INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Microbial adhesion to hydrocarbons was the subject of pioneering studies by Mudd in 1924 (24) and more extensive ones over the last 20 years (12, 24, 28, 29, 33). Cell adhesion to hydrocarbons seems to proceed mainly via proteins: in Acinetobacter sp. strain MJT/F5/199A it occurs via an acidic protein of 65 kDa, probably a glycoprotein (31), in Acinetobacter calcoaceticus RAG-1 it occurs via fimbriae (27), and in Acinetobacter sp. strain A3 (12) it occurs via two proteins of 26.5 kDa and 56 kDa. Adhesion of cells to oil droplets and cell hydrophobicity can be determined by the microbial adhesion to hydrocarbon (MATH) test (28) or by more recently developed quantitative tests such as those involving measurement of zeta potential (6) and water contact angles (26, 35).

Bacteria produce many types of biosurfactants, as has been recently reviewed (8). The studies of new Acinetobacter strains are therefore stimulating because they are good sources of new surfactants when grown on hydrocarbons.

A new n-alkane-degrading strain of Acinetobacter has recently been isolated from the Venice Lagoon (2) and classified as Acinetobacter venetianus VE-C3 (9). The present study investigates the adhesion mechanisms of this new strain during the n-alkane degradation process. To do this, the adhesion mechanisms of Acinetobacter sp. strain RAG-1 as the control strain and the newly isolated A. venetianus VE-C3 were compared with respect to their physiological differences by using molecular probes and confocal laser-scanning microscopy (CLSM). Diesel fuel containing n-alkanes or pure n-hexadecane was used as the sole carbon and energy source. The electrochemical probe, the dropping-mercury electrode (DME), was used to study in situ the surface-active constituents of bacterial cultures. The electrochemical probe responds to the adhesion of bacteria (30, 38) and n-alkane droplets and the adsorption of dissolved polymers (15, 18) in real time.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cell cultures. The restriction analysis of amplified rDNA, DNA hybridization, and GC content indicates that both VE-C3 and RAG-1 belong to the species A. venetianus (34). However, in this study we still use the species names A. venetianus VE-C3 and Acinetobacter sp. strain RAG-1 (ATCC 31012). Both strains were incubated at 28°C in a complex medium and in mineral medium. The complex medium, plate count agar (PCA), was composed of 5 g of tryptone, 2.5 g of yeast extract, 1 g of D-glucose, and 24 g of NaCl per liter of deionized water. The mineral medium had the following composition: 1.0 g of MgSO₄ · 7H₂O, 0.7 g of KCl, 2.0 g of KH₂PO₄, 3.0 g of Na₂HPO₄, 1.0 g of NH₄NO₃, and 24.0 g of NaCl per liter of deionized water. In the mineral medium, n-hexadecane or diesel fuel (2.0 g liter¹) was the sole carbon and energy source. The diesel fuel (Esso Italiana) for diesel engine vehicles is composed of a mixture of n-alkanes (C₁₂ to C₂₈) with traces of aromatics (<30 ppm polycyclic aromatic hydrocarbon PAH) and less than 1% total additives; it has a density of 0.830 g cm³ at 15°C and a viscosity of 2.0 to 4.5 mm² s¹ at 40°C (32). The diesel fuel was filtered through a 0.2-µm-pore-size Teflon filter (Sartorious) for sterilization and particle removal. Batch cultures of 50 and 250 ml were grown in flasks with continuous shaking (280 to 300 rpm) in a gyratory water bath shaker (G76; New Brunswick) or in a rotary drum.

Biomass determination. The cell biomass of the two strains was determined from the protein concentration by the method of Bradford (5). The protein absorbance was determined at 595 nm with a UV-visible spectrophotometer (UV160; Shimadzu). The standard curve of bovine serum albumin was used for calibration. The coefficient of variation in five replicate analyses was 3.1%.

O₂ consumption rates. The strains were grown overnight in 250-ml flasks containing 50 ml of mineral medium with 2 g of diesel fuel liter¹. Then 2.5 ml of each culture (1% inoculum) was transferred to 250 ml of fresh mineral medium in duplicate and incubated at 28°C in a gyratory water bath shaker. The inoculum contained 24 ± 0.9 and 22 ± 0.6 µg of proteins ml¹ in RAG-1 and VE-C3, respectively. At different times, the O₂ consumption was determined with a biological oxygen monitor (5300; Yellow Springs Instruments) equipped with a Clark probe. Analyses were carried out with 3-ml aliquots. Calculations were based on oxygen concentrations in air-saturated water at 28°C and on protein concentrations (5). The changes in pH during the experiment were not significant.

Fluorescent molecular probes for CLSM. Image analysis was performed in both strain cultures by CLSM (MRC-500 instrument; Bio-Rad Microscience Division) equipped with a Nikon Microphot microscope. Two different fluorescent molecular probes were used: the lectin concanavalin A (ConA) from Canavalia ensiformis (Jack bean), labelled with fluorescein isothiocyanate (FITC) (Sigma), and Nile Red (Nile Blue A oxazone), (Sigma). The ConA has an affinity for glucose and mannose residues, whereas Nile Red is a fluorochrome specific for neutral lipids. This staining method was described previously (1). The distributions of the two fluorescent molecular probes in specimens were observed by CLSM.

MATH tests. The two strains were grown in flasks containing 50 ml of PCA complex medium in a gyratory shaker for 18 h. The cells were harvested by centrifuging at 3,000 × g for 15 min., washed twice with deionized water, and suspended in phosphate-buffered saline (pH 7.2) to obtain a final absorbance at 600 nm (A₆₀₀) of 0.4 to 0.6 as reported by Rosenberg et al. (28). The MATH tests were performed after 0, 12, and 36 h of incubation in mineral medium containing 2 g of diesel fuel liter¹. Aliquots (3 ml) of each suspension were distributed in seven vials to which 0.15 ml of n-hexadecane was added to extract the hydrophobic cells. The A₆₀₀ of the aqueous phase (A_t) was measured after a given vortexing time, and cell concentrations were expressed with respect to the initial absorbance, A₀, as log (A_t/A₀ × 100).

Adhesion studies with the DME. The electrochemical technique is based on the current-time dependence (chronoamperometry) during oxygen reduction at the DME. This is a modification of a widely used polarographic technique for measurements of surface-active organic matter in aquatic environments (3, 7, 14, 20, 23, 37, 40, 41). Adsorption of organic molecules and submicron particles to the DME causes a decrease in the oxygen reduction current. This decrease is proportional to the extent of surface coverage of the mercury drop by the organic film. On the other hand, adhesion of oil microdroplets results in well-defined attachment signals (39) on a millisecond time scale (Fig. 1A). The amplitudes of attachment signals reflect particle size and the force of adhesion; the signal frequency depends on the particle abundance in the medium.

View larger version (26K): [in this window] [in a new window]   FIG. 1.   (A) Adhesion of an oil droplet and its spreading to form a film at the electrode. The electrical signal (transient current) is caused by displacement of the double-layer charge (1) from the contact area AC. (B) Adhesion of bacterial cells and formation of molecular contact with the electrode.

1

1

2

Electrophoretic mobilities. The electrophoretic mobilities of cell suspensions were measured in 0.1 M NaCl electrolyte and in seawater by using an automated microelectrophoresis instrument (S3000; PenKem). Zeta potentials were computed by the method of Hunter (13).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Although recent data suggest that VE-C3 and RAG-1 belong to the same new species, A. venetianus (9, 34), the two strains have different physiological behaviors in the presence of diesel fuel as the sole carbon and energy source (Fig. 2). Both strains consumed O₂ when grown in mineral medium in the presence of diesel fuel (2 g liter¹), but their growth rates (Fig. 2A) and protein (biomass) production levels (Fig. 2B) were different. RAG-1 started consuming O₂ after a 2-h lag phase, reaching the highest rate (0.5 nmol of O₂ min¹ mg of protein¹) after 6 h. This maximum value was followed by a drop to 0.03 nmol of O₂ min¹ mg of protein¹ (Fig. 2A). VE-C3 had a longer lag phase (4 h) (Fig. 2A). The O₂ consumption rate increased to about 0.3 nmol min¹ mg of protein¹ in 8 h and remained almost constant throughout the experiment (28 h). Protein production did not parallel O₂ consumption rates in either strain (Fig. 2B), and there was a more significant delay in biomass formation, measured as total proteins, for VE-C3 (21 h).

View larger version (20K): [in this window] [in a new window]   FIG. 2.   (A) Oxygen consumption rates determined with Clark's probe in cultures of Acinetobacter sp. strain RAG-1 () and A. venetianus VE-C3 () grown in mineral medium containing 2 g of diesel fuel liter1. (B) Protein determination of Acinetobacter sp. strain RAG-1 () and A. venetianus VE-C3 () grown in mineral medium containing 2 g of diesel fuel liter1.

These physiological differences may be due to different mechanisms of adhesion to diesel fuel as the carbon source. An in situ investigation of cell interaction with diesel fuel was performed by CLSM with the fluorescent lectin ConA-FITC and the fluorochrome Nile Red to image the CPS of VE-C3 and the neutral lipid moiety of emulsan molecules of RAG-1, respectively. Observations were made at constant time intervals during cell growth in mineral medium amended with diesel fuel at 28°C.

RAG-1 produces emulsan (11), which reduces the surface tension of diesel fuel. In the light transmission mode (Fig. 3a), the surface of diesel fuel drops was observed to break upon contact with the hydrophobic layer. When the diesel fuel drops were further broken down to microdroplets of around 100 µm in diameter (Fig. 3b), they were surrounded by bacteria and consumed as a carbon source. RAG-1 cells became highly fluorescent (Fig. 3d) due to Nile Red, which was bound to emulsan exuded by the cells. After 24 h of growth, the drops of diesel fuel broke down into even smaller free microdroplets about the same size as the bacteria or less (Fig. 3c). This might explain the peak and the rapid decrease in O₂ consumption by RAG-1 cells grown with diesel fuel as the sole carbon source (Fig. 2A). RAG-1 cells no longer adhered to the diesel fuel residue when the microdroplets became repulsive due to the coating emulsan, a strong polyanionic bioemulsifier, and the microdroplets were dispersed in the medium without attached bacteria (Fig. 3c).

View larger version (135K): [in this window] [in a new window]   FIG. 3.   Colonization of diesel fuel by Acinetobacter sp. strain RAG-1. (a) Decrease in surface tension of part of a diesel fuel drop colonized by RAG-1, observed with an optical microscope in the transmission mode after a 3-h incubation at 28°C in mineral medium with diesel fuel (2 g liter1). (b) Colonization of a diesel fuel droplet (diameter, 56 µm) by RAG-1, showing bacterial adhesion at the rim and on top, observed in the transmission mode after an 8-h incubation at 28°C. (c) Repellent diesel fuel microdroplets of different sizes (from 10 to less than 0.75 µm) dispersed in the medium without attached bacteria after a 24-h incubation. (d) Same image as in panel b but observed by CLSM in the fluorescence mode, formed by 16 overlapped images scanned every 0.8 µm for a total depth of 12.8 µm. This sample was stained with Nile Red fluorochrome for the neutral lipid moiety of emulsan.

The behavior of VE-C3 cells was different (Fig. 4). In the presence of diesel fuel, this strain formed cell-to-cell aggregates and then adhered to the surface of diesel fuel drops (Fig. 4a) by its capsular polysaccharide (CPS). This term is generally accepted for a polysaccharide-based polymer anchored to the proteins or lipids of outer membranes. The CPS production was stimulated by diesel fuel, and the cells became partly fluorescent after 6 h of incubation with ConA-FITC. This molecular probe binds specifically to glucose and mannose residues of CPS (Fig. 4b). After 12 h of incubation, the diesel fuel droplets were completely colonized by VE-C3 cells. This had the effect of locally decreasing the surface tension, so that the diesel fuel formed sharp and indented droplet shapes (Fig. 4c). In the fluorescence mode, all cells produced CPS and the surface of the diesel fuel showed bacterial polysaccharide "footprints" (22) (Fig. 4d). At the end of the experiment (28 h), the bacterial cells were still attached to the microdroplets, producing larger mixed aggregates of diesel fuel and microbial cells (Fig. 4f). In the fluorescence mode, the microdroplets were seen to be "glued" in the middle of the bacterial aggregate by the ConA-positive CPS (Fig. 4g).

View larger version (176K): [in this window] [in a new window]   FIG. 4.   Colonization of diesel fuel by A. venetianus VE-C3. (a) Light microscopy in the transmission mode, showing aggregates of cells (arrows) before adhesion to the diesel fuel surface after a 6-h incubation at 28°C in mineral medium with 2 g of diesel fuel liter1. (b) The same image as in panel a but observed by CLSM in the fluorescence mode, obtained by overlapping 20 images scanned every 0.6 µm for a total depth of 12 µm. VE-C3 was stained with ConA-FITC to show polysaccharide residues of glucose and mannose in CPS. Only a fraction of the cells seen in panel a (arrows) have a fluorescent CPS after 6 h of incubation. (c) Transmission mode, showing that the surface tension of a diesel fuel drop colonized by strain VE-C3 decreases and the bacteria at the rim and on top produce elongated and indented shapes after a 12-h incubation. (d) The same image as in panel c but in the fluorescence mode, with the ConA-FITC distribution imaged by CLSM. VE-C3 cells with CPS smear the elongated rim of the diesel fuel drop showing "polysaccharide footprints." (f) Transmission mode, showing a diesel fuel droplet with a diameter of about 20 µm, with many smaller microdroplets embedded in a microbial aggregate. After a 28-h incubation at 28°C, VE-C3 cells are still attached to diesel fuel droplet. (g) The same image as in panel f but in fluorescence mode with the ConA-FITC distribution imaged by CLSM. The aggregate of cells and diesel fuel microdroplets is glued together by a thick polysaccharide matrix excreted by the bacteria.

In cell-to-surface adhesion experiments, performed by the MATH test, VE-C3 cells appeared to be hydrophilic when grown in a complex medium but became hydrophobic when incubated in mineral medium with diesel fuel as the sole carbon and energy source (Fig. 5A). RAG-1 cells were always hydrophobic, even when grown in a complex medium without diesel fuel (Fig. 5B). Table 1 shows that diesel fuel-uninduced VE-C3 cells had higher negative electrophoretic mobility (0.81 m² s¹ V¹) and more negative zeta potential (10.5 mV¹) in seawater than did RAG-1 cells, in line with the MATH test results. Therefore, the two strains differ in their interfacial properties.

View larger version (21K): [in this window] [in a new window]   FIG. 5.   (A) Cell hydrophobicity of A. venetianus VE-C3 at different times, measured by the MATH test: 0 h () preincubated in PCA medium and then transferred to mineral medium with 2 g of diesel fuel liter1, then incubated for 12 h () and 36 h () in mineral medium with 2 g of diesel fuel liter1. (B) MATH test for Acinetobacter sp. strain RAG-1 under the same conditions as in panel A incubated for 0 h (), 12 h (), and 36 h ().

The adhesion of uninduced cells grown in mineral medium in the absence of diesel fuel was determined in organic-free electrolyte by analyzing current-time curves in terms of surface coverage and film formation time. We compared the adhesion behavior at an inert hydrophobic surface, the mercury electrode, by analyzing the amperometric curves recorded in the suspensions of RAG-1 and VE-C3 cells in terms of surface coverage (Fig. 6A) and film formation time (Fig. 6B). The extent of surface coverage recorded over a broad range of cell densities showed that both strains established rapid molecular contact with the surface. RAG-1 was more efficient in covering the surface despite being smaller. Full-surface coverage (100%) was reached at a cell density of 6 × 10⁷ ml¹ for RAG-1 and 1.8 × 10⁸ ml¹ for VE-C3, respectively. Consequently, the film formation rate was also higher in RAG-1 suspensions (Fig. 6B). However, at higher cell densities (>10⁸/ml) a striking difference between two strains could be identified in the dynamics of film formation. For RAG-1 strain, the film formation time leveled off at _lim = 500 ms, and for VE-C3, _lim dropped to zero. The situation where _lim is zero is typical of films formed by the adsorption of dissolved biopolymers. When _lim is greater than zero, this generally indicates that there is a rate-limiting surface process involved in film formation. This means that biofilm formation by VE-C3 is governed by the adsorption of CPS, which is faster than the coalescence of cell-spreading zones, the rate-limiting surface process in film formation by RAG-1 cells. This difference in _lim between RAG-1 and VE-C3 reflects a difference in the flexibility of their outer membranes and cell-to-cell interactions. Cell-to-cell adhesion is an important characteristic for VE-C3. In mineral medium containing diesel fuel, up to 4,000 cells per aggregate were counted.

View larger version (18K): [in this window] [in a new window]   FIG. 6.   Adhesion of uninduced VE-C3 and RAG-1 cells at the electrode from their suspensions in 0.1 M NaCl solution. The percentage of surface coverage (A) and the film formation time (B) are plotted as a function of cell density.

The adhesion response of bacteria in the oil-degrading cultures was studied simultaneously with the adhesion of dispersed oil droplets and the adsorption background of surface-active degradation products (Fig. 7). The amperometric curves recorded on three consecutive mercury drops reflect the size distribution of diesel fuel droplets and their abundance in RAG-1 cultures (Fig. 7B), in VE-C3 cultures (Fig. 7C) and in the control experiment (Fig. 7A) after 3 days of growth.

View larger version (15K): [in this window] [in a new window]   FIG. 7.   Electrochemical signals in the diesel fuel-degrading cultures of RAG-1 (curve B) and VE-C3 (curve C) after 3 days of growth. The control (curve A) is an uninoculated dispersion.

The average signal frequency and film formation times were obtained by the analysis of 50 current-time curves (Table 2). RAG-1 culture showed a significant increase of surface-active material smaller than <1 µm and no decrease in abundance of dispersed fuel droplets in the medium, but the fuel droplet decreased in size, showing that a significant fraction of newly produced surface-active material in RAG-1 cultures consists of emulsan and submicron fuel droplets formed by the action of the exocellular emulsan.

In VE-C3 cultures, a significant decrease in the attachment frequency of oil droplets was already found at the end of day 1 of growth (Table 2). After day 3 of growth, a significant increase in the rate of film formation was observed ( = 0.75 ± 0.08 s in VE-C3 and 0.9 ± 0.15 s in RAG-1), indicating a new production of surface-active material. The constant decrease in the abundance of fuel droplets dispersed in the medium was accompanied by an increase in the content of surface-active material (mostly bacteria). This remarkable difference between the two strains was even more pronounced when n-hexadecane was used instead of diesel fuel (16).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Different modes of cell adhesion to droplets of diesel fuel (a mixture of n-alkanes C₁₂ to C₂₈) and n-hexadecane are identified for two Acinetobacter strains, which helps to explain their biodegradation behavior. In this study, we demonstrated by imaging that strain VE-C3 produced a ConA-positive CPS to adhere to diesel fuel.

How does an insoluble alkane microdroplet become available as a carbon source for VE-C3 cells surrounded by hydrophilic CPS? The model proposed for Pseudomonas oleovorans growth on alkanes involves the transfer of outer membrane lipopolysaccharides to the alkane droplet, thus solubilizing the alkane material (36). However, extraction of these molecules would damage the outer cell membrane and ultimately cause cell damage and death. Therefore, an alternative mechanism of emulsification must be operating. We propose incorporation of alkane nanodroplets in the CPS as a more realistic mechanism for the continuing growth of VE-C3, based on the following observations. (i) The CPS of VE-C3 is capable of forming a stable dispersion of diesel fuel nanodroplets in the hydrophilic polymer matrix (Fig. 4g). This image suggested that a boundary layer surrounding the fuel droplets prevented them from coalescing (2). Electrochemical probe experiments clearly demonstrated that VE-C3 cells with a CPS establish molecular contact with the hydrophobic surface of mercury.

VE-C3 cells, which have a hydrophilic capsule, are therefore capable of attaching to the alkane droplet through the CPS. The three-dimensional network of the CPS hydro-gel entangles nanometer-sized fuel droplets, which do not coalesce, and stabilizes the emulsion without reducing the interfacial tension (10). This finding disagrees with other reports (4, 21), where it was demonstrated that CPS hinders attachment to hydrophobic solid surfaces.

The formation of a composite material is not a common concept in bacterial interaction with substrates but is a well-known phenomenon in materials science. The physical contact of two materials, hydrocarbons and CPS, at the nanometer level confers new properties on the mixture without the formation of chemical bonds. This interaction results in firm cell-to-substrate attachment, whereas the interactions in aggregated cells are of the cell-to-cell type and are mediated by glycoproteins, such as adhesin-like molecules.

The increasing hydrophobicity of VE-C3 measured by the MATH test is therefore most probably induced by incorporation of diesel fuel droplets into the CPS matrix, facilitating mass transfer of hydrocarbons from the medium to the cells. This incorporation mechanism is further supported by recent studies on polysaccharide giant aggregates from the northern Adriatic (17, 19). (The giant aggregate, 3 m in size, was a free-floating gelatinous formation sampled by a scuba diver at a depth of 15 m in August 1997.) The n-hexadecane could be incorporated in this hydrophilic gel to form a hydrophobic material of higher viscosity than the original gel. A gel prepared from dextran (molecular weight, 5 × 10⁷) in seawater had a similar capacity to incorporate n-hexadecane droplets and form a material of higher viscosity.