INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Biological nitrogen fixation is one of the most important processes in the natural environment: it is the major pathway for the reduction of dinitrogen molecules from air to give ammonia and subsequently glutamine and other nitrogen-containing molecules. Diazotrophic Azospirillum lipoferum enters into associative symbiosis with the roots of several cereals (rice, maize, wheat, and sorghum), tomato, legumes, etc. (16). Under conditions of low oxygen tension and ammonia limitation, bacteria fix nitrogen both in association with roots and in the free-living state (35). A. lipoferum is considered as a potential phytostimulator, since it significantly increases plant growth. However, this is mostly due to production of phytohormone-like substances (16), and only a small amount of the fixed nitrogen is transferred from the bacterium to the plant (4).

Lectins are proteins that recognize and reversibly bind to specific sugar chains of glycosylated molecules (polysaccharides, glycoproteins, etc.) (22). The sugar chains on the cell surfaces play important roles as signals in various cell-to-cell recognition events (reviewed in reference 26). During binding to their cognate cell surface receptors, lectins can cross-link and aggregate these receptors, which can, in turn, lead to a variety of biological responses (reviewed in reference 34).

Lectins exposed on plant roots may contribute to the contact with bacteria in the rhizosphere, due to their ability to distinguish between sugar moieties of the bacterial cell wall. There is evidence that root lectins of leguminous plants are involved in the recognition and subsequent binding to rhizobia (13, 23). Bacterial lectin-binding receptors (cell wall polysaccharide antigens) have been found on the cell surface of Rhizobium trifolii, and the involvement of nod genes in the regulation of their biosynthesis was reported (8). However, the molecular basis of the interaction between the lectins of gramineous plants and Azospirillum, during the process of root association, is unknown. Wheat germ agglutinin (WGA) (29) is the best studied of the closely related grass lectins. WGA is present on whole surface of wheat seedlings and on root tips of adult wheat plants (37). It may therefore be a specific attachment site for the putative receptors (including capsular glycoproteins, polysaccharides, fimbriae, and flagella) of the bacterium and could contribute to bacterial adhesion to the root surface, leading to colonization of wheat roots by Azospirillum (reviewed in reference 9).

WGA is a protein of two identical 18-kDa subunits, with two primary and two secondary independent sugar-binding sites (1, 38). It exhibits sugar binding specificity for two types of N-acetylated sugars: for terminal (14)-linked N-acetyl-D-glucosamine dimers (GlcNAc_di) and for (23)- and (26)-linked terminal N-acetylneuraminic acid (NeuNAc) residues. The primary sugar binding sites link either GlcNAc_di or NeuNAc residues (but never both at the same time), and the secondary binding sites only recognize GlcNAc_di (49, 50).

The regulation of nitrogen fixation in A. lipoferum is less well documented than that in A. brasilense. The regulation is both transcriptional and posttranslational (19) in both species. In A. brasilense, expression of the nifHDK operon (the structural genes of the nitrogenase enzyme) is positively controlled by NifA. nifA is expressed under conditions both compatible and incompatible with nitrogen fixation (30). NifA activity is modulated by the P_II protein (encoded by glnB), the intracellular signal transmitter, in response to the nitrogen status of the cell. glnB is clustered with glnA, the structural gene of glutamine synthetase (GS), required for ammonia assimilation. glnBA is expressed under control of three nitrogen-regulated promoters, glnBp1, glnBp2, and glnAp, active in nitrogen access, nitrogen fixation, and ammonia assimilation, respectively (reviewed in reference 11). In A. lipoferum, the nifHDK operon under NifA-like control has been identified (17), and the glnB gene, contiguous with glnA, has been described. However, the P_II protein has not yet been characterized (reviewed in reference 15).

In this report, we describe the specificity of the recognition event between WGA and the WGA-binding receptor on the cell surface of A. lipoferum. We also report evidence of an enhanced nitrogen fixation capacity of the bacterium, as a consequence of this interaction. Several target genes of the WGA-induced stimulus (nifH, nifA, and glnB) were identified by enhanced expression of the corresponding promoter-lacZ fusions. A preliminary characterization of the putative capsular WGA-binding receptor is reported.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, plasmids, media, and growth conditions. A. lipoferum SpBr17R is rifampin-resistant derivative of SpBr17, the wild-type strain (46). SpBr17R::Tn2706 (Rif^r Cm^r) is a WGA-nonbinding (WGA) mutant strain. A. brasilense Sp7 (46), Azotobacter vinelandii UW136 (14), and Klebsiella pneumoniae UNF122 (24) are wild-type strains. pAW1142::Tn2706 (Cm^r Tc^r) (40) was used as a transposon donor for mutagenesis. pRK2013 (Km^r) is a helper plasmid for conjugative transposon transfer (18). pAB358 (nifH-lacZ Tc^r Km^r), pAB576 (nifA-lacZ Tc^r Km^r) (30), pAB904 (glnB-lacZ Tc^r), and pAB912 (glnA-lacZ Tc^r) (12) carry A. brasilense promoter-lacZ transcriptional fusions. pAB53 (Km^r) (30) and pAB914 (Tc^r) (12) carry the nifA and glnB genes, respectively, of A. brasilense. Complete medium for Escherichia coli and A. lipoferum was Luria-Bertani broth (42). Minimal medium for A. lipoferum (39) contains 29 mmol of KH₂PO₄, 26 mmol of K₂HPO₄, 37 mmol of malic acid, 75 mmol of NaOH, 20 mmol of NH₄Cl, 1.7 mmol of NaCl, 0.81 mmol of MgSO₄, 0.13 mmol of CaCl₂, 0.01 mmol of MnSO₄, 0.02 mmol of Fe₂SO₄, 0.02 mmol of Na₂EDTA, 0.004 mmol of Na₂MoO₄, and 4 mmol of biotin. Cell numbers of A. lipoferum cultures were measured by determining the number of CFU per milliliter in plating assays. Chloramphenicol, tetracycline, rifampicin, and kanamycin were used at concentrations of 10, 20, 20, and 50 µg/ml, respectively.

Random Tn2706 mutagenesis. pAW1142::Tn2706 was introduced into A. lipoferum SpBr17R by triparental plate mating using the pRK2013 helper plasmid. A WGA strain was selected from among the fluorescein isothiocyanate (FITC)-WGA-stained Cm^r Rif^r transconjugants for lack of fluorescence by epifluorescence microscopy (10).

Differential lectin binding assay. Binding of plant lectins to A. lipoferum was tested with FITC-labeled lectins (Sigma) (Table 1). Three-day-old cells, grown on solid minimal medium, were suspended in phosphate-buffered saline (PBS), incubated with lectins (40 µg of lectin/10⁸ cells/ml) for 1 h at 30°C, and then washed in PBS. Lectin binding was quantified (Perkin-Elmer MPF 44B fluorimeter) by measuring mean epifluorescence at 525 nm (E₅₂₅) (10⁷ cells/ml). WGA-Neu and WGA-Glc were obtained by incubation of WGA with a molar excess of NeuNAc or GlcNAc at 30°C for 1 h, and unbound ligand was removed by molecular sieving on Sephadex G-25 gels.

Detection of capsular WGA-binding components. Capsules of 3-day-old cells, grown on solid minimal medium, were solubilized by shaking bacteria in 20 mmol of HEPES-1% Triton X-100 (pH 7.0) for 18 h at 4°C. Protease inhibitors (protease inhibitors set; Boehringer) were added. Capsule-free cells were removed by centrifugation at 8,000 rpm for 20 min at 4°C and discarded. The supernatant was concentrated with a Centricon-10 centrifugal concentrator (molecular weight [MW] cutoff, 10,000), and the protein content was measured as previously described (33). Proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (10% gel) as described by Laemmli (28). Gels were stained with Coomassie brilliant blue R-250. Proteins were transferred to a nitrocellulose membrane and subsequently tested for WGA binding with digoxigenin (DIG)-conjugated WGA (DIG Glycan Differentiation kit; Boehringer).

Isolation of capsular WGA-binding components by WGA affinity chromatography. Capsular fraction was prepared as described above. The solubilized capsular fraction was applied to WGA-agarose matrix (Vector Laboratories, Burlingame, Calif.) and washed with 10 mmol of HEPES-150 mmol of NaCl-0.1 mmol of CaCl₂ (pH 7.5). The bound of material was eluted with 500 mmol N-acetyl-D-glucosamine-10 mmol HEPES-150 mmol NaCl (pH 3.0). Flowthrough and eluted fractions were analyzed by SDS-PAGE (10% gel) (28). Gels were stained with Coomassie brilliant blue R-250.

Assay of nitrogenase activity in the presence of lectins. Cultures were subjected to 2 h of nitrogenase derepression (in nitrogen-free medium at 30°C under N₂-O₂ [99.5:0.5]), and lectin (1 µg/10⁷ cells/ml) and acetylene (10% [vol/vol]) were then added anaerobically. After a further 4 h of incubation (or various times for kinetic experiments), nitrogenase activity was measured by gas chromatography as ethylene production (3).

Transfer of nifH-lacZ, nifA-lacZ, glnBA-lacZ, and glnA-lacZ fusions of A. brasilense into A. lipoferum and -galactosidase assay under lectin stimulus. pAB358, pAB576, pAB904, and pAB912 were transferred into A. lipoferum wild-type and WGA strains by conjugation (45). Transconjugants were selected on minimal medium containing appropriate antibiotics. Nitrogenase was derepressed as described above, and -galactosidase activity (36) was measured 4 h after addition of lectins.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Isolation of a WGA strain of A. lipoferum. A WGA strain defective for FITC-WGA binding was isolated from a Tn2706-mutagenized population of A. lipoferum SpBr17R. FITC-WGA-stained preparations of the wild-type and mutant strains are shown in Fig. 1A and B. In the wild-type cells, binding of FITC-WGA produced fluorescence mostly coupled, without polarity, to the cell capsules. Up to 96% of the cell population was positively stained. In contrast, the WGA strain showed an FITC-WGA dark phenotype, with less than 1% of the cell population showing dim fluorescence. The mutant strain showed wild-type calcofluor staining (data not shown), indicating the presence of intact (13)- and (14)-linked glucans on the cell surface (48).

View larger version (48K): [in this window] [in a new window]   FIG. 1.   Binding of WGA to A. lipoferum cells (A and B) and effect of WGA stimulus on nitrogen fixation (C and D). (A and B) FITC-WGA-stained wild-type (A) and WGA (B) cells visualized by epifluorescence microscopy (approximately 3 × 102). (C and D) Time course of nitrogenase activity in the presence of WGA. After 2 h of nitrogenase derepression in nitrogen-free medium at 0.5% of oxygen, WGA (1 µg/ml) and acetylene (10% [vol/vol]) were added (time zero). Nitrogenase activity (nanomoles of C2H4/107 cells/milliliter) in the presence () and absence () of WGA. ×, growth curve (growth is expressed in 107 cells/milliliter). The standard deviation was less than 12% of the value in each case. The reported values are means of three independent assays.

Stimulation of nitrogenase activity by WGA. The effect of purified WGA in vitro on nitrogen fixation of the wild-type and WGA strains was investigated (Fig. 1C and D). In the absence of WGA, a similar optimum for nitrogen fixation was determined at 0.5% oxygen, both in the wild type and in the WGA strain. Under lower oxygen tensions, the nitrogen-fixing capacity decreased. The optimum oxygen requirement for nitrogen fixation was not modified by addition of WGA (data not shown).

Involvement of terminal sugar residues of the cell surface in the stimulation of nitrogen fixation by lectin. Binding of lectins to the bacterium is essentially mediated by terminal sugar residues of the corresponding receptors on the cell surface. Terminal sugar residues on the surface of the wild-type and WGA cells were surveyed by a differential lectin binding assay, using FITC-labeled lectins with different sugar-binding specificities (Table 1). All lectins tested except UEAI, were able to attach to the wild-type cells, as assessed by epifluorescence microscopy, and no agglutination of cells was observed (at 1 µg of lectin/10⁷ cells/ml) (data not shown). Consequently, terminal GlcNAc_di, terminal (23)- and/or (26)-linked NeuNAc, GalNAc, and/or -lactosyl-, -mannosyl-, and/or -glucosyl- but not L-fucosyl-(12)-linked D-galactose (Gal) residues are present on the cell surface. The specific fluorescence intensities (E₅₂₅/10⁷ cells/ml) of the cells treated with WGA, SNA, and MAA were comparable (Table 2). The fluorescence levels obtained with DSA and WGA-Neu were 23 and 66% lower, respectively, than that with WGA. No fluorescence was observed with WGA-Glc, as expected. The cell surface of the WGA strain was similarly probed with FITC-lectins. WGA, WGA-Neu, WGA-Glc, and DSA were not able to bind to the cells, whereas SNA and MAA showed a level of binding comparable to that obtained with the wild-type strain (Table 2).

View this table: [in this window] [in a new window]   TABLE 2.   Lectin-binding capacities of wild-type and WGA strains

View larger version (32K): [in this window] [in a new window]   FIG. 2.   Nitrogenase activity and expression of N-regulated genes in the wild-type strain in the presence of lectins. Lectins (at 1 µg/107 cells/ml) were added after 2 h of nitrogenase derepression. Nitrogenase activity (nanomoles of C2H4/107 cells/milliliter) and -galactosidase activity (Miller units/milligram of protein) were measured 4 h after addition of WGA. Error bars indicate the standard deviation of the mean. The standard deviation of nitrogenase activity was less than 13% of the value in each case. The values are means of three independent experiments.

Stimulation of nifH, nifA, glnB, and glnA expression by lectin binding. Next, we tested whether the lectin stimuli exert a regulatory effect through modulation of the transcription of nifH, nifA, glnB, and glnA genes. Promoter fusions nifH-lacZ, nifA-lacZ, glnBA-lacZ, and glnA-lacZ (plasmids pAB358, pAB576, pAB904, and pAB912, respectively) (12, 30) were introduced into the wild-type and WGA strains. The -galactosidase activities of the plasmid-borne fusions were monitored upon stimulation with WGA, WGA-Neu, DSA, SNA, and MAA, under conditions of nitrogen fixation. For the wild-type strain (Fig. 2), the lectins linked to GlcNAc residues of the cell surface (WGA, WGA-Neu, and DSA) increased the expression of nifH-lacZ, nifA-lacZ and glnBA-lacZ fusions. -Galactosidase activities obtained with WGA, WGA-Neu, and DSA stimuli were 2.9-, 1.8-, and 1.8-fold (nifH-lacZ), 5.7-, 3.3-, and 5.2-fold (nifA-lacZ), and 2.6-, 1.5-, and 2.1-fold (glnBA-lacZ), respectively, higher than the corresponding unstimulated -galactosidase activities. Expression of glnA-lacZ was not enhanced by any lectin stimuli. Neither of the NeuNAc-specific lectins (SNA and MAA) had an effect on the transcription of any fusion. No increased transcription of the promoter fusions by the lectin stimuli was detected in the WGA strain, as expected (data not shown). These results are in agreement with those obtained for the stimulation of nitrogenase activity (Fig. 2). In a control experiment, extra copies of functional nifA and glnB genes (provided by pAB53 and pAB914, respectively) (12, 30) were introduced into the wild-type strain (Table 4). In the absence of WGA stimulus, nitrogenase activity was 2.1 (nifA) and 2.4 (glnB) times higher than in the parental strain with only a single chromosomal copy of the corresponding genes. Addition of WGA led to no further increase. Significantly, the overexpressed nifA and glnB led to levels of nitrogenase activity comparable to that induced by the WGA stimulus in the parental strain. Similarly, in the nifA and glnB merodiploid derivatives of the WGA strain, a level of nitrogen-fixing capacity comparable to the wild-type background level was produced, both in the presence and in the absence of WGA (Table 4).

View this table: [in this window] [in a new window]   TABLE 4.   Nitrogen fixation in the presence of WGA in wild-type and WGA strains carrying plasmid-borne extra copies of nifA (pAB53) and glnB (pAB914) genes

Detection and isolation of capsular WGA-binding components. Observation of WGA-stained wild-type cells under an epifluorescence microscope (Fig. 1A) (10) showed that WGA was mainly attached to the cell capsules. We attempted to identify the capsular proteins including the WGA-binding components. The solubilized capsular material was separated by SDS-PAGE (Fig. 3A). The protein profiles of the wild-type and WGA strains were comparable. Six distinct polypeptides were revealed. On the basis of its intensity, the 32-kDa band might correspond to a doublet of two closely comigrating polypeptides. The 45- and >116-kDa polypeptides are comparable in size to the 45- and 110-kDa protein components of the lateral and polar flagella of A. brasilense, respectively (5). A Western blot of the protein patterns was probed with DIG-conjugated WGA (Fig. 3B). WGA-binding signal was detected only in the wild-type and not in the WGA preparation. This single WGA-binding component was identified as a 32-kDa polypeptide.

View larger version (19K): [in this window] [in a new window]   FIG. 3.   Detection and isolation of the putative WGA-binding receptor. (A and B) Migration patterns of capsular proteins in SDS-PAGE. (A) Protein profiles; (B) Western blot probed with DIG-labeled WGA. Lanes: a, wild-type strain; b, WGA strain. (C) Isolation of the receptor protein by WGA affinity chromatography. Protein migration patterns were determined by SDS-PAGE of unfractionated capsule (lanes u), flowthrough (lanes f), and eluted fractions (lanes e2 and e3) of capsule preparation from the wild-type (wt) and WGA strains; lane M, protein molecular weight standards (in thousands).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In this report we demonstrate that two GlcNAc-specific lectins, WGA and DSA, are able to stimulate the nitrogen-fixing capacity of A. lipoferum, presumably after formation of a complex with the corresponding receptor on the bacterium cell surface. The stimulus, which led to the up-regulation of nitrogen fixation, is absent when the lectin-bacterium interaction is prevented, as in the WGA strain.

Moreover, the stimulus is efficient with plant-associative bacteria, such as A. lipoferum and A. brasilense, but not with the free-living diazotrophs Azotobacter vinelandii and K. pneumoniae. Interestingly, not only WGA but also other lectins of a few additional plants, colonized by Azospirillum (16), bind also to NAcGlc (41).

The WGA stimulus elicited an increased expression of nifH-lacZ, nifA-lacZ, and glnBA-lacZ fusions and consequently a higher nitrogenase activity (Fig. 2). Presumably, the stimulus from the WGA-receptor complex enhances the transcription of nifA. The elevated level of NifA could, in turn, increase nifH expression. Indeed, extra copies of nifA, introduced into wild-type cells, were also able to enhance the nitrogenase activity (Table 4). Similarly, the WGA-receptor-elicited signal would also trigger the transcription of glnB. An increased level of the P_II protein would improve the conversion of excess NifA protein to its active form. Accordingly, when glnB was overexpressed in the wild-type cells, nitrogenase activity was enhanced (Table 4). Note that the plasmid-borne glnA promoter did not respond to the WGA stimulus (Fig. 2). However, the surplus ammonia, produced by WGA-triggered nitrogen fixation, could be assimilated by an elevated level of GS (12), since the glnBA mRNA is synthesized in larger amounts. Both an enhanced nitrogen fixation, induced by the WGA stimulus, and subsequently an increased level of glutamine synthetase (GS) have been reported in A. brasilense (2).

No information is available about lectin-induced bacterial signalling pathways in plant-bacterium associations. In mammalian cells, the binding of WGA triggers cellular recognition, glucose transport and lipolysis, cell growth inhibition, control of morphological states and activation of transcriptional factors, and several lectin-induced signal transducing pathways have been reported (7, 43, 51). The sugar residues of the lectin binding receptors are believed to transmit the signal (20, 27).

The binding specificity of the lectins is generally determined by analyzing terminal sugars of the carbohydrate side chains of the corresponding receptors (22). Our data contribute to understanding the specific chemical recognition between WGA and the putative receptor of A. lipoferum. The GlcNAc_di residues are the specific receptor components at the WGA-receptor binding interface and are responsible for the biologically active stimulus for the nitrogen fixation machinery. Although NeuNAc residues are also present on the cell surface (evidenced by the binding of SNA and MAA) (Table 2), presumably they are not required for the WGA-binding: the WGA strain binds with SNA and MAA but not WGA, WGA-Neu and DSA. Furthermore, the NeuNAc termini are not sufficient for the stimulation of nitrogen fixation, as shown in the case of the SNA-receptor and MAA-receptor interactions, both in the wild type and WGA strains (Fig. 2). Moreover, binding to the primary and secondary sugar binding sites on WGA were differently effective in elucidating the biologically active stimulus. The receptor, linked exclusively by the secondary sites of WGA (WGA-Neu-receptor interaction), produced only 38% of the total stimulatory effect (Fig. 2). This agrees with the reduced WGA-Neu-binding capacity of the bacterium (Table 2). The lower level of stimulation may be due to a poorer accessibility of the secondary binding sites in the WGA molecule for the cognate sugar residues of the receptor (49).

The number of molar GlcNAc links between the receptor and WGA modulates several WGA-triggered biological functions (32). From hemagglutination inhibition assays, we estimate that DSA possesses two GlcNAc sites (unpublished data) whereas the WGA molecule has four. Thus, the different stimulation strengths of WGA and DSA (Fig. 2) may be explained by the different number of GlcNAc moieties in the WGA-receptor and DSA-receptor interfaces.

A 32-kDa polypeptide was isolated in vitro, as a putative capsular glycoprotein of WGA-binding receptor. The defective receptor protein of the WGA strain has a molecular weight similar to that of the wild type. Thus, the structural gene encoding the receptor protein is most probably not inactivated by the mutation in the WGA strain. Possibly the glycosylation pattern of the protein is affected preventing the cognate linkage with the sugar-binding sites on the WGA molecule.

The interaction between Azospirillum and WGA may contribute to the adhesion of the bacteria to the root surface and to establishing a nitrogen-fixing association of improved efficiency with the wheat host. Presumably, the elevated nitrogen fixation capacity of the bacterium is supported by enhanced carbon and energy supply from the host plant.