Calcofluor- and Lectin-Binding Exocellular ... - Bashan Foundation

JOURNAL

OF

BACTERIOLOGY, June 1989,

p.

3504-3510

Vol. 171, No. 6

0021-9193/89/063504-07$02.00/0 Copyright X) 1989, American Society for Microbiology

Calcofluor- and Lectin-Binding Exocellular Polysaccharides of Azospirillum brasilense and Azospirillum lipoferumt MADDALENA DEL GALLO,t MAHIMA NEGI, AND CARLOS A. NEYRA* Department of Biochemistry and Microbiology, New Jersey Agricultural Experiment Station, Cook College, Rutgers University, New Brunswick, New Jersey 08903 Received 23 December 1988/Accepted 29 March 1989

Extracellular polysaccharides synthesized by Azospirillum brasilense and A. lipoferum were shown on agar plates and liquid flocculating cultures. The six strains used in this work expressed a mucoid phenotype, yielding positive calcofluor fluorescence under UV light. The calcofluor-binding polysaccharides were distributed between the capsular and exopolysaccharide fractions, suggesting exocellular localization. No calcofluor fluorescence was observed in residual cells after separation of the capsular and exopolysaccharide fractions. Cellulose content was significantly higher in flocculating than in nonflocculating cultures. Failure to induce flocculation by addition of cellulose (100 mg/ml) to nonflocculating cultures, together with the sensitivity of flocs to cellulase digestion, suggested that cellulose is involved in maintenance of floc stability. Different A. brasilense and A. lipoferum strains bound to a wheat lectin (fluorescein isothiocyanate-wheat germ agglutinin), indicating the occurrence of specific sugar-bearing receptors for wheat germ agglutinin on the cell surface. The biochemical specificity of the reaction was shown by hapten inhibition with N-acetyl-D-glucosamine. All six strains failed to recognize fluorescein isothiocyanate-soybean seed lectin under our experimental conditions. We conclude that azospirilla produce exocellular polysaccharides with calcofluor- and lectin-binding properties. Bacteria of the genus Azospirillum have received a great deal of attention in recent years, following their isolation from the roots of several forage and cereal grasses (17, 27, 41), their widespread geographic occurrence in tropical and temperate soils (14, 15, 20), and the increasing number of reports indicating positive plant response to inoculation with these bacteria (2, 8, 9, 41). Azospirilla are also known to fix nitrogen in culture media and to form diazotrophic associations with the roots of grasses (2, 29, 41). Azospirillum spp. can be readily isolated from surface-disinfected roots (2, 28, 32; R. B. Lamm, Ph.D. thesis, Rutgers University, New Brunswick, N.J., 1984), form cell aggregates on or around the roots (30, 31, 39, 42), and invade the root cortex (28, 31, 32, 41). Nevertheless, knowledge about the basic mechanisms controlling selectivity, infection, and other plantbacterium interactions involving Azospirillum spp. is limited. Recently, we have focused our attention on bacterial characteristics likely to be involved in cell-cell interactions, including production and structural organization of exocellular polysaccharides (20, 33, 34; Lamm, Ph.D. thesis). Production of these polymers is common to many bacteria, and their role in the establishment of symbiotic (3, 10, 32) or pathogenic (7, 19, 35) associations with plants has been extensively studied. Furthermore, the ability to bind to specific plant lectins and the presence of calcofluor-binding polysaccharides in rhizobia have been implicated as important determinants of bacterial attachment and successful colonization of legume roots (3, 11, 19, 32, 36, 37). Preliminary microscopic observation of corn and wheat roots inoculated with azospirilla revealed the formation of bacterial aggregates that superimpose over the attachment of individual cells (unpublished data). These initial studies also * Corresponding author. t This paper is dedicated to Professor R. H. Burris

on his 75th birthday. t Present address: Agrobiotechnology Department, ENEA, Casaccia, Rome 00060, Italy.

3504

revealed the presence of two bacterial cell types: highly motile swarming cells and a sessile type able to form clumps on or around roots. These two types have also been observed in culture studies (33, 34; Lamm, Ph.D. thesis). Transfer of nutrient broth (NB)-grown cells into minimal medium containing either gluconate or fructose as a carbon source results in loss of cell motility, accumulation of abundant poly-,-hydroxybutyrate within the cells, and production of a thick coat or capsule (33, 34). Scanning and transmission microscopy revealed that two main forms of exopolysaccharides exist on the outer surface. One form appears to be more dense and tightly bound to the cell and is referred to hereafter as the capsular polysaccharide component (CPS). The other looks lighter, extending away from the cell, and will be referred to as the exopolysaccharide component (EPS). In this paper, we report on the separation of the two distinctive exopolysaccharide fractions from Azospirillum brasilense and A. lipoferum strains. The main objective of the work was to study the relatively specific interactions between the extracellular polysaccharides and calcofluor, a stilbene dye which binds cellulose and other p-linked polysaccharides. Also, binding of specific lectins by the two Azospirillum spp. was demonstrated. (A preliminary report of this work has been presented [M. M. Del Gallo and C. A. Neyra, Proceedings of the 7th International Congress on Nitrogen Fixation 1988, p. 771].) MATERIALS AND METHODS

Bacterial strains. The six strains used for this study were A. brasilense Sp7 (ATCC 29145), Cd (ATCC 29710), and 245 and A. lipoferum 59b (ATCC 29707), Brl7 (ATCC 29709), and Col 5. These strains were originally isolated from a broad range of graminaceous plants, including maize (Brl7), wheat (59b and 245), Hypharrenia rufa (Col 5), Cynodon dactylon (Cd), and Digitaria decumbens (Sp7). Inoculum preparation and culture media. All strains were maintained on nutrient agar (Difco Laboratories, Detroit,

VOL. 171, 1989

EXOCELLULAR POLYSACCHARIDES OF AZOSPIRILLA

Mich.) slants at 30°C. The stock cultures were routinely subcultured into semisolid N-free malate medium (29). Before inoculum preparation, the cultures from the semisolid medium were transferred to either tryptic soy agar (TSA; Difco) or NB (Difco) starter cultures for 24 h at 30°C. Lectin-binding studies. Inocula from TSA starter plates were transferred into liquid minimal medium containing 37 mM malate and 19 mM NH4Cl (29) and allowed to grow overnight. The cultures were then diluted to an optical density at 545 nm of 0.2. Each strain was streaked with a 0.01-mi disposable loop on agar plates containing minimal medium supplemented with 13 mM NH4NO3, and the required carbon sources were added individually. Each carbon source was separately sterilized and added to the minimal medium at a final concentration of 55 mM fructose, glucose, or galactose; 42 mM gluconate, 67 mM arabinose, or 37 mM malic acid neutralized with NaOH. Nutrient agar or TSA plates were also included in some experiments. The cultures were incubated at 33°C for 3 days. Flocculation studies. Flocs were obtained by using a modification of the procedure outlined by Sadasivan and Neyra (33). Phosphate buffer in minimal salts medium was lowered to 0.01 M (pH 6.8), and a trace amount (20 mg/liter) of yeast extract (Difco) was added. Iron salt was filter sterilized before being added to the minimal salts medium. Fructose and KNO3 used as sources for carbon and nitrogen were filter sterilized and added individually to the medium at final concentrations of 8 and 0.5 mM, respectively. The final volume of the medium was made 1 liter (pH 6.8). The inoculum was harvested from a log-phase culture grown in NB by centrifugation at 3,000 x g for 10 min at 4°C. The pellet was washed three times with equal volumes of 0.01 M phosphate buffer (pH 6.8) and inoculated into the minimal medium to an initial optical density at 660 nm of 0.3 to 0.4. Experiments were conducted in 250-ml flasks containing 100 ml of medium. The cultures were incubated at 34°C on a Gyrotory shaker (New Brunswick Scientific Co., Inc., Edison, N.J.) at 200 rpm and harvested after overnight incubation for further characterization. Extracellular polysaccharides. Eight-day-old plates were visually characterized for slime production, and the total mucoid growth was carefully removed from the agar surface with 20 ml of 0.9% NaCl. The homogenate (saline fraction) was centrifuged at 10,000 rpm for 15 min at 4°C in a Sorvall RC-5B refrigerated super-speed centrifuge (Du Pont Instruments, Chadds Ford, Pa.). The EPS fraction was separated from the supernatant after overnight treatment with 3 volumes of chilled ethanol at 4°C. The ethanol-insoluble EPS was collected by centrifugation at 20,000 rpm for 20 min. The CPS fraction was obtained from the pellet of the first centrifugation after solubilization with phosphate-buffered saline (0.43 g of KH2PO4, 1.68 g of Na2HPO4, 7.2 g of NaCl per liter [pH 7.2]) for 7 days at 4°C (24). The suspension was then centrifuged at 8,000 rpm for 10 min to separate the residual cells. The CPS fraction contained in the supernatant was obtained by treatment with ethanol as described above. The carbohydrate and protein contents of the homogenate and the EPS and CPS fractions were determined by the anthrone (13) and Lowry (21) methods by using glucose and bovine serum albumin (Fraction V; Sigma Chemical Co., St. Louis, Mo.) as the respective standards. The CPS and EPS fractions of 3-day-old flocs was obtained by a procedure similar to that outlined above (24). Floc yield. The net fresh weight of flocs was obtained by filtering the flocs through filter paper (Whatman no. 1) and

3505

weighing the paper and the flocs after they were air dried for 30 min. Floc dispersion by cellulase. Fresh flocs were separated from the medium by low-speed centrifugation (1,000 rpm) for 10 min at 4°C, and the pellet was washed twice with 0.01 M phosphate buffer (pH 6.8) and resuspended in 10 ml of the same buffer. Floc dispersion was determined in the presence of crude cellulase (EC 3.2.1.4) from Trichoderma viridie (Type V; Sigma). Cellulase (10 mg dissolved in 2 ml of sodium citrate buffer [pH 4.6]) was added to 1 ml of floc suspension and incubated at 45°C for 6 h. Blanks for each strain were prepared by incubation with citrate buffer alone. Floc dispersion was recorded visually, and the reducing sugars liberated were estimated by the colorimetric procedure of Nelson-Somogyi (26). Another batch of flocs were treated with 10 ml of 1 N NaOH for 1 h at 100°C, and the alkali-insoluble fraction was washed and treated with cellulase as described above. Cellulose purification and assay. Flocs were separated from the medium by low-speed centrifugation and washed as described above. The washed pellet was treated for 30 min with the acetic-nitric acid reagent described by Updegraff (40). The insoluble residue was washed with deionized water and further digested for 1 h with 10 ml of 67% sulfuric acid. A proper dilution of the acidic solution was used to determine cellulose by anthrone analysis (40) with commercial cellulose (Whatman cellulose powder; W & R Balston, Ltd.) as the standard. Phase-contrast and fluorescence microscopy. A modification of the procedure outlined by Bohlool and Schmidt (6) was used for assessment of lectin binding by azospirilla. Fluorescein isothiocyanate (FITC)-labeled wheat germ agglutinin (WGA; Sigma) and FITC-labeled soybean lectin (SBL; Sigma) were used as lectins. Smears from 3-day-old cultures were air dried, heat fixed, and covered for 20 min with 0.1 ml of 100 ,ug of lectin solution per ml in phosphatebuffered saline. The stained smears were washed in phosphate-buffered saline for 15 min, covered with a cover slip, and observed for fluorescence with a Nikon Optiphot microscope equipped with an epifluorescence attachment (HBO 50 power supply and a B filter package; 410 to 485-nm interference excitation, a DM 505 dichroic mirror, and a 515-nm barrier filter) with normal optics to ascertain the association of fluorescence with individual bacterial cells, clumps of bacteria, or the polysaccharide material surrounding the cells and to estimate percentages of FITC-WGA- and FITCSBL-labeled bacteria. The biochemical specificity of FITCWGA binding to bacteria was checked by hapten inhibition with N-acetyl-D-glucosamine. A solution of N-acetyl-D-glucosamine (5 mg/ml of phosphate-buffered saline) was added to an equal volume of FITC-WGA solution for treatment of bacterial samples as described above. Slimy colonies from 3- to 8-day-old plates were observed under a phase-contrast microscope after being stained with alcian blue, which was used as a 1.0% (wt/vol) solution in 95% (vol/vol) ethanol, and carbol-fuchsin as a counterstain to reveal the presence of acidic heteropolysaccharides. Neutral ,-glucans were revealed with the fluorochrome calcofluor (Fluorescent Brightener 28; Sigma). The stain was used as a 0.1% (wt/vol) solution in 0.5 N saline (pH 6.0) to assure minimal distortion of cells. Stained preparations were examined for fluorescence under a long-wavelength UV lamp and with a Zeiss standard 14 epifluorescence microscope provided with a UV excitation filter (G 365) and a 425-nm barrier filter. A positive reaction was indicated by emission of blue-green fluorescence.

3506

DEL GALLO ET AL.

J. BACTERIOL.

TABLE 1. Carbohydrate contents of whole homogenates (saline fraction) and CPS EPS fractions of A. brasilense and A. lipoferum strains grown on gluconate agar plates

TABLE 2. Binding of FITC-WGA to A. lipoferum Col 5 grown on agar plates containing different carbon sources Carbon source

Carbohydrate (,ug/mg of protein) in: Bacterial strain

Whole homogenate

CPS

EPS

A. lipoferum 59b Br 17 Col 5

341 457 451

46 178 113

27 41 27

A. brasilense Sp 7 245 Cd

386 485 333

197 172 152

60 28 52

Flocs were also stained directly by using 0.1% (wt/vol) solutions of Congo red and trypan blue. Microscopic observations were made under a phase-contrast microscope. Staining of fresh flocs and EPS and CPS fractions with alcian blue and calcofluor was done as described above. The presence of encapsulated cells in flocs was determined by exclusion of India ink particles (16). RESULTS Slime production. All of the strains used in this study produced mucoid colonies (Muc+ phenotype) on complex (Nutrient agar and TSA) and synthetic (fructose and gluconate) media. Slimy colonies were clearly distinguishable on 72-h plates, but maximal slime accumulation did not occur until 6 to 8 days after seeding. A. brasilense Cd produced a characteristically pink to red coloration on all of the media tested. The other five strains produced white to creamy or beige colonies. Aging of the cultures for 2 weeks or more resulted in production of dark brown pigments by strains Sp7 and Sp 59b. Microscopic observations of the slimy growth revealed numerous encapsulated cells, as indicated by exclusion of India ink particles. Masses of cells were arranged in the form of a dense glycocalyx containing acidic polysaccharides, as indicated by positive staining with alcian blue. Positive calcofluor fluorescence indicated the presence of 3-glucans, possibly cellulose (44). CPS and EPS separation. The separation of both CPS and EPS components of the external surface of azospirilla was achieved by using differential centrifugation based on differential solubility properties. A quantitative estimate of the relative carbohydrate contents of six Azospirillum strains is shown in Table 1. The loosely bound EPS was readily solubilized by mild extraction with low salt, and it remained in the supernatant. The CPS fraction was separated from the pelleted cells by more prolonged extraction (1 week in phosphate-buffered saline). The CPS and EPS fractions were precipitated with 75% ethanol, suggesting that both contained high-molecular-weight acidic polysaccharides. The carbohydrate content of the CPS fraction was about two- to fourfold larger than the EPS content, regardless of the bacterial strain (Table 1). The distribution of total carbohydrates between the CPS and EPS fractions agrees well with results obtained with rhizobia (19, 24). Lectin-binding studies. Binding of azospirilla to FITCWGA produced a characteristic yellow-green fluorescence. This positive fluorescence was observed with bacteria grown on a variety of carbon sources, but gluconate was apparently

TSA Nutrient agar Gluconate Glucose Arabinose Malate Fructose Galactose

FITC-WGA binding % Stained cells Fluorescence"

>50 >50 >50 10-30 10-30 10-30 1-10 1-10

+++ + +++ +++ ++ + + +

" + + +, + +, and +, High, intermediate, and low fluorescence intensities, respectively.

the best single carbon source (Table 2). Five of the six strains, except Cd, yielded positive fluorescence with FITCWGA (Table 3). The biochemical specificity of the reaction was shown by hapten inhibition with N-acetyl-D-glucosamine, a competitive analog inhibitor. The presence of the hapten in the lectin-binding assay resulted in 90 to 100% reduction of the fluorescence associated with the cells. Most interesting, however, was the finding that none of the six strains was able to bind to SBL (Table 3). Calcofluor binding and flocculation studies. Cells of a log-phase culture grown in NB, washed, and transferred to flocculation medium containing fructose (8 mM) and KNO3 (0.5 mM) started to show cell aggregation and snowlike flakes just 2 h after inoculation. Intensive flocculation was observed after 4 to 6 h, when the cells began to sediment at the bottom of the flask. Phase-contrast microscopic observations of these early sedimenting cells (4 h after transfer) revealed pleomorphic forms (including motile cells) which were very lightly covered with exocellular polysaccharides (Fig. 1A). Continued incubation of cells for 24 to 48 h resulted in production of macroflocs that settled completely at the bottom of the flask, leaving behind a clear supernatant practically devoid of cells. Encapsulated cells obtained from 48-h-old cultures were nonmotile and fully covered with exocellular polysaccharides, as shown by negative staining with India ink (Fig. 1B). Removal of the EPS fraction allowed clear observation of the capsules surrounding individual cells. TABLE 3. Binding of FITC-WGA and FITC-SBL by A. brasilense and A. lipoferum strains grown on gluconate agar plates Bacterial strain

A. lipoferum 59b Br 17 Col 5 A. brasilense Sp 7 245 Cd

FITC-WGA % Stained cells

binding" Fluorescenceb

>50 >50 >50

+++ +++ ++

>50 10

+++ +

" All positive strains developed yellow-green fluorescence, except strain Sp 7, which developed yellow to yellow-green fluorescence. With FITC-SBL, there were no positively stained cells. b++ +, + +, and +, High, intermediate, and low fluorescence intensities,

respectively.

VOL. 171, 1989

EXOCELLULAR POLYSACCHARIDES OF AZOSPIRILLA

3507

TABLE 5. Cell yields and cellulose contents of flocculating and nonflocculating cultures of A. brasilense and A. lipoferum Bacterial strain

A. lipoferum 59 b Br 17 Col 5 A. brasilense Sp 7 245

Cd

Flocculating cultures"

Nonflocculating

cultures'

Floc yield

Cellulose

(g/liter)

Cell yield

(mg/g of floc)

(g/liter)'

(mg/g of floc)

3.27 2.72 3.01

7.64 6.43 5.00

1.61 1.23 1.12

0.09 0.04 0.08

2.34 2.16 1.87

6.41 8.10 8.02

1.66 1.93 1.79

0.11 0.08 0.06

Cellulose

In minimal medium containing fructose (8 mM) and

KNO3 (0.05 mM). (Difco). 'The cell yield was obtained after pelleting of 24-h-old NB cultures.

'

b In NB

FIG. 1. Phase-contrast microscopy of A. lipoferum cells from flocculating cultures containing fructose (8 mM) and KNO3 (0.5 mM). Panels: A, 4-h cultures; B, 48-h cultures. Capsules in 48-h cultures were visualized by negative staining with India ink. Cells are shown completely covered with EPS (B) and after removal of the EPS fraction (C). Bar, 10 ,um.

Positive staining of 48-h-old flocs with calcofluor, Congo Red, and trypan blue provided evidence for the presence of neutral P-glucans (Table 4). Positive alcian blue staining also indicated the presence of acidic heteropolysaccharides, associated with the flocs (Table 4). The relative contributions of neutral glucans (including cellulose) and acidic heteropolysaccharides to the formation and stability of flocs are TABLE 4. Characterization of flocs from A. brasilense and A. lipoferum by different staining procedures Bacterial strain

Staining with": Calcofluor

Congo red

Trypan blue

Alcian blue

++ + +

++ ++ ++

++ ++

++ ++ +

+ + +

++

ND

+ + +

A. lipoferum

59b Brl7 Col S

A. brasilense Sp 7 245 Cd + +,

good;

+,

++

fair; ND, not determined.

++

++ +

+

not known, but probably both are involved. However, addition of commercial cellulose (100 mg/ml) to the nonflocculating medium did not induce flocculation, even after 48 h (data not shown). Nonflocculating cultures grown on NB produced limited amounts of cellulose compared with flocculating cultures (Table 5), suggesting that cellulose plays a role in flocculation, possibly by helping to maintain floc stability. Incubation of flocs with crude cellulase resulted in significant dispersion of the macroflocs, accompanied by release of reducing sugars (Table 6). Treatment of the flocs with 1 N NaOH produced an alkali-stable residue that was more sensitive to cellulase, as evidenced by the higher amounts of reducing sugars liberated. All of these results, taken together, support the contention that cellulosic polysaccharides play a functional role in the flocculation process, but the exact mechanism still remains to be elucidated. Additional studies showed that calcofluor-binding polysaccharides in floc cells are localized on the surface as part of the exocellular fractions CPS and EPS (Fig. 2). Whole flocs and the CPS and EPS fractions yielded positive fluorescence with calcofluor (Fig. 2A, C, and D, respectively). The cells obtained after separation of EPS and CPS fractions failed to show any fluorescence with calcofluor (Fig. 2B). The CPS fraction looked more extended and veillike than the granular form of the EPS fraction. The relevance of this observation remains to be seen. TABLE 6. Floc dispersion and liberation of reducing sugars upon cellulase treatment of flocs and alkali-stable residue of A. brasilense and A. lipoferum Bacterial strain

Reducing sugars in whole flocs"

Reducing sugars

A. lipoferum 59 b Br 17 Col 5

83.33 102.77 55.55

220.18 217.64 218.16

A. brasilense Sp 7 245 Cd

103.78 118.88 108.88

275.78 322.53 326.20

in alkali-stable residue'"'

" Expressed as micromoles of glucose liberated per gram of flocs after 2 h of incubation at 45°C. b Dispersion was fair with all strains. Dispersion was good with all strains.

3508

DEL GALLO ET AL.

J. BACTERIOL.

FIG. 2. Microscopy of flocs from A. lipoferum 59b showing positive calcofluor fluorescence (A). No fluorescence was observed after removal of the EPS and CPS components (B). Positive calcofluor fluorescence was observed in association with the CPS (C) and EPS (D) fractions. Bar, 10 ,um.

DISCUSSION This study was initiated as part of a program to identify and characterize cell surface components of azospirilla that may be involved in cell-cell interactions and colonization of grass roots. In principle, the initial interaction between two symbionts takes place at the surfaces of the two interacting organisms and particular molecules present at the cell surface may be important in determining the outcome of the reaction. The results presented here demonstrated that both A. brasilense and A. lipoferum strains grow well on a variety of carbon and nitrogen sources and synthesize abundant extracellular polysaccharides on agar plates and liquid flocculating cultures. These polysaccharides appear to be organized in two distinct morphological forms: (i) as capsules tightly bound to the external cell surface (CPS fraction) and (ii) as slimy polysaccharides loosely attached to the capsules (EPS fraction). Even though extensive information on the role of exopolysaccharides is available for several gramnegative bacteria (3, 7, 9, 10, 32, 35, 43), it would be premature to define a specific role for either of the two fractions (CPS and EPS) in the process of infection of grass roots by azospirilla. Nonetheless, a recent report by Michiels et al. (22) seems to indicate that the exopolysaccharides of azospirilla perform similar functions and are complementary to those produced by rhizobia. They also

present evidence that DNA cloned from A. brasilense suppressed the nonmucoid phenotype of exoB and exoC mutants of Rhizobium meliloti and corrected the inability of these mutants to produce nitrogen-fixing nodules on alfalfa. R. meliloti exo mutants on calcofluor-containing media appear nonfluorescent under UV light. These mutants are considered to be deficient in one of the major acidic EPS and form empty, non-nitrogen-fixing nodules on alfalfa. In this study, all six Azospirillum strains expressed a mucoid phenotype that yielded positive fluorescence with calcofluor under UV light. Slime preparations observed under phasecontrast microscopy revealed the presence of encapsulated microflocs, filaments, or chains and individual vegetative cells. The sequence of events leading to microfloc formation has been discussed recently by Bleakley et al. (5), who indicated that A. lipoferum grown on P-hydroxybutyrate agar accumulated poly-p-hydroxybutyrate, followed by filamentation and septation, and the septating filaments later produced extensive capsular material (5). Our results are in line with those presented by Bleakley et al. In addition, we observed positive calcofluor fluorescence associated only with encapsulated cells in microflocs but not with the filaments or chains, suggesting that calcofluor-binding polysaccharides are located in the capsular fraction. The latter was further confirmed by positive staining of the isolated CPS fraction with calcofluor (Fig. 2). The results indicated that

VOL. 171, 1989

EXOCELLULAR POLYSACCHARIDES OF AZOSPIRILLA

calcofluor-binding polysaccharides are exocellular and were identified in the CPS and EPS fractions. No positive calcofluor fluorescence was associated with the residual cells after separation of the CPS and EPS fractions. Whole flocs were positively stained by calcofluor, Congo red, and trypan blue, suggesting the presence of cellulosic P-glucans (44). Cellulose is a common constituent of the external surface of many bacteria (12, 19, 25), but its occurrence in Azospirillum spp. has been suggested only recently (5, 33) and needs further confirmation. In this report, however, we have shown that the cellulose contents of flocculating cultures were significantly higher than those of nonflocculating, NBgrown cultures. Thus, cellulose might appear to be associated with the process of flocculation, probably in relation to the maintenance of floc stability, rather than playing a role in floc formation. In fact, addition of commercial cellulose (100 mg/ml of culture) to nonflocculating cultures failed to induce flocculation. On the other hand, addition of crude cellulase to a flocculating culture resulted in floc dispersion with concomitant release of reducing sugars. The physiological relevance of floc formation and floc dispersion in the rhizosphere still remains to be investigated, although some reports indicate that grass roots inoculated with azospirilla form thick aggregates of bacteria connected by cellulosic fibrillar material (30, 39, 42) on the root surface. The ability to bind to specific lectins and the presence of calcofluor-binding polysaccharides in rhizobia have been implicated as important determinants of bacterial attachment to legume roots and successful colonization (3, 11, 19, 32, 36, 37). Several lines of evidence indicate that the lectin receptor is present in the CPS component and binding to legume roots is limited to encapsulated cell forms (1, 3, 24). Nonetheless, other polysaccharides (EPS, LPS, and ,3-glucans, including cellulose) have also been implicated in the process of root colonization by rhizobia (9, 19, 25). Several authors (17, 31, 39) have also suggested that lectins are involved in the process of root colonization by azospirilla, but no experimental evidence was provided to support this contention. The results presented in this work demonstrate the abilities of different A. brasilense and A. lipoferum strains to bind to WGA, suggesting the presence of sugar-bearing receptors for WGA on the cell surface. One important finding was that binding to WGA was achieved by Azospirillum strains isolated from a wide range of plant species, and therefore the lectin reaction as an important determinant for selective colonization, as indicated for the legume-Rhizobium symbiosis (3, 4, 11, 37), cannot be directly extrapolated. Nonetheless, the binding by azospirilla to WGA, together with the inability to bind to SBL, is indicative that a certain degree of specificity for the interaction with lectins does exist. The biochemical specificity was shown by hapten inhibition by using N-acetyl-D-glucosamine. Lectins from graminae purified to date are very closely related to one another, the most studied being WGA (18, 38). Recently, Mishkind et al. (23) reported on the isolation of a wheat root lectin that is structurally and functionally similar to WGA. This finding (23), together with the results reported here, opens the possibility to use WGA as a model-type lectin to advance our knowledge about the role of this protein in the establishment of grass-bacterium associations. The findings of the present study may help to further advance our understanding of the interaction between azospirilla and grass roots and should also facilitate the design of genetic research aimed at improving the efficiency of these rhizocoenosis-involving azospirilla.

3509

ACKNOWLEDGMENTS This work was performed under New Jersey Agricultural Experiment Station project 01204, supported by state and federal Hatch Act funds, and is published under New Jersey Agricultural Experiment Station no. D-01204-1-88. M.D.G. was on leave as a visiting scientist to the Department of Biochemistry and Microbiology, Rutgers University, during part of this work and was partially supported by an Italian National Research Council special grant I. P. R. A. subproject 1 (paper 2116). M.N. is a postdoctoral fellow supported by New Jersey Agricultural Experiment Station Hatch Act funds. We thank J. Dobereiner (Empresa Brasileira de Pesquisa Agropecuaria, Brasflia, Brazil) and J. R. Milam (University of Florida, Gainesville) for supplying cultures. We also thank R. L. Tate and R. S. Triemer for the use of their microscope facilities. LITERATURE CITED 1. Bal, A. K., S. Shantharam, and S. Ratnam. 1978. Ultrastructure of Rhizobium japonicum in relation to its attachment to root hairs. J. Bacteriol. 133:1393-1400. 2. Baldani, V. L. D., J. I. Baldani, and J. Dobereiner. 1987. Inoculation of field grown wheat (Triticum aestivum) with Azospirillum spp. in Brazil. Biol. Fertil. Soils 4:37-40. 3. Bauer, W. D. 1981. Infection of legumes by rhizobia. Annu. Rev. Plant Physiol. 32:407-449. 4. Bhuvaneswari, T. V., S. G. Pueppke, and W. D. Bauer. 1977. Role of lectins in plant-microorganism interactions: binding of soybean lectin to rhizobia. Plant Physiol. 60:486-491. 5. Bleakley, B. H., M. H. Gaskins, D. H. Hubbell, and S. G. Zam. 1988. Floc formation by Azospirillum lipoferum grown on polyP-hydroxybutyrate. Appl. Environ. Microbiol. 54:2986-2995. 6. Bohlool, B. B., and E. L. Schmidt. 1974. Lectins: a possible basis for specificity in the rhizobium-legume root nodule symbiosis. Science 185:269-271. 7. Bradshaw-Rouse, J. J., M. H. Whatley, D. L. Coplin, A. Woods, L. Sequeira, and A. Kelman. 1981. Agglutination of Erwinia stewartii strains with a corn agglutinin: correlation with extracellular polysaccharide production and pathogenicity. Appl. Environ. Microbiol. 42:344-350. 8. Cohen, E., Y. Okon, J. Kigel, I. Nur, and Y. Henis. 1980. Increase in dry weight and total nitrogen content in Zea mtiays and Setaria italica associated with nitrogen fixing Azospirillun spp. Plant Physiol. 66:746-749. 9. Dazzo, F. B. 1980. Determinants of host-specificity in the rhizobium-clover symbiosis, p. 165-187. In W. E. Newton and W. H. Orme-Johnson (ed.), Nitrogen fixation, vol. II. University Park Press, Baltimore. 10. Dazzo, F. B., and W. J. Brill. 1979. Bacterial polysaccharide which binds Rhizobium trifolii to clover root hairs. J. Bacteriol. 137:1362-1373. 11. Dazzo, F. B., C. A. Napoli, and D. H. Hubbell. 1976. Adsorption of bacteria to roots as related to host specificity in the Rhizobium-clover symbiosis. Appl. Environ. Microbiol. 32:166-171. 12. Deinema, M. H., and L. P. T. M. Zevenhuizen. 1971. Formation of cellulose fibrils by gram-negative bacteria and their role in bacterial flocculation. Arch. Mikrobiol. 78:42-57. 13. Dische, Z. 1962. General color reactions. Methods Carbohydr. Chem. 1:477-479. 14. Dobereiner, J. 1978. Influence of environmental factors on the occurrence of Sprillum lipoferum in soils and roots. Ecol. Bull. NFR (Naturvetensk. Forskningsradet) 26:343-352. 15. Dobereiner, J., I. E. Mariel, and M. Nery. 1976. Ecological distribution of Spirillum lipoferum Beijerinck. Can. J. Microbiol. 22:1464-1473. 16. Duguid, J. P. 1951. The demonstration of bacterial capsules and slime. J. Pathol. Bacteriol. 63:673-685. 17. Elmerich, C. 1984. Molecular biology and ecology of diazotrophs associated with non-leguminous plants. Biotechnology 2:967-978. 18. Etzler, M. E. 1985. Plant lectins: molecular and biological aspects. Annu. Rev. Plant Physiol. 36:209-234. 19. Halverson, L. J., and G. Stacey. 1986. Signal exchange in plant-microbe interactions. Microbiol. Rev. 50:193-225.

3510

DEL GALLO ET AL.

20. Lamm, R. B., and C. A. Neyra. 1981. Characterization and cyst production of azospirilla isolated from selected grasses growing in New Jersey and New York. Can. J. Microbiol. 27:1320-1325. 21. Lowry, 0. H., M. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement with Folin phenol reagent. J. Biol. Chem. 193:265-275. 22. Michiels, K. W., J. Vanderleyden, A. P. Van Gool, and E. R. Signer. 1988. Isolation and characterization of Azospirillum brasilense loci that correct Rhizobium meliloti exoB and exoC mutations. J. Bacteriol. 170:5401-5404. 23. Mishkind, M., K. Keegstra, and B. A. Palevitz. 1980. Distribution of wheat germ agglutinin in young wheat plants. Plant Physiol. 66:950-955. 24. Mort, A. J., and W. D. Bauer. 1980. Composition of the capsular and extracellular polysaccharides from Rhizobium japonicum. Plant Physiol. 66:158-163. 25. Napoli, C., F. Dazzo, and D. Hubbell. 1975. Production of cellulose microfibrils by Rhizobium. Appl. Microbiol. 30:123131. 26. Nelson, H. 1944. A photometric adaptation of the Somogyi method for the determination of glucose. J. Biol. Chem. 153: 375-380. 27. Neyra, C. A., and J. Dobereiner. 1977. Nitrogen fixation in grasses. Adv. Agron. 29:1-37. 28. Okon, Y. 1985. Azospirillum as a potential inoculant for agriculture. Trends Biotechnol. 3:223-229. 29. Okon, Y., S. L. Albrecht, and R. H. Burris. 1976. Factors affecting growth and nitrogen fixation of Spirillum lipoferum. J. Bacteriol. 127:1248-1254. 30. Patriquin, D. G. 1982. New developments in grass-bacteria associations, p. 139-190. In N. S. Subba Rao (ed.), Advances in agricultural microbiology. Buttersworth Co. Publishers, Ltd., London. 31. Patriquin, D. G., J. Dobereiner, and D. K. Jain. 1983. Sites and processes of the association between diazotrophs and grasses. Can. J. Microbiol. 29:900-915. 32. Reed, W., G. Jane, S. Long, T. L. Reuber, and G. C. Walker. 1988. Analysis of the role of exopolysaccharides in Rhizobium symbiosis, p. 51-59. In N. T. Keen, T. Kosuge, and L. L. Walling (ed.), Physiology and biochemistry of plant microbe

J. BACTERIOL.

33.

34. 35.

36.

37. 38.

39. 40. 41.

42.

43. 44.

interactions. Proceedings of the 11th Annual Symposium in Plant Physiology, University of California. American Society of Plant Physiologists, Rockville, Md. Sadasivan, L., and C. A. Neyra. 1985. Flocculation of Azospirillum brasilense and Azospirillum lipoferum: exopolysaccharides and cyst formation. J. Bacteriol. 163:716-723. Sadasivan, L., and C. A. Neyra. 1987. Cyst production and brown pigment formation in aging cultures of Azospirillum brasilense ATCC 29145. J. Bacteriol. 169:1670-1677. Sequeira, L. 1978. Lectins and their role in host-pathogen specificity. Annu. Rev. Phytopathol. 16:453-481. Smit, G., J. W. Kijne, and B. J. J. Lugtenberg. 1987. Involvement of both cellulose fibrils and a Ca2+-dependent adhesin in the attachment of Rhizobium leguminosarum to pea root hair tips. J. Bacteriol. 169:4294-4301. Stacey, G., A. S. Paau, and W. J. Brill. 1980. Host recognition in the Rhizobium-soybean symbiosis. Plant Physiol. 66:609-614. Strosberg, D. A., D. Buffard, P. A. Kaminski, M. P. Chapot, P. W. Rossow, and A. Foriers. 1986. Lectin multigene families in leguminous and non-leguminous plants, p. 1-16. In L. M. Shannon and M. J. Chrispeels (ed.), Molecular biology and seed storage proteins and lectins. Proceedings of the 9th Annual Symposium Plant Physiology, University of California. American Society of Plant Physiologists. Waverly Press, Baltimore. Umali-Garcia, M., D. H. Hubbell, M. H. Gaskins, and F. B. Dazzo. 1980. Association of Azospirillum with grass roots. Appl. Environ. Microbiol. 39:219-226. Updegraff, D. M. 1969. Semi-micro determination of cellulose in biological materials. Anal. Biochem. 32:420424. Van Berkum, P., and B. B. Bohlool. 1980. Evaluation of nitrogen fixation by bacteria in association with roots of tropical grasses. Microbiol. Rev. 44:491-517. Whallon, J. H., G. Acker, and H. El-Khawas. 1985. Electron microscopy of young wheat roots inoculated with Azospirillum, p. 222-229. In W. Klingmuller (ed.), Azospirillum III: genetics, physiology and ecology. Springer-Verlag KG, Berlin. Whitfield, C. 1988. Bacterial extracellular polysaccharides. Can. J. Microbiol. 34:415420. Wood, J., and G. Flucher. 1978. Interaction of some dyes with cereal P-glucans. Cereal Chem. 55:952-966.