Nod Factors

Plants initiate the molecular dialogue with rhizobia by releasing flavonoids into their rhizospheres (Schultze and Kondorosi 1998; Broughton et al. 2003). These flavonoids are then taken up by the bacteria where they bind NodD proteins of the LysR family of transcriptional regulators (Broughton et al. 2000). The promoters of genes relevant for Nod-factor synthesis (nol, noe and nod genes) contain conserved 49 bp motifs called nod-boxes (Feng et al. 2003). NodD proteins bind nod-boxes as tetramers even in the absence of flavonoids, but activate nod-box controlled genes only in the presence of flavonoids (Fisher and Long 1993). This way, the timing and levels of production of Nod-factors are carefully controlled (Kobayashi et al. 2004) since overproduction of Nod-factors may provoke plant defence reactions and lead to the abortion of the infection process (Savouré et al. 1997; Hogg et al. 2002; Ramu et al. 2002). Regulation of Nod-factor production has been well studied in Rhizobium sp. NGR234 (from now on referred to as NGR234) (Fellay et al. 1995; Perret et al. 1999; Kobayashi et al. 2004). The symbiotic plasmid of NGR234 (pNGR234a) contains 19 nod-boxes, 18 of which are functional. Upon binding to flavonoids, NodD1 activates transcription of the genes required for Nod-factor synthesis (nod-genes). NodD1 also activates transcription of the ttsI and syrM2 genes. TtsI is a transcriptional regulator that activates genes controlled by tts-boxes, leading to the induction of the type three secretion system (T3SS) (Marie et al. 2004) (Section 8.5 and Fig. 1). In turn SyrM2, another transcriptional activator, modulates transcription of the nodD2 gene. In concert with TtsI, NodD2 also up-regulates transcription of genes required for the synthesis of rhamnose-rich lipo-poly-saccharides (LPS) (Marie et al. 2004) (section 8.4) and ultimately represses NodD1. In this way, the Nod-factor regulatory circuit is eventually closed (Kobayashi et al. 2004). Thus, flavonoids induce the production of Nod-factors, which are the first bacterial signals perceived by the plant and are crucial to nodulation (Relic et al. 1994a). Nod-factors are acylated lipo-chito-oligosaccharides made of 2 to 6 ^-1,4-linked N-acetyl-D-glucosamine units carrying a fatty acid chain at the non-reducing terminus (Perret et al. 2000). Some rare Nod-factors have a slightly different oligo-saccharide backbone (Demont et al. 1994; Bec-Ferté et al. 1996; Olsthoorn et al. 1998; Pacios-Bras et al. 2002). Chain length of the oligo-saccharide, the type of the acyl moiety, the nature of the fatty acid and various decorations resulting from fucosylation, sulphation, acetlyation, N-methylation, 3-, 4-, or 6-, O-carbamoylation, 6-O-glycosylation, D-arabinosylation or 2-O methylation all contribute to the great variety of Nod-factors known (Perret et al. 2000). Nod-factor structures and Nod-factor biosynthesis have been extensively reviewed (Dénarié et al. 1996; Broughton et al. 2000; Perret et al. 2000). Unfortunately, there are few obvious correlations between the diversity of Nod-factors produced by one bacterial strain and its host-range. NGR234 as one example produces a wide range of different Nod-factors (Price et al. 1992) and has an exceptionally broad host-range that includes more than 112 genera of legumes. Some NGR234 hosts have determinate nodules, others indeterminate structures with a permanent meristem. In addition, NGR234 nodulates the non-legume Parasponia andersonii (Pueppke and Broughton 1999). At the other extreme, stem nodules on Sesbania rostrata can only be induced by Azorhizobium caulinodans (Mergaert et al. 1993). R. fredii strain USDA257 nodulates an exact subset of the NGR234 hosts and excretes a subset of the Nod-factors produced by NGR234 (Pueppke and Broughton 1999). In spite of this, Nod-factor requirements cannot be assigned to specific legumes and Nod-factor structures cannot be used as a tool to predict host-ranges. For example, R. etli and R. tropici produce different Nod-factors but both efficiently nodulate Phaseolus vulgaris (Poupot et al. 1993, 1995). The quantity of Nod-factors produced also helps determine the spectrum of hosts. USDA257 produces one-fortieth the amount of Nod-factors secreted by NGR234. Variation in the amount of Nod-factors produced may depend on NodD1. If nodD1 of NGR234 is transferred into R. meliloti, it produces two times more Nod-factors than usual and gains the ability to nodulate Vigna unguiculata, normally a non-host of R. meliloti (Relic et al. 1994b). A strain of Mesorhizobium loti, NZP2213, produces a much greater variety of Nod-factors than any other M. loti strain described (Olsthoorn et al. 1998). This does not, however, result in an enlarged spectrum of hosts. Furthermore, even an artificial change in the spectrum of Nod-factors does not necessarily lead to an altered host-range. NodL of R. leguminosarum bv. viciae is an enzyme that adds an O-acetyl group to C-6 of the non-reducing N-acetylglucosamine residue of Nod-factors. Introduction of the nodL gene into M. loti E1R did not affect the host spectrum either, although the transformed M. loti strain also produced O-acetylated Nod-factors (Lopez-Lara et al. 2001). That said, some Nod-factor substituents have been reported to play roles in successful nodulation: R. meliloti, which produces sulphated Nod-factors, nodulates Medicago sativa. nodH mutants produce non-sulphated Nod-factors, and lose the ability to nodulate M. sativa, but gain the capacity to nodulate Vicia sativa subsp. nigra, which the wild-type R. meliloti strain cannot nodulate (Roche et al. 1991). For efficient stem and root nodulation on Sesbania rostrata, Azorhizobium caulinodans must produce Nod-factors which carry a D-arabinosyl or a L-fucosyl group, or both, at the reducing terminal residue. Bacterial mutants that produce Nod-factors lacking either the D-arabinosyl or the L-fucosyl group nodulate S. rostrata less efficiently, indicating that none of these decorations is strictly required for normal nodule formation, but that they act synergistically (D'Haeze et al. 2000).

T3ss Induction

Fig. 1. Model of induction of the T3SS of Rhizobium sp. NGR234 by flavonoids. The host plant secretes flavonoids (1), e.g., apigenin, that are captured by the bacterial NodDl transcriptional activator. NodDl binds to nod-boxes and hence nod-box dependent genes are transcribed (2). This starts the production of Nod-factors and, amongst others, of the transcriptional activator Ttsl. TtsI then binds to its-boxes and so T3SS related genes are transcribed (3). This includes the T3SS machinery itself (4), but also T3SS effector proteins such as NopL (J)

Fig. 1. Model of induction of the T3SS of Rhizobium sp. NGR234 by flavonoids. The host plant secretes flavonoids (1), e.g., apigenin, that are captured by the bacterial NodDl transcriptional activator. NodDl binds to nod-boxes and hence nod-box dependent genes are transcribed (2). This starts the production of Nod-factors and, amongst others, of the transcriptional activator Ttsl. TtsI then binds to its-boxes and so T3SS related genes are transcribed (3). This includes the T3SS machinery itself (4), but also T3SS effector proteins such as NopL (J)

Nod-Factor Perception

In some plants, homologous Nod-factors (i.e., Nod-factors produced by the native symbiont of the plant) can induce a variety of effects, such as root-hair curling, induction of pre-infection threads (Catoira et al. 2001), division of cortical cells and nodule morphogenesis (Dénarié et al. 1996; Bladergroen and Spaink 1998; Schultze and Kondorosi 1998; Downie and Walker 1999). Root-hair curling is provoked by depolarisation of the plasma-membrane of the root-hair, which allows an influx and accumulation of calcium at the tip (Felle et al. 1996, 1998; Gehring et al. 1997). Then, possibly mediated by G-proteins, calcium spiking is induced (Ehrhardt et al. 1996; Pringet et al. 1998). Nod-factors of NGR234 have been shown to stimulate the activity of a phospholipase C and of both heterotrimeric and monomeric G-proteins in Vigna unguiculata (Kelly and Irving 2001, 2003). This leads to rearrangements of the cytoskeleton (Miller et al. 1999; Cárdenas et al. 2003), provoking the typical shepherd's crook-like root-hair curling. Many transcripts are induced early in nodulation and the genes that encode them are commonly referred to as early nodulin genes (ENODs) (Schultze and Kondorosi 1998; Downie and Walker 1999). Expression of early nodulin genes and division of the cortical cells follows the extreme curling of the root hairs and eventually leads to nodule formation (Yang et al. 1994).

Many ENODs have been identified, but the exact function of none of them is known. ENOD40 for example, is expressed before cell division in the pericycle opposite the protoxylem poles occurs, but it is also expressed in dividing cells (Fang and Hirsch 1998; Kouchi et al. 1999). In transgenic legumes that over-express ENDO40, cortical-cell division is initiated at many places throughout the root (Charon et al. 1997). ENOD40 encodes a very short peptide, but it is more likely that the main function of ENOD40 resides in a long, non-coding RNA that appears to have a regulatory role. ENOD40 RNA causes MtRBP1 (Medicago trunculata RNA Binding Protein 1) to be relocated from the nucleus into the cytoplasm (Campalans et al. 2004). Other plant responses to Nod-factors include a plant-dependent increase in flavonoid production (Recourt et al. 1991; Schmidt et al. 1994). Nod-factors also appear to trigger their own hydrolysis by the plant (Staehelin et al. 1995; Ovtsyna et al. 2000).

How the plant perceives Nod-factors is the subject of intense research. A nodF-nodL double mutant of R. meliloti, which produces Nod-factors lacking the acetyl group at the non-reducing terminus and which contain an alternative lipid, loses the ability to nodulate M. sativa. This mutant is still able to elicit root-hair deformation and inner cortical cell division however (Ardourel et al. 1994). An analogous finding has been made with a R. leguminosarum bv. viciae nodFEMNTLO mutant inoculated on Vicia hirsuta (Walker and Downie 2000). Interestingly, nodulation by the nodFEMNTLO mutant can be partially rescued by over-expression of nodO. NodO is secreted from bacteria and makes pores in membranes (Vlassak et al. 1998). NodO re-establishes the ion-influx that the non-substituted Nod-factors of the nodFEMNTLO mutant can no longer provoke. A two-step model for Nod-factor perception has thus been proposed: first, a less stringent Nod-factor receptor prepares the plant for bacterial invasion; second, a more stringent Nod-factor receptor actually allows bacterial entry and infection thread growth (Ardourel et al. 1994). Alternatively, one Nod-factor receptor could perform both tasks by being activated first in a less stringent manner, but would subsequently need strong activation by tight binding to a Nod-factor to sustain bacterial entry and infection-thread growth. The latter model is supported by recent findings on Nod-factor receptors containing LysM domains. In 2003, the group of Jens Stougaard (University of Aarhus, Denmark) discovered what are most likely the high affinity Nod-factor receptors that have been sought for so long (Madsen et al. 2003; Radutoiu et al. 2003). Using a genetic approach in L. japonicus, they identified two genes, nfr1 and nfr5. Mutants of nfr1 or nfr5 fail to respond to Nod-factors. NFR1 and NFR5 code for two receptor kinases with a LysM motif in the extra-cellular domain. NFR5, however, lacks a kinase-activation loop in its kinase domain and it is no longer an active kinase. Upon binding to Nod-factors, NFR1 and NFR5 are predicted to form hetero-dimers via their LysM motifs. Formation of hetero-dimers may then activate the intra-cellular kinase domain of NFR1 (Parniske and Downie 2003). P. sativum has a homologue of NFR5, called SYM10 (Madsen et al. 2003). Another NFR1 homologue, LYK3 was also identified in M. trunculata (Limpens et al. 2003). Although NFR1 and NFR5 are early high-affinity receptors for Nod-factors, they also appear to be required throughout the infection process (Parniske and Downie 2003). So far, direct biochemical evidence supporting neither NFR1/NFR5-hetero-dimer formation, nor Nod-factor binding to the hetero-dimer or to either NFR1 or NFR5 has been found. LysM domains have been identified in an Escherichia coli protein, MltD, that binds peptidoglycan (Steen et al. 2003). Additionally, LysM domains have also been found in at least two chitin-binding proteins (Ponting et al. 1999). LysM domains are thus good candidates for Nod-factor binding sites. Furthermore, chitin is able to induce calcium-spiking in legumes (Pringet et al. 1998; Walker et al. 2000). Together, the very good genetic evidence and the circumstantial data inferred from structural similarities strongly suggest that NFR1 and NFR5 are Nod-factor receptors (Cullimore and Denarie 2003; Oldroyd and Downie 2004), but hard biochemical evidence for this is eagerly awaited.

Several other approaches have been used to identify Nod-factor receptors (Geurts and Bisseling 2002). A Nod-factor binding site (called NFBS1 for Nod-Factor Binding Site 1) has been characterised in root extracts of M. trunculata (Bono et al. 1995). NFBS1 has also been identified in tomato (Lycopersicon esculentum) however, which suggests that NFBS1 plays a more general role. NFBS1 would fit into the model of Ardourel et al. (1994) as the less-stringent Nod-factor receptor. Another Nod-factor binding site, NFBS2, has been identified in the microsomal fraction of Medicago varia cell suspension cultures (Niebel et al. 1997), although its affinity for Nod-factors is not higher than that of NFBS1. Another possible Nod-factor binding-protein is a lectin-nucleotide-phosphohydrolase (LNP), which was isolated from the roots of the legume Dolichos biflorus (Etzler et al. 1999). LNP is an ATPase and it's enzymatic activity increases upon binding of Nod-factors to the lectin domain of LNP. Furthermore, LNP is present on the surface of root hairs, but in presence of Nod-factors it accumulates at their tips (Day et al. 2000; Kalsi and Etzler 2000).

Several genes controlling the very early stages of symbiosis have also been cloned. M. trunculata mutants in three genes called dmi1 (does not make infections), dmi2, and dmi3 have pleiotropic effects on Nod-factor responses and are unable to establish a symbiotic association with endomycorrhizal fungi (Catoira et al. 2000; Mitra et al. 2004). This suggests that many of the early steps in symbioses with rhizobia and endomycorrhizal fungi are the same (Albrecht et al. 1999). Grafting experiments have shown that the genetic control of dmi1, dmi2 and dmi3 is determined at the root level (Ane et al. 2002). DMI1 has slight similarities to a ligand-gated cation channel of archaea

(Ane et al. 2004). DMI3 shows high similarity to calmodulin-dependent protein kinases (Levy et al. 2004). Various orthologues of DMI2 have been found: a nodulation receptor kinase (NORK) has been identified in M. sativa and, by analogy, in M. trunculata and P. sativum (Endre et al. 2002). NORK extra-cellular sequence-like (NSL) genes are widely distributed throughout the plant kingdom and it thus appears that an ancient system of NSL genes has been modified to serve as symbiotic Nod-factor receptors (Endre et al. 2002). In another study, symbiosis receptor-like kinase (SYMRK) genes have been cloned from L. japonicus and P. sativum (Stracke et al. 2002). SYMRK mutants of L. japonicus are unable to establish symbiosis with either rhizobia or endomycorrhizal fungi. In addition, an orthologue of the DMI2 gene of M. trunculata has been cloned from P. sativum (Sym19) (Schneider et al. 1999). SYMRK of L. japonicus, Sym19 of P. sativum, NORK of M. sativa and DMI2 of M. trunculata all belong to the family of leucine-rich repeat (LLR) receptor-kinases (Endre et al. 2002; Stracke et al. 2002). This suggests that these receptor-like kinases may form homo- or hetero-dimers or interact with other proteins containing LLR domains.

Was this article helpful?

0 0

Post a comment