infection of sugar cane by the nitrogen-fixing bacterium...

Journal of Experimental Botany, Vol. 45, No. 275, pp. 757-766, June 1994Journal ofExperimentalBotany

Infection of sugar cane by the nitrogen-fixing bacteriumAcetobacter diazotrophicus

E.K. James1'3, V.M. Reis2, F.L. Olivares2, J.I. Baldani2 and J. Dobereiner2

1 Department of Biological Sciences, University of Dundee, Dundee DD14HN, UK2 EMBRAPA-CNPAB, Serop&dica, Itaguai 23851-970, Rio de Janeiro, RJ, Brazil

Received 23 November 1993; Accepted 5 March 1994

Abstract

Significant nitrogen fixation has recently been demon-strated in Brazilian sugar cane {Saccharum officina-rum) cultivars known to form associations with anumber of diazotrophs, including Acetobacter diazo-trophicus, an acid-tolerant endophytic bacteriumwhich grows best on a sucrose-rich medium. In aseries of experiments, aseptically-grown sugar caneplantlets were rooted in a liquid medium and inocu-lated with A. diazotrophicus originally isolated fromfield-grown sugar cane. After 4, 7, 9, and 15 d, plantswere examined under light, scanning and transmissionelectron microscopes and the presence ofA. diazotrophicus on and within plant tissues was con-firmed by immunogold labelling. By 15 d, external bac-terial colonization was seen on roots and lower stems,particularly at cavities in lateral root junctions. Theloose cells of the root cap at root tips were a site ofentry of the bacteria into root tissues. Both at lateralroot junctions and root tips, bacteria were also seenin enlarged, apparently intact, epidermal cells. After15 d, bacteria were present in xylem vessels at thebase of the stem, many connected via mucus to spiralsecondary thickening. There was no obvious patho-genic reaction to the bacteria within the xylem. Fromthese observations, it is proposed that, under experi-mental conditions, A. diazotrophicus firstly colonizedthe root and lower stem epidermal surfaces and thenused root tips and lateral root junctions to enter thesugar cane plant where it was distributed around theplant in the transpiration stream. It is further sug-gested that the xylem vessels in the dense shoots ofmature plants are also a possible site of N2-fixation bydiazotrophs as they provide the low pO2 and energyas sucrose necessary for nitrogenase activity.

Key words: Acetobacter diazotrophicus, endophyte, infec-tion, nitrogen fixation, sugar cane.

Introduction

Recent approaches to finding/establishing 'new' N2-fixingsymbioses in non-legumes with rhizobia and other free-living diazotrophs have used auxins, enzyme treatment ofthe roots, or genetic engineering of the bacteria to encour-age non-legumes to allow entry of the bacteria and form'pseudonodules' or /raranodules (Cocking et al, 1993;Kennedy and Tchan, 1992; Christiansen-Weniger, 1992).However, an alternative to these approaches is to investi-gate natural associations between endophytic diazotrophs,such as Acetobacter diazotrophicus, and compatible non-legumes. For instance, in the sugar cane {Saccharumofficinarum)-Acetobacter association investigated here nosuch treatments were needed as A. diazotrophicus is anatural endophyte and has so far only been isolated fromsucrose-rich plants such as sugar cane, sweet potato(Ipomoea batatas), and Cameroon grass (Pennisetum pur-pur eum) which are propagated vegetatively (Dobereineret al, 1988, 1993; Paula et al, 1991; Li and MacRae,1991, 1992; Reis, Olivares and Dobereiner, unpublished).

A. diazotrophicus is one of a number of diazotrophicbacteria which have been found living inside sugar caneroots, stems and leaves; others are Herbaspirillum seroped-icae and H. rubrisubalbicans (Cavalcante and Ddbereiner,1988; Pimentel et al, 1991; DSbereiner et al, 1993; Reiset al, 1993). Some, or all, of these bacteria are held tobe responsible for the very significant N2-fixation observedin field experiments using N-balance and 15N isotopedilution techniques (Boddey et ai, 1991; Ddbereiner et al.,1993) that suggest that up to 80% of the N incorporatedinto Brazilian sugar cane cultivars may be obtained from

'To whom correspondence should be addressed. Fax: +44 382 322318.

Oxford University Press 1994

at National C

hung Hsing U

niversity Library on A

pril 11, 2014http://jxb.oxfordjournals.org/

Dow

nloaded from

http://jxb.oxfordjournals.org/


758 James et al.

B

at National C

hung Hsing U



Dow

nloaded from



biological nitrogen fixation (BNF) (Lima et al, 1987;Urquiaga et al, 1992). Recent evidence thatA. diazotrophicus may be of particular significance in thetransfer of the products of N2-fixation to host plants hasrecently come from Cohjo et al. (1993) who showed that,in a model system for in vitro studies of plant-bacteriainteractions, A. diazotrophicus transferred up to half ofthe nitrogen it fixed to its 'partner' in mixed cultures witha yeast.

However, the sites of nitrogen-fixation and transfer offixed N (and photosynthate in the opposite direction) insugar cane plants have not yet been located and noobvious symbiotic structures identified. Sugar cane plantscan grow up to 4 m in height with stems of 5 cm indiameter and this large biomass can support a highnumber of bacterial cells. This size makes the examinationof a mature plant for specific types of bacteria andmicroscopic structures very difficult. The initial infectionprocess from soil may not be observed in field-grownplants as A. diazotrophicus is an endophyte which isthought to be spread to mature plants mainly via vegetat-ive sets (DSbereiner et al, 1990, 1993; Paula et al, 1990;Reis, Olivares and DQbereiner, unpublished).

In the present study, to avoid the complications oflooking for a number of different diazotrophs in large,maturing plants, we followed the infection of aseptically-grown sugar cane plantlets by just a single endophyticspecies, one of the most common in Brazilian sugar cane,A. diazotrophicus. We show (using ultrastructural andimmunogold labelling techniques) that simply by rootingsugar cane plantlets in a liquid medium (Murashige andSkoog, 1962), the bacteria are able to colonize rootsurfaces substantially, and to penetrate the roots intercel-lularly at the root tip (behind the root cap) and at cracksin lateral root junctions. Further, we demonstrate thatA. diazotrophicus is able to colonize the xylem vesselswithout any obvious pathogenic reaction from the plantand be transported via the vascular system at least as faras the lower stem.

Materials and methods

Organisms, growth conditions, bacterial inoculation and acetylenereduction

Sugar cane (Saccharum officinarum) cultivars NA 56-79 and SP70-1143 were micropropagated according to the method of

Diazotrophic infection of sugar cane 759

Hendre et al. (1983). After culturing the plantlets in a rootingand shooting medium (made from a modified MS liquidmedium; Murashige and Skoog, 1962) for 50-60 d, the plantletswere transferred to 50 ml of the same medium, but withouthormones and with the concentration of nutrients and sucrosereduced 10-fold (i.e. from 20 g to 2 g sucrose 1"').

A. diazotrophicus strains PAL-5 (ATCC 49037) and PSP32(isolated from washed, crushed sugar cane stems) were grownfor 24 h in a medium containing (P1) : 2.0 g glucose, 1.5 gglutamic acid, 1.5 g peptone, 0.5 g K2HPO4) 0.5 g MgSO4.7H2O, 2 g yeast extract, pH 6.0. Plantlets were inoculated with0.1 ml of suspensions containing 106 to 107 bacteria and controlswere inoculated with the same medium with autoclaved bacteria.All plants were maintained at 30 °C with an irradiance of60ftmol photons m" 2s" 1 for 12 h per day. A. diazotrophicuswas isolated from the plants, and estimates made of the numberof bacteria in association with the plants at each harvest usingthe methods of Cavalcante and DSbereiner (1988) and Reis(1991).

Nitrogenase activity was measured using the acetylenereduction assay (ARA) 15 d after bacterial inoculation. Theplants were sealed into 200 ml bottles stoppered with suba sealsand 10% acetylene was injected into the bottles. After 24 h theethylene concentration in the bottles was measured on a PerkinElmer gas chromatograph fitted with a 50 cm Porapak columnand a hydrogen flame ionization detector.

Light and electron microscopy and immunogold labelling

Small pieces (1-2 mm) of roots and stems were taken fromplantlets at 4, 7, 9, and 15 d after inoculation and fixed in 5%glutaraldehyde in 50mol m~3 phosphate buffer (pH 7.0) for24 h. Material for light microscopy and transmission electronmicroscopy (TEM) was rinsed in buffer, dehydrated in anethanol series and embedded in LR White acrylic resin (LondonResin Company, UK), with 2 d infiltration, followed bypolymerization at 60 °C. Semi-thin sections (1-2/xm) andultrathin sections (<70 nm) were taken on a Reichert UltracutE microtome. Ultrathins were collected on pyroxylin-coatednickel grids and used for immunogold labelling (IGL) whereassemi-thins were collected on glass slides and either used forIGL or immediately stained with toluidine blue.

Polyclonal antiserum raised in rabbits againstA. diazotrophicus, originally isolated from stems of field-grownsugar cane using established methods (Cavalcante andDdbereiner, 1988; Boddey et al., 1991), was used in IGL. Theantiserum was tested using ELISA (enzyme-linked immunosorb-ent assay) against a wide range of bacteria which are normallyfound in sugar cane tissues or the soil (Azospirillum spp.,Herbaspirillum spp. and other Acetobacter spp.) and negligiblecross-reactions were seen at a dilution of 1:400 of the originalantiserum. Initial tests of the antiserum on sugar cane callusinoculated with A. diazotrophicus or with other bacteria,indicated that only sections from callus with A. diazotrophicusgave a positive signal with IGL.

Plate 1. (A) Scanning electron micrograph (SEM) of sugar cane (SP 70-1143) root tip at 4 d after inoculation (DAI) with Acetobacter diazotrophicusstrain PAL 5. Note the strands of 'mucus' (arrows). x2300. Bar=10(im. (B) SEM of sugar cane (SP 70-1143) root surface at 7 DAI withA. diazotrophicus. Note the non-polar attachment of bacteria on the root surface and the accumulation of bacteria in cracks. The bacteria are lessthan 1 ftm in length, x 24000. Bar= 1 /an. (C) Light micrograph of lateral root junction of sugar cane (NA 56-79) at 15 DAL The sections wereimmunogold labelled with a polyclonal antibody against A. diazotrophicus and a goat anti-rabbit 5 nm gold conjugate, followed by silverenhancement. The silver is visible as a black precipitate. Note the high concentration of precipitate covering the root epidermis and the accumulationof bacteria at the root junction (arrow), x 165. Bar = 100 ^m. (D) Light micrograph of an immunogold silver-enhanced lateral root junction ofsugar cane (NA 56-79) showing bacteria accumulating in intercellular cavities. Some bacteria are within cells (large arrow), although the latter arenot intact, x 825. Bar= 10 jim. (E) Transmission electron micrograph (TEM) of A. diazotrophicus on the surface of a sugar cane (NA 56-79) root15 DAI. Note that the bacteria are older and, consequently, larger than those at 7 DAI (see Plate IB), x 8000. Bar = 2 jtm.

at National C

hung Hsing U



Dow

nloaded from



760 James et al.

at National C

hung Hsing U



Dow

nloaded from



The buffer used for the 'blocking' stage of the IGL procedureand for diluting antibody solutions contained the following(1 ~ ' ) : phosphate-buffered saline (PBS: 0.22 g NaH2PO«, 1.19 gNa2HPO4) 8.5 g NaCl, 0.5 g NaN3, pH 7.0), 10 g bovine serumalbumin (BSA), 1 ml Tween 20. This buffer was effective inpreventing non-specific binding during IGL with rabbit poly-clonal antisera (James, Johnston and Williamson, unpublisheddata). Immunogold labelling for TEM and light microscopy(including silver-enhancement) was performed according to themethods of James et al. (1991). Briefly, sections on slides orgrids were immersed in blocking buffer for 1 h followed by an18 h incubation in the primary antibody to A. diazotrophicusdiluted 1:400 with blocking buffer. After thorough washing inPBS plus 1% Tween 20 (PBS-Tween) the sections were incubatedin 5 nm goat anti-rabbit gold (Amersham, UK), diluted 1:100in blocking buffer, for 4 h. The sections were then washed inPBS-Tween, PBS alone, and finally in distilled water. After ashort drying period, ultrathin sections for TEM on grids werestained for lOmin in uranyl acetate and 2min in lead citratebefore being viewed under a JEOL 1200EX TEM. Sections onslides for light microscopy were silver-enhanced using a 10 minexposure to the IntenSE M kit provided by Amersham (UK)and background stained in toluidine blue before viewing underan Olympus BH2 optical microscope.

In each IGL preparation, controls to recognize non-specificbinding by the gold conjugated antibody to the sections, wererun in parallel. These were (1) 'pre-immune' serum from thesame rabbit before inoculation with A. diazotrophicus, wassubstituted for the anti-A diazotrophicus primary antibody;(2) no primary antibody was used, the sections being incubatedin blocking buffer alone for this step.

Material for scanning electron microscopy (SEM) was fixedand dehydrated as for light microscopy. The ethanol was thengradually replaced by acetone and the specimens, in 100%acetone, were critical point-dried in a Biorad E 3000 criticalpoint drier before being coated in gold under a Biorad E5 200sputter coater. Specimens were viewed under a CambridgeStereoscan 200 SEM.

Results

Bacterial counts and nitrogenase activity

There were no obvious effects of inoculation by either ofthe two strains of bacteria on the growth of the plantletsunder the growth conditions used here. In addition, thecultivar of sugar cane had no significant effect on plantgrowth, or bacterial infection. However, the number ofviable bacteria (of both strains) on the roots and aerialparts had increased to over 7xl0 7 g~' fresh weight(Table 1) at 15 d after inoculation. At 15 d there wassignificant acetylene reduction from both strains of


Table 1. Number of cells of Acetobacter diazotrophicus on, andin, micropropagated sugar cane plantlets (cultivar NA 56-79)",and their acetylene reduction activity (ARA), 15 d after inocula-tion into a rooting medium containing 2 g sucrose l~' at the timeof inoculation

Data are means of three replicates.

Number of cells ARA (nmoles C2H4g"1 fresh weight of h"1 vial"1

plantlets ± standard error)

Uninoculated control 0A. diazotrophicus strain

PSP32 7xlO7

A. diazotrophicus strainPAL-5 8 x 107

0

155±62

102±48

"Estimated using the most probable number (MPN) technique withLGIP semi-solid medium (Reis, 1991).

A. diazotrophicus (Table 1) although no attempt wasmade to discern whether this nitrogenase activity was dueto endophytic bacteria or to the large numbers of bacteriaon the root and lower stem surfaces. A. diazotrophicuswas not detected in any of the control specimens andthere was no significant nitrogenase activity.

Microscopy and localization of bacteria in plant tissues

All the micrographs shown here were from sugar canecv. sp 70-1143 or NA 56-79 inoculated with A. diazo-trophicus strain PAL-5. At least five specimens fromeach treatment were examined per harvest and thesemicrographs are representative. A. diazotrophicus was notseen in, or on, any of the control specimens (resultsnot shown).

Under the SEM small (0.5-1 /im), rod-shaped bacteriawere seen covering the root surface between 4 and 7 dafter inoculation with A. diazotrophicus (Plate 1A, B).The colonization of the bacteria had increased greatly bythe time of the 9 d harvest, and even more so after 15 d(Plate 1C, E), when they had also increased in size to0.2-2.0 f/.m (Plate IE). The bacteria apparently adheredto the roots via a mucoid substance (Plate 1A, B), similarto that reported by Levanony et al. (1989) withAzospirillum brasilense on wheat (Triticum aestivurn) rootsand by Bashan et al. (1991) on a number of non-cerealcrop plants. The bacteria in the present study wererandomly orientated, with no indication of polar attach-

Plate 2. (A) Light micrograph of an immunogold silver-enhanced root tip (NA 56-79) at 15 DAI. Note the silver precipitate in the intercellularspaces around (and within) the meristematic zone (small arrows), the clumps of bacteria outside the root (double arrows) and the silver-enhancedstructures within the cells (large arrows). x413. Bar = 50fijn. (B) Light micrograph of a root tip (NA 56-79) at 15DAI. Bacteria can be seen inclusters outside the root and also within the root cap cells, some of which appear intact (arrow). Note the organelle-rich cells of the root tipmeristem (M). x2060. Bar=10^m. (C) TEM of surface of root cap cells from sugar cane plant (NA 56-79) 15 DAI showing a group of bacteriaadhering to the cells via a mucoid substance. This 'mucus' is immunogold labelled after incubation of the section in an antibody raised againstA. diazotrophicus followed by incubation in antibodies conjugated to 5 nm gold particles. Such groups of bacteria were often seen and suggestsdivision of cells. x44600. Bar = 200nm. (D) TEM of the root tip from Plate 2A. Note the 'infection thread-like' structure (large arrow) and itssimilarity to the IGL-SE structures in Plate 2A. There appears to be bacteria within this structure although nearby vesicles possibly containdisintegrated bacteria (small arrows). x74OO. Bar =

at National C

hung Hsing U



Dow

nloaded from



762 James et al.

at National C

hung Hsing U



Dow

nloaded from



ment, at first forming a monolayer on the epidermal cellsurfaces, and also collecting in cavities between cells(Plate IB). The bacteria were confirmed to beA. diazotrophicus by IGL with silver enhancement(IGL-SE) when viewed under the light microscope. By15 d there were substantial layers of bacteria adheringtightly to the surface of the root (Plate 1C, D, E) and tothe surface of the stem base (not shown). The bacteriaparticularly accumulated in cracks and cavities at lateralroot junctions with the tap root (Plate 1C, D) and werealso seen within enlarged epidermal cells at this locality(Plate ID), although these cells were not always intact.In contrast to work with Azospirillum and various cereals(Patriquin et al, 1983; Bashan and Levanony, 1989), inthis study there was no evidence of colonization of roothairs (on or within) by A. diazotrophicus.

Root tips were also extensively colonized byA. diazotrophicus (Plate 1A; 2A, B), the bacteria beinglocalized in intercellular cavities behind the root cap andwithin the meristem; the root cap, as it was sloughed off,apparently allowed entry of the bacteria. The enlargedcells at the end of the root tip and in the root tipepidermal cells were colonized by extracellularA. diazotrophicus and the bacteria were also seen withinsome of these cells (Plate 2A, B), although not all of theoccupied cells were intact. The extracellular bacteria wereapparently adhered to the root tip cells via a mucoidsubstance which was immunogold labelled (Plate 2C).These bacteria were often in groups of 3-4 (Plate 2C),similar to the Klebsiella 'micronodules' reported byBoonjawat et al. (1991) in rice (Oryza saliva) seedlingsand to the clusters of Pantoea agglomerans seen on winterwheat roots by Ruppel et al. (1992). The cells of the roottip meristem were typically rich in cytoplasm and organ-elles (Plate 2B, D), and in some of these there wereIGL-SE labelled structures (Plate 2A) which, under theTEM, appeared to be bacteria in structures similar to theinfection-thread used by Rhizobium to enter host legumecells (Plate 2D). If this is the case then the bacteriaaccumulated in intercellular spaces close to these cellsmay be the source of the bacteria in these 'infectionthreads'. However, further examination suggested thatthe bacteria in these cells were contained within vacuolesand digested (Plate 2D).

Although the epidermis of stem bases was heavily


colonized by A. diazotrophicus, the bacteria were not seenwithin intercellular spaces internal to the epidermis. Thisis similar to results obtained by inoculating wheat rootswith free-living diazotrophs such as Azospirillum(Levanony et al., 1989; Patriquin et al., 1983) whenbacteria are generally found on the epidermis of inocu-lated roots and, rarely, within the epidermal layer inintercellular spaces. However, A. diazotrophicus was seenin large numbers in xylem vessels in the vascular bundles,both in the lumen and in apparently intimate associationwith the spiral secondary thickening (Plate 3A, B). Thesebacteria were also surrounded by mucus, which wasimmunogold labelled (Plate 3B, and see also Plate 2C),suggesting that it is the mucus surrounding the bacteriawhich contains the majority of antigens to the antiserumused in this study. The mucus probably contained exopo-

. lysaccharide (EPS) which has previously been shown tobe highly immunogenic in other Gram-negative diazo-trophic bacteria (Brewin et al., 1986; Levanony et al,1989; Schloterefa/., 1992).

In none of the IGL preparations used in these experi-ments was there significant gold labelling on the controlsections incubated in either the blocking/diluting bufferor the pre-immune serum (Plate 3C, D).

Discussion

We have shown that A. diazotrophicus will rapidly colon-ize, and adhere tightly to, sugar cane roots and lowerstems and, during a 15 d period, may also enter youngroots via the loose cells of the root tip and root cap (andpossibly via cracks at lateral root junctions), and enterthe vascular system. The presence of bacteria in xylemvessels is suggestive that these are a means of transportingbacteria to other parts of the plant, particularly the shoot.In rapidly growing, nitrogen-fixing sugar cane in the fieldthe xylem vessels are an obvious means of transportingthe bacteria from roots to the stem and leaves as theplant grows, and has also been reported in bacterialpathogens of sugar cane such as that which causes ratoonstunting disease (Kao and Damann, 1980). Previous workwith A. diazotrophicus in mature, field-grown sugar canesuggests that it is particularly localized in the stem nodes(Reis, 1991), where the dense, sucrose-storing paren-chyma tissue, and the interconnected vascular traces

Plate 3. (A) Light micrograph of xylem vessels in the lower stem of a sugar cane plantlet (NA 56-79) 15 DAI showing bacteria within the vesselsand in intimate association with the spiral secondary'thickening (arrows). This section was not immunogold labelled, x 1860. Bar= 10 fxm. (B) TEMof bacteria in xylem vessel from the lower stem of a sugar cane plantlet (NA 56-79) 15 DAI. The bacteria are surrounded by a mucus-type materialimmunogold labelled after incubation in an antibody raised against A. diazotrophicus followed by incubation in antibodies conjugated to 5 nm goldparticles. Note the thick, lignified wall of the xylem vessel (X). x 80000. Bar = 200 nm. (C) TEM of bacteria from the root tip of a sugar caneplant (NA 56-79), 15 DAI. This section was incubated in 'blocking buffer", without any rabbit antisera, followed by incubation in goat anti-rabbitantibodies conjugated to 5 nm gold particles. There is no gold-Labelling of, or near, the bacteria, including the 'mucus' surrounding them (cf.Plate 2C, 3B). x 80000. Bar = 100 nm. (D) TEM of a bacterium from the root tip of a sugar cane plant (NA 56-79), 15 DAI. This section wasincubated in rabbit 'pre-immune serum' followed by incubation in goat anti-rabbit antibodies conjugated to 5 nm gold particles. There is no gold-labelling of, or near, the bacterium, including the 'mucus' surrounding it (cf. Plate 2C, 3B). x 120000. Bar= 100 nm.

at National C

hung Hsing U



Dow

nloaded from



764 James et al.

(Greulach, 1973; Clements, 1980) will presumably providea low pO2 atmosphere and a plentiful supply of energysubstrate. The nodes are also a favoured site for coloniza-tion by pathogens (Kao and Damann, 1980). However,the exact location of A. diazotrophicus in mature stemnodes has not yet been determined, although recent workby Olivares et al. (1993) has shown that anotherdiazotroph associated with sugar cane, Herbaspirillumspp., seems to occupy, and proliferate in, xylem vesselsafter inoculation into sorghum (Sorghum bicolor) andsugar cane leaves. In addition, Baldani et al. (1993) havedemonstrated substantial colonization of rice leaf xylemvessels by Azospirillum brasilense grown from seeds con-taining diazotrophs. Therefore, stem and leaf xylem mayalso be the ultimate site of diazotroph colonization insugar cane. This suggestion is supported by recent resultsfrom other groups which have shown (non-pathogenic)colonization of xylem vessels in winter wheat by thediazotroph P. agglomerans (Ruppel et al., 1992) and byA. brasilense strain 245 (Schloter, Bode and Hartmann,unpublished). In addition You et al. (1990) localizedanother endophytic diazotroph, Alcaligenes faecalis,within intercellular spaces of rice roots and within cellsnear xylem using immunofluorescence. Therefore, xylemvessels may act as both a path of infection for thesebacteria and as a niche for N2-fixation and exchange ofmetabolites between endophyte and plant. Such a systemwas also suggested by Patriquin and Ddbereiner (1978)when they observed apparent fixation (via tetrazoliumsalts) by Azospirillum spp. in xylem vessels of field-grownmaize (Zea mays), Panicum maximum and Digitariadecumbens.

Our work shows similarities to other studies of endo-phytic diazotrophs such'as A. faecalis in rice (You andZhu, 1989; You et al, 1990). Also Hurek et al. (1990)and Reinhold-Hurek and Hurek (1993) inoculated Kallargrass (Leptochloa fused) and rice with a newly-discovereddiazotroph, Azoarcus spp., and used antibodies raisedagainst this bacterium and the gusA reporter gene, tolocalize the bacteria intra- and intercellularly within theroots. These authors also reported 'infection thread-like'structures within Kallar grass inoculated with thisorganism.

The nitrogenase activity shown by ARA in this studywill have been mainly due to the bacteria associated withthe plants, either adhering to the roots and lower stem,or within the plant tissues, as it is likely that the sucroseconcentration in the rooting medium after 15 d will havebeen so reduced by the growing plants that it will havebeen too low to support nitrogen fixation in bacteria notassociated with the sugar cane (Boddey et al, 1991).However, no attempt was made to remove the externalbacteria and assess the ARA of only the 'endophytic'bacteria so, in future experiments of this nature, it maybe desirable more accurately to mimic 'field' conditions

by only measuring the nitrogenase activity of the endo-phytes by removal of the external bacteria, e.g. via shakingthe roots and lower stems in Winogradskys mineralnitrogen-free medium (Kennedy and Tchan, 1992), or byselective inhibition with oxygen.

The infection of sugar cane occurred under asepticconditions where A. diazotrophicus was the sole bacteriumpresent and, as the plantlets reduced the sucrose concen-tration in the rooting medium, maximum opportunitywas provided for the non-aggressive entry ofA. diazotrophicus (i.e. via loose cells or cracks) during the15 d. This experiment partially confirms the endophyticnature of A. diazotrophicus as, in greenhouse experimentsusing a liquid inoculum of A. diazotrophicus on sugarcane plantlets grown under non-sterile conditions, infec-tion of the plants did not occur (Reis, 1991), presumablydue to rapid death of the bacteria under the soil conditionsused (Paula et al, 1990). In addition, Paula et al. (1991)observed that A. diazotrophicus will only infect sugarcane, sweet potato and sweet sorghum via wound damage,via mycorrhizal spores or by vegetative propagation.A. diazotrophicus does not survive long in the soil, andin the field it is thought likely that it is transferred directlyfrom sugar cane sets to the emerging new plant (Paulaet al, 1990; Dobereiner et al, 1990, 1993; Reis et al,unpublished) rather than via direct infection from thesoil. However, the present study does support the sugges-tion that it is possible, under very specific conditions, thatsugar cane plantlets could be infected via direct inocula-tion (Reis et al, unpublished).

Although there has as yet been no direct connectionestablished between the observed nitrogen fixation andthe presence of A. diazotrophicus, or the other diazotrophscommonly found in sugar cane, the present study hasshown that IGL, using the antibodies that we haveproduced, is a reliable method of localizing diazotrophsin sugar cane tissue and could be extended to field-grown plants.

Acknowledgements

We thank R.M. Boddey, J.I. Sprent and J.M. Sutherland forhelpful discussions, and G. Baeta da Cruz, M. Gruber andM. Kierans for technical assistance. E. James was funded bythe World Bank through the Inter-American Institute forCo-operation in Agriculture (IICA) and by the Department ofBiological Sciences, University of Dundee. V.M. Reis wassupported by EMBRAPA and CNPq and this work constitutespart of her PhD thesis.

References

Baldani VLD, James EK, Baldani JI, Dobereiner J. 1993.Colonization of rice by the nitrogen-fixing bacteriaHerbaspirillum spp. and Azospirillum brasilense. In:Palacios R, Mora J, Newton WE, eds. New horizons innitrogen fixation. Dordrecht: Kluwer Academic Publishers,705.

at National C

hung Hsing U



Dow

nloaded from



Bashan Y, Levanony H. 1989. Factors affecting adsorption ofAzospirillum brasilense Cd to root hairs as compared withroot surface of wheat. Canadian Journal of Microbiology35, 936-44.

Bashan Y, Levanony H, Whitmoyer RE. 1991. Root surfacecolonization of non-cereal crop plants by pleomorphicAzospirillum brasilense Cd. Journal of General Microbiology137, 187-96.

Boddey RM, Urquiaga S, Reis VM, Dfibereiner J. 1991.Biological nitrogen fixation associated with sugar cane. Plantand Soil 137, 111-17.

Boonjawat J, Chaisiri P, Iimpananont J, Soontaros S,Pongsawasdi P, Chaopoogpang S, Pompattkul S,Wongwaitayakul B, Sangduan L. 1991. Biology of nitrogen-fixing rhizobacteria. Plant and Soil 137, 119-25.

Brewin NJ, Wood EA, Larkins AP, Galfre G, Butcher GW.1986. Analysis of lipopolysaccharide from root nodulebacteroids of Rhizobium leguminosarum using monoclonalantibodies. Journal of General Microbiology 132, 1959-68.

Cavalcante VA, Dfibereiner J. 1988. A new acid-tolerantnitrogen-fixing bacterium associated with sugar cane. Plantand Soil 108, 23-31.

Christiansen-Weniger C. 1992. N2-fixation by ammonium-excreting Azospirillum brasilense in auxin-induced roottumours of wheat (Triticum aestivum L.). Biology and Fertilityof Soils 13, 165-72.

Clements HF. 1980. Sugar cane crop logging and crop control:principles and practices. London: Pitman, 520.

Cocking EC, Davey MR, Kothari SL, Srivastava JS, Jing Y,Ridge RW, Rolfe BG. 1993. Altering the specificity controlof the interaction between rhizobia and plants. Symbiosis14, 123-30.

Cojho EH, Reis VM, Schenberg AC, Ddbereiner J. 1993.Interactions of Acetobacter diazotrophicus with an amylolyticyeast in nitrogen-free batch culture. FEMS MicrobiologyUtters 106, 341-6.

Dobereiner J, Pimentel JP, Olivares FL, Urquiaga S. 1990.Bacterias diazotr6ficas podem ser end6fitas e/ou fitopatogen-icas. Anais da Academia Brasileira de Ciencia 62, 319.

Ddbereiner J, Reis VM, Lazarini AC. 1988. New N2-fixingbacteria in association with cereals and sugar cane. In:Bothe H, De Bruijn FJ, Newton WE, eds. Nitrogen fixation:hundred years after. Stuttgart: Gustav Fischer, 717-22.

Ddbereiner J, Reis VM, Paula MA, Olivares F. 1993. Endophyticdiazotrophs in sugar cane, cereals and tuber plants. In:Palacios R, Mora J, Newton WE, eds. New horizons innitrogen fixation. Dordrecht: Kluwer Academic Publishers,671-6.

Greulach VA. 1973. Plant function and structure. New York:Macmillan Publishing Company.

Hendre RR, Iyer RS, Korwalm M, Kluspe SS, Mascarenhas AF.1983. Rapid multiplication of sugar cane by tissue culture.Sugar Cane 1, 5-8.

Hurek T, Reinhold-Hurek B, Van Montagu M, KeUenberger E.1990. Infection of intact roots of Kallar grass and riceseedlings by Azoarcus. In: Polinelli M, Materassi R,Vicenzini M, eds. Nitrogen fixation. Proceedings of the FifthInternational Symposium on Nitrogen Fixation With Non-legumes. Dordrecht: KJuwer Academic Publishers, 235—42.

James EK, Sprent Jl, Minchin FR, Brewin NJ. 1991. Intercellularlocation of glycoprotein in soybean nodules: effect of alteredrhizosphere oxygen concentration. Plant, Cell andEnvironment 14, 467-76.

Kao J, Damann KE. 1980. In situ localization and morphologyof the bacterium associated with ratoon stunting disease ofsugarcane. Canadian Journal of Botany 58, 310-15.

Kennedy IR, Tchan Y-T. 1992. Biological nitrogen fixation in


non-leguminous field crops: recent advances. Plant and Soil141,93-118.

Levanony H, Bashan Y, Romano B, Klein E. 1989. Ultrastrucrurallocalization and identification of Azospirillum brasilense Cdon and within roots by immunogold labelling. Plant and Soil117, 207-18.

Li R, MacRae IC. 1991. Specific association of diazotrophicAcetobacters with sugar cane. Soil Biology and Biochemistry23, 999-1002.

Li R, MacRae IC. 1992. Specific identification and enumerationof Acetobacter diazotrophicus in sugar cane. Soil Biology andBiochemistry 24, 413-19.

Lima E, Boddey RM, Ddbereiner J. 1987. Quantification ofbiological nitrogen fixation associated with sugar cane usinga 15-N aided nitrogen balance. Soil Biology and Biochemistry19, 165-70.

Murashige T, Skoog FA. 1962. Revised medium for rapidgrowth and bioassay with tobacco tissue cultures. PhysiologiaPlantarum 15, 473-97.

Olivares FL, James EK, Reis VM, Baldani VLD, Baldani JI,Ddbereiner J. 1993. Colonization of vascular tissue byHerbaspirillum spp. in sorghum and sugar cane. In:Proceedings of the Twenty Third Brazilian Congress of PlantPathology.

Patriquin DG, Dfibereiner J. 1978. Light microscopy observa-tions of tetrazolium-reducing bacteria in the endorhizosphereof maize and other grasses. Canadian Journal of Microbiology24, 734-42.

Patriquin DG, Dfibereiner J, Jain DK. 1983. Sites and processesof association between diazotrophs and grasses. CanadianJournal of Microbiology 29, 900-15.

Paula MA, Reis VM, Ddbereiner J. 1991. Interactions of Glomusclarum with Acetobacter diazotrophicus in infection of sweetpotato (Ipomoea batatas), sugarcane {Saccharum spp.), andsweet sorghum (Sorghum vulgare). Biology and Fertility ofSoibU, 111-15.

Paula MA, Reis VM, Urquiaga S, Ddbereiner J. 1990. Esporosde fungo MVA Glomus clarum como veiculo de infeccao deAcetobacter diazotrophicus. Anais da Academia Brasileira deCiencia 62, 318.

Pimentel JP, Olivares F, Pitard RM, Urquiaga S, Akiba F,Ddbereiner J. 1991. Dinitrogen fixation and infection of grassleaves by Pseudomonas rubrisubalbicans and Herbaspirillumseropedicae. Plant and Soil 137, 61-5.

Reinhold-Hurek B, Hurek T. 1993. Capacities of Azoarcus sp.,a new genus of grass-associated diazotrophs. In: Palacios R,Mora J, Newton WE, eds. New horizons in nitrogen fixation.Dordrecht: Kluwer Academic Publishers, 691—4.

Reis VM. 1991. Aspectos ecol6gicos e fisiol6gicos da bacteriafixadora de N2 Acetobacter diazotrophicus, MSc Thesis.Universidade Federal Rural Rio de Janeiro.

Reis VM, Olivares FL, James EK, Baldani Jl, Ddbereiner J.1993. Infection of sugar cane by nitrogen-fixing bacteria. In:Palacios R, Mora J, Newton WE, eds. New horizons innitrogen fixation. Dordrecht: KJuwer Academic Publishers,705.

Ruppel S, Hecht-Buchholz C, Remus R, Orrmann U, Schmelzer R.1992. Settlement of the diazotrophic, phytoeffective strainPantoea agglomerans on and within winter wheat: aninvestigation using ELISA and transmission electron micro-scopy. Plant and Soil 145, 261-73.

Schloter M, Bode W, Hartmann A. 1992. Characterization ofmonoclonal antibodies against cell surface structures ofAzospirillum brasilense Sp7 using ELISA techniques.Symbiosis 13, 37-45.

Urquiaga S, Cruz KHS, Boddey RM. 1992. Contribution of

at National C

hung Hsing U



Dow

nloaded from



766 James et al.

nitrogen fixation to sugar cane: Nitrogen-15 and nitrogen- You CB, Song W, Wang HX, Li JP, Hai WL. 1990. Associationbalance estimates. Soil Science Society of America Journal of Alcaligenes faecalis with wetland rice. In: Polinelli M,56, 105-14. Materassi R, Vicenzini M, eds. Nitrogen fixation. Proceedings

You C, Zhou F, 1989. Non-nodular endorhizospheric nitrogen of the Fifth International Symposium on Nitrogen Fixationfixation in wetland rice. Canadian Journal of Microbiology With Non-legumes. Dordrecht: Kluwer Academic35, 403-8. Publishers, 195-9.

at National C

hung Hsing U



Dow

nloaded from



infection of sugar cane by the nitrogen-fixing bacterium...

Documents