Electrifying symbiosis.

A role of ion channels in the interactions between bacterial pathogens and their animal hosts is well established. In some cases, regulation of endogenous channels in the target species is disrupted by bacterial toxins, resulting in an alteration of ion flux and a perturbation of cellular function (1). In a more extreme example, hemolytic bacteria actually insert poreforming proteins into the plasma membranes of their hosts, triggering massive loss of accumulated ionic solutes and cell death (2, 3). In plants as well, pathogenic bacteria may produce or elicit signals of an electrical nature; for example, electrical signaling has been implicated in plant responses to wounding (4). But not all plant-bacteria interactions are pathogenic. One example of a beneficial interaction can be found in the symbiotic relationship between legumes and rhizobia bacteria of the genera Rhizobium, Bradyrhizobium, andAzorhizobium (reviewed in refs. 5 and 6). This symbiosis results in the formation of a root nodule, within which the bacteria reduce atmospheric nitrogen into forms accessible to the plant and the plant provides organic carbon substances that sustain the bacterial cells. It has been speculated that this symbiotic relationship evolved from one that was initially pathogenic (7), and so it should perhaps not be surprising to learn that recent studies now implicate electrical signaling and ion channels in the communication processes that initiate nodule formation. The communication between legume and bacteria is quite specific, as with rare exceptions just one strain of rhizobia will successfully form nodules on a given legume species. Host-bacteria recognition begins with flavonoids exuded from the roots, which serve as chemoattractants and growth enhancers of particular bacterial strains. In many bacterial species, flavonoids also enable activation of nod genes by a transcription factor encoded by nodD. Nodulation (nod) genes are bacterial genes involved in the nodulation response; some of these (e.g., nodABC) are conserved in all rhizobia, whereas others (e.g., nodE) are found only in certain strains and help to confer host specificity. The flavonoid requirement is not universal, so flavonoids are viewed as only broad range determinants of host specificity. The ultimate determination of host specificity resides in bacterial compounds synthesized by nod gene-encoded enzymes and secreted into the rhizosphere. These compounds, called Nod factors, are Nacetylglucosamines, and the side chain composition of these substituted oligosaccharides determines the host range of the bacteria. The mechanism by which a given Nod factor is recognized only by the appropriate host species is still a mystery. However, a few years ago it was shown that Nod factor NodRm-IV(S) of Rhizobium meliloti causes membrane depolarization in root hair cells of the host plant, alfalfa, but not in root hairs of a nonlegume, tomato (8). In theory, mechanisms of membrane depolarization include cation uptake, anion efflux, and inhibition of the hyperpolarizing H+-ATPase found on the plant plasma membrane. In practice, the ionic basis of this depolarization has yet to be defined. Nevertheless, the results of Ehrhardt et al. (8) provided the first evidence that electrical signaling is important in this highly evolved interaction. Now, data recently published in this journal by Downie's group (9) provide a glimpse into one mechanism by which bacteria may affect the membrane potential and electrical status of the host cells. The study of Sutton, Lea, and Downie (9) concerns the legumes pea (Pisum) and vetch (Vicia) and their nodulation by Rhizobium leguminosarum biovar viciae. Downie and Surin's (10) previous assays of nodulation capacity by various R. leguminosarum biovar viciae mutants set the stage for the present experiments and are one example of how the powerful tools of molecular biology have been utilized to dissect the genetics of nitrogen fixation. A deletion mutant of R. leguminosarum biovar viciae lacking seven host-specifying nod genes, nodEFLMNTO, produces a Nod factor that has a C18:1 fatty acid side chain rather than the C18:4 fatty acid that is normally synthesized. This mutant is unable to form nodules (10). Nodulation on pea and Vicia hirsuta is partially restored by plasmids carrying nodEF, and this is reasonable because the products of these genes are implicated in biosynthesis of the C18:4 fatty acid (11). However, surprisingly, nodulation capacity on V. hirsuta is also partially restored to the deletion mutant by a plasmid carrying nodO, even though mutation of nodO in an otherwise wild-type bacteria does not affect nodulation (10, 12). The nodO gene does not appear to be involved in the synthesis of Nod factor; rather, this gene encodes a secreted Ca2+-binding protein with homology to the bacterial pore-forming hemolysins (13). The structural similarity of NodO to hemolysins inspired Sutton et al. (9) to test the pore-forming properties of this rhizobial product. Indeed, when Sutton et al. (9) purified NodO and inserted the protein into artificial bilayers, they found that it formed a cation-selective ion channel. These researchers speculate that NodO permits cation uptake in vivo, thereby producing or enhancing the membrane depolarization apparently associated with successful bacteria-plant interactions. Sutton et al. (9) further speculate that the channel formed by NodO may be Ca21 permeable, thus mediating increases of cytosolic Ca2+ and initiation of Ca21dependent signal transduction cascades, including gene regulation (14). Elevation of Ca2+ in the bilayer system also decreased conductance of NodO channels, and this could perhaps serve as a negative feedback mechanism in vivo, preventing intracellular Ca2+ concentrations from reaching cytotoxic levels. To build on the pioneering in vitro study of Sutton et al. (9), a key experiment now is to demonstrate that NodO does in fact insert into legume membranes. Since it is known that bacterial hemolysins insert only in bilayers of certain lipid composition (2), it is critical to perform such in vivo experiments. One way to assay for insertion would be to use the method of Ehrhardt et al. (8) to monitor for changes in membrane potential upon application of purified NodO. If NodO does trigger membrane depolarization, this effect may increase the efficacy of, or increase the plant's sensitivity to, an abnormally weak signal caused by aberrant Nod factors, thereby allowing the nodulation signal to be successfully transmitted. This may explain why NodO can partially complement the nodEFLMNTO mutant, for example. Because NodO is not necessary for nodulation when other nod genes are intact, it seems likely that R. leguminosarum biovar viciae Nod factor itself (and possibly Nod factors of other species as well) causes membrane depolarization by an alternate route, not requiring the cation channel of NodO. A more speculative explanation would be that the host plant integrates two types of nodulation signals: a depolarizing