores) of a suspension of F. oxysporum f. sp. cu- S2), indicating that A. thaliana can recognize NFs in an bense race 4 were injected into the upper left quarter of a AtCERK1-dependent manner. given leaf. As shown in Figure 2(a,b), cerk1-2 plants Transgenic cerk1-2 lines expressing LjNFR1–AtCERK1 expressing either one or both chimeras did not show and/or LjNFR5–AtCERK1 were further used to test whether increased ROS formation as determined by staining the expressed chimeras mediate expression of chitin- with 30,30-diaminobenzidine (DAB). Accordingly, the trans- responsive genes in response to NFs. Transcript levels in formed plants were susceptible to F. oxysporum infection leaves were determined by qRT-PCR 2 h after infiltration in a similar way to the non-transformed cerk1-2 mutant as with NFs or water. When treated with NFs, the cerk1-2 line analyzed by the quantification of leaf chlorosis. In contrast, expressing both chimeras showed increased transcript wild-type (Col-0) plants and the cerk1-2 mutant comple- levels of the examined genes. However, no obvious stimu- mented with the full-length AtCERK1 gene showed an latory effects were seen for treatments with chitin, chitooli- apparent increased resistance to F. oxysporum, which was gosaccharides and control material from NGRDnodABC. accompanied by strong ROS accumulation and weak leaf Transcript levels also remained low in NF-infiltrated leaves chlorosis. As determined by trypan blue staining, no obvi- of cerk1-2 lines expressing LjNFR1–AtCERK1 or LjNFR5– ous cell death was detected 24 h post-inoculation in any of AtCERK1 alone, although a slight increase was observed the inoculated plants. compared to the non-transformed cerk1-2 mutant. Simi- Quantitative reverse transcription (qRT)-PCR was further larly to the ROS response, the NF treatment also induced performed to determine transcript levels of chitin-responsive transcript levels of the chitin-responsive genes in wild-type

Figure 2. Generation of reactive oxygen species (ROS), leaf chlorosis and defense gene expression in Fusarium oxysporum inoculated leaves of Ara- bidopsis thaliana cerk1-2 derivatives expressing chimeric lysin motif (LysM) domain receptor genes. Genotype abbreviations: cerk1-2, mutant of A. thaliana ecotype Columbia with mutated AtCERK1 gene; LjNFR1–AtCERK1, cerk1-2 expressing LjNFR1– AtCERK1; LjNFR5–AtCERK1, cerk1-2 expressing LjNFR5–AtCERK1; LjNFR1– AtCERK1 + LjNFR5–AtCERK1, cerk1-2 co-expressing LjNFR1–AtCERK1 and (a) LjNFR5–AtCERK1; Col-0, A. thaliana ecotype Columbia (wild-type); AtCERK1 complemented, cerk1-2 expressing AtCERK1 (under the control of the CaMV 35S promoter). (a) Fusarium oxysporum inoculated leaves stained with 30,30-diaminobenzidine (DAB) for ROS generation (upper pictures) or with trypan blue to visualize cell death (lower pictures) 24 h post-inoculation. For comparison, Col-0

wild-type plants were mock-inoculated with 10 mM MgSO4. Col-0 leaves treated with 2 M salicylic acid (SA) (harvested 3 h later) served as a positive control for cell death symptoms. (b) Chlorotic areas (expressed as a percentage of total leaf areas) 7 days post-inoculation. For comparison, mock-inoculated Col-0 plants were included in the analysis. Columns represent means SE from 60 leaves per genotype from three independent repeats. (c) Expression analysis of indicated defense-related genes in leaves by (b) quantitative RT-PCR. Leaves were harvested 2 h after inoculation of F. oxysporum. Mock-inoculated Col-0 plants were used as a reference. Data indicate relative expression levels (means SE) from three independent biological replica (three RNA extractions; n = 3).

chimeras can perceive NFs and mediate defense signaling, resulting in ROS generation and the induction of defense genes.

Pre-treatment of the cerk1-2 line co-expressing LjNFR1– AtCERK1 and LjNFR5–AtCERK1 with Nod factors results in resistance to F. oxysporum

To examine effects of pre-treatment with NFs on fungal infection, the different A. thaliana lines were ﬁrst treated with NFs or NGR234 bacteria and then inoculated with F. oxysporum. Leaf chlorosis was used as a measure of disease severity. Compared with the non-transformed cerk1-2 mutant, leaves of the cerk1-2 line co-expressing LjNFR1–AtCERK1 and LjNFR5–AtCERK1 showed nearly no (c) chlorosis when pre-treated with NFs. Pre-treatment with water or control material from NGRDnodABC showed no obvious effects and chlorotic disease symptoms were clearly visible (Figure 4a,b). Reduced leaf chlorosis was seen when the cerk1-2 line expressing both chimeras was treated with NGR234 bacteria prior to fungal inoculation, whereas NGRDnodABC bacteria or mock inoculation with

MgSO4 did not prevent formation of chlorotic disease symptoms (Figures 4a,b). Pre-treatment with chitin or chitooligosaccharides also did not reduce leaf chlorosis in the cerk1-2 line expressing both chimeras (Figure S3a). Similar to the non-transformed cerk1-2 mutant, cerk1-2 derivatives expressing a single chimeric receptor developed chlorotic disease symptoms that were not reduced by rhizobial pre- inoculation (Figure S3b,c). Wild-type (Col-0) plants were included in the experiment as controls and showed no visi- (Col-0) plants and in the cerk1-2 mutant complemented ble leaf chlorosis. with AtCERK1 (Figure 3c). Taken together, these ﬁndings Furthermore, F. oxysporum-inoculated leaves pre-trea- indicate that the LjNFR1–AtCERK1 and LjNFR5–AtCERK1 ted with NFs from NGR234 were stained with trypan blue

Figure 3. Generation of reactive oxygen species (ROS) and defense gene expression in Ara- bidopsis thaliana cerk1-2 derivatives expressing chimeric lysin motif (LysM) domain receptor genes after treatment with Nod factors (NFs) from Rhizobium sp. NGR234. For comparison, leaves were treated with (a) water, control material from NGRDnodABC, chitooligosacchrides and chitin. Genotype abbreviations are the same as for Figure 2. (a) Leaves of different genotypes stained either with 30,30-diaminobenzidine (DAB) or trypan blue 48 h after treatment with indicated chemicals. (b) Quantiﬁcation of ROS generation in leaves of different genotypes by a luminol-dependent chemiluminescence assay (at 30, 60, 90, and 120 min after treatment). Data shown in arbi- trary units (means SE) were obtained from a total of 20 leaves per genotype (from three independent experiments, n = 3). (c) Quantitative RT-PCR expression analysis for indicated defense genes in leaves of different genotypes. Plants were harvested 2 h after treatment with indicated chemicals. Data indi- (b) cate means SE from three independent RNA isolations (n = 3).

(c)

(a) (c)

(b) (d)

Figure 4. Effects of Nod factors (NFs) and NGR234 bacteria on subsequent fungal colonization in leaves of Arabidopsis thaliana cerk1-2 co-expressing LjNFR1- AtCERK1 and LjNFR5-AtCERK1.

Water, control material from NGRDnodABC,10mM MgSO4 (mock inoculation) and NGRDnodABC bacteria served as controls for the pre-treatment. Genotype abbreviations are the same as for Figure 2. (a) Examples of leaves of indicated genotypes pre-treated twice either with NFs or water prior to inoculation with Fusarium oxysporum. Photos were taken 7 days after fungal inoculation. (b) Leaf chlorosis of cerk1-2 expressing LjNFR1–AtCERK1 and LjNFR5–AtCERK1 pre-treated twice with NFs (left) and NGR234 bacteria (right) 7 days after F. oxysporum inoculation. Control plants were similarly pre-treated with water, control material from NGRDnodABC,10mM MgSO4 (mock inoculation) and NGRD- nodABC and then inoculated with the fungus. Data indicate means SE from 20 leaves. (c) Examples of hyphal structures formed in leaves of the indicated genotypes pre-treated twice with NFs from strain NGR234 and subsequently inoculated with F. oxysporum. Photographs were taken 10 days post-inoculation. Leaves were stained with trypan blue to visualize fungal structures (bar = 0.2 mm). (d) Degree of fungal colonization as quantified by the determination of hyphal structures in randomly selected square areas of the pictures taken (values expressed in percentages). Plants of indicated genotypes were pre-treated twice with NFs and harvested 7 days after fungal inoculation. Data indicate means SE from a total of 20 leaves per genotype (from three independent experiments, n = 3). to quantify differences in fungal colonization among the dif- roots (Schauser et al., 1999). Accordingly, the non-nodulat- ferent genotypes. The cerk1-2 line co-expressing LjNFR1– ing NF receptor mutant nfr1-1 (pNIN::GUS) (mutated in AtCERK1 and LjNFR5–AtCERK1 pre-treated with NFs as well LjNFR1 and carrying a NIN promoter–GUS gene construct) as wild-type (Col-0) plants showed reduced growth of does not show any GUS activity in response to either hyphae. By contrast, fungal colonization was significantly M. loti inoculation or purified NFs (Radutoiu et al., 2003). increased in cerk1-2 lines expressing only one or no Here, we tested whether nodulation signaling in L. japoni- chimera (Figure 4c,d). Hence, NFs can specifically induce cus nfr1-1 (pNIN::GUS) co-expressing the chimeras resistance to F. oxysporum in the cerk1-2 line expressing OsCERK1–LjNFR1 and OsCEBiP–LjNFR5 can be induced by both chimeras, which is in accordance with the observed chitin (Figure 1). Agrobacterium rhizogenes-transformed NF-induced defense responses in this line. hairy roots were treated with chitin or chitooligosaccharides and activity of the NIN promoter was visualized by Chitin activates nodulation signaling in L. japonicus NF GUS staining. No GUS activity was detected in roots of receptor mutants that co-express OsCERK1–LjNFR1 and OsCEBiP–LjNFR5 nfr1-1 (pNIN::GUS) expressing a single chimeric receptor or the empty vector in response to chitin or chitooctaose Expression of the NIN gene is a specific marker for activa- (Figure 5a). However, roots of nfr1-1 (pNIN::GUS) tion of NF-triggered nodulation signaling in L. japonicus co-expressing OsCERK1–LjNFR1 and OsCEBiP–LjNFR5

(a) (b) (c)

(d) (e)

Figure 5. Illustration of GUS staining reﬂecting NIN promoter activity induced by chitin (100 lgml1) or chitooctaose in hairy roots of nfr1-1 (pNIN::GUS) expressing OsCERK1–LjNFR1 and/or OsCEBiP–LjNFR5. (a) The GUS-stained hairy roots of nfr1-1 (pNIN::GUS) expressing indicated chimeric receptor 7 h after local application of 10 ll of chitin or chitooctaose 6 (10 M). For comparison, plants transformed with the empty vector pCAMBIA1305DCaMV35S were also analyzed (Control). (b) The GUS-stained hairy roots of nfr1-1 (pNIN::GUS) co-expressing OsCERK1–LjNFR1 and OsCEBiP–LjNFR5 7 h after local application of 10 ll chitin, indicated 6 1 chitooligosaccharides (10 M) or Nod factors (NFs) from Rhizobium sp. NGR234 (100 lgml ). (c) The GUS-stained hairy roots of nfr1-1 (pNIN::GUS) co-expressing OsCERK1–LjNFR1 and OsCEBiP–LjNFR5 7 h after local treatment with 10 ll chitooctaose at the indicated concentrations. 6 (d) The GUS-stained hairy roots of nfr1-1 (pNIN::GUS) co-expressing OsCERK1–LjNFR1 and OsCEBiP–LjNFR5 spot-treated with 10 ll chitooctaose (10 M) and incubated at 23°C for the indicated time period. (e) The GUS-stained hairy roots of wild-type Gifu (pNIN::GUS) plants co-expressing OsCERK1–LjNFR1 and OsCEBiP–LjNFR5 7 h after local application of 10 ll 6 1 chitin, indicating chitooligosaccharides (10 M) or NFs (100 lgml ). showed blue staining at the treatment site. In contrast to Further experiments were conducted with different chitooctaose, a local treatment with short chitooligosaccha- concentrations of chitooctaose and different incubation rides (DP = 4, 5 and 6) or NFs from Rhizobium sp. NGR234 periods before analysis by GUS staining. In roots of nfr1-1 did not induce GUS staining (Figure 5b). We also observed (pNIN::GUS) expressing both chimeras, local GUS staining –9 GUS staining for nfr5-2 (pNIN::GUS) plants (mutated in the was visible at a threshold level of 10 M chitooctaose LjNFR5 gene) co-expressing OsCERK1–LjNFR1 and Os- when plants were analyzed 7 h after the treatment CEBiP–LjNFR5 when roots were spot-treated with chitin or (Figure 5c). The incubation time with chitooctaose also chitooctaose. In contrast, such treatments did not induce inﬂuenced NIN promoter activity. Maximal GUS staining any GUS activity in roots of nfr5-2 (pNIN::GUS) transformed was reached at 6 h post-treatment, whereas an incubation with one chimera or without chimera (Figure S4a,b). As of 2 h or less resulted in no visible detection of GUS activ- expected, ‘wild-type’ plants [Gifu (pNIN::GUS)] showed ity (Figure 5d). GUS activity in response to locally applied NFs from Rhizo- For comparison, we also performed spot-treatment bium sp. NGR234, whereas a treatment with chitin or chi- experiments with Gifu (pNIN::GUS) ‘wild-type’ plants, tooligosaccharides had no visible effects (Figure S4c). Thus, which were transformed with OsCERK1–LjNFR1 and these data indicate that chitin or chitooctaose induce nodu- OsCEBiP–LjNFR5. Non-transformed Gifu (pNIN::GUS) lation signaling in hairy roots of test plants that express plants were included in the analysis. The GUS activity in both chimeras. hairy roots expressing both chimeras was locally activated

© 2014 The Authors The Plant Journal © 2014 John Wiley & Sons Ltd, The Plant Journal, (2014), 78, 56–69 64 Wei Wang et al. by NFs as well as by application of chitin or chitooctaose Similar to the NIN promoter activity analyzed in the previ- (Figure 5e). In contrast, normal ‘wild-type’ (pNIN::GUS) ous experiments, transcript levels of NSP1 and NSP2 were roots were only responsive to NFs and did not show any induced in nfr1-1 (pNIN::GUS) co-expressing OsCERK1– GUS activity when treated with chitin or chitooligosaccha- LjNFR1 and OsCEBiP–LjNFR5 in response to a chitin treat- rides (Figure S4c). These results substantiate our finding ment. Inoculation of these plants with M. loti mixed with that OsCERK1–LjNFR1 and OsCEBiP–LjNFR5 can mediate chitin resulted in a similar increase in transcripts. In con- activation of nodulation signaling in response to chitin or trast, nfr1-1 (pNIN::GUS) expressing one or no chimeras chitooctaose. did not show elevated gene expression. In wild-type plants We further examined whether a mixture of M. loti (strain inoculated with M. loti, however, NSP1 and NSP2 expres- MAFF303099) and chitin triggers nodulation responses in sion was increased, whereas the mixture containing bacte- our test plants. A bacterial suspension supplemented with ria and chitin was nearly inactive in inducing expression of chitin (100 lgml1) was added to nfr1-1 (pNIN::GUS) these genes (Figure S7a,b). Hence, chitin elicited the expressing one or both chimeras. Root-hair responses in expression of defense genes in all examined test plants GUS-stained hairy roots were microscopically examined and had a negative effect on expression of M. loti-induced 24 h after treatment (Figure S5a–d). In roots of nfr1-1 GRAS domain transcriptional regulators in wild-type (pNIN::GUS) co-expressing OsCERK1–LjNFR1 and Os- plants. CEBiP–LjNFR5, we observed balloon-like swelling of the DISCUSSION root-hair tips, which showed strong GUS staining (Figure S5e). Such root-hair responses were not seen for nfr1-1 The expression of receptors with replaced ectodomains in (pNIN::GUS) plants expressing a single chimera (Figure plants is a strategy to activate specific signaling pathways S5c,d). Mutant plants expressing both chimeras showed a by applied stimuli and thus to regulate cellular properties. similar swelling of root-hair tips and GUS staining when In our study, the chimeric receptor genes LjNFR1–AtCERK1 analyzed 7 h after chitin application (Figure S5 g,h). Histo- and LjNFR5–AtCERK1 were expressed in the cerk1-2 mutant chemical analysis revealed GUS-stained cells in the rhizo- of A. thaliana in order to induce AtCERK1-mediated plant dermis and the root cortex, whereas cells in the pericycle defense reactions by NF signals. When pre-treated with NFs and the central cylinder showed no staining (Figure S5f). or rhizobia, plants expressing both chimeras showed Although root-hair responses were visible, no infection increased resistance to F. oxysporum. Induced resistance threads or nodules were observed for nfr1-1 (pNIN::GUS) was accompanied by the formation of ROS and elevated co-expressing OsCERK1–LjNFR1 and OsCEBiP–LjNFR5 after transcript levels of chitin-responsive genes. We also show inoculation with a M. loti suspension containing chitin. that chitin or chitooctaose can trigger expression of a GUS Hence, exogenously applied chitin did not rescue the non- reporter gene driven by the NIN promoter in NF receptor nodulation phenotype of this NF receptor mutant. ‘Wild- mutants of L. japonicus that express the chimeras type’ Gifu (pNIN::GUS) plants inoculated with the mixture OsCERK1–LjNFR1 and OsCEBiP–LjNFR5. Our results indi- of bacteria and chitin formed about five nodules per plant cate that the chimeric receptors constructed in this study under the test conditions used. This prompted us to investi- were expressed as functional proteins and that the ectodo- gate to what extent defense genes are activated in our test mains of the chimeras possess recognition properties that plants. Based on previous reports (Wan et al., 2008; Nakag- are congruent to the original LysM domain receptors. Upon awa et al., 2011), qRT-PCR analysis was performed for four ligand binding, the non-modified protein kinase domains of defense gene markers (LjPRp27b, Chitinase, MPK3 and the chimeras seem to phosphorylate downstream targets, WRKY33). RNA was isolated from roots harvested 2 h after i.e. the substrates for AtCERK1 and LjNFR1, respectively. application of chitin, water (mock treatment), M. loti bacte- Arabidopsis thaliana cerk1-2 mutant plants expressing ria or a mixture of bacteria and chitin. Interestingly, test LjNFR1–AtCERK1 and LjNFR5–AtCERK1 showed defense plants [namely nfr1-1 (pNIN::GUS) expressing OsCERK1– signaling in response to NFs, whereas treatment with chi- LjNFR1 and/or OsCEBiP–LjNFR5; wild-type plants] showed tin or simple chitooligosaccharides had nearly no effect a similar pattern of defense gene expression. Compared (Figures 2 and 3). Wild-type (Col-0) A. thaliana plants, how- with the control, treatments with chitin as well as the mix- ever, seem to recognize NFs in an AtCERK1-dependent ture containing bacteria and chitin strongly induced tran- manner (Figure S2). This is reminiscent of the chitin per- script levels of the examined genes (up to 30-fold for ception system of tomato cells that can perceive NFs at Chitinase), whereas inoculation with bacteria alone nanomolar concentrations as measured by an extracellular resulted in a weak induction (about five-fold) (Figure S6). In alkalinization response (Staehelin et al., 1994). It is worth addition, we analyzed symbiotic signaling in this experi- noting in this context that defense responses induced by ment by quantifying transcripts of NSP1 and NSP2, which the flagellin peptide flg22 in A. thaliana and other plants encode GRAS domain transcriptional regulators required were partially suppressed by application of a NF from for NF-induced gene expression (Heckmann et al., 2006). Bradyrhizobium japonicum as well as by chitotetraose.

Further analysis of A. thaliana mutants revealed that the the NF-deficient mutant NGRDnodABC showed no effects observed effects depended on AtLYK3 (LysM-containing in our experiments. Remarkably, treatment with NFs was receptor-like kinase 3), whereas plants defective in more efficient than rhizobial inoculation (Figure 4b). This AtCERK1 were no different from wild-type plants in the might be due to the limited amounts of NFs produced by performed experiments (Liang et al., 2013). bacteria under the test conditions used. As NF production The NF receptor mutants nfr1-1 (pNIN::GUS)ornfr5-2 by NGR234 largely depends on legume flavonoids (Staeh- (pNIN::GUS)ofL. japonicus expressing OsCERK1–LjNFR1 elin et al., 1994), we supplemented the rhizobial inoculum and OsCEBiP–LjNFR5 remained insensitive to NFs as deter- with apigenin. Rhizobia that constitutively synthesize NFs mined by GUS staining of plants carrying the pNIN::GUS in the absence of flavonoids can be engineered (Spaink construct (Figures 5 and S4). However, GUS activity reflect- et al., 1989) and NGR234 modified in a similar way could ing nodulation signaling was induced when transformed be used in future tests. Hence, our experiments suggest hairy roots of these NF receptor mutants were treated with the possibility of using NF-producing rhizobia as biocontrol chitin or chitooctaose. Shorter chitooligosaccharides were bacteria in agriculture to enhance resistance to certain inactive in inducing nodulation signaling, indicating that pathogens in non-legumes. the chain length of the oligosaccharide is crucial for elicitor Co-expression of the NF receptor genes LjNFR1 and activity (Figures 5b,e and S4b). This is in accordance to the LjNFR5 in the non-legumes Nicotiana benthamiana and findings of previous studies, in which long-chain chitooli- leek (Allium ampeloprasum) caused a cell death response, gosaccharides were used to study OsCERK1/OsCEBiP-med- which was also observed in the absence of NFs. Analysis of iated chitin signaling in rice (e.g. Kaku et al., 2006; Iizasa mutant proteins revealed that cell death was only observed et al., 2010). when the kinase domain of LjNFR1 was functional (Nakaga- When only a single chimeric receptor was expressed, wa et al., 2011). Similarly, NF-independent rapid cell death ligand-induced responses were not observed in either the was observed in N. benthamiana co-expressing the NF A. thaliana cerk1-2 mutant or the NF receptor mutants of receptor genes MtNFP and MtLYK3 of M. truncatula (Pie- L. japonicus. These findings indicate that co-expression of traszewska-Bogiel et al., 2013). These effects are perhaps the constructed chimeric receptor pairs is essential for their due to constitutive activation of downstream MAP kinase function. We suggest that a heterodimeric interaction signaling. Indeed, over-expression of genes involved in occurs between the ectodomains of the constructed chime- MAP kinase signaling was reported to cause cell death ras. Our findings support recently published biochemical in tobacco plants (Yang et al., 2001; Zhang and Liu, 2001). data that provided evidence for the formation of a protein In our study, however, no cell death was found in the complex of the OsCERK1/OsCEBiP pair in rice (Shimizu A. thaliana cerk1-2 mutant co-transformed with LjNFR1–At- et al., 2010) and the LjNFR1/LjNFR5 pair in L. japonicus CERK1 and LjNFR5–AtCERK1, whereas ROS generation and (Madsen et al., 2011; Broghammer et al., 2012). defense gene activation were detected specifically in The correct conformational changes of receptors are response to NFs. These findings suggest that the con- essential for ligand-induced activation of their kinase structed chimeric receptors showed no or low constitutive domains. Transmembrane domains in receptors may play a kinase activity and that they were only activated upon crucial role in the ligand-dependent dimerization processes. ligand binding. Charged amino acid residues located in juxtamembrane Our experiments with L. japonicus NF receptor mutants and transmembrane regions can be particularly critical for indicate that nodulation signaling in the nfr1-1 (pNIN::GUS) the correct function of a given receptor (Constantinescu mutant co-expressing OsCERK1–LjNFR1 and OsCEBiP– et al., 2001; Davis et al., 2008). Therefore, we conserved the LjNFR5 was activated in response to chitin or chitooctaose. intracellular and transmembrane domains in our con- These findings suggest that OsCERK1–LjNFR1 interacts with structed chimeras to ensure optimal ligand-induced confor- OsCEBiP–LjNFR5, providing support for ligand-dependent mational changes and efficient activation of the kinase conformational changes of the hybrid receptors. These domain. We fused the ectodomain of OsCEBiP to the trans- results are in agreement with previous findings that the membrane/intracellular domain of LjNFR5 even though OsCERK1/OsCEBiP proteins are essential components of LjNFR5 has been described as lacking kinase activity (Mad- the chitin receptor complex in rice (Shimizu et al., 2010). sen et al., 2011). In future studies, it would be interesting to Although the constructed chimeras seem to be fully examine the expression, stability and interactions of the functional, the nfr1-1 (pNIN::GUS) mutant co-expressing constructed chimeras at the protein level. OsCERK1–LjNFR1 and OsCEBiP–LjNFR5 showed root-hair In the A. thaliana cerk1-2 mutant line expressing swelling but did not form nodules after inoculation with a LjNFR1–AtCERK1 and LjNFR5–AtCERK1, pre-treatment with mixture of M. loti and chitin. Hence, induction of nodula- NFs or rhizobia induced defense signaling and subsequent tion signaling by chitin probably induced feedback increased resistance to F. oxysporum. Nod factors seem to responses that prevented bacterial infection. A well-known be the only elicitors secreted by strain NGR234, because response is the so-called autoregulation phenomenon, in

© 2014 The Authors The Plant Journal © 2014 John Wiley & Sons Ltd, The Plant Journal, (2014), 78, 56–69 66 Wei Wang et al. which nodule formation and mycorrhization are controlled Lotus japonicus seedlings were placed on nitrogen-free half- by long-distance signals. This regulatory circuit enables strength B&D medium (Broughton and Dilworth, 1971) plates and ° the host plant to suppress further rhizobial or mycorrhizal cultivated at 24 C in a temperature-controlled growth room with a 16-h light/8-h dark cycle. Rice seedlings (O. sativa L. cv. Nippon- infection once symbiosis has been established (Ferguson bare) were kept at 30°C in a growth chamber with a 16-h light/8-h et al., 2010; Staehelin et al., 2011). Furthermore, activation dark cycle. All plasmids and bacterial strains used in this study are of chitin-induced defense reactions in hairy roots (Figure listed in Table S1. Fusarium oxysporum f. sp. cubense race 4 was S6) probably suppressed rhizobial infection. It is worth obtained from Dr Jianghui Xie (Chinese Academy of Tropical Agri- noting in this context that M. loti-inoculated wild-type cultural Sciences, Zhanjiang, China). Standard protocols were used for plasmid construction. The PCR primers are listed in Table S2. L. japonicus formed only about five nodules in the presence of chitin. This is reminiscent of effects caused by Construction of chimeric receptor genes and plant application of the flagellin peptide flg22, which resulted in transformation significantly reduced nodule formation in L. japonicus Plasmid DNAs containing the coding sequences of LjNFR1 (Lopez-Gomez et al., 2012). Future work is required to (AJ575248) and LjNFR5 (AJ575255) from L. japonicus were kindly examine chitin perception and the effect of chitin-induced provided by Dr Zhong-Ming Zhang (Huazhong Agricultural Univer- defense gene activation on nodulation signaling. The chitin sity, Wuhan, China). The AtCERK1 (AB367524) gene from A. thali- receptor genes of L. japonicus have not been identified but ana ecotype Col-0 as well as OsCERK1 (AK111766) and OsCEBiP (AB206975) from O. sativa were PCR-cloned into plasmids using probably exist among the 17 sequenced LysM domain cDNA derived from total plant RNA isolated with TRIZOL reagent receptor genes (Lohmann et al., 2010). (Invitrogen, http://www.invitrogen.com/). In conclusion, our work shows that non-legumes can be For construction of the chimeric receptor gene LjNFR1– modified to initiate defense reactions in response to NFs AtCERK1, the ectodomain of AtCERK1 (amino acid residues 1–234) instead of chitin. This opens the possibility of using harm- was replaced by the ectodomain of LjNFR1 (amino acid residues – less rhizobia as biocontrol bacteria. We also found that 1 224). Similarly, the ectodomain of AtCERK1 was replaced by the ectodomain of LjNFR5 (amino acid residues 1–234) to generate NF-insensitive L. japonicus mutants expressing receptors LjNFR5–AtCERK1 (see Figure 1). The chimeras were then inserted with altered ectodomains are able to initiate nodulation into pRT104 (Topfer€ et al., 1987), which carries the CaMV 35S pro- signaling in response to chitin. Although rhizobial infection moter and a poly-A signal. Finally, all the expression cassettes was blocked in these mutants, such engineered plants were cloned into pCAMBIA1305.1 (http://www.cambia.org), which could be used in future to study the relationship between contains a GUS gene under the control of the CaMV 35S promoter. The plasmids (see Table S1) were then introduced into defense reactions and nodule formation. Furthermore, A. tumefaciens EHA105 by electroporation. Transgenic A. thaliana plants expressing chimeric LysM domain receptors could plants (cerk1-2 derivatives) were obtained by the floral-dip method be constructed to examine to what extent the establish- and selected on 1% (w/v) agar containing MS medium (Duchefa, ment of symbiosis with mycorrhizal fungi is stimulated by http://www.duchefa-biochemie.nl/) and 50 mg l 1 kanamycin application of NFs, related fungal Myc factors or chitin. (Zhang et al., 2006). – Indeed, induction of symbiotic signaling by application of For construction of OsCERK1 LjNFR1, the ectodomain of LjNFR1 (amino acid residues 1–224) was replaced by the ectodomain of NFs promoted mycorrhization in various legumes (Xie OsCERK1 (amino acid residues 1–237). Similarly, OsCEBiP–LjNFR5 et al., 1995, 1997; Olah et al., 2005). was constructed by replacing the ectodomain of LjNFR5 (amino acid residues 1–246) with the ectodomain of OsCEBiP (amino acid EXPERIMENTAL PROCEDURES residues 1–334). After insertion into pRT104, the expression cassettes with CaMV 35S promoters were cloned into pCAM- Biological material, plasmids and primers BIA1305DCaMV35S, which is a pCAMBIA1305 derivative lacking the CaMV 35S promoter upstream of the GUS gene sequence (no The chitin receptor mutant cerk1-2 of Arabidopsis thaliana ecotype constitutive GUS expression in transformed plants). Plasmids Columbia (Col-0) (SALK_007193C; donated by Dr Joseph Ecker, (Table S1) were then transferred into A. rhizogenes LBA9402 by The Salk Institute for Biological Studies, La Jolla, CA, USA) was electroporation. Transgenic hairy roots on L. japonicus mutant obtained from the Arabidopsis Biological Resource Center at Ohio plants (grown on Fahraeus medium plates) were obtained accord- State University (http://abrc.osu.edu/). This mutant contains a ing to previously described procedures (Chabaud et al., 2006; Nak- T-DNA insertion that renders AtCERK1 non-functional as described agawa et al., 2011). previously (Miya et al., 2007). Homozygous plants were identified by RT-PCR with gene-specific primers (Table S2). Arabidopsis thali- Chitin, chitooligosaccharides and NFs ana plants were grown under long-day conditions (16-h light/8-h dark, 90–110 lEm2 sec1,22°C). Mutants of the L. japonicus (eco- Details of the chitin, chitooligosaccharides, NFs from Rhizobium D type Gifu) nfr1-1 (pNIN::GUS), nfr5-2 (pNIN::GUS) and Gifu (pNIN:: sp. strain NGR234 and the control material from NGR nodABC GUS) expressing the GUS gene uidA under the control of the sym- (Price et al., 1992) used in this work are described in Methods S1. biotic NIN gene promoter (Radutoiu et al., 2003) were obtained Treatment of A. thaliana plants from Dr Jens Stougaard (University of Aarhus, Aarhus, Denmark) and confirmed by nodulation tests (non-nodulation phenotype) Four-week-old A. thaliana plants were used for all treatments with with M. loti strain MAFF303099 (Japan Collection of Microorgan- chemicals and inoculation experiments as described in Methods isms; RIKEN BioResource Center, http://www.brc.riken.jp/inf/en/). S2.

Treatment of L. japonicus plants Methods S1. Chitin, chitooligosaccharides and Nod factors. Methods S2. Treatment of Arabidopsis thaliana plants. Lotus japonicus genotypes with a pNIN::GUS construct were Methods S3. Treatment of Lotus japonicus plants. locally treated with NFs, chitooligosaccharides, chitin and M. loti Methods S4. Cell death and measurements of reactive oxygen strain MAFF303099 as summarized in Methods S3. species. Methods S5. Quantitative reverse transcription (qRT)-PCR. Cell death and ROS measurements REFERENCES Methods used for cell death staining with trypan blue and for Akamatsu, A., Wong, H.L., Fujiwara, M. et al. (2013) An OsCEBiP/Os- ROS measurement with luminol or DAB are described in Methods CERK1-OsRacGEF1-OsRac1 module is an essential early component of S4. chitin-induced rice immunity. Cell Host Microbe, 13, 465–476. Albert, M., Jehle, A.K., Mueller, K., Eisele, C., Lipschis, M. and Felix, G. Quantitative reverse transcription (qRT)-PCR (2010) Arabidopsis thaliana pattern recognition receptors for bacterial elongation factor Tu and flagellin can be combined to form functional Quantitative RT-PCR was performed as summarized in Methods chimeric receptors. J. Biol. Chem. 285, 19035–19042. S5. The primers used are listed in Table S2. Arrighi, J.F., Barre, A., Ben Amor, B. et al. (2006) The Medicago truncatula lysine motif-receptor-like kinase gene family includes NFP and new nod- ACKNOWLEDGEMENTS ule-expressed genes. Plant Physiol. 142, 265–279. We are grateful to Dr Jens Stougaard (University of Aarhus, Aar- Boller, T. and Felix, G. (2009) A renaissance of elicitors: perception of hus, Denmark) for L. japonicus mutants and Dr Zhong-Ming Zhang microbe-associated molecular patterns and danger signals by pattern-recognition receptors. Annu. Rev. Plant Biol. 60, 379–406. (Huazhong Agricultural University, Wuhan, China) for plasmids Broghammer, A., Krusell, L., Blaise, M. et al. (2012) Legume receptors per- containing LjNFR1 and LjNFR5 genes. We also thank Dr Jianghui ceive the rhizobial lipochitin oligosaccharide signal molecules by direct Xie (Chinese Academy of Tropical Agricultural Sciences, Zhanjiang, binding. Proc. Natl Acad. Sci. USA, 109, 13859–13864. China) for F. oxysporum and Dr William J. Broughton (University Broughton, W.J. and Dilworth, M.J. (1971) Control of leghaemoglobin of Geneva, Geneva, Switzerland) for strain NGRDnodABC.We synthesis in snake beans. Biochem. J. 125, 1075–1080. thank Qi Sun, Li-Ming Liang, Feng Yang and Dr Christian Wagner Brutus, A., Sicilia, F., Macone, A., Cervone, F. and De Lorenzo, G. (2010) A (Sun Yat-sen University) for their help with various aspects of this domain swap approach reveals a role of the plant wall-associated kinase work. This study was supported by the National Basic Research 1 (WAK1) as a receptor of oligogalacturonides. Proc. Natl Acad. Sci. – Program of China (973 program, no. 2010CB126501), by USA, 107, 9452 9457. Chabaud, M., Boisson-Dernier, A., Zhang, J., Taylor, C.G., Yu, O. and Barker, the National Natural Science Foundation of China (grant D.G. (2006) Agrobacterium rhizogenes-Mediated Root Transformation. 31070219), by the Guangdong Natural Science Foundation (grant The Samuel Roberts Noble Foundation, Ardmore, OK, USA: The Medica- 10251027501000014), by the Science Foundation of the State Key go truncatula Handbook. Laboratory of Biocontrol (grant SKLBC201123) and by the Guang- Constantinescu, S.N., Keren, T., Socolovsky, M., Nam, H.S., Henis, Y.I. and dong Key Laboratory of Plant Resources (grant plant01k19). Lodish, H.F. (2001) Ligand-independent oligomerization of cell-surface erythropoietin receptor is mediated by the transmembrane domain. Proc. SUPPORTING INFORMATION Natl Acad. Sci. USA, 98, 4379–4384. Davis, T.L., Walker, J.R., Loppnau, P., Butler-Cole, C., Allali-Hassani, A. and Additional Supporting Information may be found in the online Dhe-Paganon, S. (2008) Autoregulation by the juxtamembrane region of version of this article. the human ephrin receptor tyrosine kinase A3 (EphA3). Structure, 16, Figure S1. Transcript detection of constructed chimeric LysM 873–884. domain receptor genes in Arabidopsis thaliana and Lotus japoni- Faulkner, C., Petutschnig, E., Benitez-Alfonso, Y., Beck, M., Robatzek, S., Lipka, cus mutants by RT-PCR. V. and Maule, A.J. (2013) LYM2-dependent chitin perception limits molecular flux via plasmodesmata. Proc. Natl Acad. Sci. USA, 110,9166–9170. Figure S2. 30,30-Diaminobenzidine staining of Arabidopsis thaliana Felix, G., Regenass, M. and Boller, T. (1993) Specific perception of sub- wild-type (Col-0) and cerk1-2 mutant leaves after treatment with nanomolar concentrations of chitin fragments by tomato cells: induction D Nod factors from NGR234 or control material from NGR nodABC. of extracellular alkalinization, changes in protein phosphorylation, and Figure S3. Chlorotic areas in Fusarium oxysporum inoculated establishment of a refractory state. Plant J. 4, 307–316. leaves of Arabidopsis thaliana cerk1-2 mutant derivatives after Ferguson, B.J., Indrasumunar, A., Hayashi, S., Lin, M.H., Lin, Y.H., Reid, D.E. various pre-treatments. and Gresshoff, P.M. (2010) Molecular analysis of legume nodule develop- Figure S4. Hairy roots of Lotus japonicus nfr5-2 (pNIN::GUS) ment and autoregulation. J. Integr. Plant Biol. 52,61–76. co-expressing OsCERK1–LjNFR1 and/or OsCEBiP–LjNFR5 spot- Genre, A., Chabaud, M., Balzergue, C. et al. (2013) Short-chain chitin 2+ treated with various chemicals and stained for GUS activity. oligomers from arbuscular mycorrhizal fungi trigger nuclear Ca spiking in Medicago truncatula roots and their production is enhanced by strigo- Figure S5. Root-hair swelling and GUS activity in hairy roots of – – lactone. New Phytol. 198, 190 202. Lotus japonicus nfr1-1 (pNIN::GUS) co-expressing OsCERK1 Gust, A.A., Biswas, R., Lenz, H.D. et al. (2007) Bacteria-derived peptidogly- – LjNFR1 and OsCEBiP LjNFR5 in response to a mixture of chitin cans constitute pathogen-associated molecular patterns triggering innate and Mesorhizobium loti or chitin alone. immunity in Arabidopsis. J. Biol. Chem. 282, 32338–32348. Figure S6. Quantitative RT-PCR analysis of defense genes in Gust, A.A., Willmann, R., Desaki, Y., Grabherr, H.M. and Nurnberger,€ T. hairy roots of Lotus japonicus nfr1-1 (pNIN::GUS) expressing (2012) Plant LysM proteins: modules mediating symbiosis and immunity. chimeric receptors after application of chitin and/or Mesorhizo- Trends Plant Sci. 17, 495–502. bium loti. Hamel, L.P. and Beaudoin, N. (2010) Chitooligosaccharide sensing and Figure S7. Quantitative RT-PCR analysis of NSP1 and NSP2 downstream signaling: contrasted outcomes in pathogenic and beneficial plant–microbe interactions. Planta, 232, 787–806. genes in hairy roots of Lotus japonicus nfr1-1 (pNIN::GUS) He, Z., Wang, Z.Y., Li, J., Zhu, Q., Lamb, C., Ronald, P. and Chory, J. (2000) expressing chimeric receptors after application of chitin and/or Perception of brassinosteroids by the extracellular domain of the recep- Mesorhizobium loti. tor kinase BRI1. Science, 288, 2360–2363. Table S1. Strains and plasmids used in this study. Heckmann, A.B., Lombardo, F., Miwa, H., Perry, J.A., Bunnewell, S., Table S2. Primers used in this study. Parniske, M., Wang, T.L. and Downie, J.A. (2006) Lotus japonicus

nodulation requires two GRAS domain regulators, one of which is func- Nakagawa, T., Kaku, H., Shimoda, Y., Sugiyama, A., Shimamura, M., Takan- tionally conserved in a non-legume. Plant Physiol. 142, 1739–1750. ashi, K., Yazaki, K., Aoki, T., Shibuya, N. and Kouchi, H. (2011) From Hussain, A.A., Jiang, Q., Broughton, W.J. and Gresshoff, P.M. (1999) Lotus defense to symbiosis: limited alterations in the kinase domain of LysM japonicus nodulates and fixes nitrogen with the broad host range Rhizo- receptor-like kinases are crucial for evolution of legume-Rhizobium bium sp. NGR234. Plant Cell Physiol. 40, 894–899. symbiosis. Plant J. 65, 169–180. Iizasa, E.I., Mitsutomi, M. and Nagano, Y. (2010) Direct binding of a plant Olah, B., Briere, C., Becard, G., Denari e, J. and Gough, C. (2005) Nod factors LysM receptor-like kinase, LysM RLK1/CERK1, to chitin in vitro. J. Biol. and a diffusible factor from arbuscular mycorrhizal fungi stimulate lateral Chem. 285, 2996–3004. root formation in Medicago truncatula via the DMI1/DMI2 signalling Indrasumunar, A., Kereszt, A., Searle, I., Miyagi, M., Li, D., Nguyen, C.D., pathway. Plant J. 44, 195–207. Men, A., Carroll, B.J. and Gresshoff, P.M. (2010) Inactivation of dupli- Oldroyd, G.E. (2013) Speak, friend, and enter: signalling systems that cated Nod factor receptor 5 (NFR5) genes in recessive loss-of-function promote beneficial symbiotic associations in plants. Nat. Rev. Microbiol. non-nodulation mutants of allotetraploid soybean (Glycine max L. Merr.). 11, 252–263. Plant Cell Physiol. 51, 201–214. Perret, X., Staehelin, C. and Broughton, W.J. (2000) Molecular basis of Indrasumunar, A., Searle, I., Lin, M.H., Kereszt, A., Men, A., Carroll, B.J. symbiotic promiscuity. Microbiol. Mol. Biol. Rev. 64, 180–201. and Gresshoff, P.M. (2011) Nodulation factor receptor kinase 1a controls Petutschnig, E.K., Jones, A.M., Serazetdinova, L., Lipka, U. and Lipka, V. nodule organ number in soybean (Glycine max L. Merr). Plant J. 65, (2010) The lysin motif receptor-like kinase (LysM-RLK) CERK1 is a major 39–50. chitin-binding protein in Arabidopsis thaliana and subject to Kaku, H., Nishizawa, Y., Ishii-Minami, N., Akimoto-Tomiyama, C., Dohmae, chitin-induced phosphorylation. J. Biol. Chem. 285, 28902–28911. N., Takio, K., Minami, E. and Shibuya, N. (2006) Plant cells recognize Pietraszewska-Bogiel, A., Lefebvre, B., Koini, M.A., Klaus-Heisen, D., Tak- chitin fragments for defense signaling through a plasma membrane ken, F.L., Geurts, R., Cullimore, J.V. and Gadella, T.W. (2013) Interaction receptor. Proc. Natl Acad. Sci. USA, 103, 11086–11091. of Medicago truncatula lysin motif receptor-like kinases, NFP and LYK3, Ke, D., Fang, Q., Chen, C., Zhu, H., Chen, T., Chang, X., Yuan, S., Kang, H., produced in Nicotiana benthamiana induces defence-like responses. Hong, Z. and Zhang, Z. (2012) The small GTPase ROP6 interacts with PLoS One, 8, e65055. NFR5 and is involved in nodule formation in Lotus japonicus. Plant Phys- Price, N.P.J., Relic, B., Talmont, F., Lewin, A., Prome, D., Pueppke, S.G., iol. 159, 131–143. Maillet, F., Denari e, J., Prome, J.C. and Broughton, W.J. (1992) Broad-- Kishi-Kaboshi, M., Okada, K., Kurimoto, L. et al. (2010) A rice fungal host-range Rhizobium species strain NGR234 secretes a family of carba- MAMP-responsive MAPK cascade regulates metabolic flow to antimicro- moylated, and fucosylated, nodulation signals that are O-acetylated or bial metabolite synthesis. Plant J. 63, 599–612. sulphated. Mol. Microbiol. 6, 3575–3584. Kishimoto, K., Kouzai, Y., Kaku, H., Shibuya, N., Minami, E. and Nishizawa, Radutoiu, S., Madsen, L.H., Madsen, E.B. et al. (2003) Plant recognition of Y. (2010) Perception of the chitin oligosaccharides contributes to disease symbiotic bacteria requires two LysM receptor-like kinases. Nature, 425, resistance to blast fungus Magnaporthe oryzae in rice. Plant J. 64, 343– 585–592. 354. Schauser, L., Roussis, A., Stiller, J. and Stougaard, J. (1999) A plant regulator Kouzai, Y., Kaku, H., Shibuya, N., Minami, E. and Nishizawa, Y. (2013) controlling development of symbiotic root nodules. Nature, 402, 191–195. Expression of the chimeric receptor between the chitin elicitor receptor Shimizu, T., Nakano, T., Takamizawa, D. et al. (2010) Two LysM receptor CEBiP and the receptor-like protein kinase Pi-d2 leads to enhanced molecules, CEBiP and OsCERK1, cooperatively regulate chitin elicitor sig- responses to the chitin elicitor and disease resistance against Magnapor- naling in rice. Plant J. 64, 204–214. the oryzae in rice. Plant Mol. Biol. 81, 287–295. Shinya, T., Motoyama, N., Ikeda, A., Wada, M., Kamiya, K., Hayafune, M., Liang, Y., Cao, Y., Tanaka, K., Thibivilliers, S., Wan, J., Choi, J., Kang, C., Kaku, H. and Shibuya, N. (2012) Functional characterization of CEBiP and Qiu, J. and Stacey, G. (2013) Nonlegumes respond to rhizobial Nod CERK1 homologs in Arabidopsis and rice reveals the presence of differ- factors by suppressing the innate immune response. Science, 341, ent chitin receptor systems in plants. Plant Cell Physiol. 53, 1696–1706. 1384–1387. Soyano, T., Kouchi, H., Hirota, A. and Hayashi, M. (2013) Nodule inception Limpens, E., Franken, C., Smit, P., Willemse, J., Bisseling, T. and Geurts, R. directly targets NF-Y subunit genes to regulate essential processes of (2003) LysM domain receptor kinases regulating rhizobial Nod root nodule development in Lotus japonicus. PLoS Genet. 9, e1003352. factor-induced infection. Science, 302, 630–633. Spaink, H.P., Okker, R.J., Wijffelman, C.A., Tak, T., Goosen-de Roo, L., Pees, Liu, B., Li, J.F., Ao, Y. et al. (2012a) Lysin motif–containing proteins LYP4 E., van Brussel, A.A. and Lugtenberg, B.J. (1989) Symbiotic properties of and LYP6 play dual roles in peptidoglycan and chitin perception in rice rhizobia containing a flavonoid-independent hybrid nodD product. innate immunity. Plant Cell, 24, 3406–3419. J. Bacteriol. 171, 4045–4053. Liu, T., Liu, Z., Song, C. et al. (2012b) Chitin-induced dimerization activates Staehelin, C., Granado, J., Muller,€ J., Wiemken, A., Mellor, R.B., Felix, G., a plant immune receptor. Science, 336, 1160–1164. Regenass, M., Broughton, W.J. and Boller, T. (1994) Perception of Rhizo- Lohmann, G.V., Shimoda, Y., Nielsen, M.W. et al. (2010) Evolution and reg- bium nodulation factors by tomato cells and inactivation by root chitin- ulation of the Lotus japonicus LysM receptor gene family. Mol. Plant ases. Proc. Natl Acad. Sci. USA, 91, 2196–2200. Microbe Interact. 23, 510–521. Staehelin, C., Xie, Z.P., Illana, A. and Vierheilig, H. (2011) Long-distance Lopez-Gomez, M., Sandal, N., Stougaard, J. and Boller, T. (2012) Interplay transport of signals during symbiosis: are nodule formation and myco- of flg22-induced defence responses and nodulation in Lotus japonicus. J. rrhization autoregulated in a similar way? Plant Signal. Behav. 6, Exp. Bot. 63, 393–401. 372–377. Madsen, E.B., Madsen, L.H., Radutoiu, S. et al. (2003) A receptor kinase Topfer,€ R., Matzeit, V., Gronenborn, B., Schell, J. and Steinbiss, H.H. (1987) gene of the LysM type is involved in legume perception of rhizobial A set of plant expression vectors for transcriptional and translational signals. Nature, 425, 637–640. fusions. Nucleic Acids Res. 15, 5890. Madsen, E.B., Antolin-Llovera, M., Grossmann, C. et al. (2011) Autophos- Wan, J., Zhang, S. and Stacey, G. (2004) Activation of a mitogen-activated phorylation is essential for the in vivo function of the Lotus japonicus protein kinase pathway in Arabidopsis by chitin. Mol. Plant Pathol. 5, Nod factor receptor 1 and receptor-mediated signalling in cooperation 125–135. with Nod factor receptor 5. Plant J. 65, 404–417. Wan, J., Zhang, X.C., Neece, D., Ramonell, K.M., Clough, S., Kim, S.Y., Sta- Maillet, F., Poinsot, V., Andre, O. et al. (2011) Fungal lipochitooligosaccha- cey, M.G. and Stacey, G. (2008) A LysM receptor-like kinase plays a criti- ride symbiotic signals in arbuscular mycorrhiza. Nature, 469,58–63. cal role in chitin signaling and fungal resistance in Arabidopsis. Plant Miya, A., Albert, P., Shinya, T., Desaki, Y., Ichimura, K., Shirasu, K., Naru- Cell, 20, 471–481. saka, Y., Kawakami, N., Kaku, H. and Shibuya, N. (2007) CERK1, a LysM Wan, J., Tanaka, K., Zhang, X.C., Son, G.H., Brechenmacher, L., Nguyen, receptor kinase, is essential for chitin elicitor signaling in Arabidopsis. T.H. and Stacey, G. (2012) LYK4, a lysin motif receptor-like kinase, is Proc. Natl Acad. Sci. USA, 104, 19613–19618. important for chitin signaling and plant innate immunity in Arabidopsis. Mueller, K., Bittel, P., Chinchilla, D., Jehle, A.K., Albert, M., Boller, T. and Plant Physiol. 160, 396–406. Felix, G. (2012) Chimeric FLS2 receptors reveal the basis for differential Willmann, R., Lajunen, H.M., Erbs, G. et al. (2011) Arabidopsis lysin-motif flagellin perception in Arabidopsis and tomato. Plant Cell, 24, 2213–2224. proteins LYM1 LYM3 CERK1 mediate bacterial peptidoglycan sensing

and immunity to bacterial infection. Proc. Natl Acad. Sci. USA, 108, phosphorylated by the chitin receptor and mediates rice immunity. Cell 19824–19829. Host Microbe, 13, 347–357. Xie, Z.P., Staehelin, C., Vierheilig, H., Wiemken, A., Jabbouri, S., Broughton, Yang, K.Y., Liu, Y. and Zhang, S. (2001) Activation of a mitogen-activated W.J., Vogeli-Lange,€ R. and Boller, T. (1995) Rhizobial nodulation factors protein kinase pathway is involved in disease resistance in tobacco. Proc. stimulate mycorrhizal colonization of nodulating and nonnodulating soy- Natl Acad. Sci. USA, 98, 741–746. beans. Plant Physiol. 108, 1519–1525. Zhang, S. and Liu, Y. (2001) Activation of salicylic acid–induced protein Xie, Z.P., Muller,€ J., Wiemken, A., Broughton, W.J. and Boller, T. (1997) Nod kinase, a mitogen-activated protein kinase, induces multiple defense factors and tri-iodobenzoic acid stimulate mycorrhizal colonization and responses in tobacco. Plant Cell, 13, 1877–1889. affect carbohydrate partitioning in mycorrhizal roots of Lablab purpure- Zhang, X., Henriques, R., Lin, S.S., Niu, Q.W. and Chua, N.H. (2006) Agro- us. New Phytol. 139, 361–366. bacterium-mediated transformation of Arabidopsis thaliana using the ﬂo- Yamaguchi, K., Yamada, K., Ishikawa, K. et al. (2013) A receptor-like ral dip method. Nat. Protoc. 1, 641–646. cytoplasmic kinase targeted by a plant pathogen effector is directly

Supporting Information: Figures (Wei Wang, Zhi-Ping Xie and Christian Staehelin. Functional analysis of chimeric LysM domain receptors mediating Nod factor-induced defense signaling in Arabidopsis thaliana and chitin-induced nodulation signaling in Lotus japonicus)

Figure S1. Examples for transcript detection of chimeric LysM domain receptors in transformed A. thaliana and L. japonicus mutant plants by RT-PCR. The 18S rRNA gene was used as internal control. Used primers are listed in Table S2. (a) Detection of LjNFR1-AtCERK1 and LjNFR5-AtCERK1 transcripts in transformed A. thaliana cerk1-2 mutant plants with primers specific for the ectodomains of LjNFR1 and LjNFR5. (b) Detection of OsCERK1-LjNFR1 and OsCEBiP-LjNFR5 transcripts in transformed hairy roots of the L. japonicus nfr1-1 (pNIN::GUS) mutant with primers specific for the ectodomains of OsCERK1 and OsCEBiP. Wang et al. (Figures S1-S7) ̶ 2

Figure S2. DAB staining of A. thaliana wild-type (Col-0) and cerk1-2 mutant plants after application of NFs. Purified NFs (100 μg/ml) from Rhizobium sp. NGR234 or corresponding control material from NGRΔnodABC were infiltrated into the upper left quarter of a given leaf. (a) Leaves of wild-type plants stained with DAB 24 h after infiltration with NFs or control material from NGRΔnodABC. (b) Leaves of the cerk1-2 mutant stained with DAB 24 h after infiltration with NFs or control material from NGRΔnodABC. Wang et al. (Figures S1-S7) ̶ 3

Figure S3. Quantification of chlorotic areas induced by F. oxysporum in leaves of A. thaliana cerk1-2 mutant derivatives after pretreatment with NFs, chitin, chitooligosaccharides or rhizobia (Rhizobium sp. strain NGR234 and mutant NGRΔnodABC). Data indicate means ± SE from 20 leaves per genotype. Genotype abbreviations are the same as for Figure 2. (a) Leaf chlorosis in cerk1-2 co-expressing LjNFR1-AtCERK1 and LjNFR5-AtCERK1 pretreated twice with chitin or indicated chitooligosaccharides prior inoculation with F. oxysporum. (b) Leaf chlorosis of indicated genotypes pretreated twice with water, NFs from NGR234 or control material from NGRΔnodABC prior inoculation with F. oxysporum (data for cerk1-2 co-expressing LjNFR1-AtCERK1 and LjNFR5-AtCERK1 are shown in Figure 4).

(c) Leaf chlorosis of indicated genotypes pretreated twice with 10 mM MgSO4, with NGR234 bacteria or with NGRΔnodABC bacteria prior inoculation with F. oxysporum (data for cerk1-2 co-expressing LjNFR1-AtCERK1 and LjNFR5-AtCERK1 are shown in Figure 4). Wang et al. (Figures S1-S7) ̶ 4

Figure S4. GUS staining reflecting NIN promoter activity in hairy roots of nfr5-2 (pNIN::GUS) expressing OsCERK1-LjNFR1 and/or OsCEBiP-LjNFR5. For comparison, non-transformed Gifu (pNIN::GUS) plants were also analyzed. Plants were spot-treated with 10 μL of indicated chemicals (100 μg/ml chitin; chitooligosaccharides at 10-6 M; 100 μg/ml NFs from Rhizobium sp. NGR234). Plants were harvested 7 h later and then stained for GUS activity. (a) Hairy roots of nfr5-2 (pNIN::GUS) expressing either OsCERK1-LjNFR1 or OsCEBiP-LjNFR5 alone spot-treated with chitin or chitooctaose. Hairy roots transformed with the empty vector pCAMBIA1305ΔCaMV35S served as an additional control. (b) Hairy roots of nfr5-2 (pNIN::GUS) co-expressing OsCERK1-LjNFR1 and OsCEBiP-LjNFR5 after treatment with indicated chemicals. (c) Non-transformed roots of Gifu (pNIN::GUS) after treatment with indicated chemicals.

Wang et al. (Figures S1-S7) ̶ 5

Figure S5. Root-hair swelling and GUS activity in hairy roots of L. japonicus nfr1-1 (pNIN::GUS) co-expressing OsCERK1-LjNFR1 and OsCEBiP-LjNFR5 in response to a mixture of chitin (100 μg/ml) and M. loti strain MAFF303099 (panels a-e) or chitin alone (panels f-h). For comparison, other genotypes were also included. Plants were stained for GUS activity at the time of harvest. (a) Typical root-hair curling of a L. japonicus wild-type (Gifu) plant 24 h post inoculation (bar = 30 μm). (b) Root-hairs of the nfr1-1 (pNIN::GUS) mutant 24 h post inoculation (bar = 50 μm). (c) Root-hairs of nfr1-1 (pNIN::GUS) expressing OsCEK1-LjNFR1 24 h post inoculation (bar = 30 μm). (d) Root-hairs of nfr1-1 (pNIN::GUS) expressing OsCEBiP-LjNFR5 24 h post inoculation (bar = 30 μm). (e) Balloon-like swelling and strong GUS staining at root-hair tips in nfr1-1 (pNIN::GUS) co-expressing OsCEK1-LjNFR1 and OsCEBiP-LjNFR5 24 h post inoculation (bar = 30 μm). (f) Cross-section of a hairy root of nfr1-1 (pNIN::GUS) co-expressing OsCERK1-LjNFR1 and OsCEBiP-LjNFR5. Plants were harvested 24 h after treatment with chitin. In total, 30 hairy roots were analyzed in triple and all showed a similar GUS staining pattern (bar = 30 μm). (g) Hairy root of nfr1-1 (pNIN::GUS) co-expressing OsCERK1-LjNFR1 and OsCEBiP-LjNFR5 harvested 24 h after treatment with chitin (bar = 30 μm). (h) Ballon-like swelling and GUS staining at root-hair tips of nfr1-1 (pNIN::GUS) co-expressing OsCEK1-LjNFR1 and OsCEBiP-LjNFR5. Plants were harvested 24 h after treatment with chitin (bar = 30 μm). Wang et al. (Figures S1-S7) ̶ 6

Figure S6. qRT-PCR analysis of defense genes in hairy roots of L. japonicus nfr1-1 (pNIN::GUS) expressing indicated chimeras. Roots were treated with water (mock inoculation), 1 ml chitin (100 μg/ml), a 2 ml suspension of M. loti strain MAFF303099 or a mixture of bacteria and chitin. Plants were harvested 2 h after the treatment. Wild-type plants (Gifu) were analyzed in a similar way. Data indicate means ± SE from three independent RNA isolations (n=3). Used primers specific for LjPRp27b (a), Chitinase (b), MPK3 (c) and WRKY33 (d) are listed in Table S2.

Wang et al. (Figures S1-S7) ̶ 7

Figure S7. qRT-PCR analysis of NSP1 and NSP2 genes in hairy roots of L. japonicus nfr1-1 (pNIN::GUS) expressing indicated chimeras after application of chitin and/or M. loti. The experiment was performed as described in the legend to Figure S6. Used primers specific for NSP1 (a) and NSP2 (b) are listed in Table S2. Wang et al. (Supporting Information: Tables S1 and S2) ̶ 1

Supporting Information: Tables (Wei Wang, Zhi-Ping Xie and Christian Staehelin. Functional analysis of chimeric LysM domain receptors mediating Nod factor-induced defense signaling in Arabidopsis thaliana and chitin-induced nodulation signaling in Lotus japonicus)

Table S1. Strains and plasmids used in this study.

Strain or plasmid Re levant characteristics Reference

Escherichia coli DH5α ΔlacU169 (φ80lacZsupE44ΔM15) hsdR17 GIBCO BRL, recA1 endA1 gyrA96 thi-1 relA1 Bethesda, MD, USA

Rhizobium sp. NGR234 Rhizobium (Sinorhizobium, Ensifer) sp. strain Trinick, 1980 NGR234 isolated from Lablab purpureus (Rifr derivative)

NGRΔnodABC nodABC deletion mutant of NGR234; no Nod Price et al., 1992 factor synthesis (Rifr, Spr)

Agrobacterium tumefaciens

EHA105 A derivative of EHA101 lacking the kanamycin Hood et al., 1993 marker (Kanr)

Agrobacterium rhizogenes LBA9402 Harboring the root-inducing (Ri) plasmid Hooykaas et al., pRil855 (Rifr) 1977

Fusarium oxysporum f. sp. cubense race 4 Pathogenic fungus proliferating in the xylem Hennessy et al., and causing leaf wilting, especially in banana 2005

Plasmids pYES2 Yeast protein expression vector with a Invitrogen, galactose-inducible promoter (GAL1) , 2μ California, USA origin and a URA3 gene as selectable marker, (Ampr)

Wang et al. (Supporting Information: Tables S1 and S2) ̶ 2 pYES2-LjNFR1 pYES2 derivative containing the LjNFR1 gene Kindly provided by of Lotus japonicus (ecotype Gifu) Prof. Zhong-Ming Zhang, Huazhong Agricultural University, Wuhan, China

pMD18-T High-efficiency TA cloning vector constructed Takara, Ohtsu, Japan from pUC18 with a TA cloning site and a URA3 gene as selectable marker (Ampr)

pMD18-LjNFR5 pMD18-T derivative containing the LjNFR5 Kindly provided by gene of Lotus japonicus (ecotype Gifu) Prof. Zhong-Ming Zhang, Huazhong Agricultural University, Wuhan, China

pBluescript II SK (+) High copy number ColE1-based phagemid Stratagene (Agilent (Ampr) Technologies, Shanghai, China) pSK-LjNFR1 pBluescript II SK (+) derivative containing the This work LjNFR1 gene of Lotus japonicus (ecotype Gifu) PCR-amplified with primers 1 and 2 using pYES2-LjNFR1 as template (Ampr)

pSK-LjNFR5 pBluescript II SK (+) derivative containing the This work LjNFR5 gene of L. japonicus (ecotype Gifu) PCR-amplified with primers 3 and 4 using pMD18-LjNFR5 as template (Ampr)

pSK-AtCERK1 A 1854-bp fragment containing the complete This work ORF of AtCERK1 PCR-cloned into pBluescript II SK (+) amplified with primers 5 and 6 from cDNA derived from Arabidopsis thaliana (Col-0) leaf RNA (Ampr)

pSK-OsCERK1 pBluescript II SK (+ ) derivative carrying a This work 1875-bp fragment containing the complete ORF of OsCERK1 amplified with primers 7 and 8 from cDNA derived from rice (Oryza sativa L. cv. Nipponbare) root RNA (Ampr)

Wang et al. (Supporting Information: Tables S1 and S2) ̶ 3 pSK-OsCEBiP pBluescript II SK (+) derivative carrying a This work 1035-bp fragment containing the complete ORF of OsCEBiP amplified with primers 9 and 10 from cDNA derived rice (O. sativa L. cv. Nipponbare) root RNA (Ampr) pRT104 Binary vector containing the cauliflower Töpfer et al., 1987 mosaic virus (CaMV) 35S promoter and a poly-(A) signal (Ampr) pRT-LjNFR1(ED) pRT104 derivative containing a 672-bp This work fragment encoding the extracellular domain of LjNFR1 PCR-amplified from pSK-LjNFR1 using primers 1 and 11 (Ampr)

pRT-LjNFR1(ED)- pRT-LjNFR1(ED) derivative carrying a 1152-bp This work AtCERK1(TM+ID) fragment PCR-amplified from pSK-AtCERK1 using primers 6 and 12 (Ampr)

pRT-LjNFR5(ED) pRT104 derivative containing a 738-bp This work fragment encoding the extracellular domain of LjNFR5 PCR-amplified from pSK-LjNFR5 using primers 3 and 13 (Ampr)

pRT-LjNFR5(ED)- pRT-LjNFR5(ED) derivative carrying an This work AtCERK1(TM+ID) additional 1152-bp fragment PCR-amplified from pSK-AtCERK1 using primers 6 and 12 (Ampr) pRT-OsCERK1(ED) pRT104 derivative containing a 711-bp This work fragment encoding the coding region of the extracellular domain of OsCERK1 PCR-amplified from pSK-OsCERK1 using primers 7 and 14 (Ampr)

pRT-OsCERK1(ED)- pRT-OsCERK1(ED) derivative carrying an This work LjNFR1(TM+ID) additional 1194-bp fragment PCR-amplified from pSK-LjNFR1 using primers 2 and 15 (Ampr) pRT-OsCEBiP(ED) pRT104carrying derivative a 1035-bp fragment This work PCR-amplified from pSK-OsCEBiP using primers 9 and 18 (Ampr)

Wang et al. (Supporting Information: Tables S1 and S2) ̶ 4 pRT-OsCEBiP(ED)- pRT-OsCEBiP(ED) derivative carrying an This work LjNFR5(TM+ID) additional 1050-bp fragment amplified from pSK-LjNFR5 using primers 4 and 19 (Ampr)

pCAMBIA1305.1 A binary vector with a pBR322 ori for high http://www. copy replication in E. coli and a broad host cambia.org range ori for low copy, stable replication in Agrobacterium (Kanr)

pCAMBIA-LjNFR1(ED)- pCAMBIA1305 derivative containing the This work AtCERK1(TM+ID) expression cassette of pRT-LjNFR1(ED)- AtCERK1(TM+ID) inserted at the HindIII site (Kanr) pCAMBIA-LjNFR5(ED)- pCAMBIA1305 derivative containing the This work AtCERK1(TM+ID) expression cassette of pRT-LjNFR5(ED)- AtCERK1(TM+ID) inserted into the SalI-SacI sites (Kanr)

pCAMBIA-LjNFR1(ED)- pCAMBIA-LjNFR1(ED)-AtCERK1(TM+ID) This work AtCERK1(TM+ID)+ derivative with the expression cassette of LjNFR5(ED)-AtCERK1(TM+ID) pRT-LjNFR5(ED)-AtCERK1(TM+ID) inserted into the SalI-SacI sites (Kanr)

pCAMBIA1305ΔCaMV35S pCAMBIA1305 derivative digested with This work HindIII and SpeI and religated; lacks the CaMV 35S promoter of the GUS sequence; no GUS expression in transformed plants (Kanr) pCAMBIAΔ-OsCERK1(ED)- pCAMBIA1305ΔCaMV35S derivative This work LjNFR1(TM+ID) containing the expression cassette of pRT-OsCERK1(ED)-LjNFR1(TM+ID) PCR-amplified using primers 16 and 20 and inserted at the SalI site (Kanr) pCAMBIAΔ-OsCEBiP(ED)- pCAMBIA1305ΔCaMV35S derivative This work LjNFR5(TM+ID) containing the expression cassette of pRT-OsCEBiP(ED)-LjNFR5(TM+ID) PCR-amplified with primers 16 and 20 and inserted at the SacI site (Kanr)

Wang et al. (Supporting Information: Tables S1 and S2) ̶ 5 pCAMBIAΔ-OsCERK1(ED)- pCAMBIAΔ-OsCERK1(ED)-LjNFR1(TM+ID) This work LjNFR1(TM+ID)+ derivative containing the expression cassette of OsCEBiP(ED)-LjNFR5(TM+ID) pRT-OsCEBiP(ED)-LjNFR5(TM+ID) PCR-amplified with primers 17 and 21 and inserted at the SacI site (Kanr)

Ampr, Kanr, Rifr, Spr – resistance to ampicillin, kanamycin, rifampin, and spectinomycin, respectively.

References cited in Table S1

Trinick, M.J. (1980) Relationships amongst the fast-growing rhizobia of Lablab purpureus, Leucaena leucocephala, Mimosa spp., Acacia farnesiana and Sesbania grandiflora and their affinities with other rhizobial groups. J. Appl. Microbiol. 49, 39-53. Price, N.P., Relić B., Talmont, F., Lewin, A., Promé, D., Pueppke, G.S., Maillet, F., Dénarié, J., Promé, J.C. and Broughton, W.J. (1992). Broad-host-range Rhizobium species strain NGR234 secretes a family of carbamoylated, and fucosylated, nodulation signals that are O-acetylated or sulphated. Mol. Microbiol. 6, 3575-3584. Hood, E.E., Gelvin, S.B., Melchers, L.S. and Hoekema, A. (1993) New Agrobacterium helper plasmids for gene transfer to plants. Transgenic Res. 2, 208-218. Hooykaas, P.J.J., Klapwijk, P.M., Nuti, M.P., Schilperoort, R.A. and Rörsch, A. (1977) Transfer of the Agrobacterium tumefaciens Ti plasmid to avirulent agrobacteria and to Rhizobium explanta. J. Gen. Microbiol. 98, 477-484. Töpfer, R., Matzeit, V., Gronenborn, B., Schell, J. and Steinbiss, H.H. (1987) A set of plant expression vectors for transcriptional and translational fusions. Nucleic Acids Res. 15, 5890. Hennessy, C., Walduck, G., Daly, A. and Padovan, A. (2005) Weed hosts of Fusarium oxysporum f. sp. cubense tropical race 4 in northern Australia. Australas. Plant Path. 34, 115-117. Wang et al. (Supporting Information: Tables S1 and S2) ̶ 6

Table S2. Primers used in this study.

No. Sequence (5’ to 3’) RE site* Description

1 Forward CCCTCGAGATGAAGCTAAAAACT XholI Amplification of the 1866-bp coding region of GGTCTACT LjNFR1 from L. japonicus cDNA; for 2 Reverse CGGAATTCTTATCTCACAGACAG EcoRI construction of pSK-LjNFR1 TAAATTTATG

3 Forward CCCTCGAGATGGCTGTCTTCTTTC XholI Amplification of the 1788-bp coding region of TTACCTCT LjNFR5 from L. japonicus cDNA; for 4 Reverse CGGAATTCTTAACGTGCAGTAAT EcoRI construction of pSK-LjNFR5 GGAAGTCAC

5 Forward CGGAATTCATGAAGCTAAAGATT EcoRI Amplification of the 1854-bp coding region of TCTCTAATC AtCERK1 from A. thaliana cDNA; for 6 Reverse GCTCTAGACTACCGGCCGGACAT XbaI construction of pSK-AtCERK1 AAGACTG

7 Forward CCCTCGAGATGGAAGCTTCCACC XholI Amplification of the 1875-bp coding region of TCCCTC OsCERK1 from O. sativa cDNA; for 8 Reverse CGGAATTCCTATCTCCCGGACATT EcoRI construction of pSK-OsCERK1 AGGTTG

9 Forward CCCTCGAGATGGCGTCGCTCACC XholI Amplification of the 1071-bp coding region of GCCGCCCT OsCEBiP from O. sativa cDNA; for 10 Reverse CGGAATTCTCAAAGGAAACAGAT EcoRI construction of pSK-OsCEBiP AATGATCAAC

11 Reverse CGGAATTCTCTGTGGTACAAGGG EcoRI Amplification of a 672-bp fragment encoding AACATAG the extracellular domain of LjNFR1 from pSK-LjNFR1 (combined with primer 1); for construction of pRT-LjNFR1(ED)

12 Forward CGGAATTCATTGCTGGTATAGTTA EcoRI Amplification of a 1152-bp fragment encoding TAGGAGT the transmembrane and intracellular domains of AtCERK1 from pSK-AtCERK1 (combined with primer 6); for construction of pRT-LjNFR1(ED)+AtCERK1(TM+ID) and pRT-LjNFR5(ED)+AtCERK1(TM+ID)

Wang et al. (Supporting Information: Tables S1 and S2) ̶ 7

13 Reverse CGGAATTCCAGAAGATGAATGCT EcoRI Amplification of a 738-bp fragment encoding GCTTTTC the extracellular domain of LjNFR5 from pSK-LjNFR5 (combined with primer 3); for construction of pRT-LjNFR5(ED)

14 Reverse GGGGTACCAGAAGCTCCCTTTCC KpnI Amplification of a 711-bp fragment encoding TGGTGAT the extracellular domain of OsCERK1 from pSK-OsCERK1 (combined with primer 7); for construction of pRT-OsCERK1(ED)

15 Forward GGGGTACCACCGCAGGTCTAGCT KpnI Amplification of a1194-bp fragment encoding AGTGGTGC transmembrane and intracellular domains of LjNFR1 from pSK-LjNFR1 (combined with primer 2); for construction of pRT-OsCERK(ED)-LjNFR1(TM+ID)

16 Forward GCGTCGACGGCATGCCTGCAGGT SalI Amplification of the expression cassette from CAACAT pRT-LjNFR5(ED)-AtCERK1(TM+ID); for 17 Reverse CGAGCTCGGCAGGTCACTGGATT SacI construction of pCAMBIA-LjNFR1(ED)- TTGGTT AtCERK1(TM+ID) +LjNFR5(ED)- AtCERK1(TM+ID)

18 Reverse GGGGTACCGGAGATAACAGACAT KpnI Amplification of a 1035-kb fragment encoding GCTCCAC the extracellular domain of OsCEBiP from pSK-OsCEBiP (combined with primer 9); for construction of pRT-OsCEBiP(ED)

19 Forward GGGGTACCGTTATACTTGGTATTA KpnI Amplification of the 1050-bp fragment encoding CCCTGGGA the transmembrane and intracellular domains of LjNFR5 from pSK-LjNFR5 (combined with primer 4); for construction of pRT-OsCEBiP(ED)- LjNFR5(TM+ID)

20 Reverse GCGTCGACGCAGGTCACTGGATT SalI Amplification of the expression cassette from TTGGTT pRT-OsCERK1(ED)-LjNFR1(TM+ID) (combined with primer 16); for construction of pCAMBIAΔ-OsCERK1(ED)-LjNFR1(TM+ID)

21 Forward CGAGCTCGGGCATGCCTGCAGGT SacI Amplification of the expression cassette from CAACAT pRT-OsCEBiP(ED)+LjNFR5(TM+ID) (combined with primer 17); for construction of pCAMBIAΔ-OsCERK1(ED)-LjNFR1(TM+ID) +OsCEBiP(ED)-LjNFR5(TM+ID)

Wang et al. (Supporting Information: Tables S1 and S2) ̶ 8

22 Forward TCCACGAGCACACGGTTCCAG Gene-specific primers for detection of the 23 Reverse GACGAAAAGAGAGTGGATG cerk1-2 mutant 24 Reverse CCCATTTGGACGTGAATGTAGAC AC

25 Forward CGTACGAGCTACAGTCGTCCAAG qRT-PCR; amplification of a Nit4 fragment CC from A. thaliana cDNA 26 Reverse CCTTTAGCGGTACGAGAACCAAT AGC

27 Forward CGTTCTAGCTTCTCCAACCACAG qRT-PCR; amplification of a WRKY33 fragment GA from A. thaliana cDNA 28 Reverse GTGGTCGGAGCAGAAACTCCTGA TG

29 Forward GTGGCAAATCCCGGCAGTGTTC qRT-PCR; amplification of a WRKY53 fragment 30 Reverse GGCGATGATGACTCTCGCTAGAA from A. thaliana cDNA CC

31 Forward GGCTACCACATCCAAGGAAGGC qRT-PCR, amplification of a 18S rRNA AG fragment from A. thaliana cDNA 32 Reverse CTGGAATTACCGCGGCTGCTG

33 Forward GAGGTCGAGCTTTGTTGAGG qRT-PCR; amplification of a NSP1 fragment 34 Reverse ATTCCCATCCAGCTTCCAC from A. thaliana cDNA (Heckmann et al., 2006; Murakami et al., 2006)

35 Forward CATCGACTCCATGATTGACG qRT-PCR; amplification of a NSP2 fragment 36 Reverse GGTTGTTGTTGTCGTGGTTG from L. japonicus cDNA (Heckmann et al., 2006; Murakami et al., 2006)

37 Forward ATGCAGATCTTTTGTGAAGAC qRT-PCR; amplification of a Ubiquitin 38 Reverse ACCACCACGGAAGACGGAG fragment from L. japonicus cDNA (Flemetakis et al., 2000)

39 Forward GCAGTGTCTGATTAGTTAGTCTC qRT-PCR; amplification of a Chitinase A fragment from L. japonicus cDNA (Nakagawa 40 Reverse GAAGTGATCACATACAGTACTCA et al., 2011) AC

41 Forward CATGGTTATGATGTTACTGCTCG qRT-PCR; amplification of a LjPRp27b T fragment from L. japonicus cDNA (Nakagawa 42 Reverse GATCACTATAACCAGTCCTCATC et al., 2011) T Wang et al. (Supporting Information: Tables S1 and S2) ̶ 9

43 Forward CACCCTTGCGTAGAGAGTTTACT qRT-PCR; amplification of a MPK3 fragment GATGTC from L. japonicus cDNA (Wan et al., 2008) 44 Reverse GTTGACGAGGATATTGAGGAAG TTGTCTG

45 Forward AGTTGTGGTTCAGACCACCAGTG qRT-PCR; amplification of a WRKY33 ACATTG fragment from L. japonicus cDNA (Wan et al., 46 Reverse ACCCCATTGAGTTTCCAAACCCT 2008) GATGAG

*RE site: restriction enzyme site

References cited in Table S2

Nakagawa, T., Kaku, H., Shimoda, Y., Sugiyama, A., Shimamura, M., Takanashi, K., Yazaki, K., Aoki, T., Shibuya, N. and Kouchi, H. (2011) From defense to symbiosis: limited alterations in the kinase domain of LysM receptor-like kinases are crucial for evolution of legume-Rhizobium symbiosis. Plant J. 65, 169-180. Wan, J., Zhang, X. C., Neece, D., Ramonell, K. M., Clough, S., Kim, S. Y., Stacey, M. G. and Stacey, G. (2008) A LysM receptor-like kinase plays a critical role in chitin signaling and fungal resistance in Arabidopsis. Plant Cell, 20, 471-481. Heckmann, A.B., Lombardo, F., Miwa, H., Perry, J.A., Bunnewell, S., Parniske, M., Parniske, M., Wang, T.L. and Downie, J.A. (2006) Lotus japonicus nodulation requires two GRAS domain regulators, one of which is functionally conserved in a non-legume. Plant Physiol., 142, 1739-1750. Flemetakis, E., Kavroulakis, N., Quaedvlieg, N. E., Spaink, H. P., Dimou, M., Roussis, A. and Katinakis, P. (2000) Lotus japonicus contains two distinct ENOD40 genes that are expressed in symbiotic, nonsymbiotic, and embryonic tissues. Mol. Plant Microbe Interact.13, 987-994. Murakami, Y., Miwa, H., Imaizumi-Anraku, H., Kouchi, H., Downie, J.A., Kawaguchi, M. and Kawasaki, S. (2006) Positional cloning identiﬁes Lotus japonicus NSP2, a putative transcription factor of the GRAS family, required for NIN and ENOD40 gene expression in nodule initiation. DNA Res. 13, 255–265. Wang et al. (Supporting Methods S1 to S5) ̶ 1

Supporting Experimental Procedures (Wei Wang, Zhi-Ping Xie and Christian Staehelin. Functional analysis of chimeric LysM domain receptors mediating Nod factor-induced defense signaling in Arabidopsis thaliana and chitin-induced nodulation signaling in Lotus japonicus)

Methods S1. Chitin, chitooligosaccharides and NFs. Crab shell chitin was purchased from Sigma (St Louis, MO, USA). Purified chitooligosaccharides

[chitotetraose, (GlcNAc)4; chitopentaose, (GlcNAc)5; chitohexaose (GlcNAc)6] were supplied from Seikagaku Kogyo Co. (Tokyo, Japan). Chitooctaose (GlcNAc)8 was obtained by re-acetylation of chitosan oligosaccharides (molecular weight <1300; Nicechem. Industrial Co. Ltd; Shanghai, China) and HPLC purification. Briefly, 200

mg of chitosan oligosaccharides were dissolved in 25 ml of saturated NaHCO3 and re-acetylated by adding 12 ml acetic anhydride at room temperature for 40 min. The solution was then adjusted to pH 3.0 with acetic acid and desalted with Dowex 50 (H+). The remaining non-charged chitooligosaccharides were precipitated with 70% ethanol (-20°C) and then purified under isocratic conditions on a Bondapak Amino

(NH2) HPLC column (10 m 8×100 mm Radial-Pak Cartridge, Waters, Milford, MA; 67% (v/v) acetonitrile/water as mobile phase). Chitooligosaccharides from Seikagaku Kogyo Co were used as standards. NFs from Rhizobium sp. strain NGR234 and control fractions from the NF-deficient mutant NGRΔnodABC (Price et al., 1992) were prepared as described previously (Staehelin et al., 1994). Dried preparations were first re-suspended in dimethyl sulfoxide (DMSO) and then with water. The final concentration was adjusted to 100 μg/ml (containing 0.71% (v/v) DMSO).

Methods S2. Treatment of A. thaliana plants. Fully expanded leaves of 4-weeks-old A. thaliana plants were used for all treatments and inoculation experiments. For fungal inoculation, a F. oxysporum suspension (ca. 1×103 spores) was injected by a 1-ml syringe (without needle) into the upper left quarter of a given leaf (three leaves per plant). Leaves treated with an equivalent amount of 10 mM MgSO4 were used as a control. Similar injection treatments were done with NFs from NGR234 (100 μg/ml), control material from NGRΔnodABC, chitin (100 μg/ml) and chitooligosaccharides (10-6 M). Inoculated plants were then placed in a growth chamber at 22°C with a 12-h light-dark cycle and maintained at 100% relative humidity. Leaves were photographed and chlorotic areas (expressed as percentage for a whole leaf) were quantified using Image J software (http://rsb.info.nih.gov/ij/). Leaf material was used for determination of ROS activity, cell death staining and analysis of gene expression by qRT-PCR. To test whether prior applied NFs show an effect on a subsequent infection of F. oxysporum, A. thaliana plants were pretreated twice with NFs (24 and 48-h before fungal inoculation). Whole leaves were sprayed with the NF preparation from Rhizobium sp. NGR234 at a final concentration of 100 μg/ml supplemented with 0.01% of Silwet L-77 using a 50-ml sprayer. Control plants were similarly treated with corresponding material from NGRΔnodABC. In addition, rhizobial suspensions 9 (strains NGR234 or NGRΔnodABC; re-dissolved in 10 mM MgSO4; 1×10 CFU/ml) supplemented with 0.01% of Silwet L-77 were sprayed onto whole A. thaliana leaves. Control plants were sprayed with water or with 10 mM MgSO4 supplemented with Wang et al. (Supporting Methods S1 to S5) ̶ 2

0.01% of Silwet L-77. Harvested leaves were photographed to quantify leaf chlorosis and the degree of F. oxysporum infection. Fungal structures were visualized by trypan blue staining according to Ramonell et al. (2005) and quantified by determining the presence of fungal structures in randomly selected square areas of the taken pictures (values expressed in percentages). Data were acquired from three independent experiments and 20 leaves were selected for each experiment.

Methods S3. Treatment of L. japonicus plants. One month after transformation with A. rhizogenes, defined areas of transgenic hairy roots of L. japonicus with a pNIN::GUS construct were locally treated with 10 μl NFs from NGR234 (100 μg/ml), with 10 μl chitooligosaccharides [(GlcNAc)4; (GlcNAc)5; (GlcNAc)6; (GlcNAc)8] at a final concentration of 10-6 M or with 10 μl chitin (100 μg/ml). Where indicated, different concentrations of chitooctaose (10-6 to 10-10 M) and incubation periods (0 to 7 h) were chosen. Whole roots were incubated in a GUS staining solution for 4 h at room temperature and then photographed (Kosugi et al., 1990). Where indicated, L. japonicus plants were treated with 2 ml of Mesorhizobium loti strain MAFF303099 (1×106 CFU/ml), bacteria combined with 1 ml chitin (100 μg/ml) or with chitin alone. For cryosectioning, GUS-stained roots were fixed and embedded into “optimal cutting temperature compound” (Tissue-Tek, VWR Scientific, Bridgeport, NJ, USA). Sections (8-20 μm) were analyzed with a Leica DMLB or Imager Z1 Zeiss microscope (Leica, Nanterre, France).

Methods S4. Cell death and ROS measurements. For visualization of cell death, A. thaliana leaves were briefly boiled in a 50% trypan blue solution (5 ml lactic acid, 10 ml 50% (v/v) glycerol, 1 mg trypan blue, and 5 ml phenol) and then destained overnight with chloral hydrate (10% w/v) (Koch and Slusarenko, 1990). Leaves injected with 2 M salicylic acid (SA) were used as a positive control for cell death. Generation of ROS by excised A. thaliana leaf disks was measured according to Keppler et al. (1989) using 20 μM luminol and 1 μg of horseradish peroxidase. Chemiluminescence was measured with a Tecan Infinite M200 microplate reader (Tecan, Groedig, Austria). To estimate produced H2O2 in situ, A. thaliana leaves were vacuum-infiltrated with 1 mg/ml of 3’,3’-diaminobenzidine (DAB), then incubated for 2-3 h at room temperature and finally cleared in 90% ethanol (Seligman et al., 1968).

Methods S5. Quantitative reverse transcription (qRT)-PCR. Total RNA isolated with the Trizol reagent (Invitrogen, Carlsbad, CA, U.S.A.) and incubated with RNase-free DNase (Takara, Kyoto, Japan) was used for synthesis of cDNA with M-MLV reverse transcriptase (Takara, Kyoto, Japan). qRT-PCR was performed using the LightCycler® 480 SYBR Green I Master Mix in a LightCycler® 480 System (Roche Diagnositcs, Mannheim, Germany). Each PCR reaction (20 μl) consisted of 2 μl of cDNA template, 1 μM of each primer (see Table S2), and 10 μl of the Lightcycler® SYBR Green Master Mix. The real-time PCR consisted of a 3-min denaturation step at 95ºC and 45 following cycles (95ºC for 30 s, 60ºC for 10 s and 72ºC for 20 s). All reactions were carried out in triplicates. The housekeeping gene 18S rRNA served as an internal control to normalize the values for transcript abundance among different samples.

Wang et al. (Supporting Methods S1 to S5) ̶ 3

References cited in Supporting Experimental Procedures

Keppler, L.D., Baker, C.J. and Atkinson, M.M. (1989) Active oxygen production during a bacteria-induced hypersensitive reaction in tobacco suspension cells. Phytopathology 79, 974-978. Koch, E. and Slusarenko, A. (1990) Arabidopsis is susceptible to infection by a downy mildew fungus. Plant Cell 2, 437-445. Kosugi, S., Ohashi, Y., Nakajima, K. and Arai, Y. (1990) An improved assay for β-glucuronidase in transformed cells: methanol almost completely suppresses a putative endogenous β-glucuronidase activity. Plant Sci. 70, 133-140. Price, N.P.J., Relic, B., Talmont, F., Lewin, A., Promé, D., Pueppke, S.G., Maillet, F., Dénarié, J., Promé J.C. and Broughton, W.J. (1992) Broad-host-range Rhizobium species strain NGR234 secretes a family of carbamoylated, and fucosylated, nodulation signals that are O-acetylated or sulphated. Mol. Microbiol. 6, 3575-3584. Ramonell, K., Berrocal-Lobo, M., Koh, S., Wan, J., Edwards, H., Stacey, G. and Somerville, S. (2005) Loss-of-function mutations in chitin responsive genes show increased susceptibility to the powdery mildew pathogen Erysiphe cichoracearum. Plant Physiol. 138, 1027-1036. Seligman, A.M., Karnovsky, M.J., Wasserkrug, H.L. and Hanker, J.S. (1968) Nondroplet ultrastructural demonstration of cytochrome oxidase activity with a polymerizing osmiophilic reagent, diaminobenzidine (DAB). J. Cell Biol. 38, 1-14. Staehelin, C., Granado, J., Müller, J., Wiemken, A., Mellor, R.B., Felix, G., Regenass M., Broughton, W.J. and Boller, T. (1994) Perception of Rhizobium nodulation factors by tomato cells and inactivation by root chitinases. Proc. Natl. Acad. Sci. USA 91, 2196-2200.