- Top of page
- Materials and Methods
The soil bacterium Sinorhizobium meliloti is able to establish a nitrogen-fixing symbiosis with its host plants alfalfa (Medicago sativa) and the model legume Medicago truncatula. This symbiotic interaction leads to the formation of a completely new organ, the so-called root nodule that hosts the rhizobia intracellularly. Inside the nodules, they are able to fix atmospheric nitrogen and convert it to ammonia, which is then provided to the plant (Hirsch, 1992).
Since plants encounter not only beneficial microbes, but also have to cope with phytopathogenic microorganisms, they have developed a set of defence mechanisms that enables them to identify pathogens by so-called PAMP (pathogen-associated molecular patterns) and start suitable defence reactions, such as the induction of a hypersensitive response (HR) or the production of reactive oxygen species, as it was lately reviewed by Montesano et al. (2003). By contrast, in symbiosis, a complex molecular dialogue between the two interacting partners regulates plant defence reactions and thus enables symbiosis (Djordjevic et al., 1987).
In addition to plant flavonoids (Brewin, 1991) and rhizobial nodulation factors (Roche et al., 1991), bacterial surface saccharides such as exopolysaccharides (EPS) and lipopolysaccharides (LPS) form another important class of signal molecules for an effective symbiosis (Niehaus & Becker, 1998). These molecules are most likely involved in the infection process during establishment of the symbiosis. An indication of the significance of LPS as a signal molecule could be deduced from bacterial mutants defective in LPS biosynthesis. Various rhizobial mutants with alterations in LPS were defective in symbiotic interactions with their host plants (Carlson et al., 1987; Lagares et al., 1992; Campbell et al., 2003). lpsB mutants of S. meliloti exhibit an altered LPS and fail to establish an effective symbiosis with M. truncatula (Niehaus et al., 1998). The symbiotic defect of the lpsB mutant was correlated with typical indications for plant defence reactions like the production of reactive oxygen species. By contrast, on M. sativa an effective symbiosis was established by this mutant, indicating a different specific recognition of the LPS by alfalfa than by M. truncatula. Recently, the role of rhizobial LPS in symbiosis was partly characterized by Albus et al. (2001), who showed that the isolated LPS was able to suppress the elicitor-induced oxidative burst in cell suspension cultures of the host plant M. sativa. Interestingly, in the nonhost plant Nicotiana tabacum (tobacco) they exhibit an opposite effect and elicit the plant pathogen defence comparable to that of LPS isolated from the phytopathogen Xanthomonas campestris pv. campestris (S. Braun and A. Meyer, pers. comm.).
In this work, we examined in more detail, which parts of the S. meliloti lipopolysaccharides are recognized by host and nonhost plants. For this purpose, we compared the suppressor, and elicitor activity respectively, of the wild-type LPS with a core- (Sm6963) (Lagares et al., 1992) and a Lipid A-mutant (acpXL) (Sharypova et al., 2003) and pure Lipid A obtained by hydrolysis of wild-type LPS after elicitation in M. sativa, M. truncatula and N. tabacum cell cultures.
By using LPS of S. meliloti mutants and chemically obtained substructures the biological active part of the molecule could be defined in M. sativa and M. truncatula cell cultures.
- Top of page
- Materials and Methods
During the establishment of the root nodule symbiosis a massive infection of the host plant takes place. In the case of infection by wild-type rhizobia, no obvious signs of plant defence are visible. A closer look at the curled root hair and the initiation of the infection thread revealed a moderate oxidative burst at the primary site of infection (Santos et al., 2001). In addition, excessive infection of the host plant is prevented by defence-like abortion of infection threads inside the root hair (Vasse et al., 1993). Obviously, plant defence reactions are involved in the early steps of this beneficial plant–microbe interaction. Plant mutants that defeat an infection by rhizobia (Mitra & Long, 2004) and rhizobial mutants defective in EPS (Niehaus et al., 1993) or in LPS (Niehaus & Becker, 1998) biosynthesis induce the plant defence system. These findings indicate that an active suppression of plant defence could take place during the establishment of this symbiosis. Evidence supporting this concept came from the observation that isolated S. meliloti lipopolysaccharides were able to suppress an elicitor-induced oxidative burst in cell cultures of M. sativa, but not in nonhost cell cultures of N. tabacum (Albus et al., 2001).
To further characterize the suppressor activity of the S. meliloti LPS, a cell suspension culture of the model legume M. truncatula was established and selected for responsiveness against the general elicitor invertase (Fath & Boller, 1996). Simultaneous application of invertase and S. meliloti 2011 wild-type LPS led to the production of less hydrogen peroxide in the M. truncatula cell culture than application of invertase alone. Interestingly, the general level of suppressor activity of the S. meliloti LPS was much lower compared with the M. sativa cell culture. One explanation for these results may be that M. truncatula is a Mediterranean legume and therefore not a perfect host plant that S. meliloti strain 2011 encounters in its natural environment. This would indicate that the biologically active LPS contributes to the host-specificity, even though it was published that the S. meliloti LPS carries not enough structural information to be strain-specific (Reuhs et al., 1998).
The next step in characterizing the biological activity of the S. meliloti LPS was the identification of the suppressor-active part of the LPS molecule. Two S. meliloti LPS mutants (6963 and L994) were examined for suppressor activity in cell suspension cultures of M. sativa and M. truncatula. Mutant 6963 carries a defect in the lpsB gene and produces LPS with a (probably) changed core region (Lagares et al., 1992; Kanipes et al., 2003) that generates a delayed Fix+ phenotype in M. sativa and a Fix− phenotype in M. truncatula (Niehaus et al., 1998). Isolated LPS from this mutant acted as a suppressor in M. sativa and in M. truncatula cell suspension cultures. The Fix+ phenotype on M. sativa and the Fix− phenotype on M. truncatula (Niehaus et al., 1998) is not as clearly reflected on the cell culture level. Nevertheless, the suppressor effect was always weaker in M. truncatula cell cultures compared with M. sativa, as was reported earlier in this study for the wild-type LPS.
From this observation two functions can be proposed for the rhizobial LPS: one early function, most probably during the infection of the root hairs, suppressing the oxidative burst, and a late function involved in the endocytotic uptake of rhizobia in the infected nodule cells.
The same suppressor activity was obtained for the LPS from S. meliloti mutant L994, which carries a mutation in the acpXL gene. This leads to the loss of the long-chain fatty acid in the lipid part of the LPS and induces a delayed Fix+ phenotype in symbiosis (Sharypova et al., 2003). Lipopolysaccharides of the mutant L994 displayed the same level of suppression in cell cultures of M. sativa and M. truncatula as wild-type LPS. This result is interesting, because it implies that the most noticeable characteristic feature of rhizobial LPS, the hydroxylated C28 fatty acid (Hollingsworth & Carlson, 1989; Bhat et al., 1994), is not necessary for the recognition of the LPS by host plants in order to suppress the production of reactive oxygen species. It is possible that the delayed nodulation phenotype results from a beneficial effect of the C28 fatty acid in later stages of symbiosis, such as the release of the bacteria from the infection thread or uptake into the plant cells, or in the establishment of functional symbiotic membranes, as reported previously for the acpXL mutant of Rhizobium leguminosarum in pea (Vedam et al., 2004).
Since both mutants analysed so far in this study did not show differences in their ability to suppress the oxidative burst reaction in the host plants M. sativa and M. truncatula, the most basic form of LPS, Lipid A, was used for further experiments. The chemically released Lipid A moiety of the LPS was also able to suppress the production of reactive oxygen species with the same efficiency as the complete LPS in both Medicago cell cultures tested. A solvent (DMSO) was necessary to dilute the hydrophobic Lipid A and to increase the bioavailability of the molecule. To date, this recognition of bacterial Lipid A has only been known for endotoxin in the animal immune system (Galanos et al., 1986; Ulmer et al., 2002).
It would be interesting for future studies to compare the suppressor activity of LPS and Lipid A of S. meliloti with other LPS and Lipid A species from the Rhizobiaceae and other, non-symbiotic Enterobacteriaceae such as Escherichia coli.
To elucidate, whether the structural features recognized by the host plants are specific for the rhizobia-legume symbiosis, the effects of LPS and Lipid A (Fig. 4) from S. meliloti were tested for elicitor activity in cell suspension cultures of N. tabacum. The isolated entire lipopolysaccharides induced a potent and long-lasting oxidative burst in the cell cultures as reported by Albus et al. (2001), but the addition of Lipid A did not result in the formation of extra hydrogen peroxide. Obviously, Lipid A represents no elicitor-active structure. This finding also implies that at least a part of the carbohydrate LPS substructure is necessary for elicitation.
Figure 4. Chemical structure of the Sinorhizobium meliloti Lipid A with additional 3-deoxy-d-manno-2-octulosonic acid (KDO) residues. Shown here is the structure of the S. meliloti Lipid A as it was published recently. It consists of two glucosamines, two phosphates, five fatty acids and a hydroxy-butyrate. In this image two additional KDO residues were added to the Lipid A structure (Kanipes et al., 2003).
Download figure to PowerPoint
For plants, it is of vital importance to recognize invading pathogens in order to start suitable defence reactions, but at some point in evolution it became more beneficial to host invading nitrogen-fixing bacteria instead of defending against them (Sharifi, 1983; Quispel, 1998). This hypothesis is substantiated by the finding that M. truncatula expresses two symbiosis-specific nodulins that share homologies to pathogenicity-related genes (Gamas et al., 1998). In this work, it was shown that even the basic Lipid A structure shows the same biological activity in host plant cell cultures as the LPS, but was completely inactive in the non-host plant tobacco. These results indicate a clear difference in the recognition of LPS in pathogenicity and symbiosis in plants, generating an even more complex picture of LPS perception and signalling in plants than already suggested. In this background, it becomes more comprehensible that M. sativa and M. truncatula might possess two different perception systems for LPS: one for the recognition of pathogenic Gram-negative bacteria and another for the symbiosis-specific recognition of rhizobia.