Colonization of sugarcane by Acetobacter diazotrophicus is inhibited by high N-fertilization Luis E. Fuentes-Ram| Lrez a , Jesu L s Caballero-Mellado a , Jorge Sepu L lveda b , Esperanza Mart| Lnez-Romero a; * a P. de Ecolog| La Molecular y Microbiana, Centro de Investigacio Ln sobre Fijacio Ln de Nitro Lgeno, Universidad Nacional Auto Lnoma de Me Lxico, Apdo. Postal 565-A, Cuernavaca, Morelos, Me Lxico b Unidad de Microscop| La Electro Lnica, Instituto de Fisiolog| La Celular, Universidad Nacional Auto Lnoma de Me Lxico, Me Lxico DF, Me Lxico Received 9 September 1998; received in revised form 9 December 1998; accepted 9 December 1998 Abstract Acetobacter diazotrophicus is a nitrogen-fixing endophytic bacterium, originally isolated from sugarcane. Its colonizing ability was evaluated in high and low N-fertilized sugarcane plants by inoculating stem-cuts with a L-glucuronidase marked A. diazotrophicus strain. Bacterial quantification by the most probable number technique showed a severe decrease of A. diazotrophicus cells in plants fertilized with high levels of nitrogen. The inoculated strain was detected inside low N-fertilized sugarcane plants by histochemical staining of L-glucuronidase and scanning electron microscopy. A. diazotrophicus was found mainly inside cortical cells of stems and inside xylem vessels. No L-glucuronidase activity was observed in non-inoculated plants. High nitrogen fertilization of fields might be a threat to maintaining naturally occurring endophytic associations. z 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords : Acetobacter ; Auxin ; Endophyte ; Nitrogen-¢xing bacteria ; Sugarcane ; Xylem 1. Introduction The sugarcane crop is vegetatively propagated by use of stems and this plant produces large amounts of biomass which demand a massive input of nu- trients, especially N and K [1]. In almost all coun- tries where this crop is cultivated, a common agricul- tural practice is to apply 250 kg N or more per Ha. Nevertheless, Brazilian farmers have used amounts of fertilizer that do not adequately cover the theoret- ical loss of nitrogen occurring when the plants are harvested. Surprisingly, these crops do not show ni- trogen de¢ciencies, and their response to the addition of nitrogen fertilizer is usually negligible [2]. Conse- quently, biological nitrogen ¢xation (BNF) has been suggested to contribute to the nutrition of sugarcane plants [3]. In fact, experiments using 15 N isotope dilution or N balance methods gave evidence that BNF provided an important proportion of the nitro- gen requirements of di¡erent sugarcane varieties [4,5]. Di¡erent N 2 -¢xing bacteria, such as Enterobacter cloacae, Erwinia herbicola, Klebsiella pneumoniae, 0168-6496 / 99 / $20.00 ß 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII:S0168-6496(98)00125-1 * Corresponding author. Tel.: +95 (73) 131697; Fax: +95 (73) 175581; E-mail: [email protected]FEMS Microbiology Ecology 29 (1999) 117^128
Transcript
Colonization of sugarcane by Acetobacter diazotrophicus isinhibited by high N-fertilization
Luis E. Fuentes-Ram|èrez a, Jesuès Caballero-Mellado a, Jorge Sepuèlveda b,Esperanza Mart|ènez-Romero a;*
a P. de Ecolog|èa Molecular y Microbiana, Centro de Investigacioèn sobre Fijacioèn de Nitroègeno, Universidad Nacional Autoènoma de Meèxico,Apdo. Postal 565-A, Cuernavaca, Morelos, Meèxico
b Unidad de Microscop|èa Electroènica, Instituto de Fisiolog|èa Celular, Universidad Nacional Autoènoma de Meèxico, Meèxico DF, Meèxico
Received 9 September 1998; received in revised form 9 December 1998; accepted 9 December 1998
Abstract
Acetobacter diazotrophicus is a nitrogen-fixing endophytic bacterium, originally isolated from sugarcane. Its colonizingability was evaluated in high and low N-fertilized sugarcane plants by inoculating stem-cuts with a L-glucuronidase marked A.diazotrophicus strain. Bacterial quantification by the most probable number technique showed a severe decrease of A.diazotrophicus cells in plants fertilized with high levels of nitrogen. The inoculated strain was detected inside low N-fertilizedsugarcane plants by histochemical staining of L-glucuronidase and scanning electron microscopy. A. diazotrophicus was foundmainly inside cortical cells of stems and inside xylem vessels. No L-glucuronidase activity was observed in non-inoculatedplants. High nitrogen fertilization of fields might be a threat to maintaining naturally occurring endophyticassociations. z 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rightsreserved.
The sugarcane crop is vegetatively propagated byuse of stems and this plant produces large amountsof biomass which demand a massive input of nu-trients, especially N and K [1]. In almost all coun-tries where this crop is cultivated, a common agricul-tural practice is to apply 250 kg N or more per Ha.Nevertheless, Brazilian farmers have used amountsof fertilizer that do not adequately cover the theoret-
ical loss of nitrogen occurring when the plants areharvested. Surprisingly, these crops do not show ni-trogen de¢ciencies, and their response to the additionof nitrogen fertilizer is usually negligible [2]. Conse-quently, biological nitrogen ¢xation (BNF) has beensuggested to contribute to the nutrition of sugarcaneplants [3]. In fact, experiments using 15N isotopedilution or N balance methods gave evidence thatBNF provided an important proportion of the nitro-gen requirements of di¡erent sugarcane varieties[4,5].
Di¡erent N2-¢xing bacteria, such as Enterobactercloacae, Erwinia herbicola, Klebsiella pneumoniae,
0168-6496 / 99 / $20.00 ß 1999 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.PII: S 0 1 6 8 - 6 4 9 6 ( 9 8 ) 0 0 1 2 5 - 1
Azotobacter vinelandii, Paenibacillus polymyxa (for-merly Bacillus polymyxa), Herbaspirillum seropedi-cae, Herbaspirillum rubrisubalbicans and Acetobacterdiazotrophicus colonize the sugarcane rhizosphereand inner tissues [3,6,7]. These bacteria, and possiblyother diazotrophs not yet isolated could contributeto BNF in this plant. It appears that when the stalksare sown, they carry endophytic bacteria that mayspread inside the plant after budding. A. diazotrophi-cus has been suggested to be an endophytic contrib-utor of nitrogen to this crop, as it ¢xes nitrogen inculture medium under acidity levels and sugar con-centrations that resemble those inside the plant[6,8,9]. It has been reported that the frequency ofisolation of A. diazotrophicus from sugarcane plantsdiminishes in relation to the amounts of N-fertiliza-tion used in the ¢elds [10,11]. Caballero-Mellado etal. [12] found that Brazilian isolates were geneticallymore diverse than Mexican ones and suggested thatthis could be related to the di¡erence in nitrogenfertilization levels between the two countries, insuch a manner that the application of more fertilizercaused a diminished diversity. In the nitrogen-¢xingsymbiosis of Rhizobium and legumes, high nitrogenfertilization abolishes nodulation or, when applied toexisting nodules, nitrogen ¢xation. It was thereforeof interest to evaluate if the supposedly nitrogen-¢x-ing A. diazotrophicus-sugarcane association was sim-ilarly a¡ected by nitrogen.
By using microscopic techniques in root-inocu-lated sterile plantlets, James et al. [13] detected A.diazotrophicus colonizing the root intercellularspaces, and the interior of root epidermal cells.They proposed that A. diazotrophicus could be dis-tributed from the base of the stem to other organsvia the stem xylem vessels, since they also detectedxylem colonization in the basal region of the stalk.In non-inoculated sugarcane plants, Dong et al. [14]isolated A. diazotrophicus from apoplastic £uid, thatincludes £uid from various locations, such as cellwalls, intercellular spaces, and xylem sap [15].
The aim of this work was to examine the e¡ects ofnitrogen fertilization on the endophytic colonizationof the sugarcane by inoculating an A. diazotrophicusgusA marked strain. In addition, we attempted toclarify the location of A. diazotrophicus inside theplant.
2. Materials and methods
2.1. Bacterial strains, plasmids and growth conditions
The bacterial strains and plasmids used in thisstudy are listed in Table 1. The Escherichia colistrains were cultured in Luria^Bertani medium at37³C. When necessary kanamycin (Km), streptomy-cin (Sm), spectinomycin (Sp), and nalidixic acid
FEMSEC 1009 8-6-99 Cyaan Magenta Geel Zwart
Table 1Strains and plasmids
Strain or plasmid Relevant characteristicsa Reference
E. coliCMK Nals, Sms [41]
A. diazotrophicusUAP 5541 Wild-type, without plasmids, able to ¢x N2 in vitro, common clone by multilocus assay, Nalr [10,18]
pRG960SD Smr, Spr cosmid, IncP, Mob�, promoter-less gusA with a Shine and Dalgarno sequence [17]
pBI426 Apr, Kmr vector with gusA-NPTII expressed from a double 35S CaMV virus promoter plus aleader sequence from alfalfa mosaic virus, also gusA-NPTII expressed from unidenti¢ed regionin di¡erent Gram-negative bacteria
[16]
pRGS561 Smr,Spr, pRG960SD derivative with gusA-NPTII constitutive expression in A. diazotrophicus This work
(Nal) were added at ¢nal concentrations of 25, 60,60, and 15 Wg ml31, respectively. LGI medium [6]was used for growing A. diazotrophicus strains, andN-free semi-solid LGI [6] for the isolation of theinoculated strains. For triparental conjugations,MESMA medium with the following compositionwas used (g l31) : yeast extract, 2.7; glucose, 2.7;mannitol, 1.8; MES (Sigma, St. Louis, MO), 4.4;K2HPO4, 4.81; KH2PO4, 0.65; Bromothymol blue,0.025; and agar, 12; pH 6.7. For measuring the GUSactivity of A. diazotrophicus, cells were grown in LGIbroth, containing (g l31) : K2HPO4, 0.2; KH2PO4,0.6; MgSO4W7H2O, 0.2; CaCl2W2H2O, 0.02;FeCl3W6H2O, 0.01; Na2Mo4W2H2O, 0.002; and(NH4)2SO4, 0.617; plus the carbon source previously¢lter-sterilized (sucrose, fructose, glucose or gluco-nate, 1.5%), pH 5.5. GUS activity was also deter-mined in cells growing with sucrose (10%) added tosemi-solid SUCMES, containing (g l31) : MES 4.4,K2HPO4 4.81, KH2PO4 0.65, (NH4)2SO4 1.48, andagar 1.5, pH 6.7.
2.2. Plasmids and strain construction
To construct a gusA marked strain, a DNA frag-ment from pBI426 [16] carrying an alfalfa mosaicvirus leader sequence and a double 35S CaMV pro-moter fused to gusA-NPTII, was inserted into thebroad-host range plasmid pRG960SD digested withEcoRI^HindIII [17]. The resulting construct(pRGS561) was conjugatively mobilized from E.coli to A. diazotrophicus strain UAP 5541 [10,18]by triparental mating using E. coli HB 101/pRK2013 as a helper. A. diazotrophicus transconju-gants were selected on MESMA plates containingNal (15 Wg ml31) and Sm (45 Wg ml31).
2.3. Fluorogenic L-glucuronidase assays
The L-glucuronidase (GUS) activity of A. diazotro-phicus UAP 5541 carrying pRGS561 was tested invitro with a £uorogenic assay as described by Je¡er-son [19]. The wild-type strain A. diazotrophicus UAP5541 and its derivative were grown with di¡erentcarbon sources at 1.5% concentration (sucrose at1.5 and 10%). Bacterial cells were resuspended inextraction bu¡er containing 50 mM sodium phos-phate, pH 7.0; 10 mM L-mercaptoethanol; 10 Mm
Na2EDTA; 0.1% N-lauroylsarcosine; and 0.1% Tri-ton X-100. Bacterial extracts were incubated at 37³Cin MUG bu¡er, consisting of 1 mM 4-methylumbel-liferyl-D-glucuronide (Sigma, St. Louis, MO) in ex-traction bu¡er. Aliquots were removed every 5 minfor 30 min. The reaction was stopped by mixingaliquots with 0.2 M Na2CO3. For each assay, a cal-ibration curve was preformed with 100 nM 4-meth-ylumbelliferone (MU) in extraction bu¡er. Fluores-cence determinations were performed with a TKO100 £uorometer (Hoefer Scienti¢c Instruments, SanFrancisco, CA), at wavelengths of 365 nm (excita-tion) and 460 nm (emission).
2.4. Inoculation and growth of plants
Adult stalks of the sugarcane varieties Z MEX5532, MEX 57-473, RD 75-11, MY 55-14, and RB76-5418, regenerated from tissue cultures and subse-quently grown in experimental ¢elds, were kindlysupplied by R. Meèndez-Salas (Instituto Nacional deInvestigaciones Forestales Agr|ècolas y Pecuarias, Za-catepec, Mexico). This plant material was selectedfor the colonization experiments as it was shown tolack endogenous L-glucuronidase activity.
A. diazotrophicus strains were introduced insidesugarcane stems (setts) prior to their budding so asto resemble the A. diazotrophicus^sugarcane relation-ship under natural conditions. Setts having one nodewere inoculated with bacteria suspended in water.The setts were previously dehydrated at 45³C for8^10 h. Approximately 107 bacterial cells were ino-culated per plant. The setts were planted in sterilehumid vermiculite/perlite mixture (1:1), incubatedin a greenhouse at 28³C, and watered with sterilewater until they began to bud. After budding, 50 mlof MS modi¢ed mineral solution [20] was addedweekly per plant, for ten times maximum. Themineral solution contained the following: 1.5 mMMgSO4W7H2O, 100 WM H3BO3, 30 WMZnSO4W7H2O, 100 pM CuSO4W5H2O, 5.3 WM KI,105 pM CoCl2W6H2O, 1 WM Na2MoO4W2H2O, 100mM MnSO4WH2O, 3 mM CaCl2W2H2O, 4.1 mMNa2EDTA, 6.7 mM FeSO4W7H2O, and 1.3 mM po-tassium phosphate, pH 6.0. Nitrogen fertilizer wassupplied every 2 weeks, with the high and low treat-ments consisting of 11 and 0.56 mmol of NH4NO3
per plant, respectively. The plants used for histo-
chemical GUS assays, inoculated strain isolation andscanning electron microscopy were collected 1, 2, 3,5 or 7 months after sprouting.
2.5. Histochemical L-glucuronidase analysis
The histochemical assay was done as recom-mended by Je¡erson and Wilson [21]. Each plantsample from all the varieties tested was asepticallyseparated into two subsamples. Sections of stem androots from one subsample were incubated in the fol-lowing bu¡er: 2 mM X-Gluc (Biosynth, Staad, Swit-zerland), previously dissolved in DMSO, 100 mMsodium phosphate pH 7.0, 0.5 mM Triton X-100,0.5 mM K3Fe(CN)6, 0.5 mM K4Fe(CN)6W3H2O,10 mM Na2EDTA, and 2 nM Na2S2O5. Sectionsfrom the other subsample were incubated in thesame bu¡er supplemented with 0.02% NaN3. Thesections were incubated for 24 h at 30³C and usedfor microscopical analysis.
2.6. Scanning electron microscopy
Plant samples were ¢xed in 3% glutaraldehyde in100 mM sodium phosphate pH 7.0, washed with100 mM sodium phosphate pH 7.0, ¢xed in 1% os-mium tetroxide, washed again with phosphate bu¡erand predehydrated with increasingly concentratedethanol (from 30 to 99%). The specimens were dehy-drated to critical point and gold coated. Observa-tions were carried out in a JSM-5410LV (Jeol, To-kyo, Japan) scanning electron microscope.
2.7. Optical microscopy
At each sampling time (10 days, 1, 2, 3, 5 and 7months after sprouting) stems and roots of twoplants assayed for GUS activity were observed underlow magni¢cation. Two sections from each stem ofplants without GUS activity and two stem sectionswith GUS activity were selected for examination athigher magni¢cations. Samples were ¢xed in glutar-aldehyde, washed in 100 mM sodium phosphate pH7.0 and predehydrated with ethanol as describedabove. Samples were immersed in propylene oxide,and in 1:1 propylene oxide-Eponate 12 resin (Pelco,Redding, CA) mixture, and embedded in Eponate12. Polymerization was carried out overnight at
60³C and 1.5 Wm sections were used for observation.From the ¢xed samples three subsections were takenfrom two di¡erent non-inoculated and two inocu-lated plants which had been grown under two nitro-gen fertilizer doses.
2.8. A. diazotrophicus re-isolation
A. diazotrophicus was isolated from inoculatedsugarcane plants as described previously [10]. Smallpieces from the plant samples to be assayed histo-chemically for GUS activity were aseptically sepa-rated and crushed for inoculation in LGI semi-solidmedia. After 6 days at 30³C the bacterial growth wasstreaked on LGI plates and incubated at the sametemperature for 5 days.
2.9. Quanti¢cation of A. diazotrophicus cells
From sugarcane plantlets cv. Z MEX 5532 (1 and2 months after spouting), A. diazotrophicus cells werequanti¢ed by the most probable number method(MPN). Root and stem samples obtained from sur-face sterilized sugarcane plants were ¢nely maceratedand resuspended in a chilled sucrose solution (1%).Serial dilutions were inoculated by triplicate in LGIsemi-solid media containing cycloheximide (150 Wgml31) and incubated at 30³C for 5^6 days. Diazo-trophs were enriched by incubating under similarconditions in the same media. Positive growth ofA. diazotrophicus was determined by acidi¢cationand formation of the typical pellicle [6]. Numbersof bacteria were normalized to fresh weight of tissue.The presence of A. diazotrophicus was veri¢ed bymorphology in LGI plates. In addition, Sm resist-ance in MESMA plates and GUS activity were con-¢rmed in isolates from plants inoculated with UAP5541 carrying pRGS561. The plasmid was puri¢edfrom 15 colonies and observed by ethidium bromidestaining.
2.10. Experimental design
Two di¡erent experiments were performed to eval-uate the e¡ect of the nitrogen on A. diazotrophicuscolonization. For the ¢rst experiment, plants (cv. ZMEX 5532) were used 30 days after sprouting. Theplants for the ¢rst experiment consisted of 20 canes
grown with the low nitrogen dose and 24 canes withthe high nitrogen dose. The di¡erence between thenumber of A. diazotrophicus in canes grown underlow and high nitrogen fertilization 30 days aftersprouting were tested with Student's t-test. For thesecond experiment, plants 60 days after sprouting(cv. Z MEX 5532) were used. For this determination,four inoculated plants grown with the low nitrogenfertilization and four grown with the high nitrogenfertilization were processed, as well as eight non-in-oculated controls, four of which were grown underthe high and four under the low fertilization condi-tions.
Additionally, colonization was detected by re-iso-lating A. diazotrophicus from inoculated plants 1, 2,3, 5 and 7 months after sprouting. The varieties usedin this experiment were Z MEX 5532, MEX 57-473,RD 75-11, MY 55-14, and RB 76-5418.
3. Results
3.1. Colonization of sugarcane by A. diazotrophicus
The inoculated strains could be recovered andidenti¢ed from low N-fertilized sugarcane varietiesZ MEX 5532, MEX 57-473, RD 75-11, MY 55-14,and RB 76-5418 at 1, 2, 3, 5 and 7 months aftersprouting. Isolated bacteria produced the typical A.diazotrophicus growth in semi-solid LGI and showedthe expected colony morphology in LGI plates. Thenumbers of A. diazotrophicus that endophyticallycolonized sugarcane stems or roots (cv. Z MEX5532) fertilized with di¡erent nitrogen quantitiesshowed signi¢cant di¡erences (Table 2). Di¡erencesin A. diazotrophicus colonization was also seen in60-day plants. The numbers in low N-fertilizedplants were 2.9U103 and 1.2U103 colony forming
units (CFU) per g of fresh weight in the stem andin the roots, respectively, while in the high N-fertil-ized plants, A. diazotrophicus was not detected by themost probable number method. Even in older plants(up to 7 months) of ¢ve di¡erent cultivars (Z MEX5532, MEX 57-473, RD 75-11, MY 55-14, and RB76-5418), A. diazotrophicus could not be isolatedfrom inoculated plants maintained under high nitro-gen fertilization, while isolation was always success-ful from plants maintained under low nitrogen fer-tilization (results not shown).
No A. diazotrophicus isolates were obtained fromnon-inoculated plants. An estimate of the total num-ber of bacteria endophytically colonizing the rootsand stems of sugarcane was in a range from 4U106
to 40U106 CFU per g of plant fresh tissue. Thesevalues were obtained from the bacterial growth inthe semi-solid selective medium LGI. Possibly agreat proportion of these bacteria were nitrogen-¢xers, since the nitrogen in semi-solid LGI mediumcomes only from impurities in the components. Dif-ferent types of bacteria were isolated in LGI mediafrom non-inoculated plants. None of them corre-sponded morphologically to A. diazotrophicus insemi-solid LGI or LGI plates and we did not tryto de¢ne their taxonomic status.
3.2. Localization of A. diazotrophicus in planta
To follow plant colonization by A. diazotrophicus,a GusA� strain was obtained by using a DNA frag-ment with the alfalfa mosaic virus leader and thedouble CaMV promoter fused to gusA [16]. Thestrain showed GUS activity in vitro when grownwith carbon sources, such as gluconic acid, fructose,glucose and sucrose (results not shown).
GUS activity was detected in stems, but only inplants with low nitrogen fertilization (varieties: Z
FEMSEC 1009 8-6-99 Cyaan Magenta Geel Zwart
Table 2A. diazotrophicus cell numbers colonizing inoculated sugarcane plants at 30 days after sproutinga
Stems Roots
Low Nb 5.7U102 A (1.75U102) 2.7U102 A (1.28U102)High Nc 0.5U102 B (0.4U102) 60.3U102 B
Superscript A and B mean that di¡erence in bacterial numbers is signi¢cant (Ps 0.05).aCFU per g of fresh weight determined by most probable number counting technique.bMean of 20 plants (S.E.M. in parentheses).cMean of 24 plants (S.E.M. in parentheses).
Fig. 2. A. diazotrophicus-like cells inhabiting cells of sugarcane stems. Sections of 5-month-old plants grown with low N-fertilization. (A)Section of sucrose storage parenchymatous cells located near to the stem cortex. (B) Section of a vascular bundle, a tracheary element(black asterisk) is surrounded by parenchyma cells (white asterisks).
MEX 5532, MEX 57-473, RD 75-11, MY 55-14, andRB 76-5418). Only in assays without NaN3 wasGUS activity detected. In stem tissues, the highestGUS activity was present in the cortex and the vas-cular bundle (xylem vessels and apparently alsophloem sieve tubes), (Fig. 1A,C,D). GUS activitywas never observed in non-inoculated plants (notshown). Scanning microscopy of stem samples fromplants inoculated with the strain UAP 5541 carryingpRGS561 showed bacterial cells adhering to thewalls of plant cells (Fig. 2). These bacteria were mor-phologically indistinguishable from typical A. diazo-trophicus rods (0.6U2 Wm). The vascular bundle andits surrounding cells were most abundantly colonizedby these cells. In the non-inoculated controls, it wasalso possible to observe bacterial cells (not shown),but they were clearly di¡erent in size and shape fromthe A. diazotrophicus cells.
4. Discussion
The internal colonization of sugarcane by A. dia-zotrophicus in plants maintained with low and highnitrogen fertilizer doses was evaluated with a gusASmr strain by re-isolation of the strain, con¢rmationof its identity by histochemical staining for L-glucur-onidase and antibiotic resistance, and by scanningelectron microscopy.
We used plants originally regenerated from tissueculture because they lacked endogenous GUS activ-ity. GUS activity was obtained from stems of di¡er-ent sugarcane varieties coming from agricultural¢elds, and it is most probably of bacterial origin,as described in Dioscorea [22]. Successful coloniza-tion was observed in all the sugarcane varieties used,but the inoculated strain and the GUS activity wereonly detected in plants grown under low levels ofnitrogen fertilization. The lack of A. diazotrophicuscolonization in the presence of high supplied nitro-gen explains the low presence and low frequency ofisolation of A. diazotrophicus from sugarcane plantsgrown in ¢elds with high nitrogen fertilization levels,reported previously by Fuentes Ram|èrez et al. [10]and Muthukumarasamy [11]. In experiments withthe grasses Miscanthus sinensis, M. sacchari£orusand Spartina pectinata, Kirchhof et al. [23] foundsimilar results by quantifying the cell numbers of
the diazotrophic endophytic community inhabitingplants fertilized and unfertilized with nitrogen. Thise¡ect of nitrogen on colonization might not be uni-versal since, for instance, Herbaspirillum rubrisubal-bicans behaves as a pathogen in susceptible sugar-cane cultivars grown in countries where high levelsof nitrogen fertilization are used [7]. We detected102^103 CFU per g of fresh tissue of A. diazotrophi-cus endophytically colonizing low N-fertilized 30-day-old sugarcane plants. In roots of non-inoculatedsugarcane plants, Reis et al. [24] found 104^106 CFUof A. diazotrophicus per g of fresh tissue. The quan-titative di¡erence found between that report andours could be related to the inoculation processthat we used. In addition, the estimate of Reis etal. [24] could also include super¢cially adhering cellsin addition to endophytic ones, and the plant culti-var might also have an in£uence [15]. In a study ofthe association of A. diazotrophicus with di¡erentcultivars of sugarcane, da Silva et al. [25] suggestedthat the A. diazotrophicus population is sensitive tothe plant genotype. They observed that only in oneof their cultivars the A. diazotrophicus populationincreased during the time of the study (15 months),but in the other ones, they did not detected any trendin the bacterial numbers.
The e¡ect observed on the A. diazotrophicus pop-ulation colonizing sugarcane does not seem to be adirect negative e¡ect of the fertilizer on the bacteria.We did not detect negative e¡ects on wild-type andGusA� A. diazotrophicus strains growing in culturemedia supplied with high nitrogen levels (10 mMNO3
3), and at the same nitrogen concentration theGusA� construct also expressed L-glucuronidase ac-tivity. Thus, it is more probable that the physiolog-ical state of the plant is altered by the nitrogen, andthis subsequently a¡ects its association with the en-dophyte. Pelaez Abellan et al. [26] observed that su-crose synthesis is reduced in sugarcane leaves by ap-plication of NO3
3 in a highly productive variety andincreased sucrose synthesis in a variety with low pro-ductivity. In complete sugarcane plants, high nitratedoses were associated with a decrease in the concen-trations of sucrose in the leaves and of the reducingsugars and sucrose in the stem [27].
A. diazotrophicus is commonly found in roughlythe same numbers in sugarcane roots, stems andleaves under ¢eld conditions [6,28,29]. In the present
study, the inoculated strains were recovered fromstem and root samples, and we also detected GUSactivity in the stems, but not in roots, in spite of thesimilar bacterial numbers recovered from both or-gans. We cannot explain the lack of activity in roottissue, especially considering the similar numbers ofA. diazotrophicus cells inhabiting this organ relativeto the numbers inside the stem.
In this study, the expression of GUS activity wasonly used as a qualitative reporter of the location ofthe inoculated strain. Nevertheless, the lack of GUSactivity in NaN3 treated samples suggests that the A.diazotrophicus population in the sugarcane was lowerthan the limit of detection of the assay that we pre-viously determined in vitro (2U105 CFU per cm3 oftissue). That population-size indicator supported thedata that was obtained by MPN technique. This es-timation does not consider that X-Gluc di¡usioncould be limited by the plant cell walls. Anotherpossibility is that a low proportion of bacterial cellsretain the plasmid, but this does not seem to be thecase, since we found plasmid maintenance to behigher than 95% after 7 months inside the plant(not shown). The absence of a bacterial growth in-hibitor (NaN3) in the histochemical incubationsprobably allowed an increase in the A. diazotrophicuspopulation and the detection of GUS activity. Lowpopulation densities of this bacterium in sugarcaneare expected as there is no evidence of specializedplant structures which harbor high concentrationsof bacteria ([13^15], and this work). It is probablethat A. diazotrophicus is almost equally distributedinside several structures throughout the bulk of theplant, and that population growth is limited, forsome unknown reason, in the sucrose-rich tissues.Higher GUS activity was observed in stem xylemvessels, phloem sieve tubes and cortex, and by scan-ning microscopy A. diazotrophicus-like rods werefound in cells of the stem cortex, adhering to theinner cell wall. We presumed those cells to be A.diazotrophicus because of their physical similarity tocells grown in culture, and since they were foundonly in A. diazotrophicus inoculated plants andwere more abundant in GUS positive sections. Wedo not know if the plant cells seemingly colonized byA. diazotrophicus were damaged or alive. By immu-nogold labeling, James et al. [13] detected A. diazo-trophicus inside cells from the cortex of sugarcane
plantlet roots, and inside the xylem vessels fromthe base of the stem. They suggested that the rootxylem could be the route for stem and leaf xyleminfection. Our work supports previous results fromanother group [13,15] in that the cavities formed bythe xylem secondary wall are one of the preferen-tially colonized microhabitats. Dong et al. [14] haveproposed that A. diazotrophicus is found in the in-tercellular spaces of the stem storage parenchyma,where there are plentiful nutrients [30]. James et al.[13] also suggested that the stem xylem and the leafxylem might be the ¢nal colonized environments in-side the cane. In addition, we present evidence thatA. diazotrophicus could colonize the stalk corticaltissue as well as its xylem. From the xylem, the bac-terium might preferentially migrate to the cells ofselected tissues, such as the cortex. This ultimatedistribution may provide a more favorable environ-ment for the endophyte and its N2-¢xing activity,considering that the xylem apoplastic £uid is almostdevoid of carbon sources [31]. Some pathogenic andmildly pathogenic bacteria of sugarcane also colonizeand survive in the xylem elements, and from theretranslocate to other places [32,33]. In another work,Dong et al. [34] claimed that the xylem vessels werean improbable colonization site for A. diazotrophi-cus, since after introducing this bacterium insidethe stem, the plant reacted by producing substancesthat may have clogged the vessels. Nevertheless, theirexperiment probably did not re£ect the natural asso-ciation between A. diazotrophicus and sugarcane,since they made their observations on plants thatwere recently stressed by wounds. Moreover, asthey inoculated stems by submerging their cut endsfor several days in a growing bacterial suspension,with the consequent release of metabolic products,the defense reaction observed might have been ex-pected with any bacteria.
Dong et al. [34] asserted that the xylem vessels ofsugarcane were discontinuous, preventing the trans-port of A. diazotrophicus through the xylem. We pre-sume that even if the xylem vessels are limited intheir ability to translocate particulate material, A.diazotrophicus, and probably other species adaptedto this environment, could induce plant morpholog-ical changes, such as formation of continuous ves-sels, by releasing plant growth regulators. It has beenpreviously shown that A. diazotrophicus produces
auxins in a minimal culture medium [10]. The hy-pothesis of the role of A. diazotrophicus, is basedon the observation that during the course of xylemelement formation, the walls that separate adjacentvessel cells are hydrolyzed in a process that seems tobe controlled by the presence of auxins [35].
The promoter used for expression of gusA isknown to be active in eukaryotic tissues. Neverthe-less, a DNA fragment that includes a duplicatedCaMV promoter plus a leader sequence of AMV(alfalfa mosaic virus) was shown to induce high L-glucuronidase activity under di¡erent conditions inA. diazotrophicus. The bacterial recognition of eu-karyotic promoter sequences might not be entirelysurprising as it is known that some plant plastidpromoters share consensus sequences with 335310 bacterial promoters [36]. Moreover, at leastone eukaryotic transcription factor (TFIID) isknown to show high similarity with bacterial c-fac-tors [37]. Particularly, the most similar region be-tween TFIID and the c-factors has been suggestedto interact with DNA, binding to the eukaryoticTATA box in TFIID, or to single stranded DNAand to 310 bacterial promoters, in the bacterial fac-tors. In addition, the presence of a leader sequencecould enhance the translation of the gusA transcript,as has been observed with mRNA in di¡erent Gram-negative bacteria [38].
Under the conditions used here, sugarcane plantsup to 8 month of age showed no di¡erences in devel-opment when inoculated with A. diazotrophicus.From preliminary results in our laboratory, no nitro-genase expression was detected in planta from an A.diazotrophicus strain containing a nifH^gusA fusion,nevertheless we do not discard the possibility of ben-e¢cial e¡ects of the bacteria in plants grown underother conditions, as have been reported by Sevilla etal. [39]. Azoarcus sp., another endophytic diazotrophhas also been located inside root cortical cells ofKallar grass and inoculated rice [40]. In this associ-ation, the authors found some bene¢cial e¡ect onbiomass and protein content in rice plants inoculatedwith this bacterium.
Endophytic relationships are becoming an interest-ing ¢eld for studying plant-bacteria interactions andtheir study is still at an initial phase. Two threats tothe naturally occurring endophytic associations arethe high N-fertilization levels used in the modern
agriculture, and the now common use of tissue cul-ture to propagate pathogen-free sugarcane. Bothpractices will probably eliminate diazotrophic bacte-ria, as reported in this work, and sugarcane pro-ducers should be aware of this situation.
Acknowledgments
We are grateful to J. Esp|èritu, G. Guerrero and A.Leija for their help with micrographs and to E.K.James for advice on vegetal anatomy. We are alsograteful to R. Bustillos Cristales and M. A. Rogelfor technical help, J. Mart|ènez for statistical dataanalysis, D. Romero for his continuous and valuablecontributions, G. Carrillo, T. Laeremans, M. Pardoand P. Mavingui for their valuable comments, M.Dunn and M. Hynes for their critical reading ofthe manuscript and E. Cinta and M.L. Tavera fortheir library services. We acknowledge an anony-mous referee. L.E.F. was supported by ConsejoNacional de Ciencia y Tecnolog|èa (CONACyT).This work was in part supported by grants fromCONACyT (400343-5-1848N), UNAM-DGAPA(IN209496) and from PADEP-UNAM (030350 and030391).
References
[1] Wood, R.A. (1990) The roles of nitrogen, phosphorus andpotassium in the production of sugarcane in South Africa.Fert. Res. 26, 89^98.
[2] Ruschel, A.P. and Vose, P.B. (1984) Biological nitrogen ¢xa-tion in sugar cane. In: Current Developments in BiologicalNitrogen Fixation (Subba Rao, N., Ed.), pp. 221^235. Ed-ward Arnold, London.
[3] Ruschel, A.P. (1981) Associative N2-¢xation by sugar cane.In: Associative N2-Fixation, Vol. 2 (Vose, P.B. and Ruschel,A.P., Eds.), pp. 81^90. CRC Press, Baton Rouge.
[4] Urquiaga, S., Cruz, K.H.S. and Boddey, R.M. (1992) Contri-bution of nitrogen ¢xation to sugar cane: nitrogen-15 andnitrogen-balance estimates. Soil Sci. Soc. Am. J. 56, 105^114.
[5] Vose, P.B., Ruschel, A.P., Victoria, R.L. and Matsui, E.(1981) Potential N2-¢xation by sugarcane, Saccharum sp. insolution culture, I. E¡ect of NH4
nitrogen-¢xing bacterium associated with the sugarcane. PlantSoil 108, 23^31.
[7] Olivares, F.L., Baldani, V.L.D., Reis, V.M., Baldani, J.I. andDoëbereiner, J. (1996) Occurrence of the endophytic diazo-trophs Herbaspirillum spp. in roots, stems, and leaves, pre-dominantly of Gramineae. Biol. Fertil. Soils 21, 197^200.
[8] Gillis, M., Kersters, K., Hoste, B., Janssens, D., Kroppen-stedt, R.M., Stephan, M.P., Teixeira, K.R S., Doëbereiner, J.and de Ley, J. (1989) Acetobacter diazotrophicus sp. nov., anitrogen-¢xing acid bacterium associated with sugarcane. Int.J. Syst. Bact. 39, 361^364.
[9] Stephan, M.P., Oliveira, M., Teixeira, K.R.S., Mart|ènez-Drets, G. and Doëbereiner, J. (1991) Physiology and dinitrogen¢xation of Acetobacter diazotrophicus. FEMS Microbiol. Lett.7, 67^72.
[10] Fuentes-Ram|èrez, L.E., Jimeènez-Salgado, T., Abarca-Ocam-po, I.R. and Caballero-Mellado, J. (1993) Acetobacter diazo-trophicus, an indoleacetic acid producing bacterium isolatedfrom sugarcane cultivars of Meèxico. Plant Soil 154, 145^150.
[11] Muthukumarasamy, R. (1995) Endophytic nitrogen-¢xingbacteria from sugarcane in India. In: Proceedings of the In-ternational Symposium on Sustainable Agriculture for TheTropics: the Role of Biological Nitrogen Fixation, pp. 216^217. Rio de Janeiro, Brazil.
[12] Caballero-Mellado, J., Fuentes-Ram|èrez, L.E., Reis, V.M. andMart|ènez-Romero, E. (1995) Genetic structure of Acetobacterdiazotrophicus populations and identi¢cation of a new genet-ically distant group. Appl. Environ. Microbiol. 61, 3008^3013.
[13] James, E.K., Reis, V.M., Olivares, F.L., Baldani, J.I. andDoëbereiner, J. (1994) Infection of sugarcane by the nitrogen-¢xing bacterium Acetobacter diazotrophicus. J. Exp. Bot. 45,757^766.
[14] Dong, Z., Canny, M.J., McCully, M.E., Regla Reboredo, M.,Fernandez Cabadilla, C., Ortega, E. and Rodeès, R. (1994) Anitrogen-¢xing endophyte of sugarcane stems: a new role forthe apoplast. Plant Physiol. 105, 1139^1147.
[15] James, E.K. and Olivares, F.L. (1998) Infection and coloniza-tion of sugar cane and other graminaceous plants by endo-phytic diazotrophs. Crit. Rev. Plant Sci. 17, 77^119.
[16] Datla, R.S.S., Hammerlindl, J.K., Pelcher, L.E., Selvaraj, G.and Crosby, W.L. (1990) A bi-functional gene fusion betweenneomycin phosphotransferase and L-glucuronidase: a broadspectrum genetic marker for plants. J. Cell Biochem. Suppl.14E, 279 (Abstr.).
[17] Van den Eede, G., Deblaere, R., Goethals, K., Van Montagu,M. and Holsters, M. (1992) Broad host range and promoterselection vectors for bacteria that interact with plants. Mol.Plant-Microb. Interact. 5, 228^234.
[18] Caballero-Mellado, J. and Mart|ènez-Romero, E. (1994) Lim-ited genetic diversity in endophytic sugarcane bacterium Ace-tobacter diazotrophicus. Appl. Environ. Microbiol. 60, 1532^1537.
[20] Murashige, T. and Skoog, F. (1962) A revised medium forrapid growth and bioassays with tobacco tissue cultures. Phys-iol. Plant 15, 473^497.
[21] Je¡erson, R.A. and Wilson, K. (1988) The GUS gene fusionsystem. In: Plant Molecular Biology Manual (Gelvin, S. andSchilperoot, R., Eds.), pp. 1^22. Kluwer Academic, Dor-drecht.
[22] Toër, M., Mantell, S.C. and Ainsworth, C. (1992) Endophyticbacteria expressing L-glucuronidase cause false positives intransformation of Dioscorea species. Plant Cell Rep. 11,452^456.
[23] Kirchhof, G., Reis, V.M., Baldani, J.I., Eckert, B.,Doëbereiner, J. and Hartmann, A. (1997) Occurrence, physio-logical and molecular analysis of endophytic diazotrophic bac-teria in gramineous energy plants. Plant Soil 194, 45^55.
[24] Reis, V.M., Olivares, F.L. and Doëbereiner, J. (1994) Improvedmethodology for isolation of Acetobacter diazotrophicus andcon¢rmation of its endophytic habitat. World J. Microbiol.Biotechnol. 10, 401^405.
[25] da Silva, L.G., Reis V.M., dos Reis, F.B., Jr. and Boddey,R.M. (1995) Ontogenic variation of diazotrophic bacteria insugar cane (Saccharum spp. ) tissues. In: Proceedings of theInternational Symposium on Sustainable Agriculture for TheTropics: the Role of Biological Nitrogen Fixation, pp. 234^235. Rio de Janeiro, Brazil.
[26] Pelaez Abellan, I., de Armas Urquiza, R., Valadier, M.-H.and Champigny, M.-L. (1994) Short-term e¡ect of nitrate oncarbon metabolism of two sugar cane cultivars di¡ering intheir biomass production. Phytochemistry 36, 819^823.
[27] da Silveira, J.A.G. and Crocomo, O.J. (1991) Efeitos de nitro-geênio e potaèssio na utilizac°aìo de nitrato e distribuc°aìo de car-boidratos em cana-de-ac°uècar. Pesq. Agr. Bras. 26, 247^257.
[28] Doëbereiner, J., Reis, V.M. and Lazarini, A.C. (1988) New N2
¢xing bacteria in association with cereals and sugar cane. In:Nitrogen ¢xation: hundred years after (Bothe, H., de Bruijn,F.J. and Newton, W.E., Eds.), pp. 717^722. Gustav Fischer,Stuttgart.
[29] Li, R. and MacRae, I.C. (1992) Speci¢c identi¢cation andenumeration of Acetobacter diazotrophicus in sugarcane. SoilBiol. Biochem. 24, 413^419.
[30] Welbaum, G.E. and Meinzer, F.C. (1990) Compartmentationof solutes and water in developing sugarcane stalk tissue.Plant Physiol. 93, 1147^1153.
[31] Welbaum, G.E., Meinzer, F.C., Grayson, R.L. and Thorn-ham, K.T. (1992) Evidence for and consequences of a barrierto solute di¡usion between the apoplast and vascular bundlesin sugarcane stalk tissue. Aust. J. Plant Physiol. 19, 611^623.
[32] Kao., J. and Damann, K.E., Jr. (1980) In situ localization andmorphology of the bacterium associated with ratoon stuntingdisease of sugarcane. Can. J. Bot. 58, 310^315.
[33] Olivares, F.L., James, E.K., Baldani, J.I. and Doëbereiner, J.(1997) Infection of mottled stripe disease-susceptible and re-sistant sugar cane varieties by the endophytic diazotroph Her-baspirillum. New Phytol. 135, 723^737.
[34] Dong, Z., McCully, M.E. and Canny, M.J. (1997) Does Ace-tobacter diazotrophicus live and move in the xylem of sugar-cane stems? Anatomical and physiological data. Ann. Bot. 80,147^158.
[35] Esau, K. (1977) Anatomy of Seed Plants, 2nd edn., 550 pp.John Wiley and Sons, New York.
[36] Cornelissen, M. and Vandewiele, M. (1989) Nuclear transcrip-tional activity of the tobacco plastid psbA promoter. NucleicAcid Res. 17, 19^29.
[37] Horikoshi, M., Wang, C.K., Fujii, H., Cromlish, J.A., Weil,P.A. and Roeder, R.G. (1989) Cloning and structure ofa yeast gene encoding a general transcription initiationfactor TFIID that binds to the TATA box. Nature 341,209^303.
[38] Gallie, D.R. and Kado, C.I. (1989) A translational enhancerderived from tobacco mosaic virus is functionally equivalentto a Shine-Dalgarno sequence. Proc. Natl. Acad. Sci. USA 86,129^132.
[39] Sevilla, M., Lee, S., Brockschnieder, D., de Olivera, A., Bal-dani, I. and Kennedy, C. (1998) Studies on Acetobacter diazo-trophicus : analysis of nif and related genes and contributions
to sugarcane nutrition. In: Biological Nitrogen Fixation forthe 21st Century (Elmerich, C., Kondorosi, A. and Newton,E., Eds.), pp. 383^384. Kluwer Academic, Dordrecht.
[40] Hurek, T., Reinhold-Hurek, B., Van Montagu, M. and Kel-lenberger, E. (1994) Root colonization and systemic spreadingof Azoarcus sp. strain BH72 in grasses. J. Bacteriol. 176,1913^1923.
[41] Bien, M., Ste¡en, H. and Schulte-Frohlinde, D. (1988) Repairof plasmid pBR322 damaged by Q-irradiation or by restrictionendonucleases using di¡erent recombination pro¢cient E. colistrains. Mutat. Res. 194, 193^205.
[42] Figurski, D.H. and Helinski, D.R. (1979) Replication of anorigin-containing derivative of plasmid RK2 dependent on aplasmid function provided in trans. Proc. Natl. Acad. Sci.USA 73, 1648^1652.
