Journal of Experimental Botany, Vol. 52, No. 357, pp. 747-760,
April 15, 2001
© 2001 Oxford University Press
Original Papers |
Further observations on the interaction between sugar cane and Gluconacetobacter diazotrophicus under laboratory and greenhouse conditions1
2 Department of Biological Sciences, University of Dundee, Dundee DD1 4HN, UK
3 Setor de Citologia Vegetal, Lab. Biologia Celular e Tecidual, Centro de Biociências e Biotecnologia, Universidade Estadual do Norte Fluminense, Campos dos Goytacazes, RJ 28015-620, Brazil
4 EMBRAPA-Agrobiologia, km 47, Seropédica, Rio de Janeiro, 23851-970, Brazil
Received 22 June 2000; Accepted 12 October 2000
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Sugar cane (Saccharum spp.) variety SP 70-1143 was inoculated with Gluconacetobacter diazotrophicus strain PAL5 (ATCC 49037) in two experiments. In experiment 1 the bacteria were inoculated into a modified, low sucrose MS medium within which micropropagated plantlets were rooted. After 10 d there was extensive anatomical evidence of endophytic colonization by G. diazotrophicus, particularly in lower stems, where high numbers of bacteria were visible within some of the xylem vessels. The identity of the bacteria was confirmed by immunogold labelling with an antibody raised against G. diazotrophicus. On the lower stems there were breaks caused by the separation of the plantlets into individuals, and at these ‘wounds’ bacteria were seen colonizing the xylem and intercellular spaces. Bacteria were also occasionally seen entering leaves via damaged stomata, and subsequently colonizing sub-stomatal cavities and intercellular spaces. A localized host defence response in the form of fibrillar material surrounding the bacteria was associated with both the stem and leaf invasion. In experiment 2, stems of 5-week-old greenhouse-grown plants were inoculated by injection with a suspension of G. diazotrophicus containing 108 bacteria ml-1. No hypersensitive response (HR) was observed, and no symptoms were visible on the leaves and stems for the duration of the experiment (7 d). Close to the point of inoculation, G. diazotrophicus cells were observed within the protoxylem and the xylem parenchyma, where they were surrounded by fibrillar material that stained light-green with toluidine blue. In leaf samples taken up to 4 cm from the inoculation points, G. diazotrophicus cells were mainly found within the metaxylem, where they were surrounded by a light green-staining material. The bacteria were growing in relatively low numbers adjacent to the xylem cell walls, and they were separated from the host-derived material by electron-transparent ‘haloes’ that contained material that reacted with the G. diazotrophicus antibody.
Key words: Sugar cane, Gluconacetobacter diazotrophicus, endophytic bacteria, nitrogen fixation, immunogold labelling.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Applications of mineral N fertilizer to Brazilian sugar cane (interspecific hybrids of Saccharum sp.) fields are typically much lower than those used in other countries, and they do not appear to be sufficient to compensate for N losses from annual harvesting (Ruschel, 1981
; Boddey et al., 1995). It has long been hypothesized that the apparent deficit in N inputs is made up by biological nitrogen fixation (BNF) (Patriquin et al., 1980; Ruschel, 1981). Studies using long-term N-balances and 15N isotope dilution (Lima et al., 1987; Urquiaga et al., 1992) have shown that some Brazilian sugar cane varieties (e.g. SP 70-1143, SP 79-2312, CB 45-3, Krakatau) may actually fix up to 70% of their N requirements. Although the specific micro-organisms responsible for the fixation have yet to be determined (see reviews by James and Olivares, 1998; James, 2000), the roots, stems, leaves, and trash of Brazilian varieties contain substantial numbers (up to 107 g-1 fresh weight) of diazotrophic bacteria, such as Acetobacter diazotrophicus (recently renamed Gluconacetobacter diazotrophicus; Yamada et al., 1997, 1998; Franke et al., 1999) and Herbaspirillum spp. (Boddey et al., 1995, 2000; Reis et al., 1994; Olivares et al., 1996; Baldani et al., 1997; Döbereiner et al., 1995). It has been suggested that as these bacteria are endophytes, and hence live within the plant tissues, they can fix N2 more efficiently than diazotrophs that remain in the rhizosphere or on the rhizoplane (Patriquin et al., 1983; Döbereiner et al., 1995). This may be due to the plant directly providing photosynthate for their nutrition, but also because the interior of a plant could provide a low O2 environment which is necessary for the expression of the O2-sensitive enzyme, nitrogenase. In addition, endophytic bacteria will not have to compete with other soil microbes for scarce resources (Sprent and James, 1995; Hallmann et al., 1997; Reinhold-Hurek and Hurek, 1998a, b; James and Olivares, 1998). In return for providing this niche, it may be that the bacteria provide fixed N and/or plant growth-promoting compounds to the host, in this case sugar cane (Sevilla and Kennedy, 2000).
Although endophytic diazotrophs, such as G. diazotrophicus, have been isolated from all parts of the sugar cane plant, their exact location has still to be established (James and Olivares, 1998). In an earlier study, sterile-grown plantlets of variety NA 56-79 were inoculated with G. diazotrophicus, and the plants were subsequently examined under the optical and electron microscopes (James et al., 1994). After 4 d the bacteria had colonized the surface of the roots, with the lateral root junctions and the root tips being the most preferred sites. After a further 11 d, G. diazotrophicus was observed within the plants, specifically within the xylem vessels at the base of the stems. The bacteria in the xylem were confirmed to be G. diazotrophicus using immunogold labelling. Since that study was published, further micrographs showing xylem vessels colonized by G. diazotrophicus have been presented (Döbereiner et al., 1995; James and Olivares, 1998; Reis et al., 1999; Fuentes-Ramirez et al., 1999). However, results have been presented that contrast with these studies (Dong et al., 1994). These authors suggested that the intercellular spaces in the sucrose storage parenchyma of the stems of mature sugar cane stalks were the most likely location of a symbiosis with G. diazotrophicus as (1) the intercellular spaces contain an acidic solution rich in sucrose (up to 13%), and G. diazotrophicus grows and fixes N2 well when cultured in a low pH media (pH 5.5) containing high sucrose (10–12%; Cavalcante and Döbereiner, 1988; Reis et al., 1994), (2) G. diazotrophicus (1.1x104 cells ml-1) could be isolated from ‘apoplastic fluid’ obtained by centrifuging stem pieces; this fluid is rich in sucrose (11%) and consists of fluid from the apoplast (i.e. intercellular spaces, cell walls and xylem lumina), with up to 20% of its volume comprising xylem sap, and (3) electron microscopy failed to reveal any bacteria within the vascular tissue, but a few unidentified bacteria were observed within intercellular spaces.
In a later paper (Dong et al., 1997), it was further suggested that the xylem of sugar cane stems was an unlikely location for ‘symbiotic’ G. diazotrophicus as (a) it is very low in sucrose (0–9%; Hawker, 1965; Bull et al., 1972; Welbaum et al., 1992), (b) at the stem nodes, the xylem (particularly in varieties resistant to the xylem-dwelling pathogen Clavibacter xyli subsp. xyli) contains convolutions and discontinuities (Teakle et al., 1977; Gillaspie and Teakle, 1989; Harrison and Davis, 1988) that would not allow dye and paint particles of a size similar to G. diazotrophicus to pass freely through to the next internode, and (c) when the ends of cut stalks were placed within liquid cultures of G. diazotrophicus the bacteria entered the xylem and elicited a ‘most violent’ reaction from the plant involving the production of a gum that stained bright red with toluidine blue; this appeared to arrest the movement of the bacteria and eventually kill them. Furthermore, Dong et al. (Dong et al., 1997) were of the opinion that the micrographs presented by James et al. (James et al., 1994) contained artefacts and that the ‘xylem vessels’ that the latter authors purported to show containing G. diazotrophicus were actually dead lignified xylem parenchyma cells in which the bacteria had been moved during preparation for microscopy.
In order to resolve these differences, two sets of experiments are performed. (1) Given the above criticisms of the study of James et al. (1994), the interaction between G. diazotrophicus and micropropagated sugar cane plantlets was examined in more detail using the methodology of Reis et al. (Reis et al., 1999). One of the aims of this infection study was to confirm the data presented previously (James et al., 1994; Döbereiner et al., 1995; James and Olivares, 1998), and show that xylem vessels are indeed a possible site of colonization by G. diazotrophicus. (2) To examine the infection of stems and leaves of greenhouse-grown sugar cane plants that had been directly inoculated with high numbers of G. diazotrophicus. This was done in order to determine the preferred location of the bacteria in older plants, and also to see if there was any host defence reaction and/or disease symptoms when they are forcibly introduced in large numbers. This is a method commonly used by sugar cane pathologists to determine resistance or susceptibility to bacteria, such as Xanthomonas albilineans and X. campestris pv. vasculorum, the xylem-dwelling agents of leaf scald disease (LSD) and ‘gumming disease’, respectively (Ricaud and Ryan, 1989; Rott et al., 1997; Ricaud and Autrey, 1989). This inoculation technique has been used successfully to examine the infection of sugar cane and sorghum leaves by Herbaspirillum spp (Olivares et al., 1997; James et al., 1997).
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Organisms, growth conditions, bacterial inoculation, and counts
Experiment 1
Micropropagated sugar cane plantlets (cv. SP 70-1143) were grown for 60 d and then inoculated with G. diazotrophicus strain PAL-5 according to the procedure previously used (Reis et al., 1999
). Briefly, the plantlets were separated into individuals and transferred to 50 ml tubes containing a modified MS medium, without vitamins or hormones, and with the concentrations of salts and sucrose reduced 10-fold, for example, from 20 mg sucrose l-1 to 2 mg l-1. The plantlets were then inoculated with 0.1 ml suspensions of G. diazotrophicus containing 108 cells ml-1 (controls were inoculated with sterile distilled water or autoclaved bacteria). Plants were harvested at 10 d after inoculation, and pieces of roots, leaves and stems from five inoculated and five control plants were taken for microscopical analysis. Whole plantlets were surface-sterilized for 5 min in 1% chloramine T and Most Probable Number (MPN) counts of bacteria were made according to the methods of Reis et al. (Reis et al., 1994, 1999). Dry weights were determined according to Reis et al. (Reis et al., 1999). Acetylene reduction assays were performed according to the methodology of James et al. (James et al., 1994) on five plants from each treatment after they were surface-sterilized.
Experiment 2
Greenhouse-grown sugar cane plants (cv. SP 70-1143) were inoculated at 5 weeks after germination with 1 ml of a liquid culture of G. diazotrophicus strain PAL-5 growing in LGI medium, resuspended in sterilized water and adjusted to 108 cells ml-1. The inoculation method, and the point of inoculation (the leaf ‘pocket’ at the base of the stem), was the same as that used previously (Pimentel et al., 1991; Olivares et al., 1997) with Herbaspirillum spp. Five replicate plants were inoculated, and five control plants were inoculated with sterile water, with the exuded excess fluid in both cases being immediately mopped up with sterile cotton wool. At 7 d after inoculation, leaves with visible inoculation points were taken for microscopy and bacterial counts. The leaves were prepared for the counts according to the method of Olivares et al. (Olivares et al., 1997), and the MPN of G. diazotrophicus cells were determined as in Experiment 1.
Microscopy and immunogold labelling
Samples from Experiment 1 were fixed for 24 h in 5% glutaraldehyde in 50 mM phosphate buffer (pH 6.8). They were then dehydrated in an ethanol series before being embedded in LR White resin (Agar Aids, UK) (James et al., 1994). Pieces of leaves from Experiment 2 were sampled around the inoculation point and up to 4 cm above the inoculation point, and these were fixed and embedded as above. Optical and transmission electron microscopy (TEM) were performed on the samples (James et al., 1994; Olivares et al., 1997). The 1–2 µm sections used in the optical microscopy were stained in 1% toluidine blue O in an aqueous solution of 1% sodium tetraborate (borax), except for those sections that were used for immunogold labelling which were left unstained. Immunogold labelling for optical microscopy (using silver-enhancement) and TEM used the methods of James et al. (James et al., 1994). Two antibodies were used in the immunogold analysis: a polyclonal antibody raised in a rabbit against G. diazotrophicus strain PRJ2 (Silva, 1999; Boddey et al., 2000), and a polyclonal antibody raised against the Fe(NifH)-protein of nitrogenase from Rhodospirillum rubrum (a gift from PW Ludden, Madison, Wisconsin, USA). In cross-reaction tests using enzyme-linked immunosorbent assays (ELISA) the antibody against strain PRJ2 gave a very strong reaction with G. diazotrophicus strain PAL-5, but gave little or no reaction with other bacteria commonly associated with sugar cane, including Herbaspirillum and Azospirillum spp., as well as with other members of the Acetobacteriaceae (Silva, 1999). Pre-immune sera from the same rabbits before they were immunized were used as negative controls in the immunogold analyses, and additional controls used substitution of normal rabbit serum (Sigma) for the primary antibody, and omission of the primary antibody (Olivares et al., 1997). Five nm gold particles conjugated to goat anti-rabbit antibodies were used for silver-enhancement, and 15 nm particles were used for TEM (both from Amersham, UK). The IntenSE M silver enhancement kit from Amersham (UK) was used according to the instructions of the manufacturer.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Experiment 1: infection of micropropagated plants
At 10 d after inoculation none of the plantlets from either treatment exhibited any macroscopically visible disease symptoms (data not shown). The dry weights of the G. diazotrophicus-inoculated plants (68.7±21.2 mg), were greater than the controls (50.0±16.1 mg), but owing to high variation the differences were not significant (values given are ±se, n=4). There was no acetylene reduction activity by surface-sterilized plants from any treatment.
There was a slight red colour on most of the G. diazotrophicus-inoculated plants at breaks/wounds in the base of the stem caused by the separation of the plantlets into individuals; the red coloration was not seen on the control plants (data not shown). After surface-sterilization there were 1.8x107±0.26x107 cells of G. diazotrophicus g-1 fresh weight in the inoculated plantlets, but none in the controls. The surfaces of the roots were colonized by the bacteria in a manner similar to that reported previously (James et al., 1994; Reis et al., 1999), with the bacterial cells surrounded by a layer of mucus (not shown). As in these studies, the bacteria accumulated at lateral root junctions and colonized damaged epidermal cells (not shown), but in the present study there were no indications that they had actually penetrated beyond the root epidermis. However, in contrast to the roots, using immunogold silver-enhancement to highlight the presence of G. diazotrophicus, there was clear evidence for internal colonization of the stems, particularly within the vascular tissue (Fig. 1a, c). In transverse sections taken from approximately 5 mm above the broken/wounded regions on the lower stems of five inoculated plantlets, the proportion of infected vascular bundles ranged from 4–36%, with a mean of 16.1% (±6.3 s.e.). This infection generally took the form of one or two of the metaxylem vessels being almost completely filled with bacteria (Fig. 1a, c). No bacteria were apparent within control plants (Fig. 1d). The bacteria in the infected xylem vessels (Fig. 1a, c), as well as those within some of the intercellular spaces (not shown), and leaf sheath epidermal cells (not shown) reacted strongly with the G. diazotrophicus antibody, with the antibody recognizing the surface of the bacteria, and also ‘mucus’ that was produced by them (Fig. 1e, f).
|
At the sites of breaks/wounds in the lower stem there was considerable colonization and invasion by bacteria of the wound surface, particularly within broken cells (Fig. 2a
). Further into the wound sites, intercellular and intracellular bacteria (Fig. 2b) could be seen, and these were immunogold labelled with the anti-G. diazotrophicus antibody (Fig. 2c; not shown). The intracellular bacteria were observed only in cells that appeared to be dead or damaged, without intact cytoplasm or organelles (Fig. 2c; not shown). Some xylem vessels in this region (Fig. 2a, b, d) were densely occluded with a material that stained pink with toluidine blue (not shown), and these vessels did not contain bacteria. However, bacteria were observed within adjacent vessels that were much less densely occluded with pink-staining material, and the bacteria within them were usually surrounded by a loosely-fibrous matrix (Fig. 2d). Deeper into the stem cortex, and away from the wound sites, bacteria were seen within apparently unoccluded xylem vessels (Fig. 2e). Some of the non-senescent xylem-dwelling bacteria were recognized, albeit sparsely, by an antibody raised against the Fe-protein of nitrogenase (Fig. 2f). No intercellular bacteria were observed within the stem cortex away from the wound sites (Fig. 2e), and none of the intercellular or xylem-dwelling bacteria within the wound sites labelled with the anti-nitrogenase antibody (not shown).
|
Although they were not usually seen within leaves, in some specimens bacteria were observed entering via stomata, and the latter appeared to be damaged by this invasion process (not shown). These bacteria subsequently colonized the sub-stomatal cavities and adjoining intercellular spaces, within which a host-derived matrix surrounded the bacteria (Fig. 3b
). Both the bacteria and the matrix reacted with the G. diazotrophicus antibody in immunogold assays (Fig. 3c), but none of them reacted with the antibody against the Fe-protein of nitrogenase (not shown).
|
Experiment 2: infection of mature leaves
The leaves were examined at 12 h intervals over the first 2 d after inoculation and there were no signs of a HR, i.e. water-soaked necrotic regions around the inoculation point. At 7 d after inoculation, G. diazotrophicus-inoculated leaves contained 1.1x105 G. diazotrophicus cells g-1 fresh weight, whereas the controls had no detectable bacteria. No macro symptoms of disease were visible on leaves from either treatment (not shown), and no disease symptoms appeared in the leaves over the subsequent 4 weeks. Although the lack of symptoms made the bacteria difficult to localize under the microscope, close to the points of inoculation of both control and G. diazotrophicus-treated leaves, there were signs of degradation at the point of inoculation itself (not shown) as well as in adjacent vascular bundles (Fig. 4a). In the latter, there was distortion of the metaxylem elements that were closest to the inoculation point and also gumming of the vessels by a material (Fig. 4a) that stained light pink with toluidine blue (not shown). In the G. diazotrophicus-inoculated (but not the control leaves; not shown), there was an accumulation of (green-staining) material in the intercellular spaces adjacent to the occluded vascular bundles (Fig. 4a) and, under the TEM, a few bacteria were observed within this material (not shown). However, most bacteria, even some in the form of micro-colonies, were seen within the metaxylem (Fig. 4a), again surrounded by pink-staining material (not shown).
|
Further from the point of inoculation (up to 4 cm), bacteria were observed in only 3 out of 10 leaves that were sectioned. In these samples, bacteria were visible only within vascular tissue (Fig. 4b
, c, d). The bacteria in these sections labelled strongly with the anti-G. diazotrophicus antibody (Figs 4b, 5a, c, d), whereas the serial sections used as immunogold controls did not label (Fig. 5b). Although some of the bacteria were in the protoxylem (Figs 4c, 5a), most of them were in the metaxylem vessels, often arranged in a monolayer immediately adjacent to the vessel walls (Figs 4b, c, d, 5b). A blue-green gum had accumulated in the colonized metaxylem vessels (Fig. 4c, d; data not shown) and, under the TEM, this was revealed to be a fibrous matrix, probably of plant origin, that surrounded the bacteria (Fig. 5b–d). This material was less distinct in the colonized protoxylem (Fig. 5a). Interestingly, the bacteria were separated from the fibrous matrix by a distinct, electron-transparent ‘halo’ around each bacterium. When the sections were immunogold labelled with the G. diazotrophicus antibody, gold particles that were not actually attached to the bacteria themselves, were prominent within these haloes (Fig. 5a, d). Under the TEM, the appearance of the bacteria ranged from senescent to obviously healthy (Fig. 5a–c), and some were dividing (Fig. 5d). No bacteria were observed in the control leaves (not shown).
|
None of the bacteria in sections from Experiment 2 gave a significant immunogold reaction with the antibody raised against the Fe-protein of nitrogenase (not shown). Controls for the immunogold labelling also gave no significant signal (Figs 1f
, 5b).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Infection of micropropagated plantlets
Micropropagated plantlets are being used increasingly by Brazilian farmers as a means of obtaining disease-free plants (Reis et al., 1999
). However, this also means that unlike the traditional method of vegetative propagation from seed pieces (settes) there may not be transmission from generation to generation of possibly beneficial bacteria, such as G. diazotrophicus. Therefore, the plantlets need to be inoculated, and a rooting medium has been devised (based upon a modified MS medium) (Reis et al., 1999) by which this can be done in order to maximize the number of G. diazotrophicus within the plantlets over a short period (7 d) before transplanting into the field/greenhouse. This medium, which was also used in the present study, is sufficiently low in sucrose (2 g l-1) that the growth of the G. diazotrophicus within it is not so great that it becomes detrimental to the plants. On the other hand, there is still enough sucrose to allow the bacteria to multiply slowly and, as they consume it, the bacteria appear to be induced to invade most of the plantlets within 7–10 d after inoculation (Reis et al., 1999; Sevilla et al., 1998; Olivares and James, 2000). This invasion process appears not to be detrimental to the health of the plantlets and may even confer some growth benefits upon them (Sevilla et al., 1998; Reis et al., 1999; this study). At this early stage of the interaction, the number of G. diazotrophicus within the surface-sterilized plantlets appears to be quite high (107–108 g-1 fresh weight; James et al., 1994; Reis et al., 1999; Olivares and James, 2000; this study), although it should be noted that this number almost certainly includes many surface-dwelling bacteria that have survived the disinfestation process via tight adherence to plant surfaces within ‘mucus’ and/or a preference for colonizing cracks and crevices (see James et al., 1994; Reis et al., 1999 for more details). Moreover, previous studies have shown that these high ‘internal’ numbers are rarely sustained beyond the early stages of the interaction. When the inoculated plantlets are transferred to pots under greenhouse conditions with high irradiances the concentration of bacteria within them decreases considerably, being of the order of only 104–105 g-1 fresh weight by 30 d after inoculation (Paula et al., 1991; Reis et al., 1999; Sevilla et al., 1998). Therefore, it would appear that G. diazotrophicus behaves similarly to most other non-pathogenic, endophytic bacteria (Hallmann et al., 1997), i.e. its numbers are diluted as the plants grow and the bacteria are dispersed.
The present study has shed further light upon the invasion processes illustrated by previous studies (James et al., 1994; Reis et al., 1999). These studies suggested that G. diazotrophicus first colonized the root surfaces and then infected the roots via lateral root junctions and/or root tips, and subsequently entered the root vascular system from whence they were translocated to the lower stem in the xylem. Although the present study has not confirmed the infection of roots by G. diazotrophicus, high numbers of the bacteria were observed within the xylem vessels in the stems of all the inoculated plantlets, thus supporting the possibility of xylem translocation from the roots. However, in addition to the possibility of infection at lateral root junctions, the present study suggests that there were at least two other potential sites of infection: wounds caused when the plantlets were separated into individuals at the start of the inoculation experiments (Fig. 2), and stomata (Fig. 3). In both these locations the bacteria elicited a localized host defence response in the form of a polymeric matrix material that surrounded them. As G. diazotrophicus is generally regarded as a non-pathogen (Baldani et al., 1997; Sevilla and Kennedy, 2000) this host defence response was particularly interesting, and it could be that the very low irradiances used in micro-propagation (Reis et al., 1999), as well as their very young physiological age, possibly so weaken the plantlets that it makes them more susceptible to aggressive infection, even from non-pathogens. Indeed, it is possible that G. diazotrophicus from the wound sites were able subsequently to infect healthy xylem vessels deeper within the cortices of the stem bases (Fig. 2e, f), and these may actually be the source of the bacteria illustrated in Fig. 1a–c, rather than (or in addition to) bacteria translocated from the roots (James et al., 1994; Reis et al., 1999; Olivares and James, 2000).
Finally, as with H. rubrisubalbicans in sorghum and sugar cane leaves (James et al., 1997; Olivares et al., 1997), some non-senescent G. diazotrophicus cells were shown to express the Fe-protein of nitrogenase within the xylem (Fig. 2f). The fact that there was no acetylene reduction activity with these plants however, suggests that either the nitrogenase enzyme was not active or else activity was too low to be detected by this assay. This contrasts with the study of James et al. (James et al., 1994) which showed some nitrogenase (acetylene reduction) activity with inoculated plantlets, although it was impossible to determine if this activity was due to the endophytic or the epiphytic bacteria (or a combination of both). Therefore, although the present study has shown some very limited expression of nitrogenase protein by xylem-dwelling G. diazotrophicus, it is clear that further studies are needed to determine if this bacterium can actually fix N2 within the plants (Fuentes-Ramirez et al., 1999; James, 2000).
Infection of inoculated leaves
Experiment 2 has shown that direct inoculation of sugar cane stems with G. diazotrophicus will result in the subsequent colonization of the leaves by the bacteria, without any macroscopically visible symptoms, and without eliciting a HR. This is similar to the response of sugar cane cv. SP 70-1143 to inoculation with the diazotrophic pathogen, Herbaspirillum rubrisubalbicans (Olivares et al., 1997). As a HR is normally elicited after plants are infected with incompatible phytopathogens (Sequeira et al., 1977; McKhann and Hirsch, 1994), this demonstrates that G. diazotrophicus and H. rubrisubalbicans have compatibility with sugar cane leaves. This is not surprizing, as both G. diazotrophicus (Li and MacRae, 1992; Reis et al., 1999; Boddey et al., 2000; Muthukumarasamy et al., 1999) and H. rubrisubalbicans (Olivares et al., 1996; Boddey et al., 2000; Muthukumarasamy et al., 1999) have been routinely isolated from sugar cane leaves in field and greenhouse studies. However, this contrasts with the behaviour of another important sugar cane endophytic diazotroph, Herbaspirillum seropedicae. Although this non-pathogenic bacterium can be isolated from roots and stems, it cannot be isolated from sugar cane leaves (Olivares et al., 1996; Boddey et al., 2000), and it will elicit a HR after being inoculated into them (Olivares et al., 1997).
As with H. rubrisubalbicans (Olivares et al., 1997), the present study suggests that the vascular tissue is the most likely location for G. diazotrophicus within sugar cane leaves. On the other hand, unlike leaves infected with H. rubrisubalbicans, those infected with G. diazotrophicus contained relatively few bacteria, and these were largely confined to the walls of the metaxylem vessels. Moreover, and again in contrast to H. rubrisubalbicans (Olivares et al., 1997; James et al., 1997; James and Olivares, 1998), the G. diazotrophicus cells generally did not form microcolonies within the leaf vessels (except in the damaged vessels close to the points of inoculation; Fig. 4a), and nor did they express the nitrogenase Fe-protein (although see Fig. 2f of infected stems). Distinctive electron-transparent ‘haloes’ separating the host defence gum from the bacteria were a consistent feature of the leaf colonization by G. diazotrophicus and they appeared to mark the boundary between the bacteria and the host. The fact that they were also often a site of accumulation of immunogold particles suggests that the bacteria within them were releasing immunoreactive material, most likely exopolysaccharide (EPS) (James et al., 1994; James and Olivares, 1998). A similar interaction was reported with Herbaspirillum spp. in sugar cane and sorghum leaves, where bacterial microcolonies that had formed within the host defence material that filled the vessels had very distinct boundaries between them and the plant gums (Olivares et al., 1997; James et al., 1997; James and Olivares, 1998). Electron-transparent haloes surrounding xylem-colonizing bacteria are a common phenomenon, for example, they have also been observed in sugar cane infected by C. xyli subsp. xyli (Gillaspie and Teakle, 1989), and in tomato (Lycopersicon esculentum) infected by Pseudomonas (Ralstonia) solanacearum (Vasse et al., 1995).
Although it is known that H. rubrisubalbicans is both compatible with sugar cane, and a mild pathogen, causing symptoms of ‘mottled stripe disease’ on leaves of susceptible varieties (Pimentel et al., 1991; Olivares et al., 1997), the observation from the present study (Experiments 1, 2) that G. diazotrophicus may also be slightly ‘pathogenic’ when inoculated in high numbers is new. It is also rather controversial, as this bacterium has hitherto been regarded as a purely symptomless endophyte, causing no harm to the host (James et al., 1994; Dong et al., 1994; Döbereiner et al., 1995; Baldani et al., 1997). On the other hand, there were no macroscopically visible symptoms, indicating that infection by G. diazotrophicus is different from that by ‘genuine’ vascular pathogens of sugar cane, such as H. rubrisubalbicans, C. xyli subsp. xyli, X. albilineans, and X. campestris pv. vasculorum. These pathogenic bacteria may accumulate within the vessels in very large numbers (up to 1010 CFU g-1 fresh wt. in the case of X. albilineans; Rott et al., 1997), and consequently elicit a much more aggressive host defence reaction. Also, in susceptible varieties, they may cause long-term damage, or even plant death (Kao and Damann, 1980; Harrison and Davis, 1988; Gillaspie and Teakle, 1989; Ricaud and Ryan, 1989; Ricaud and Autrey, 1989; Pimentel et al., 1991; Olivares et al., 1997; Rott et al., 1997). In contrast to these pathogens, with G. diazotrophicus all evidence of ‘pathogenicity’ was at the microscopical level, and mainly took the form of gum production in the vessels. The only evidence of actual damage to the plants was distorted metaxylem vessels and the degraded bundle sheath cells immediately adjacent to the points of inoculation. As the latter damage was also seen with the control plants inoculated with sterile water (not shown), much of it is likely to be a simple wound response brought about by the injection procedure. Similarly, the violent response observed when cut sugar cane stalks were immersed in liquid cultures of G. diazotrophicus (Dong et al., 1997) is also likely to be largely the result of massive (fatal) wounding caused by the stems being cut (Fuentes-Ramirez et al., 1999).
Therefore, although it was shown that G. diazotrophicus can provoke a mild, localized host defence response when inoculated into sugar cane, it cannot properly be termed a phytopathogen as all of the numerous studies with field-grown plants have isolated it only from plants showing no macroscopically visible symptoms (Cavalcante and Döbereiner, 1988; Li and MacRae, 1992; Fuentes-Ramirez et al., 1993; Muthukumarasamy et al., 1999). Moreover, it does not produce a HR with sugar cane (this study), or in a standard pathogenicity test with tobacco leaves (FL Olivares, unpublished data). The most likely reason for the limited ‘pathogenicity’ shown by G. diazotrophicus in the present study is that high numbers of the bacteria were ‘forced’ into the plants, either by inoculation of wounded plantlets (Experiment 1) or by leaf injection (Experiment 2). Indeed, the limited host defence response observed is similar to that which has been observed previously when plants have been inoculated with high numbers (e.g. 108 cells ml-1) of non-pathogenic bacteria, for example, heat-killed pathogenic Pseudomonas spp and live Escherischia coli (Sequeira et al., 1977; McKhann and Hirsch, 1994), and hence suggests that any bacteria, if inoculated into a plant in sufficiently high numbers, will elicit some host defence response in order to control and suppress their numbers.
Does Gluconacetobacter diazotrophicus live in the xylem or the intercellular apoplast of sugar cane?
Although Experiment 1 has confirmed that G. diazotrophicus will readily colonize the xylem of inoculated sugar cane plantlets, as these experiments were conducted with very young plants and under unusual growth conditions (e.g. low irradiances; James et al., 1994; Reis et al., 1999; Sevilla et al., 1998) the pattern of localization may not be representative of that in mature plants. On the other hand, Experiment 2 has shown that the xylem is the principal location of the bacteria in fully-expanded leaves of inoculated greenhouse-grown plants (albeit with a slight host-defence response), and this may be more representative of the field situation. This location is strongly supported by a recent study of mature plants infected by ß-glucuronidase marked G. diazotrophicus, in which it was reported that the xylem and the xylem parenchyma (and possibly the phloem) were the only observed sites of colonization (Fuentes-Ramirez et al., 1999). Therefore, although the possibility of G. diazotrophicus living in other locations (e.g. the intercellular apoplast; Dong et al., 1994; Figs 2b, c, 3a, b, this study) is not ruled out, considering that Reis et al. (Reis et al., 1994) and Caballero-Mellado et al. (Caballero-Mellado et al., 1995) have also isolated the bacteria from the xylem sap of field-grown plants, it must be concluded that at least part of the G. diazotrophicus population in sugar cane resides in the xylem.
This is in strong contrast to Dong et al. (Dong et al., 1994, 1997), who concluded that the intercellular apoplast of the sucrose storage tissue was the only probable location for the bacteria in mature stems, and that it would be ‘most unlikely’ for them to live within the xylem (see Introduction for more details). However, there are a number of flaws in this argument. Firstly, there is no convincing anatomical evidence for the presence of G. diazotrophicus within the intercellular spaces of mature stems. Dong et al. did not confirm the identity of the bacteria in their micrographs via immunological or molecular methods (Dong et al., 1994) and, therefore, given the diverse microflora within sugar cane (James and Olivares, 1998; Fuentes-Ramirez et al., 1999), it is impossible to say with any certainty that the bacteria were G. diazotrophicus. Secondly, the ‘apoplastic fluid’ obtained from stems by centrifugation also contained up to 20% xylem sap (v/v) (Dong et al., 1994), and hence the bacteria isolated from it may also have come from the xylem as well as the intercellular spaces. Finally, although Dong et al. (Dong et al., 1994, 1997) have insisted that the sucrose-rich stem apoplast is where G. diazotrophicus must live due to its ‘need’ for a high concentration (10–12%) of sucrose to support ‘symbiotic’ N2 fixation (see Introduction), not only is this insistence not supported by convincing anatomical evidence (see above), it is also based on two misconceptions.
(1) There is little evidence, as yet, that G. diazotrophicus is actually an N2-fixing symbiont of sugar cane, or that it even expresses active nitrogenase in planta (James and Olivares, 1998; Fuentes-Ramirez et al., 1999; this study). Indeed, recent studies suggest that any beneficial effects it may have on plant growth are more likely to be via mechanisms other than N2 fixation, such as production of indole-acetic acid (IAA) (Fuentes-Ramirez et al., 1993; Sevilla et al., 1998; Bastián et al., 2000).
(2) Gluconacetobacter diazotrophicus does not have a requirement for high levels of sucrose. Reis and Döbereiner (Reis and Döbereiner, 1998) have shown that it will grow and fix N2 in media containing only 1% sucrose, and therefore concentrations at the lower end of the range commonly found within sugar cane xylem (0–9%; Dong et al., 1997, and references therein) could provide sufficient carbon to support the relatively low populations of the bacteria that are commonly isolated from the aerial parts of the plant (c.<102–104; Paula et al., 1991; Dong et al., 1994; Reis et al., 1999; Fuentes-Ramirez et al., 1999; dos Reis et al., 2000).
|
Acknowledgments |
|---|
We thank RM Boddey, J Döbereiner, LE Fuentes-Ramirez, C Kennedy, JA Raven, JI Sprent, and M Sevilla for helpful discussions, and G Baeta da Cruz, M Gruber and M Kierans for technical assistance. EK James was funded by the World Bank through the Inter-American Institute for Co-operation in Agriculture (IICA), and FL Olivares by the Brazilian National Research Council (CNPq).
|
Notes |
|---|
1 This paper is dedicated to the memory of Joanna Döbereiner (1924 to 2000), the discoverer of Gluconacetobacter diazotrophicus.
5 Present address and to whom correspondence should be sent: Centre for High Resolution Imaging and Processing, MSI/WTB Complex, School of Life Sciences, University of Dundee, Dundee DDI 5EH, UK. Fax: +44 1382 345893. E-mail: e.k.james{at}dundee.ac.uk
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Baldani JI, Caruso L, Baldani VLD, Goi SR, Döbereiner J.1997. Recent advances in BNF with non-legume plants. Soil Biology and Biochemistry 29, 911–922.
Bastián F, Rapparini F, Baraldi R, Piccoli P, Bottini R.2000. Inoculation with Acetobacter diazotrophicus increases glucose and fructose content in shoots of Sorghum bicolor (L.) Moench. Symbiosis 27, 147–156.
Boddey RM, de Oliveira OC, Urquiaga S, Reis VM, Olivares FL, Baldani VLD, Döbereiner J.1995. Biological nitrogen fixation associated with sugar cane and rice: contributions and prospects for improvement. Plant and Soil 174, 195–209.
Boddey RM, da Silva LG, Reis VM, Alves BJR, Urquiaga S.2000. Assessment of bacterial nitrogen fixation in grass species. In: Triplett EW, ed. Prokaryotic nitrogen fixation: a model system for analysis of a biological process. UK: Horizon Scientific Press, 705–726.
Bull TA, Gayler KR, Glasziou KT.1972. Lateral movement of water and sugar across xylem in sugarcane stalks. Plant Physiology 49, 1007–1011.
Caballero-Mellado J, Fuentes-Ramirez LE, Reis VM, Martinez-Romero E.1995. Genetic structure of Acetobacter diazotrophicus populations and identification of a new genetically distant group. Applied and Environmental Microbiology 61, 3008–3013.[Abstract]
Cavalcante VA, Döbereiner J.1988. A new acid-tolerant nitrogen-fixing bacterium associated with sugarcane. Plant and Soil 108, 23–31.
Döbereiner J, Baldani VLD, Reis VM.1995. Endophytic occurrence of diazotrophic bacteria in non-leguminous crops. In: Fendrik I, del Gallo M, Vanderleyden J, de Zamaroczy M, eds. Azospirillum VI and related microorganisms. Berlin, Heidelberg: Springer-Verlag, 3–14.
Dong Z, Canny MJ, McCully ME, Roboredo MR, Cabadilla CF, Ortega E, Rodes R.1994. A nitrogen-fixing endophyte of sugarcane stems. Plant Physiology 105, 1139–1147.[Abstract]
Dong Z, McCully ME, Canny MJ.1997. Does Acetobacter diazotrophicus live and move in the xylem of sugarcane stems? Anatomical and physiological data. Annals of Botany 80, 147–158.
Franke IH, Fegan M, Hayward C, Leonard G, Stackebrandt E, Sly LI.1999. Description of Gluconacetobacter sacchari sp. nov., a new species of acetic acid bacterium isolated from the leaf sheath of sugar cane and from the pink sugar-cane mealy bug. International Journal of Systematic Bacteriology 49, 1681–1693.
Fuentes-Ramírez LE, Caballero-Mellado J, Sepúlveda J, Martínez-Romero E.1999. Colonization of sugarcane by Acetobacter diazotrophicus is inhibited by high N-fertilization. FEMS Microbiology Ecology 29, 117–128.
Fuentes-Ramírez LE, Jimenez-Salgado T, Abarca-Ocampo IR, Caballero-Mellado J.1993. Acetobacter diazotrophicus, an indoleacetic acid producing bacterium isolated from sugarcane cultivars of Mexico. Plant and Soil 154, 145–150.
Gillaspie AG, Teakle DS.1989. Ratoon stunting disease. In: Ricaud C, Egan BT, Gillaspie AG, Hughes CG, eds. Diseases of sugarcane. Amsterdam: Elsevier, 59–80.
Hallmann J, Quadt-Hallmann A, Mahaffee WF, Kloepper JW.1997. Bacterial endophytes in agricultural crops. Canadian Journal of Microbiology 43, 895–914.
Harrison NA, Davis MJ.1988. Colonization of vascular tissues by Clavibacter xyli subsp. xyli in stalks of sugarcane cultivars differing in susceptibility to ratoon-stunting disease. Phytopathology 78, 722–727.
Hawker JS.1965. The sugar content of cell walls and intercellular spaces in sugar cane stems and its relation to sugar transport. Australian Journal of Biological Science 18, 959–969.
James EK.2000. Nitrogen fixation in endophytic and associative symbiosis. Field Crops Research 65, 197–209.
James EK, Olivares FL.1998. Infection and colonization of sugar cane and other graminaceous plants by endophytic diazotrophs. Critical Reviews in Plant Sciences 17, 77–119.
James EK, Olivares FL, Baldani JI, Döbereiner J.1997. Herbaspirillum, an endophytic diazotroph colonizing vascular tissue in leaves of Sorghum bicolor L. Moench. Journal of Experimental Botany 48, 785–797.
James EK, Reis VM, Olivares FL, Baldani JI, Döbereiner J.1994. Infection of sugar cane by the nitrogen-fixing bacterium Acetobacter diazotrophicus. Journal of Experimental Botany 45, 757–766.
Kao J, Damann KE.1980. ‘In situ' localization and morphology of the bacterium associated with ratoon stunting disease of sugar cane. Canadian Journal of Botany 58, 310–315.
Li R, MacRae IC.1992. Specific identification and enumeration of Acetobacter diazotrophicus in sugar cane. Soil Biology and Biochemistry 24, 413–419.
Lima E, Boddey RM, Döbereiner J.1987. Quantification of biological nitrogen fixation associated with sugar cane using a 15N-aided nitrogen balance. Soil Biology and Biochemistry 24, 413–419.
McKhann HI, Hirsch AM.1994. Does Rhizobium avoid the host response? In: Dangl JL, ed. Bacterial pathogenesis of plants and animals. Berlin, Heidelberg: Springer-Verlag, 139–162.
Muthukumarasamy R, Revathi G, Lakshminarasimhan C.1999. Influence of N fertilization on the isolation of Acetobacter diazotrophicus and Herbaspirillum spp. from Indian sugarcane varieties. Biology and Fertility of Soils 29, 157–164.
Olivares FL, Baldani VLD, Reis VM, Baldani JI, Döbereiner J.1996. Occurrence of the endophytic diazotrophs Herbaspirillum spp. in roots, stems and leaves predominantly of Gramineae. Biology and Fertility of Soils 21, 197–200.
Olivares FL, James EK.2000. Endophytic establishment of diazotrophic bacteria in sugar cane plants. In: Pedrosa FO, Hungria M, Yates MG, Newton WE, eds. Nitrogen fixation: from molecules to crop productivity. Dordrecht: Kluwer, 413–414.
Olivares FL, James EK, Baldani JI, Döbereiner J.1997. Infection of mottled stripe disease susceptible and resistant varieties of sugar cane by the endophytic diazotroph Herbaspirillum. New Phytologist 135, 723–737.
Patriquin DG, Gracioli LA, Ruschel AP.1980. Nitrogenase activity of sugar cane propagated from stem cuttings in sterile vermiculite. Soil Biology and Biochemistry 12, 413–417.
Patriquin DG, Döbereiner J, Jain DK.1983. Sites and processes of association between diazotrophs and grasses. Canadian Journal of Microbiology 29, 900–915.
Paula MA, Reis VM, Dobereiner J.1991. Interactions of Glomus clarum with Acetobacter diazotrophicus in infection of sweet potato (Ipomoea batata), sugarcane (Saccharum sp.) and sweet sorghum (Sorghum vulgare). Biology and Fertility of Soils 11, 111–115.
Pimentel JP, Olivares FL, Pitard RM, Urquiaga S, Akiba F, Döbereiner J.1991. Dinitrogen fixation and infection of grass leaves by Pseudomonas rubrisubalbicans and Herbaspirillum seropedicae. Plant and Soil 137, 61–65.
Reinhold-Hurek B, Hurek T.1998a. Interactions of gramineous plants with Azoarcus spp. and other diazotrophs: identification, localization, and perspectives to study their function. Critical Reviews in Plant Sciences 17, 29–54.
Reinhold-Hurek B, Hurek T.1998b. Life in grasses: diazotrophic endophytes. Trends in Microbiology 6, 139–144.[ISI][Medline]
Reis dos FB, Silva da LG, Reis VLM, Döbereiner J.2000. Ocorrência de bactérias diazotróficas em diferentes genótipos de cana-de-açúcar. Pesquisa Agropecuaria Brasileira 35, 985–994.
Reis VM, Olivares FL, Döbereiner J.1994. Improved methodology for isolation of Acetobacter diazotrophicus and confirmation of its endophytic habitat. World Journal of Microbiology and Biotechnology 10, 101–104.
Reis VM, Döbereiner J.1998. Effect of high sugar concentration on nitrogenase activity of Acetobacter diazotrophicus.