PerspectivePLANT SCIENCES

How Legumes Select Their Sweet Talking Symbionts

See allHide authors and affiliations

Science  24 Oct 2003:
Vol. 302, Issue 5645, pp. 575-578
DOI: 10.1126/science.1091269

Legumes such as soybean, pea, peanut, and alfalfa are able to fix nitrogen because of the bacterial symbionts (rhizobia) that inhabit nodules on their roots. The amount of ammonia produced by rhizobial fixation of nitrogen rivals that of the world's entire fertilizer industry. Consequently, this symbiotic relationship between legumes and rhizobia is of great agronomic and ecological importance. Signals from rhizobial bacteria, called Nod factors (lipochitooligosaccharides), are crucial for initiating the symbiotic response of legumes. This response leads to recognition of bacteria by root-hair cells, curling of root hairs, growth of infection threads, and finally the formation of root nodules (nodulation). The molecular events establishing this symbiosis are the subject of three reports, one by Limpens et al. (1) on page 630 of this issue and the other two by Stougaard's group (2, 3) in a recent issue of Nature. These three papers identify genes of legumes that encode receptors involved in Nod factor recognition, leading to the initiation of infection by rhizobia. These receptors belong to a family of plant transmembrane receptor-like serine/threonine kinases (RLKs) whose extracellular regions contain LysM domains (LysM-RLKs).

Plants often establish symbiotic relationships with specific bacteria and fungi in order to acquire the nutrients they need to support growth and development. The plant symbiotic relationships most studied are those involving rhizobial bacteria and mycorrhizal fungi (4, 5). Although these symbiotic relationships may not, at first sight, appear to have much in common, both bacterial and fungal symbiotic partners are able to trigger the plant host genetic program that permits localized infection and controlled growth (4, 5).

Rhizobial access to the root cortex where nodule formation begins depends on initial recognition of the bacteria by epidermal root-hair cells, followed by root-hair curling and the formation of infection threads (see the figure). Nod factors released by rhizobia are required to establish the earliest steps of the legume-bacterial symbiotic relationship. These molecules are tetramers and pentamers of the carbohydrate chitin, which are attached to a fatty acid chain (5). Species-specific chemical substitutions in the sugars, combined with structural differences in the fatty acid moiety, imbue Nod factors with specificity for plant hosts. At pico-to micromolar concentrations, these molecules initiate many symbiotic responses in the root cells of the plant host, including the rapid influx of calcium ions and calcium spiking, expression of specific nodulin genes, alterations in epidermal cell morphology, and mitosis in cortical root cells (5) (see the figure).

Discriminating root hairs.

(A) Symbiotic rhizobial bacteria secrete Nod factors that are perceived by two transmembrane LysM-RLKs, NFR1 and NFR5. The activated (putative) NFR1/NFR5 heterodimer initiates rapid calcium influx and swelling of root-hair tips, which are early events in the plant symbiotic response that might be specific for bacterial symbionts. Activation of NFR1/NFR5 is also required to activate the putative receptor complex, NORK/DMI1, which results in plant symbiotic responses to both bacterial and fungal symbionts. The NORK/DMI1 complex may also be involved in direct recognition of rhizobial and mycorrhizal signals. (B) Rhizobial bacteria entrapped in a curling root hair. In order for rhizobia to enter root hairs and to initiate formation of infection threads and nodulation, the Nod factors they release must be recognized by highly specific plant receptors. The LysM-RLKs LYK3 and LYK4 of M. truncatula are involved in Nod factor recognition. (C) NFR5 has three LysM domains and its kinase domain does not contain an activation loop. NFR1, LYK3, and LYK4 have two LysM domains, and their kinase domains contain an activation loop. SP, signal peptide; TM, transmembrane domain; AL, activation loop in the kinase domain. [Data from (13, 6, 7)]

CREDIT: KATHARINE SUTLIFF/SCIENCE

Legume genes required for Nod factor signal transduction and nodule formation have been characterized, but no Nod factor receptors have been identified to date. Interestingly, some of the genes characterized so far—called SYM or DMI genes, depending on the legume species—are also involved in establishing symbiotic relationships with mycorrhizal fungi (4, 6). Thus, both bacterial and fungal symbionts share common steps in their symbiotic signaling pathways (4, 6). Legumes with mutations in the common SYM genes still show some responses to Nod factors, including rapid calcium influx and swelling of root-hair tips. This suggests that Nod factors are “perceived” by one or more plant receptors acting upstream of the common SYM/DMI pathway (2, 6, 7).

Two different genetic strategies have been used to identify plant genes encoding Nod factor receptors. The first approach analyzes legume mutants that cannot respond to rhizobial Nod factors and consequently do not form nodules, yet still form symbiotic relationships with mycorrhizal fungi. This strategy has led to the identification of Sym10 in pea (8), NFP in Medicago truncatula (7), and in the new work by Stougaard's group, NFR1 and NFR5 in Lotus japonicus (2, 3). The second strategy is to exploit natural allelic variation involved in recognizing the specific structural features of different Nod factors. This strategy has led to the characterization of an allele in pea (SYM2A) that permits infection only by rhizobia producing a particular Nod factor (1, 5).

In the new studies, the investigators use two model legumes to facilitate molecular characterization of Nod factor receptor genes (13). Limpens, Bisseling, and colleagues identify a cluster of seven genes (LYK1-7) in M. truncatula that are syntenic with the SYM2A region of pea (1). Using RNA interference (RNAi), Limpens et al. demonstrate that the LYK3 and LYK4 genes are required for Nod factor-induced infection of root-hair cells by rhizobia. In their complementary pair of studies, Stougaard and co-workers show that the NFR1 and NFR5 genes encode receptors required for Nod factor recognition in L. japonicus (2, 3).

The NFR5 gene of L. japonicus contains a single exon encoding an RLK with an extracellular region containing three LysM domains and an intracellular kinase domain lacking the typical activation loop (3, 9). This gene appears to be orthologous to pea SYM10 and M. truncatula NFP genes (3, 7, 8). The NFR1 gene of L. japonicus has 12 exons and encodes an RLK containing two LysM domains in the extracellular part and a kinase domain with an activation loop (2, 9); this receptor shares only about 30% identity with NFR5 (2). The M. truncatula LYK3 and LYK4 genes have a similar 12- exon structure encoding proteins with two LysM domains and a kinase domain very similar to that of NFR1 (1) (see the figure). Five genes encoding putative LysM-RLKs have been identified in the model plant Arabidopsis thaliana (10), and related genes have been found in rice and other plants. This suggests that the LysM-RLKs of symbiotic legumes have evolved from a preexisting class of plant LysM-RLKs.

Do the newly identified genes actually encode Nod factor receptors? LysM domains were originally identified as 40- to 50-residue sequences present in lysins and other bacterial enzymes that degrade bacterial cell walls. These domains are commonly found in bacteria, are absent in archaea, and show a patchy distribution in eukaryotes. Most LysM proteins contain one to three LysM domains (11). The LysM domains of the bacterial autolysin produced by Lactococcus lactis interact with the glycan part of peptidoglycans (12). In eukaryotes, LysM domains are found in some chitinases and in a variety of proteins whose functions are, for the most part, unknown. The phenotype of the legume mutants described in the new studies (13), coupled with our current knowledge of LysM domains, strongly suggests that the symbiotically active LysM-RLKs—NFR1, NFR5, and LYK3—are receptors for rhizobial Nod factors. It is tempting to speculate that nonsymbiotic LysM-RLKs may be important for the “perception” of oligosaccharide ligands involved in plant defense and development.

In L. japonicus, both NFR1 and NFR5 are required for early Nod factor responses, which suggests that they may operate as a heterodimer (2). In pea and M. truncatula, a complete defect in Nod factor responses has only been observed with mutations in genes orthologous to NFR5; the lack of null mutations in a second locus (orthologous to NFR1) might be due to functional redundancy. In M. truncatula both LYK3 and LYK4 are required for Nod factor-dependent infection of root-hair cells (1). Thus, the formation of heterodimers could be a general feature of Nod factor receptors in legumes.

In pea, vetch, Medicago sativa, and M. truncatula, the structural requirements for Nod factors may be more stringent at the later stages of root-hair cell entry involving the formation of infection threads than for induction of earlier symbiotic responses. Hence, there may exist two types of Nod factor receptors—one for “entry” and one for “signaling” (1, 5). In support of this notion is the phenotype of SYM2A mutant pea plants and of the M. truncatula LYK3 and LYK4 RNAi transgenic lines. The phenotype of these plants is characterized by a block in the formation of infection threads when the plants are inoculated with rhizobial mutants producing altered Nod factors; earlier symbiotic responses appear not to be affected (1).

Future studies should rapidly tell us whether the LysM-containing region of symbiotic LysM-RLKs is directly involved in Nod factor binding and recognition. In addition, we need to know whether the LysM-RLKs need to be assembled into heterodimers in order to bind ligand and activate downstream targets, whether the same heterodimer is both the “entry” and “signaling” receptor, and whether the same two dimer partners are active in the different stages of the symbiotic response and infection.

How do putative Nod factor receptors participate in symbiotic signaling pathways? The NFR1 and NFR5 Nod factor receptors act upstream of the common SYM pathway. Legumes seem able to interact with rhizobia through the specialized ability of certain LysM-RLKs to recognize rhizobial Nod factors and to transduce this signal through the common SYM pathway. One of the SYM pathway genes has been cloned (SYMRK = NORK) and encodes an RLK with leucine-rich repeats in the extracellular region. In M. truncatula, mutations in NORK and in another SYM gene called DMI1 have the same phenotype, which suggests that the encoded proteins form a complex (6) that recognizes bacterial and fungal symbiotic signals. It is not known how the Nod factor receptor activates the NORK-DMI1 complex, nor how the common SYM pathway leads to two different symbiotic programs.

The identification of a key role for LysM-RLKs in Nod factor perception is a landmark in our understanding of plant-microbe interactions and in oligosaccharide signaling in plants. The development of genomic approaches in model legumes has been instrumental in this success and will benefit future genome-wide analyses of symbiotic gene expression programs in plants. Gene transfer of the LysM-RLK nodulation genes into nonlegumes will establish whether Nod factor responses can be triggered in these plants and how far the process of nodule formation can go. The finding that LysM domains are involved in the perception of well-defined oligosaccharide ligands should facilitate the dissection of this common protein domain. It may also lead to identification of ligands that activate plant orphan receptors so far identified only in genome sequencing projects.

References and Notes

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.

Navigate This Article