PerspectivePlant Science

Infectious Heresy

See allHide authors and affiliations

Science  01 Jun 2007:
Vol. 316, Issue 5829, pp. 1296-1297
DOI: 10.1126/science.1143870

Leguminous plants, such as peas and soybeans, enter into a symbiotic relationship with soil bacteria called rhizobia. For years it has been the accepted wisdom that Nodulation (Nod) factors secreted by rhizobia enable them to infect a legume and initiate formation of nodules on the host plant's roots. Within these nodules, the bacteria convert free nitrogen to ammonia, which the plant uses for its growth. On page 1307 in this issue, Giraud et al. (1) provide evidence that overturns this orthodoxy. They determined that the genomes of two strains of legume-nodulating rhizobia do not contain genes that are necessary for the synthesis of Nod factors. This means that these bacteria must have an alternative way of initiating the dialogue that results in legume nodulation.

The symbiosis between legumes and rhizobia is an agriculturally important relationship because such legumes grow well without added nitrogen fertilizer. Nod factors activate a plant signaling pathway that induces oscillations in the concentration of intracellular calcium (called calcium spiking). Calcium spiking triggers the expression of genes that are required for nodule morphogenesis in roots (2). Nodulation is intimately linked with the establishment of threadlike structures that convey rhizobia into nodule cells (see the figure).

The two rhizobia (ORS278 and BTAi1) sequenced by Giraud et al. are unusual because they induce nodule formation on both stems and roots of some legumes. Stem nodulation can occur on some legumes that undergo periods of flooding, which enables colonization of stems by rhizobia. Another oddity is that these strains engage in photosynthesis and contain a cluster of photosynthesis-related genes that are similar to those found in closely related photosynthetic bacteria. However, the real surprise is that neither strain contains nodA or nodC genes, which are essential for the synthesis of Nod factors. NodC produces chitin oligomers from uridine 5'-diphosphate-N-acetylglucosamine, and NodA transfers a long-chain acyl group onto the chitin molecules to form a core Nod-factor structure, which can then be modified by other nod gene products (3).

Getting, or not getting, the Nod.

Nod factors produced by symbiotic bacteria infect the legume host and initiate the formation of nitrogen-fixing nodules (left). Nod factor-independent nodulation may result from rhizobial colonization through cracks in the root epidermis (right). Increased sensitivity to a cytokinin-like signal may allow part of the normal Nod-factor signaling cascade (middle) to be bypassed, leading to nodulation and infection in the absence of Nod factors.


How do ORS278 and BTAi1 infect and induce development of nitrogenfixing nodules on their hosts, the aquatic legumes Aeschynomene sensitiva and Aeschynomene indica? Not only are these bacteria unusual, but so are these legumes. To demonstrate this, Giraud et al. used a nod-gene containing Bradyrhizobium strain (ORS285), which nodulates other legumes in addition to A. sensitiva. Mutation of the nod genes in ORS285 does not affect nodulation of A. sensitiva but does block nodulation of the other legumes. This shows that unlike the other legumes, A. sensitiva can respond to a rhizobial signal other than Nod factor.

How does A. sensitiva induce nodule development without Nod-factor signaling? An important clue comes from recent work on legume mutants that spontaneously form nodules in the absence of any bacteria or Nod factors. Constitutive activation of lotus histidine kinase 1 (LHK1), a cytokinin hormone receptor (4) that is required for nodulation (5) in the legume Lotus japonicus, results in spontaneous formation of nodules in the absence of rhizobia or Nod factors. This activated cytokinin receptor appears to bypass several of the early steps in the Nod factor-activated signaling cascade, including the gene products required for calcium spiking (4). Other evidence supporting a nodulation role for cytokinins is that a strain of Sinorhizobium meliloti that lacks Nod factor but is genetically engineered to secrete a cytokinin, can initiate nodule morphogenesis (6). Also, treatment of certain legumes with cytokinins induces the expression of several genes that are expressed during normal nodule development (7).

To identify the signal that ORS278 and BTAi1 use to initiate nodulation, Giraud et al. screened for mutants of ORS278 that could not induce nodule formation in A. sensitiva. Although no mutants completely defective for nodulation were found, some were greatly impaired for nodulation. Several of these mutants had defective purine biosynthesis. Because plant cytokinins are derived from the purine adenine, this suggested that the bacterial mutants might produce smaller amounts of a cytokinin (or cytokinin precursor) that initiates nodulation. Supporting a possible role for such a signal during conventional nodulation is the observation that other rhizobia carrying mutations in purine biosynthesis are defective for infection of nodules in common legumes (8). Furthermore, proteins similar to a previously unrecognized plant cytokininactivating enzyme, encoded by the rice LONELY GUY gene, may be present in agrobacteria (9). However, in the absence of a mutant that is completely nodulation defective, Giraud et al. cannot formally distinguish between a signal that activates the entire nodulation signaling pathway or a signal that bypasses the signaling pathway by stimulating production of a cytokinin.

If Nod factors are not required for nodulation in some legumes, why should the Nod factor- induced nodulation pathway be so predominant? Formation of nitrogen-fixing nodules requires rhizobial infection, and this requires entrapment of the bacteria. During conventional infection, Nod factors induce roothair deformation, leading to bacterial entrapment. Subsequent growth of infection threads requires modification of Nod factors (as specified by nod genes), and probably increased concentrations of Nod factors. In some legumes, however, bacteria gain entry to roots through cracks in the epidermis, often at the sites where lateral roots emerge. In another well-studied stem-nodulating legume (Sesbania rostrata), this type of infection shows much less specificity for Nod-factor structure, although it does require some type of Nod factor (10). The mode of infection of A. sensitiva and A. indica by ORS278 and BTAi1 shows similarities to that seen in S. rostrata. Such entry through cracks may allow bacteria to accumulate and form a signaling center (10) that can induce nodulation and infection (see the figure). Nodule development could occur if a cytokinin-type signal accumulates in such a signaling center and bypasses the early Nodfactor signaling pathway. However, induction of infection-thread growth without Nod factors is highly unusual. Therefore, for both nodule and infection-thread development to occur in this system implies that A. sensitiva and A. indica may enhance intercellular colonization by bacteria and/or be unusually sensitive to some bacterially made signal. Whether this signal is a cytokinin or not remains to be established.


View Abstract

Navigate This Article