A Gliding Bacterium Strain Inhibits Adhesion and Motility of Another Gliding Bacterium Strain in a Marine Biofilm

Abstract

Two species of gliding bacteria were isolated from a marine biofilm. They were described and identified as members of the genus Cytophaga. One of them (RB1057) produced an extracellular inhibitor of colony expansion of the other (RB1058). The inhibitor was characterized as a glycoprotein with an apparent molecular mass of 60 kDa. It inhibited RB1058 adhesion to and gliding on substrata. Motility and adhesion of several other aquatic gliding bacteria were not measurably affected by this agent.

In aquatic habitats, much of the diverse microbial flora is found in biofilms. There is increasing interest in the structure and biology of these communities (8, 12, 21). It is likely that there is competition for space and nutrients among biofilm bacteria. Maintenance of individual species on substrata would be facilitated by the production of antimicrobial agents such as bacteriocins that target their competitors. In a biofilm, the same end would be achieved if one species were able to inhibit the adhesion and/or spreading of a competitor. Such interspecies interactions, termed ammensalism or bacterial interference, are likely to be common, but few have been described previously (for examples, see references 2, 7, 15, and 17).

The presence of gliding bacteria in aquatic environments (for examples, see references 4 and 24), their ability to adhere to various substrata (for example, see reference 6), and the fact that temporary adhesion is a requisite for function of their motility machinery all suggest that gliding bacteria are likely to be members of microbial films. During a study of the role of extracellular polymers in the adhesion and motility of marine gliding bacteria, we isolated several strains from a benthic biofilm in an inlet of Buzzards Bay, Mass., on HSM medium (yeast extract [0.05%], tryptone [0.2%], and marine salts [3% Instant Ocean], pH 7.5). Serendipitous coculture of two of these isolates revealed that expansion of the colony margins of RB1058 was inhibited when it was ≤2 mm from the periphery of a colony of RB1057 (Fig. 1). This led to the hypothesis that the latter produces an extracellular, diffusible inhibitor of adhesion and/or motility of the former.

FIG. 1.

Margins of RB1057 (bottom) and RB1058 (top) colonies >2 mm apart and <2 mm apart (left and right panels, respectively).

Characterization of strains.

Both RB1057 and RB1058 are obligate aerobes and require marine salts for growth. Temperature maxima for growth of both strains were 39°C; both grew at 4°C. Growth rate constants in HSM at 25°C were 1.1 and 1.2 doublings/h for RB1057 and RB1058, respectively. Extracellular protease(s) was produced by both; agarase, amylase, cellulase, and chitinase activities were not detected for either.

Colonies of both strains were pale yellow as viewed by overhead illumination or transillumination. Absorption spectra of acetone extracts of colonies from both isolates demonstrated peaks at 481 and 453 nm and a shoulder at 430 nm, typical of a mixture of carotenoid pigments (18). Addition of KOH did not produce a bathochromic spectral shift, indicating the absence of flexirubins (25). When viewed under oblique lighting against a dark background, colonies of these isolates were distinctively iridescent, a characteristic of some cytophagas (1, 1a, 24, 25a). RB1057 demonstrated red iridescence; RB1058 was green.

The thermal denaturation temperatures of DNA purified from RB1057 and RB1058 (22) indicated that both have a moles percent guanine-cytosine content of 30, in keeping with their tentative identification as members of the genus Cytophaga (24).

Biochemical characterization of the RB1057 inhibitor.

The inhibitor was assayed by applying minute droplets of exponential-phase culture of RB1058 around the periphery of wells in HSM agar plates with a 20-μl pipetter. Wells were loaded with 50 μl of solutions containing the inhibitor. After 7 h of incubation at 25°C, the spots of RB1058 were examined by stereomicroscopy for inhibition of the centrifugal movement of flares of gliding bacteria (Fig. 2). In determining the titer of the agent, the lowest concentration that produced detectable inhibition was termed the effective concentration.

FIG. 2.

Agar well assay of inhibitor activity. Droplets (<1 μl) of log-phase RB1058 cells were spotted (center) adjacent to a 3-mm-diameter well (lower left) in HSM agar containing the supernatant of an RB1057 culture and incubated for ∼7 h at 25°C.

The inhibitor was concentrated under N₂ with an Amicon PM10 filter from 10,000 × g supernatant fractions of 3-day static cultures of RB1057 grown in HSM medium at 25°C. Low-level activity was detected in log-phase cultures, too. The concentrate was clarified by centrifugation at 100,000 × g for 75 min. The supernatant was further concentrated and applied to a Q-Sepharose anion-exchange column (Pharmacia Hi trap Q; 0.7 by 2.5 cm). The column was flushed with 20 ml of 0.05 M NaCl in Tris buffer (0.02 M, pH 7.5) followed by a 50 to 200 mM NaCl linear gradient in the same buffer at a flow rate of 1 ml/min with a Pharmacia fast protein liquid chromatography (FPLC) system. Approximately 80% of the well assay activity was recovered in fractions 24 to 30, which comprise a 214 nm-absorbing peak (Fig. 3A). On a sodium dodecyl sulfate–10% polyacrylamide gel (19), the active fractions resolved as a band with an apparent molecular mass of 60 kDa (Fig. 3B). The intensity of staining of the band from each fraction correlated with the well assay activity of that fraction. Analysis of peak fractions (24 to 30) by high-performance size exclusion chromatography with a Zorbax SE250 column (100 mM NaPO₄ buffer, pH 6.8, mobile phase) yielded a 60.5-kDa peak in each, the size of which corresponded to its well assay titer (data not presented).

FIG. 3.

(A) FPLC fractionation of 100,000 × g supernatant fraction of concentrated RB1057 culture supernatant (10,000 × g) with a 0.05 to 0.2 M NaCl gradient in 0.02 M Tris buffer (pH 7.5). (B) FPLC fractions from a concentrate of the RB1057 inhibitor run on a sodium dodecyl sulfate–10% polyacrylamide gel electrophoresis gel and stained with Coomassie blue (19). Molecular mass standards (in kilodaltons, shown at left) are in the leftmost lane. Well assay titers are the reciprocals of the dilutions giving the effective concentrations.

The isoelectric point of the inhibitor was determined to be 4.9 with Bio-Rad ampholytes in a polyacrylamide gel with a Bio-Rad model 111 mini-isoelectric focusing cell. Colorimetric assays of the anion-exchange purified material demonstrated the presence of protein (Bio-Rad protein assay reagent [3]) and carbohydrate (phenol-sulfuric acid assay [10]) with a mass ratio of 1:1. Activity was decreased >10-fold by incubation at 70°C for 20 min. Activity was also lost by incubation at pH 2 and 10 or in 5 M guanidine-HCl and by periodate oxidation. Amino acid analysis of the pooled peak FPLC fractions was performed by the methods described by Henderson et al. (13). The proportion of tryptophan residues was determined based on the ratio of absorbance at 294 nm to that at 206 nm (14, 28) (Table 1). Edman degradation revealed that the N terminus of the polypeptide component is blocked.

TABLE 1.

Amino acid composition of the RB1057 inhibitor

Amino acid	No. of residues/moleculea
Ala	11
Arg	7
Asp-Asn	44
Cys	5
Glu-Gln	20
Gly	9
His	4
Ile	31
Leu	21
Lys	12
Met	1
Phe	7
Pro	7
Ser	27
Thr	24
Trp	10
Tyr	7
Val	10

^aBased on a composition of 50% polypeptide–50% polysaccharide. 

Carbohydrate analysis was carried out on FPLC-purified material that had been hydrolyzed in 4 M HCl at 100°C (23). The monosaccharides were separated with a Dionex high-performance anion-exchange chromatography–pulsed amperometric detection system on a CarboPac PA1 column with 16 mM NaOH solvent for neutral sugars and 150 mM Na-acetate–100 mM NaOH for acidic sugars. Monosaccharides were detected with a pulsed amperometric detector with a gold electrode. Identifications were made based on elution times of standard sugars. The approximate compositional ratio of monosaccharides in the inhibitor was determined from the areas under the peaks in the elution profile and comparison with peaks from known molar quantities of standards of the same sugars. Five neutral sugars and no acidic sugars were detected (Table 2).

TABLE 2.

Monosaccharide composition of the RB1057 inhibitor

Monosaccharide	Molar ratioa
Galactosamine	1.0
Galactose	1.7
Glucosamine	2.0
Glucose	2.7
Mannose	4.0

^aApproximate molar ratio of monosaccharide. The value for galactosamine was arbitrarily set as 1.0. 

Biological characterization of the inhibitor.

Since margins of inhibited colonies increased in thickness, indicative of bacterial growth, we concluded that inhibition of RB1058 colony spreading was not the result of antibiotic action (Fig. 1). Additionally, when RB1058 was cocultured with RB1057, the viability of RB1058 was unaffected. Also, incubation of the bacteria in concentrated inhibitor (40× effective concentration) did not affect viability.

Alternatively, the observed inhibition of colony spreading could result from inhibition of bacterial adhesion to the substratum or of gliding motility. In wet mounts, RB1058 cells quickly adhered to and glided on both glass surfaces. They pivoted in place on one pole, a characteristic of the cytophagas (for example, see reference 20). To assess the effect of the inhibitor on bacterial behavior, we made video recordings of Nomarski microscopic images. Dilutions of the inhibitor were considered to be active at concentrations at which ≥50% of previously motile cells were inhibited from gliding, defined as translocation parallel to the long axis of the cell over a linear distance of >1 cell length. The presence of the inhibitor at twice the effective concentration resulted in back-and-forth movements parallel to the bacterial long axis ≤1/2 cell length in either direction. Adhesion of the affected bacteria became more tenuous, and they exhibited shaky, sideways motion.

Kinetics of inhibition of RB1058 motility were examined by videomicroscopic tracking of bacteria over a 20-min period. At 2.5- or 5-min intervals, translocation of all the bacteria in a microscopic field was measured. Those gliding unidirectionally over a distance of >1 cell length during a 30-s period were recorded as motile. At t₀ (within 1 min of addition of the inhibitor and preparation of the wet mount), all of the cells observed in the microscopic field were motile at all concentrations of inhibitor tested (Fig. 4). We observed an approximately exponential loss of motility over time; the rate of loss increased with the concentration of the inhibitor. An apparent decrease in the tenacity of adhesion over time was also observed.

FIG. 4.

Kinetics of inhibition of RB1058 gliding motility by the RB1057 inhibitor. The percentage of bacteria demonstrating gliding was determined at the indicated time intervals after addition of the RB1057 inhibitor at various concentrations. ▵, effective concentration of the inhibitor; ○ and ◊, two and four times effective concentration, respectively; □, no inhibitor.

Direct and simple quantitative bacterial adhesion assays employing microtiter plates have been reported elsewhere (9, 26). We were unable to demonstrate an effect of the inhibitor by a modification of the Shea and Williamson assay (26). As an alternative, we developed a microcapillary adhesion assay. Acid-washed, rectangular cross-section capillaries (0.1-mm path length, 5-μl volume; Microslides; Vitro Dynamics Inc.) were loaded with suspensions of RB1058 in HSM (A₅₄₀ of 0.1) and incubated for 5 min at ambient temperature. Sterile HSM was then pumped through the capillary with a peristaltic pump at a flow rate of 1 ml or 200 capillary volumes/min for 2 min. This left an average of ∼30 to 40 bacteria adherent to the upper and lower capillary surfaces per microscopic field examined at ×400 magnification. When the inhibitor at the effective concentration or higher was added to the bacterial suspension prior to capillary loading, the number of adherent bacteria after flushing was reduced by ≥5-fold.

The effect of the inhibitor on bacteria associated with substrata having surface wettabilities distinct from that of untreated glass was also examined. Glass was derivatized with organosilanes or was treated in a muffle furnace at 482°C for 2 h (6, 11). Wettability assays were performed by the standard harmonic mean (SHM) method in which drop diameters of methanol-water mixtures on the surface in question are measured and reduced to a single number on a 0 to 100 scale of increasing wettability (11). Gerhart et al. (11) reported a strong linear relationship between the SHM value and γ^P, the combined polar components of solid surface tension generated from contact angle measurements. Videomicroscopy confirmed our earlier observations (6); RB1058 was less firmly adherent to the derivatized surfaces as the surface wettability increased. On octadecyldimethylaminopropyltrimethoxysilane (ODAPQ; SHM = 16), the bacteria adhered to the glass rapidly and more firmly than they did to untreated glass (SHM = 49). Gliding typically commenced after a lag of approximately 5 min; in contrast, no lag was detectable on untreated glass. In the presence of inhibitor at the effective concentration, the bacteria appeared to adhere firmly to the derivatized glass but demonstrated no gliding. Addition of inhibitor at four times the effective concentration resulted in tenuous adhesion, again characterized by side-to-side shaky movement, and no translocation.

On muffled glass (SHM ≥ 83), RB1058 adhered weakly but maintained the capacity to glide. In the presence of the inhibitor at the effective concentration, bacteria appeared to be even more tenuously adherent and were subject to sidewise dislocation in response to flow of the suspending medium. No gliding was observed.

Another manifestation of both the adhesive properties of the cytophaga cell surface and the gliding motility machinery is the ability of these bacteria to bind and translocate microscopic particles (for example, see reference 20). At a concentration of inhibitor identical to that which effects inhibition of motility on untreated glass (i.e., twice the effective concentration), very limited binding of 0.38-μm-diameter polystyrene latex microspheres was observed. Microspheres that did bind appeared to do so tenuously and were translocated on the bacterial surface only briefly before detaching.

Reversibility of the inhibitory effect was examined by incubating RB1058 in the presence of inhibitor at 20 times the effective concentration. After the treated bacteria were subjected to two cycles of centrifugation and resuspension in fresh growth medium, they demonstrated poor adhesion and no gliding.

The ability of RB1057 to produce this agent may provide it with a competitive advantage in colonizing substrata. To test this hypothesis, we performed mixed-culture experiments under conditions designed to approximate the habitat of these bacteria. Sand, oyster shell, or glass beads was the substratum submerged in marine salts with 1/1,000 and 1/100 the nutrient concentration of HSM. In the HSM/1,000 medium, the growth rate constant of RB1057 was 0.86; that of RB1058 was 0.75. Preliminary results indicated that RB1057 has no competitive growth advantage over RB1058, either on the substrata tested or in the aqueous phase in either dilute medium. We found the inhibitor to be present at a low level in cultures of RB1057 growing in HSM/100.

We also assayed the effect of the inhibitor on other aquatic gliding bacteria from both marine and freshwater habitats by the agar well assay and by direct microscopic observation on ODAPQ-derivatized slides. Neither adhesion nor motility of RB1057, the producer of the inhibitor, was detectably affected by the inhibitor at 10 times the concentration found to block gliding of RB1058. Similarly, there was no effect on two other marine cytophagas, one of which was isolated from the same biofilm as RB1058, and three freshwater gliding bacteria (Cytophaga sp. strain U67, Cytophaga johnsonae, and Flexibacter columnaris). In the capillary adhesion assay, RB1057 and several other gliding bacteria were unaffected by the inhibitor at ≥16 times its effective concentration.

The inhibitor’s specificity for RB1058 might be explained by the dissimilarity of its cell surface to that of several other gliding bacteria, each of which demonstrates an array of polypeptides that are accessible to radioiodination with an immobilized iodination catalyst (27). In contrast, the RB1058 surface has one predominant surface-exposed ∼50-kDa polypeptide (reference 5 and unpublished results) which may be the target of the inhibitor. Characterization of the inhibitor’s interaction with this putative target might facilitate our understanding of the mechanisms of adhesion and motility of this bacterium.

Assuming that the inhibitor is produced in the marine biofilms inhabited by RB1057, the biosynthetic cost of synthesis and export of a 60-kDa glycoprotein would be significant and would argue for a significant role(s) in the biology of its producer. If this agent does indeed function as an inhibitor of adhesion and motility in such biofilms, its high molecular weight would prove advantageous over a low-molecular-weight secondary metabolite in that it would more likely be retained in the proximity of the producing bacteria by being trapped within the extracellular polymeric matrix of the biofilm and by diffusing relatively slowly through the biofilm’s channels (8).

Marine bacteria produce a variety of products with useful properties (for example, see reference 16). These studies suggest that it may be productive to screen other marine biofilm bacteria for the production of other antiadhesion agents. A broad-spectrum agent might prove useful in the prevention of biofouling.