3.1. Nod factor-induced root hair formation, a biologically relevant assay. Upon addition of 10"9 M PI Nod factors (Mergaert et al. 1997), bushes of axillary root hairs were observed, specifically at bases of lateral roots, at the sites where root nodules develop upon inoculation with A. caulinodans ORS571 (Ndoye et al. 1994). The efficiency of Nod factor-induced formation of axillary root hairs was high, since over 70% of the bases of lateral roots exhibited this phenomenon, when inspected five days after inoculation. Nod factors did not only induce the formation of axillary root hairs, but also root hair deformation. Most root hairs present in bushes of axillary root hairs showed moderate distortions, including tip swellings, branches, and corkscrews. Shepherd's crooks were not observed.
To determine whether root hair formation is also induced during nodule initiation, roots were inoculated with the ORS571 wild-type strain and scored for root hair formation at bases of lateral roots. As early as 15 hpi the presence of axillary root hair bushes was obvious. At 24 hpi, a similar phenotype was observed, and at 36 hpi, swelling of lateral root bases was apparent, indicative for nodule primordium formation. The root hair formation, occurring during nodule initiation, was Nod factor-dependent, since ORS571-V44, a nodA mutant that is unable to synthesize Nod factors, did not induce the formation of axillary root hairs. These observations suggested that Nod factor-induced root hair formation is a biologically relevant assay to study Nod factor-induced responses during nodule initiation by ORS571.
3.2. Ethylene synthesis and perception required for Nod factor-induced root hair formation and nodule initiation. The root hair formation assay was used to test the effect of inhibitors of ethylene synthesis (AVG, AOA, AIB, Co2+) and perception (Ag+, 2,5-norbornadiene) on Nod factor-induced root hair formation. Inhibitors were added prior to Nod factors. All inhibitors tested completely blocked induction of Nod factor-induced axillary root hair formation. Treatment of roots with exogenous ethylene induced the formation of root hairs. These appeared as bushes of axillary root hairs, but, in contrast to application of Nod factors, an enhanced number of root hairs was also detected on the primary root and lateral roots, only straight but not deformed root hairs could be observed, and the efficiency was ten-fold lower.
Some of the ethylene synthesis and perception inhibitors were tested for their influence on root nodule development on S. rostrata. When added two days prior to ORS571, all inhibitors tested blocked nodulation completely, scored approximately 10 days post inoculation. No swellings could be observed macroscopically. Controls for the effect of those inhibitors on flavonoid secretion, bacterial growth, and Nod synthesis and production were performed.
To investigate whether the Nod factor-induced ethylene production is only needed for nodule initiation or also during later stages of nodule development, the number of root nodules was counted during 14 days and compared between roots that were treated with Ag+, added at different time points in respect to the wild type. When Ag+ ions were added 1 or 2 days prior to the wild-type strain, or at the same moment, a complete block of nodule initiation was noticed. When the inhibitor was added 1 or 2 days after, approximately 3 and 7 nodules were formed, respectively, and the number did not increase in time. These were probably nodules that had been initiated during the time period at which the ethylene perception inhibitor was not yet present. This observation was in sharp contrast to the kinetics of nodule formation by ORS571 alone, that exhibited a continuous increase in nodule number within the period of inspection, suggesting a continuous process of nodule initiation. A similar experimental set up was achieved for AO A, an inhibitor of ethylene synthesis, and a so far identical graph was obtained. Thus, Nod factor-induced ethylene production is only required for nodule initiation.
3.3. Nod factor-induced root hair formation and nodulation require hydrogen peroxide. A proposed way for the synthesis of hydrogen peroxide in plants involves an NADPH oxidase complex and super oxide dismutase, the function of which is inhibited by DPI and DDC, respectively. These inhibitors, together with ascorbic acid, a scavenger of hydrogen peroxide, blocked Nod factor-induced root hair formation. Exogenous hydrogen peroxide induced the formation of axillary root hairs, that were, similar to ethylene-induced root hairs, not deformed. The efficiency of hydrogen peroxide-induced root hair formation was approximately five times lower than when ethylene was applied.
All compounds used above to inhibit the generation of reactive oxygen species also blocked nodulation when added two days prior to the wild-type strain. Thus both the ethylene production and the generation of reactive oxygen species are needed for Nod factor-induced root hair formation and nodulation. When ethylene was added to roots that were treated with DPI, bushes of axillary root hairs were observed. Although this observation is preliminary, it suggests that an oxidative burst may precede the ethylene production.
3.4. Morphological effects of Ag+ and DPI on ORS571-induced root nodule initiation. To be able to evaluate the effect of Ag+ and DPI on invasion and nodule development, a morphological approach was applied. Bases of lateral roots were imbedded in Technovit, semi-thin sections were made in a transversal direction, and stained with toluidine blue.
Initially, morphological aspects of lateral root bases of mock-inoculated roots, harvested 2, 6, and 10 dpi, were studied. Sections of material harvested at these three time points showed similar features. The vascular bundle is surrounded by cortical tissue that consists of typical, highly vacuolated cells, and big intercellular spaces. A few of these cortical cells stained strongly with toluidine blue. Since almost the entire cytoplasm and vacuole were greenish to blue colored, clearly distinguishable from neighboring cells, these cells will be further referred to as 'blue cells'. At positions in the lateral root bases, located close to the primary root, no clear epidermis was present, in contrast to locations further away from the primary root. It seems that, besides remainders of the epidermis, outer cortical tissue at bases of lateral roots is in direct contact with the environment. Most of these peripheral outer cortical cells, and occasionally also epidermal cells, have a swollen and/or enlarged appearance, which may explain the presence of bulge-like structures, observed under the binocular. In some cases, outer cortical cells seem to elongate thereby separating two neighboring covering peripheral cells.
As soon as 24 hpi with the wild-type strain, a nodule primordium, characterized by a focus of dividing cells and young cells with relatively small vacuoles, has been developed, opposite of a proto-xylem pole. Infection pockets were present in the outer cortical cell layers. Sections through developing root nodules, at 48 hpi, are characterized by the presence of nodule primordia with a shape of an open basket. Infection pockets are positioned deeper in the nodule tissues and infection threads, originating from the infection pockets, guide bacteria to the nodule primordia. Within the infection center, many intracellular infection threads were noticed.
The presence of either Ag+ or DPI had a tremendous effect on nodule invasion and the initiation of nodule development. Apart from few bacteria present in the vicinity of outer cortical cells, neither invasion nor infection pocket formation could be observed, at six dpi. There were no signs of nodule primordium formation. These sections resembled those through mock-inoculated lateral root bases.
3.5. Morphology of Nod factor-induced pseudo-nodules. Inoculation of S. rostrata roots with 10"8 M Nod factors led to the formation of small swellings or pseudo-nodules at lateral root bases (Mergaert et al. 1993). Such pseudo-nodules were harvested 4, 6, 8, 9 and 10 dpi. Pseudo-nodules harvested at these time points showed similar features. At four dpi, the total number of cells present in a transversal section is higher than that of mock-inoculated controls, suggesting that cell division must have occurred. This is also suggested by the nicely organized appearance of cells in the inner cortex. Cell division was more clear in a pseudo-nodule harvested at nine dpi, exhibiting regions of small cells with relatively small vacuoles. A remarkable observation was the enhanced number of blue cells, compared to controls. A differential staining procedure was used to distinguish nuclei of dying cells from those of healthy cells (Kosslak et al. 1997). By this approach, the nucleolus of nuclei of healthy cells is generally dark blue stained, whereas the rest of the nucleus is lightly stained. Chromatin condensation, occurring within the nucleus of dying cells, caused that the demarcation between nucleolus and nucleus is less clear (Kosslak et al. 1997). The nucleus of healthy cells within Nod factor-induced pseudo-nodules was pink colored and the nucleolus dark blue. However, in blue cells, the whole nucleus became dark blue, suggesting chromatin condensation and/or DNA fragmentation. Strikingly, blue cells within Nod factor-induced pseudo-nodules were often present in the vicinity of large schizo-lysogenic holes, containing pieces of partially degraded cell walls. These schizo-lysogenic holes, resembled a sort of empty infection pocket-like structures and were never observed in sections through mock-inoculated lateral root bases. The observation of remainders of cell walls strongly suggested that a process of cell death must have occurred before.
When Ag+ ions were added, prior to 10"8 M Nod factors, none of the features characterizing Nod factor-induced pseudo-nodules could be observed, suggesting a complete block of root hair formation, cell division, and cell death processes leading to the formation of schizo-lysogenic holes.
3.6 Morphology of lateral root bases treated with exogenous ethylene. Lateral root bases treated with exogenous applied ethylene were harvested at 6 dpi and sectioned. Almost identical features were observed as those present within Nod factor-induced pseudo-nodules. Many blue cells were present, most often in the neighborhood of large schizo-lysogenic holes, that again contained remainders of plant cell walls. These schizo-lysogenic holes were so big that they were at some locations even in contact with the environment. This confirmed previous observations that ethylene may participate in the Nod factor signaling pathways.
3.7. In situ localization of hydrogen peroxide during initiation of stem nodule development. Because of the strong indications that hydrogen peroxide may be involved and required for nodule initiation on S. rostrata, hydrogen peroxide was localized during stem nodule initiation. We have opted to use the histochemical assay based on the reaction of hydrogen peroxide with CeCU to produce electron-dense insoluble precipitates of cerium perhydroxides, Ce[0H]200H and Ce[0H]300H (Bestwick et al. 1997). Stems were inoculated with ORS571 and material was harvested at 50 and 70 hpi, immediately fixed or stained with CeC13, and imbedded for electron microscopy.
At 50 hpi, it seems that the oxidative burst, represented by the generation of hydrogen peroxide, arises. Outer cortical cells showed dark regions, corresponding to the presence of hydrogen peroxide at the level of the cytoplasmic membrane and the cell wall. An outer cortical cell, located one cell layer beneath the peripheral outer cortical cells, is almost completely surrounded by hydrogen peroxide, localized in its cell wall and also in intercellular spaces. This particular cell showed signs of death including vacuole fragmentation and a loss of cell shape rigidity. No hydrogen peroxide could be observed in the nodule primordia.
At 70 hpi, large infection pockets were observed containing massive proliferating bacteria. Parts of cell walls of outer cortical cells in the direct vicinity of the infection pockets showed a strong staining, suggesting a massive production of hydrogen peroxide. The cytoplasmic membrane is strongly stained and hydrogen peroxide seems to diffuse from the cytoplasmic membrane toward the peripheral part of the cell wall. Bacteria that are close to these cell walls are imbedded in a matrix containing high concentrations of hydrogen peroxide. Remarkably, the bacterial body seemed to be protected against these high concentrations of hydrogen peroxide by the presence of a thick layer of low electron dense material, most probably EPSs, forming a clear barrier. Cells producing high amounts of hydrogen peroxide often show the onset of vacuole fragmentation and loosening of the cytoplasmic membrane from the cell wall, whereas the neighboring cells do not. No hydrogen peroxide was observed in nodule primordia. At 70 hpi, infection threads were present, and bacteria within infection thread were imbedded in an electron-dense matrix that contained high amounts of hydrogen peroxide. Hydrogen peroxide was also detected within cell walls of some intracellular infection threads. A longitudinal section through an intercellular infection thread revealed the presence of hydrogen peroxide within the matrix in which bacteria are imbedded, and illustrated again the possible role of the EPS layer to protect invading bacteria against those high concentrations of hydrogen peroxide.
3.8. A model for nodule initiation on S. rostrata. Nod factors probably induce an oxidative burst generating ROS, including hydrogen peroxide. ROS may subsequently induce a local ethylene production, which may eventually enhance ROS production. The preliminary observation that ethylene-induced axillary root hair formation was not inhibited in the presence of DPI suggests this sequence. This was also in agreement with our genetic approach by which it was demonstrated that the expression of both a peroxidase gene and an ACC synthase gene is induced very early in the symbiotic interaction. Interestingly, the peroxidase gene was expressed in cells surrounding infection pockets. Furthermore, a GA-20-oxidase, involved in giberellin synthesis, was expressed in cells that are going to form infection pockets, and nodulation and Nod factor-induced root hair formation was blocked in the presence of giberellin synthesis inhibitors. Possibly, giberellins may sensitize cells for hydrogen peroxide (Bethke, Jones 2001). Ethylene and/or ROS may be involved in the induction of a local plant cell death, required for infection pocket formation, a prerequisite for nodule initiation. Blocking the oxidative burst or ethylene production caused in addition abolishment of nodule primordium formation. Supposedly, ROS and/or other secondary signals may influence hormone landscapes or other factors.
Bestwick CS et al. (1997) Plant Cell 9, 209-221
D'Haeze W et al. (1998) Mol. Plant-Microbe Interact. 11, 999-1008
Mergaert P et al. (1993) Proc. Natl. Acad. Sci. USA 90, 1551-1555
Mergaert P et al. (1997) Mol. Plant-Microbe Interact. 10, 683-687
Ndoye I et al. (1994) J. Bacteriol. 176, 1060-1068
We would like to acknowledge G. Van De Sype, S. Pagnotta, and D. Hérouart (Université de Nice Sophia Antipolis, Nice, France) for help with hydrogen peroxide localization.
Was this article helpful?