ULTRASTRUCTURE AND CYTOCHEMISTRY OF IN VITRO …means of light and transmission electron microscopy with gold cytochemistry. MATERIALS AND METHODS Fungal and bacterial isolates and

SUMMARY

The present studies were undertaken to investigatethe interaction of the bacterial antagonists Bacilluscereus X16 and B. thuringiensis 55T with Fusa-rium roseum var. sambucinum, the causal agent of pota-to dry rot. On wounded potato tubers, both bacilli ef-fectively suppressed the development of Fusarium dryrot. Nevertheless, confrontation of the fungal pathogenwith the antagonists on nutrient agar revealed that B.cereus X16, induced a strong visible inhibition zonebut B. thuringiensis 55T did not. Light microscopy ofthe interaction regions after confrontation with B.cereus X16 on nutrient agar generally showed appar-ently intact fungal cells with densely stained proto-plasm, indicating that the inhibition was due to stasisrather than toxicity. By contrast, in presence of B.thuringiensis, Fusarium cells appeared markedly dam-aged, with partial to complete cell wall disintegrationand disorganization and generally complete loss of pro-toplasm. These observations suggest that, contrary toour expectations, antibiotic and chitinolytic activitiesof B. thuringiensis 55T may be highly significant in theparasitism of the pathogen. Confrontation in liquidmedium revealed that both bacteria completely inhibit-ed the germination of Fusarium macroconidia and re-sulted in their destruction. This suggests that, to pro-duce their fungitoxic and hydrolytic effects, the twobacteria should make intimate contact with Fusariumcells.

Key words: Bacillus, biocontrol, Fusarium dry rot,microscopy, cytochemistry.

Corresponding author: M. ChérifFax: 216.-1.-799 391E-mail: [email protected]

Abbreviations: B = bacterium; CY = cytoplasm;F = fungus; N = nucleus; NA = nutrient agar; S = sep-tum; TEM = transmission electron microscopy;VA = vacuole; W = wall; WA = wall apposition;WT = wall thickenings.

INTRODUCTION

Fusarium spp. are responsible for dry rot of potatotubers and important crop losses under traditionaland cold storage conditions (Tivoli et al., 1985; Daa-mi-Remadi and El Mahjoub, 1996; Carnegie et al.,1998; Chérif et al., 2000). Of the different Fusariumspecies studied, F. roseum var. sambucinum appears tobe the most dominant and the most destructive fungusprevailing on stored potatoes (Boyd, 1972; Schisler etal., 1997; Chérif et al., 2000). This pathogen is com-monly reported to be responsible for losses higherthan 25%, especially under traditional storage, whereenvironmental conditions are particularly conduciveto dry rot development (Daami-Remadi and ElMahjoub, 1996; Chérif et al., 2000). All commonlygrown potato cultivars are susceptible to Fusarium dryrot (Pawlak et al., 1987; Schisler et al., 1997), andmanagement of the disease with chemical measureshas proved impractical, mainly due to the appearanceof fungicide-resistant strains, concern over the pres-ence of chemical residues in the food chain, and prob-lems of environmental pollution (Kawchak et al.,1994; Secor et al., 1994).

Among alternatives being explored, use of microbialagents has shown significant potential (Benhamou etal., 1997; El-Ghaouth et al., 1998). Currently differentspecies of microorganism are registered for commercialuse against fungal plant pathogens in Europe and inthe United States. These include fungi (Gliocladiumvirens G.21; Trichoderma harzianum KRL-AG2; Candi-

Journal of Plant Pathology (2002), 84 (2), 83-93 Edizioni ETS Pisa, 2002 83

ULTRASTRUCTURE AND CYTOCHEMISTRY OF IN VITRO INTERACTIONSOF THE ANTAGONISTIC BACTERIA BACILLUS CEREUS X16

AND B. THURINGIENSIS 55T WITH FUSARIUM ROSEUM VAR. SAMBUCINUMM. Chérif1, N. Sadfi1, N. Benhamou2, A. Boudabbous3, A. Boubaker1, M.R. Hajlaoui4 and Y. Tirilly5

1 Laboratore de Phytopathologie, Institut National Agronomique de Tunisie, 43 Avenue Charles Nicolle,1082 Cité Mahrajène, Tunis, Tunisie

2 Recherches en Sciences de la Vie et de la Santé, Université Laval, Pav. Charles-Eugène Marchand, Quebec G1k 7P4, Canada.3 Laboratore de Microbiologie, faculté des Sciences de Tunis, 1060 Campus Universitaire, Tunisie

4 Laboratoire de Cryptogamie-Bactériologie, Institut National de la Recherche Agronomique de Tunisie, Rue Hédi Karray,2049 Ariana, Tunisie

5 Ecole Supérieure de Microbiologie et Sécurité Alimentaire de Brest, Technopôle Brest-Iroise, 29 280 Plouzané, France

da oleophila 182), gram-negative bacteria(Pseudomonas fluorescens EG1053, P. syringea ESC-10and ESC-11, Burkolderia cepacia) and gram-positivebacteria such as Bacillus subtilis GB03 and MBI 600(Kim et al., 1997; El-Ghaouth et al., 1998). Bacillusspecies, as a group offer several advantages over othergram-negative bacteria, including longer shelf life be-cause of their ability to form endospores and thebroad-spectrum activity of their antibiotics (Fiddmanand Rossall, 1995; Kim et al., 1997). Bacillus speciesare generally soil-inhabitating or exist as epiphytes andendophytes in the spermosphere and rhizosphere(Walker et al., 1998). They are also found in many en-vironments as their survival is aided by their ability toform endospores resistant to UV irradiation, dessica-tion, heat and organic solvents. Some Bacillus species,such as B. thuringiensis and B. sphaericus, are ento-mopathogenic as they produce toxins effective againstthe larvae of a wide range of insects. The success of B.thuringiensis as an insecticide has promoted researchand development for additional Bacillus-based prod-ucts. There is a growing list of reports of postharvestfungal disease control with Bacillus species, includingB. subtilis for control of green mold, sour rot, and stemend rot on citrus (Singh and Deverall, 1984), brownrot on stone fruit (Utkhede and Sholberg, 1986), andseveral diseases of apple (Sholberg et al., 1995) andavocado (Korsten et al., 1997).

Recently, 83 Bacillus isolates, obtained from saltysoils, and five B. thuringiensis strains were screened inour laboratory for ability to suppress growth of F. ro-seum var. sambucinum in vitro and to control dry roton potato tubers (Sadfi et al., 2001). Among these iso-lates, strains X16 of B. cereus and 55T of B.thuringiensis showed high levels of control in vivo.Nevertheless, while B. cereus inhibited the growth ofthe fungal pathogen in vitro by forming inhibitionzones in dual cultures, B. thuringiensis (55T) failed toproduce such inhibition zones and seemed to be inef-fective on agar plates. According to the literature,Bacillus spp. may cause their antagonistic effectsagainst fungal pathogens by antibiosis, nutrient com-petition, site exclusion, parasitism and/or induced re-sistance (Kehlenbeck et al., 1994; Muninbazi andBullerman, 1998; Walker et al., 1998). The currentstudy was undertaken to investigate more closely theinteraction of B. cereus (X16) and B. thuringiensis(55T) with F. roseum var. sambucinum in vitro bymeans of light and transmission electron microscopywith gold cytochemistry.

MATERIALS AND METHODS

Fungal and bacterial isolates and growth conditions.The isolate of F. roseum var. sambucinum used was ob-tained from infected potato tubers with typical symp-toms of Fusarium dry rot and reported to be virulent onpotato tubers (Chérif et al., 2000). The fungal pathogenwas grown on potato-dextrose agar (PDA) medium at25°C.

Isolate X16 of B. cereus was selected among a collec-tion of 83 Bacillus spp. isolated from samples of saltysoils collected from different locations in the south ofTunisia (Sadfi et al., 2001). Strain 55T of B. thuringien-sis was selected among six strains kindly provided byDr. A. Boudabbous from the laboratory of Microbiolo-gy of the Faculté des Sciences de Tunis. Bacterial iso-lates were maintained on slants of nutrient agar (NA,Oxoid) at 4°C and subcultured at two-month intervals.

Antagonistic activity on solid medium. Two-day-oldcultures of B. cereus X16 and B. thuringiensis 55T werestudied in their interaction with F. roseum var. sam-bucinum on NA in 9-cm Petri plates using a dual cul-ture technique. Bacteria were streaked across the centerof each agar plate with a loopful of pure bacterial cul-ture. Two 5-mm disks cut from a 7-day-old culture of F.roseum var. sambucinum, were placed 2.5 cm apart oneach side of the bacteria. The plates were then incubat-ed at 25°C. Mycelial samples were collected from theinteraction region 2 to 6 days after inoculation andprocessed for light and transmission electron mi-croscopy (TEM).

Antagonistic activity in liquid medium. These testswere performed in 250 ml-Erlenmeyer flasks containing100 ml of sterile nutrient broth (NB). Fungal spore sus-pensions and bacterial suspensions were obtained byflooding 10-day-old PDA cultures of F. roseum var.sambucinum and 2-day-old NA cultures of Bacillus spp.,respectively with sterile distilled water. The protago-nists were added to the flasks containing NB and ad-justed to obtain a final concentration of 106 CFU ml-1 ofthe bacteria and 105 spores ml-1 of the fungus. Bacterialand fungal concentrations were determined, respective-ly by dilution plating and counting with a heamocy-tometer. The flasks were then incubated at 25°C and100 rev min-1 on a rotary shaker. Samples were collect-ed from 2 to 6 days after incubation. Fungal and bacter-ial cells were pelleted by centrifugation at 7000 g for 15min. The pellets were embedded in 2% water agar andprocessed for light microscopy and TEM.

In order to determine the effect of the bacterial an-tagonists on the fungal mycelium, the same procedure

84 Potato dry rot: antagonism of Bacillus spp. in vitro Journal of Plant Pathology (2002), 84 (2), 83-93

was adopted with the only exception that bacteria wereadded to the liquid medium 5 days after inoculationwith fungal macroconidia. By that time, fungal sporesgerminated and developed a dense mycelium.

Antagonistic activity on potato tubers. Potato tuberscv. ‘Spunta’ were surface-sterilized by soaking in 2%aqueous sodium hypochlorite for 10 min. They werethen thoroughly rinsed, dried by using sterile filter pa-pers, and then wounded by removing a plug 3 mm indiameter and 3 mm in depth with a sterile cork-borer.B. cereus X16 and B. thuringiensis 55T isolates wereused after, respectively, 24 h and 48 h of culturing at106 CFU ml-1, and bacterized potato wounds received20 µl of bacterial suspension. Potato wounds were chal-lenged with the pathogen immediately after bacterial in-oculation and received 20 µl of a conidial suspensionadjusted to 105 spores ml-1. The treated wounds weresealed with scotch tape and potato tubers were placedin plastic bags to maintain a high humidity and then in-cubated at 20°C for one week. For each treatment 20potato tubers were assayed and the experiment was re-peated at least twice.

Light microscopy and TEM. Before fixation andprocessing for TEM, samples were always examined bylight microscopy and photographed with a Zeiss Ax-ioskop microscope (Carl Zeiss, Thorwood, NY). Sam-ples collected from the interaction regions of solid dualcultures, from liquid cultures, as well as samples frompure cultures of the fungus and each bacterium, werefixed in 2% (v/v) glutaraldehyde in 0.1 M cacodylatebuffer, pH 7.2, for 2 h at room temperature andovernight at 4°C. Samples were rinsed in three changesof cacodylate buffer and post-fixed in 1% (w/v) osmi-um tetroxide in the same buffer. They were then dehy-drated in a graded series of ethanol and embedded inEpon 812 resin. For light microscopy, thin sections(0.25 to 0.5 µm) were collected on microscope slidesand stained with ethylene blue. For TEM, ultrathin sec-tions were collected on 200-mesh nickel grids coatedwith Formvar and stained with uranyl acetate and leadcitrate. Grids were examined with a JEOL 1200 EXtransmission microscope (JEOL, Tokyo) operating at80 kV.

For each treatment, samples were collected from atleast four replicates to obtain a representative samplingof the interactions. For each sample, more than 10 thinor ultrathin sections were examined by light mi-croscopy and TEM, respectively.

Cytochemical labeling. Colloidal gold suspensionwas prepared as described by Grandmaison et al.

(1988). A lectin with N-acetylglucosamine bindingspecificity, was used for localizing fungal chitin accord-ing to a previously described procedure (Benhamou,1989; Chérif and Benhamou, 1990). Because of its lowmolecular weight, this lectin could not be directly com-plexed to colloidal gold. It was used in a two-step pro-cedure, using ovomucoid as a second step reagent.Ovomucoid was conjugated to gold at pH 5.4.

Sections were first incubated on a drop of WGA [25µg ml-1 in phosphate- buffered saline (PBS), pH 7.4]for 60 min at room temperature, rinsed with PBS, andtransferred to a drop of ovomucoid gold-complex for30 min at room temperature. After washing with PBSand rinsing with double-distilled water, sections werecontrasted with uranyl acetate and lead citrate, and ex-amined by TEM.

Specificity of labeling was assessed by the followingcontrols: (i) incubation with WGA to which was previ-ously added N,N’,N’’ triacetyl-chitotriose (1 mg ml-1 inPBS); (ii) incubation with WGA, followed by unlabeledovomucoid and finally by ovomucoid gold-complex;and (iii) direct incubation with the ovomucoid gold-complex, the lectin step being omitted.

Reagents. Tetrachloroauric acid was purchased fromBDH chemicals, Montreal. All other reagents for elec-tron microscopy were obtained from JBEM company,Pointe-Claire, QB, Canada.

RESULTS

Macroscopic observations. Coinoculation of F. rose-um var. sambucinum with bacterial antagonists on NArevealed that B. thuringiensis 55T was apparently unableto inhibit growth of the fungus after 5 days of incubation(Fig. 1A). By that time, however, the fungus was stronglyinhibited by B. cereus X16, with the appearance of inhi-bition zones (Fig. 1B). While the first contact between B.thuringiensis 55T and the pathogen was observed by 3days after coinoculation, intimate contact between F. ro-seum var. sambucinum and B. cereus X16 was never ob-served even after many weeks of incubation.

Potato wounds inoculated with F. roseum var. sam-bucinum alone started to show typical dry rot symptomsby the third day of incubation at 20°C. By the seventhday, all potato tubers infected with the pathogen werediseased and showed lesions of more than 20 mm in di-ameter (Fig. 1C).

Application of both Bacillus isolates to potatowounds before challenge with the pathogen, significant-ly reduced development of dry rot (Fig. 1D and E). Tu-bers treated with B. cereus X16 and inoculated with the

Journal of Plant Pathology (2002), 84 (2), 83-93 Chérif et al. 85

pathogen showed no symptoms of infection after 7 daysat 20°C (Fig. 1D). Although some small, brownish le-sions could be seen in potatoes bacterized with B.thuringiensis 55T and inoculated with the pathogen af-ter 7 days of incubation (Fig. 1E), their frequency anddiameter never reached those in the infected controls.Tubers inoculated with either B. cereus X16 (Fig. 1F) orB. thuringiensis 55T alone were free of disease symp-toms.

Light microscopy observations. Mycelial samplesfrom pure cultures of F. roseum var. sambucinum re-vealed densely stained hyphal cells (Fig. 2A). After 3days of confrontation, examination of sections from theedge of Fusarium colonies in contact with the inhibitionzones caused by B. cereus X16 on NA generally showedapparently intact fungal cells with densely stained pro-toplasm (Fig. 2B). A relatively low percentage of Fusa-rium cells showed signs of cell damage, even after amuch longer period of exposure to B. cereus X16. Bycontrast, in presence of B. thuringiensis 55T, where

contact between the two protagonists was macroscopi-cally observed, Fusarium cells appeared markedly dam-aged, as evidenced by disorganization of the cytoplasmand generally complete loss of protoplasm (Fig. 2C, D).

In pure liquid cultures (NB), fungal macroconidiagerminated after 24 h of incubation (Fig. 3A) and gavea very dense mycelium with numerous newly formedmacroconidia 5 days later (Fig. 3B). Coinoculation withfungal spores and either B. cereus X16 or B. thuringien-sis 55T completely inhibited the germination of macro-conidia after 24h (Fig. 3C, and D) and 6 days (Fig. 3E-G) of incubation. By the end of the experiment, theBacillus spp. had abundantly colonized the medium andfungal macroconidia, resulting in destruction of a highpercentage of macroconidia, which appeared brightlystained (Fig. 3E-G). In some instances, bacterial cells ofB. thuringiensis 55T seemed to be inside macroconidia(Fig. 3G, arrow). This was never observed with B.cereus X16. Nevertheless, this bacterial isolate causedsevere damage to the macroconidia in liquid medium(Fig. 3F).


Fig. 1. Effect of the bacterial antagonistsB. thuringiensis 55T and B. cereus X16on (A ,B) mycelial growth of thepathogen F. roseum var. sambucinum indual cultures, after 5 days of incubationon nutrient agar medium, and on (C-F)dry rot development on wounded potatotubers cv. ‘Spunta’, after 7 days of inoc-ulation. (A) B. thuringiensis 55T and thefungal pathogen in dual culture. (B) B.cereus X16 and the fungal pathogen indual culture. (C) Control potato tuberinoculated with the pathogen F. roseum.var. sambucinum alone. (D) Coinocula-tion with B. cereus X16. (E) Coinocula-tion with B. thuringiensis 55T. (F) Inoc-ulation with B. cereus X16 alone.

Ultrastructural and cytochemical observations

Interaction of B. cereus X16 with the pathogen. Thetime course study of fungal growth in dual cultures inpresence of B. cereus X16 revealed that colonization ofthe NA medium by Fusarium ceased 48-72h after con-frontation. Up to that time, the majority of Fusarium hy-phae were free of any sign of alteration, as judged by theintegrity of the cytoplasm and cell wall (Fig. 4A). From 3to 6 days after inoculation, some hyphae showed variousdegrees of cytoplasm disintegration, leading in some cas-es to complete depletion of their cytoplasm (Fig. 4B).Nevertheless, hyphae showing such severe alterationwere very rarely observed and most cells of F. roseumvar. sambucinum displayed apparently well-preserved or-ganelles and cytoplasm. Labeling with the WGA/ovo-mucoid-gold complex revealed that even hyphae withadvanced cytoplasm alteration preserved the integrity oftheir cell walls, which were intensely and regularly la-beled with gold particles (Fig. 4C). In many instances,Fusarium cells developed cell wall thickenings, which

were also intensively labeled (Fig. 4D). By contrast,coinoculation of Fusarium and B. cereus X16 in liquidmedium (NB), resulted in much more damage to fungalcells (Fig. 4E, F). Under these conditions, the contactbetween the pathogen and B. cereus X16 was more inti-mate and direct (Fig. 4E). Six days after inoculation,most fungal cells appeared severely damaged and themain characteristic of these cells was partial to completewall disintegration associated with depletion of cyto-plasm contents (Fig. 4E). This observation was con-firmed by labeling with the WGA/ovomucoid-gold com-plex, which showed that some fungal cells were reducedto traces that could only be identified by the presence ofgold particles in remaining wall debris (Fig. 4F).

Interaction of B. thuringiensis 55T with the pathogen.Mycelial samples from the edge of growing colonies ofF. roseum var. sambucinum examined two days aftercoinoculation with B. thuringiensis 55T on NA, re-vealed generally dense hyphal cells with apparently pre-served cell walls, nucleus and organelles (Fig. 5A).


Fig. 2. Light microscope observations of F.roseum var. sambucinum cells from pure cul-tures (A), from the edge of Fusarium coloniesin contact with the inhibition zone caused byB. cereus X16 (B), or from fungal colonies incontact with B. thuringiensis 55T (C, D), after3 days of incubation on nutrient agar plates.F = Fungus.

Alterations in wall labeling with the WGA/ovomucoid-gold complex started to appear by 3 days after inocula-tion, at which time the two protagonists were very closeto each other (Fig. 5B). These wall-labeling alterationswere always associated with abundant presence of goldparticles in NA medium (Fig. 5B, arrows). At this stage,fungal nuclei and vacuoles often showed obvious signsof malformation, as exemplified by contortion of theirmembranes (Fig. 5B, and C).

In many instances, fungal cells responded to B.thuringiensis 55T attack by the accumulation of cellwall appositions and thickenings (Fig. 5D). Such appo-sitions were intensely labeled with the WGA/ovomu-coid-gold complex indicating high accumulation ofchitin at these sites (Fig. 5D). As early as 3 days after in-oculation, significant alterations of the fungal cytoplasmwere readily discernable, such as retraction, aggregationand organelle disintegration (Fig. 5C). By that time,fungal cell walls generally showed good structural

preservation. Past this stage, from 4 to 6 days after inoc-ulation, most Fusarium cells became affected (Fig. 6A),with pronounced damage, as judged by total loss ofprotoplasm (Fig. 6A-C) and partial or complete walldisintegration (Fig. 6B), leading to empty hyphal shellswith many holes, corresponding to locally digested ar-eas in the cell walls (Fig. 6C1, and C2). Such cell walldegradation was generally accompanied by the releaseof gold particles, which were scattered over the sur-rounding agar medium (Fig. 6C2)

Examination of ultrathin sections from samples ofFusarium macroconidia and mycelium confronted withB. thuringiensis 55T in liquid medium (NB) revealedthe same events of fungal cytoplasm disintegration andcell wall degradation (Fig. 6D), with the exception thatalterations started to appear by 24 h after inoculation,thus earlier than for solid cultures, where the first con-tact between the two protagonists was not observed be-fore 3 days of incubation.


Fig. 3. Light microscope observations of F.roseum var. sambucinum cells after incubationof macroconidia in liquid medium (nutrientbroth) in pure cultures (A, B) or in presence ofthe bacterial antagonists B. cereus X16 (C, E,F) and B. thuringiensis 55T (D, G). (A) Pureculture Fusarium after 24 h of incubation. (B)Pure culture of Fusarium after 7 days of incu-bation. (C) Coinoculation with B. cereus X16after 24h of incubation. (D) Coinoculationwith B. thuringiensis 55T after 24h of incuba-tion. (E, F) Coinoculation with B. cereus X16after 6 days of incubation. (G) Coinoculationwith B. thuringiensis 55T after 6 days of incu-bation. A and C-F: Bar = 20 µm; B: Bar = 60µm; G: Bar = 10 µm.

Specificity of labeling with the WGA/ovomucoid-gold complex was assessed by the negative results ob-tained with all control tests including the previous ad-sorption of the WGA with N,N’,N’’-triacetylchitotriose(data not shown).

DISCUSSION

Our results based on ultrastructural observationsand cytochemical localization of N-acetylglucosamineresidues showed that contrary to our expectations, an-tibiotic and chitinolytic activities of B. thuringiensis 55Tmay be highly significant in the parasitism of F. roseumvar. sambucinum. In fact, although in vitro antagonismtests performed on NA showed no apparent inhibitionof Fusarium growth in presence of B. thuringiensis 55T,microscopic observations revealed that this antagonistwas very destructive to fungal hyphae, on both solid

and liquid media. Moreover, although apparently inef-fective in vitro on NA, B. thuringiensis 55T effectivelyinhibited dry rot in vivo on wounded tubers. This con-firms the idea that Bacillus isolates unable to form inhi-bition zones on solid medium are not necessarily inca-pable of killing the pathogen in vitro and inhibiting dis-ease development in vivo.

Confrontation of B. cereus X16 with the pathogen onsolid medium (NA) demonstrated that Fusarium hy-phae appeared generally intact with a very low percent-age of damaged cells, indicating that the inhibition zoneobserved exerted a fungistatic rather than a fungitoxicrole. These conclusions were confirmed by ultrastruc-tural observations, which generally revealed fungal hy-phae with preserved cytoplasm, organelles and cellwalls. Even the few pathogen cells showing damagedprotoplasm preserved the integrity of their cell walls asevidenced by the intense and regular labeling for chitin.By contrast, confrontation of the pathogen with B.


Fig. 4. Transmission electron micrographs ofF. roseum var. sambucinum cells grown on nu-trient agar (A-D) or in liquid medium (nutri-ent broth, E, F) in presence of the antagonisticbacterium Bacillus cereus X16. (A) After 48 hof incubation. (B) After 3 days of incubation.(C) Labeling with the WGA/ ovomucoid-goldcomplex of fungal hyphae incubated for 6days in presence of the antagonist. (D) Fungalcell with a cell wall thickening, intensively la-beled WGA/ ovomucoid-gold complex. (E) Inliquid medium, bacterial cells are in close con-tact with the pathogen by the sixth day ofincubation. (F) Fusarium hyphae showinghighly degraded Cell walls after labelingwith the WGA/ovomucoid-gold complex.Bar = 1 µm. B = Bacterium; CY = Cytoplasm;F = Fungus; S= Septum; W = Wall; WT =Wall thickenings.

thuringiensis 55T resulted in serious damage to a highpercentage of hyphae, associated with a series of degra-dation events, including alteration of chitin macromole-cules and visible cell wall disruption, retraction of theplasmalemma, distortion of the nucleus, generalizeddisorganization of the cytoplasm, and, ultimately, com-plete loss of protoplasm. Fungal cell wall disintegrationand cytoplasm disorganization were observed as soon as3 days after confrontation, just before first contact be-tween the two protagonists was established. This sug-gests that fungitoxic compounds and hydrolytic en-zymes produced by B. thuringiensis 55T diffuse a shortdistance into the agar to cause the observed distur-bances in advance of physical contact.

Cell wall appositions and thickenings in Fusariumcells were observed in presence of both bacteria. Suchdeposits contained large amounts of chitin and aresimilar to wall appositions observed in the cells of dif-ferent fungal pathogens submitted to the activity of an-tagonistic fungi (Benyagoub et al., 1998; Benhamou etal., 1999) and to treatment with fungicides (Robertsonand Fuller, 1990), chitosan (Benhamou, 1992) or 2-de-oxy-D-glucose (El-Ghaouth et al., 1997). The mecha-nisms that control the process of chitin deposition at

such sites and the exact biological function played bythese deposits are still unclear. Nevertheless, from theliterature the most accepted explanations indicate thatthe deposited material may be laid down as newly syn-thesized molecules via deregulation of fungal mem-brane-bound enzymes involved in the synthesis ofstructural compounds (Benhamou et al., 1999). Ac-cordingly, the massive accumulation of these structuralcompounds as abnormal wall-like deposits reflects adefense strategy elaborated by the fungal pathogen forpreventing penetration of mycoparasites and fungitox-ic compounds. Whatever the role played by the de-posited material, our Bacillus antagonists, when in di-rect contact with the pathogen, i.e. in liquid medium,were able to circumvent such barriers and cause severedamage to fungal cells, which were rapidly reduced toempty shells.

The differences observed between B. cereus X16 andB. thuringiensis 55T on NA, relative to the percentageof fungal cells affected, cell wall degradation and theextent of cytoplasm disintegration, were not observedin liquid medium. In fact, B. cereus X16 was as effectiveas B. thuringiensis 55T and caused extensive cell walldisruption and cytoplasm disorganization of the


Fig. 5 A-D. Transmission electron micro-graphs of F. roseum var. sambucinum hyphaegrown on nutrient agar medium in presence ofthe antagonistic bacterium B. thuringiensis55T. Fusarium cells after 2 days (A) and 3 days(C) of incubation. (B, D) Fusarium cells after 3days of incubation labeled with WGA/ ovo-mucoid-gold complex. In D, a fungal cell re-sponding to bacterial attack by the formationof a cell wall apposition. Bar = 0.5 µm. CY =Cytoplasm; N = nucleus; VA = vacuole; W =Wall; WA = Wall apposition.

pathogen. This suggests that in order to exert its fungi-toxic and hydrolytic effects, B. cereus X16 must be inintimate contact with the Fusarium cells. The absenceof inhibition zones on solid medium after application ofB. thuringiensis 55T may be explained by the fact thatthe fungal pathogen can counteract fungistatic effectsand therefore reach intimate but deadly interactionwith the bacterium. In contrast, fungistatic compoundsproduced by B. cereus X16 may halt fungal spread butavoid the lethal effects of close contact.

When intimate contact between the protagonists isachieved, fungal cell wall degradation by hydrolytic en-zymes produced by both antagonists seems to play amajor role in the outcome of their interaction with thepathogen. In a recent study, we have reported that B.cereus X16 and B. thuringiensis 55T exhibited strongchitinolytic activity as determined by the formation ofclearing zones on chitin agar, the release of reducingsugars from colloidal chitin and by the release of p-ni-trophenol (pNP) from dimeric, trimeric and tetramericchromogenic chito-oligosaccharides, thus showing that

these antagonists are able to produce N-acetyl-β-D-glucosaminidases, chitobiosidases and endochitinas-es (Sadfi et al., 2001). Our TEM observations revealedthat while the alteration of chitin first occurred locally,as illustrated by the release of N-acetylglucosamineresidues in the growing medium, it appeared to bemore generalized later on and at more advanced stagesof the interaction of the pathogen with the antagonisticbacilli. Fusarium cell walls were completely digested.From these observations, it can be speculated that thesynergistic and coordinated action of chitinases withother polysaccharidases, such as β-1,3-glucanases, li-pases and proteases may be an important determinantin the antagonistic process (Sivan and Chet, 1989;Chérif and Benhamou, 1990; Benhamou and Chet,1996).

Confrontation of conidia and hyphae of F. roseumvar. sambucinum to the bacilli generally resulted in se-vere alteration of their protoplasm. Such alteration maydirectly correlate with the toxic action of antifungal sub-stances (i.e. antibiotics) produced by the antagonists. In


Fig. 6. Transmission electron micrographs ofF. roseum var. sambucinum hyphae grown onnutrient agar (A-C) or in liquid medium (D) inpresence of the antagonistic bacterium B.thuringiensis 55T. (A, B) Fusarium hyphaeshowing pronounced damage after 4 days ofincubation. (C1, C2) Fusarium hyphae labeledwith the WGA/ovomucoid-gold complex.Note that cell wall degradation is accompaniedwith the release of gold particles (C2, a highermagnification of C1, arrow). (D) Confronta-tion of the pathogen with B. thuringiensis 55Tin liquid medium. Bar = 1 µm. B = Bacterium;F = Fungus; W = Wall.

fact, fungal protoplasmic alterations have been reportedin other antagonistic interactions involving antibiosis asthe main mechanism of action (Hajlaoui et al., 1993;Bélanger et al., 1995). For instance, Benyagoub et al.(1996), showed that antibiotics from Sporothrix floccu-losa caused protoplasm alteration and depletion in treat-ed pathogenic fungi by promoting modification of thelipid composition of the plasmalemma. This raises thequestion as to what extent weakening of the cell wall ofpathogenic fungi through the action of hydrolytic en-zymes of the antagonist may facilitate the diffusion oftoxic substances toward membrane receptors by in-creasing the wall permeability. A growing body of evi-dence indicates that the synergistic action of wall hydro-lases, especially chitinases, and antibiotics is required inthe antagonistic process and that alteration of host cellwalls are essential prerequisites for further antibioticdiffusion (Di Pietro et al., 1993). These findings are inline with our observations, which have demonstrated ahigh percentage of affected fungal cells with severely al-tered cytoplasm in liquid medium, where close contactbetween the fungus and B. cereus X16 was well estab-lished and cell wall alterations of the pathogen were evi-dent. Such results were not observed on solid medium,where the fungal cell walls appeared intact, and chitinbreakdown was not detected by cytochemical labeling.

ACKNOWLEDGEMENTS

This research was supported by funds from the In-ternational Foundation for Science (C/2600-2), AU-PELF-UREF (Fond International de Coopération Uni-versitaire 2000/PAS/44) and IRESA, Tunisia (ProjetFédérateur Pomme de Terre). We thank Alain Gouletfor his technical assistance.

REFERENCES

Bélanger R.R., Dufour N., Caron J., Benhamou N., 1995.Chronological events associated with the antagonisticproperties of Botrytis cinerea: Indirect evidence forsepuential role of antibiosis and parasitism. Biocontrol Sci-ence and Technology 5: 41-53.

Benhamou N., 1989. Preparation and application of lectin-gold complexes. In: Hayat M.A. (ed.). Colloidal gold: prin-ciples, methods and applications, Vol. 1, pp. 95-143. Aca-demic Press, New York, USA.

Benhamou N., 1992. Ultrastructural and cytochemical aspectsof chitosan on Fusarium oxysporum f.sp. radicis-lycopersici,agent of tomato crown and root rot. Phytopathology 82:1185-1193.

Benhamou N., Chet I., 1996. Parasitism of sclerotia of Scle-rotium rolfsii by Trichoderma harzianum: ultrastructuraland cytochemical aspects of the interaction. Phytopatholo-gy 86: 405- 416.

Benhamou N., Rey P., Picard K., Tirilly Y., 1999. Ultrastruc-tural and cytochemical aspects of the interaction betweenthe mycoparasite Pythium oligandrum and soilborne plantpathogens. Phytopathology 89: 506-517.

Benhamou N., Rey P., Chérif M., Hockenhull J., Tirilly Y.,1997. Treatment with the mycoparasite, Pythium oligan-drum, triggers the induction of defense-related reactions intomato roots upon challenge with Fusarium oxysporumf.sp. radicis-lycopersici. Phytopathology 87: 108-122.

Benyagoub M., Benhamou N., Carisse O., 1998. Cytologicalaspects of the antagonistic interaction between Mi-crosphaeropsis sp. (isolate P130A) and Venturia inaepealis.Phytopathology 88: 605-613.

Benyagoub M., Willemot C., Bélanger R.R., 1996. Influenceof a subinhibitory dose of antifungal fatty acids fromSporothrix floculosa on cellular lipid composition in fungi.Lipids 31: 1077-1082.

Boyd A.E.W., 1972. Potato storage diseases. Review of PlantPathology 51: 297-312.

Carnegie S.F., Cameron A.M., Lindsay D.A., Sharp E., Nevi-son I.M., 1998. The effect of treating seed potato tuberswith benzimidazole, imidazole and phenylpyrrole fungi-cides on the control of rot and skin blemish diseases. An-nals of Applied Biology 133: 343-363.

Chérif M., Benhamou N., 1990. Cytochemical aspects ofchitin breakdown during the parasitic action of a Tricho-derma sp. on Fusarium oxysporum f.sp. radicis lycopersici.Phytopathology 80: 1406-1414.

Chérif M., Raboudi A., Souissi S., Hajlaoui M., 2000. Sélec-tion de Trichoderma antagonistes vis-à-vis de l’agent de lapourriture des tubercules de pomme de terre Fusarium ro-seum var. sambucinum. Revue de l’I.N.A.T. 15: 115-130.

Daami-Remadi M., El Mahjoub M., 1996. Fusariose de lapomme de terre en Tunisie III. Comportement des var-iétés de pomme de terre vis-à-vis des souches locales deFusarium. Annales de l’I.N.R.A.T. 69: 113-130.

Di Pietro A., Lorito M., Hayes C.K., Broadway R.M., Har-man G.E., 1993. Endochitinases from Gliocladium virens:Isolation, chraracterization and synergistic antifungal ac-tivity in combination with gliotoxin. Phytopathology 83:308-313.

El-Ghaouth A., Charles L.W., Wisniewski M., 1998. Ultra-structural and cytochemical aspects of the biological con-trol of Botrytis cinerea by Candidate sanitoana in applefruit. Phytopathology 88: 282-291.

El-Ghaouth A., Wilson C.L., Wisniewski M., 1997. Antifun-gal activity of 2-deoxy-D-glucose on Botrytis cinerea, Peni-cillium expansum, and Rhizopus stolinifer: ultrastructuraland cytochemical aspects. Phytopathology 87: 772-779.


Fiddman P.J., Rossall S., 1995. Selection of bacterial antago-nists for the biological control of Rhizoctonia solani inoilseed rape (Brassica napus). Plant Pathology 44: 695- 703.

Grandmaison J., Benhamou N., Furlan J., Visser S.A., 1988.Ultrastructural localization of N-acethylglucosamineresidues in the cell wall of Gigaspora margarita throughoutits life cycle. Biology of the Cell 63: 89-100.

Hajlaoui M.R., Benhamou N., Bélanger R.R., 1993. Cyto-chemical study of the antagonistic activity of Sporotrix fol-culosa on rose powdery mildew, Sphaerotheca pannosa var.rosae. Phytopathology 82: 583-589.

Kawchuck L.M., Holley J.D., Lynch D.R., Clear R.M., 1994.Resistance to thiabendazole and thiophanate-methyl inCanadian isolates of Fusarium sambucinum andHelminthosporium solani. American Potato Journal 71:185-192.

Kehlenbeck H., Krone C., Oerke E.-C., Schönbeck F., 1994.The effectiveness of induced resistance on yield ofmildewed barley. Journal of Plant Diseases and Protection101: 11-21.

Kim D.-S., Cook R.J., Weller D.M., 1997. Bacillus sp. L324-92 for biological control of three root diseases of wheatgrown with reduced tillage. Phytopathology 87: 551-558.

Korsten L., De Villiers E.E., Wehner F.C., Kotzé J.M., 1997.Field sprays of Bacillus subtilis and fungicides for controlof preharvest fruit diseases of avocado in south Africa.Plant Disease 81: 455-459.

Muninbazi C., Bullerman L.B., 1998. Isolation and partialcharacterization of antifungal metabolites of Bacilluspumilis. Journal of Applied Microbiology 84: 959-968.

Pawlak A., Pavek J.J., Corsini D.L., 1987. Resistance to stor-age diseases in breeding stocks. In: Jellis G.J., RichardsonD.E. (eds.). The production of new potato varieties: Tech-nical technical advances. Cambridge University Press,New York.

Robertson R.W., Fuller M.S., 1990. Effects of the demethy-lase inhibitor, cyproconazole, on hyphal tip cells of Scle-rotium rolfsii. II. An electron microscope study. Experi-mental Mycology 14: 124-135.

Sadfi N., Chérif M., Fliss I., Boudabbous A., Antoun H.,2001. Evaluation of bacterial isolates from salty soils andBacillus thuringiensis strains for the biocontrol of Fusariumdry rot of potato tubers. Journal of Plant Pathology 83:101-118.

Schisler D.A., Slininger P.J., Bothast R.J., 1997. Effects of an-taginists cell concentration and two-strain mixtures on bio-logical control of Fusarium dry rot of potatoes. Phy-topathology 87: 177-183.

Secor G.A., Rodriguez D., Rodriguez J., Gudmestad N.C.,1994. Distribution and incidence of benzimidazole-resis-tant Fusarium sambucinum and Helminthosporium solaniisolated from potato in North America. In: BCPC Mono-graph 60: Fungicide resistance, pp. 271-274. British CropProtection Council, Ferhman, England.

Sholberg P.L., Marchi A., Bechard J., 1995. Biocontrol ofpostharvest diseases of apple using Bacillus spp. isolatedfrom stored apples. Canadian Journal of Microbiology 41:247-252.

Singh V., Deverall B.J., 1984. Bacillus subtilis as a controlagent against fungal pathogens of citrus fruit. Transactionsof British Mycological Society 83: 487-490.

Sivan C.J., Chet I., 1989. Degradation of fungal cell walls bylytic enzymes of Trichoderma harzianum. Journal of Gener-al Microbiology 135: 675-682.

Tivoli B., Jouan B., Lemarchand E., 1985. Etude comparéedes capacités infectieuses des différentes espèces ou var-iétés de Fusarium responsables de la pourriture des tuber-cules de pomme de terre. Potato Research 29: 13-32.

Utkhede R.S., Sholberg P.L., 1986. In vitro inhibition of plantpathogens by Bacillus subtilis and Enterobacter aerogensand in vivo control of two postharvest cherry diseases.Canadian Journal of Microbiology 32: 963-967.

Walker R., Powel A.A., Seddon B., 1998. Bacillus isolatesfrom the sepermosphere of peas and dwarf French beanswith antifungal activity against Botrytis cinerea and Pythi-um species. Journal of Applied Microbiology 84: 791-801.


Received 21 November 2001Accepted 15 March 2002

Documents

ULTRASTRUCTURE AND CYTOCHEMISTRY OF IN VITRO …means of light and transmission electron microscopy with gold cytochemistry. MATERIALS AND METHODS Fungal and bacterial isolates and