Dispersion of TiO₂ Nanoparticle Agglomerates by Pseudomonas aeruginosa

Abstract

Engineered nanoparticles are increasingly incorporated into consumer products and are emerging as potential environmental contaminants. Upon environmental release, nanoparticles could inhibit bacterial processes, as evidenced by laboratory studies. Less is known regarding bacterial alteration of nanoparticles, including whether bacteria affect physical agglomeration states controlling nanoparticle settling and bioavailability. Here, the effects of an environmental strain of Pseudomonas aeruginosa on TiO₂ nanoparticle agglomerates formed in aqueous media are described. Environmental scanning electron microscopy and cryogenic scanning electron microscopy visually demonstrated bacterial dispersion of large agglomerates formed in cell culture medium and in marsh water. For experiments in cell culture medium, quantitative image analysis verified that the degrees of conversion of large agglomerates into small nanoparticle-cell combinations were similar for 12-h-growth and short-term cell contact experiments. Dispersion in cell growth medium was further characterized by size fractionation: for agglomerated TiO₂ suspensions in the absence of cells, 81% by mass was retained on a 5-μm-pore-size filter, compared to only 24% retained for biotic treatments. Filtrate cell and agglomerate sizes were characterized by dynamic light scattering, revealing that the average bacterial cell size increased from 1.4 μm to 1.9 μm because of nano-TiO₂ biosorption. High-magnification scanning electron micrographs showed that P. aeruginosa dispersed TiO₂ agglomerates by preferential biosorption of nanoparticles onto cell surfaces. These results suggest a novel role for bacteria in the environmental transport of engineered nanoparticles, i.e., growth-independent, bacterially mediated size and mass alterations of TiO₂ nanoparticle agglomerates.

The large-scale production and use of nanoparticulate titanium dioxide (TiO₂) in a suite of consumer products (www.nanotechproject.org) make it likely that these nanoparticles will enter the environment via waste or erosional processes (15, 27). As with other potential pollutants, interactions of nanoparticles with environmental bacteria could result in either direct toxicity to bacteria or biotransformations which would influence nanoparticle mobility in soil and water. Studies of the effects of bacteria on nanoparticles have mainly focused on bacterial coalescence of already-dispersed nanoparticles. For example, Limbach et al. (21) observed that nanoparticulate CeO₂ was initially dispersed in wastewater but then sorbed to activated sludge bacteria. Similarly, Kiser et al. (18) described biosorption of nanoscale TiO₂ to activated sludge within a wastewater treatment plant and in laboratory batch reactors. Nanoparticles were initially dispersed and then became associated with bacterial cells or exopolymers, which would tend to decrease the mobility of individual nanoparticles.

Nanoparticle adhesion to bacterial surfaces has been observed in toxicity studies for a range of nanoparticle chemistries and bacterial species. For example, cytotoxicity of nano-CeO₂ to Escherichia coli depended on nanoparticle adhesion to the bacterial outer membrane (37). Nanoparticulate fullerene (19), ZnO (10), and CdTe quantum dot (22) toxicity to bacteria was predicated upon nanoparticle-cell contact. Nanoparticle adhesion to bacteria may be by nonspecific electrostatic interactions (as for CeO₂ [37]) or may involve other interactions with bacterial surface polymers. Jucker et al. (12) demonstrated irreversible adhesion of outer membrane polysaccharides from E. coli to nanoscale TiO₂ by formation of hydrogen bonds. The affinities of nanoparticles for bacterial membranes imply a general consequence beyond direct bacterial toxicity: cell biosorption of agglomerated or dispersed nanoparticles will strongly influence nanoparticle mobility in the environment.

Due to high surface area and interfacial energies (11), TiO₂ nanoparticles will agglomerate in aqueous environments. Agglomeration driven by electrostatics promotes settling (31) and can thus remove nanoparticles from the water column. Given the high affinity between many nanoparticles and bacteria, initially agglomerated nanoparticulate TiO₂, upon simple contact with bacteria, should tend toward dispersion due to particles preferentially sorbing onto cells. Here, a combination of high-resolution imaging, quantitative image analysis, and size-based agglomerate fractionation is used to describe bacterially mediated dispersion of TiO₂ nanoparticle agglomerates. This phenomenon is likely nonspecific to either Pseudomonas aeruginosa or TiO₂ nanoparticles and thus reveals a generalizable role for bacteria in enhancing the mobility of nanoparticles in the environment.

MATERIALS AND METHODS

Bacteria, nanoparticles, and other chemicals.

The environmental isolate P. aeruginosa PG201 (9, 32) was used. Nanoparticulate TiO₂ was industrial P25 Aeroxide (Evonik, Parsippany, NJ) (75% anatase and 25% rutile) (29). To test the dependence of some findings on nanoparticle size and morphology, laboratory-synthesized TiO₂ (23, 24) (80% anatase and 20% rutile) was also studied in cell growth experiments (see the supplemental material for synthesis details). Hereinafter, P25 Aeroxide TiO₂ is called “industrial” TiO₂. Nanoparticles were maintained in the dark at room temperature prior to experiments. Experiments were performed either in Luria-Bertani (LB) broth (Fisher Scientific, Pittsburgh, PA) prepared with nanopure water (pH 6.9, 18.2 MΩ-cm) or in marine coastal water (pH 7.7, conductivity, 51.5 mS) collected from the Carpinteria Salt Marsh in Carpinteria, CA (34°24′N, 119°31′W) (3). All chemicals were reagent grade or better (Sigma Chemical or Fisher Scientific).

Zeta potential.

Zeta potentials of bacterial cells and TiO₂ nanoparticles were measured in LB media at 25°C using a Nano-Series Zeta Sizer (Nano-ZS ZEN3600; Malvern Instruments, Worcestershire, United Kingdom) (see the supplemental material).

Bacterial growth and contact experiments.

Three types of experiments were performed under dark conditions: one where P. aeruginosa was grown for 12 h with preagglomerated industrial TiO₂ (cell growth experiment), another where preagglomerated industrial TiO₂ was exposed for 10 min to washed, exponential-phase P. aeruginosa cells (cell contact experiment), and the last, where preagglomerated industrial TiO₂ was added to cell-free 12-h culture supernatant (supernatant contact experiment). For cell growth experiments only, bacterial effects on laboratory-synthesized TiO₂ were also tested. Here, “preagglomerated” means that nanoparticles were agglomerated in sterile LB media (10 min) before bacterial inoculation. P. aeruginosa cells were grown from frozen (−80°C) stock on LB agar plates for 8 h, and then colonies were subcultured in LB media (10 ml, 200 rpm, 37°C) for 8 h; cells were determined to be mid-exponential phase at this time (data not shown). For cell growth experiments, LB medium (10 ml) containing 0.5 mg ml⁻¹ of industrial TiO₂ was inoculated with P. aeruginosa to an initial optical density at 600 nm (OD₆₀₀) of 0.01. Based on prior research (1), 0.5 mg ml⁻¹ TiO₂ was not expected to inhibit bacterial growth under dark conditions but did promote extensive nanoparticle agglomeration. Several 400-μl aliquots were removed within 5 min of inoculation and again at late exponential phase (12 h, 200 rpm, 37°C) for imaging by scanning electron microscopy.

For cell contact experiments, bacteria were grown for 12 h, then separated by centrifugation (12,000 × g, 20 min). The supernatant was removed and reserved (4°C). The cell pellet was dispersed in 5 ml of fresh LB medium, followed by 5 ml of 1.0-mg ml⁻¹ industrial TiO₂ in LB medium. For supernatant contact experiments, industrial TiO₂ nanoparticles were added to the culture supernatant at 0.5 mg ml⁻¹ and vortexed briefly. For cell contact and supernatant contact experiments, aliquots (400 μl each) were removed within 30 min for imaging.

To test whether dispersion may also occur in natural waters, contact experiments were performed in filter (0.2 μm)-sterilized marsh water. As above, P. aeruginosa was grown in, and recovered from, LB broth, then washed (3 times) with nanopure water and resuspended in marsh water (5 ml). Industrial TiO₂ (1.0 mg ml⁻¹) prepared in marsh water was added to the cell suspension and briefly vortexed.

Fractionation of agglomerates by filtration.

Abiotic and cell contact treatments were prepared in triplicate to determine the fraction of industrial TiO₂ able to pass through a porous filter. Samples were syringe filtered through membranes (Nuclepore track-etched polycarbonate membranes, 5-μm pore size; Whatman, GE Health Care), and the filter-retained dry mass (after 8 h at 105°C) was measured. The filtrates were reserved for dynamic light scattering.

Dynamic light scattering.

Agglomerate size distributions in contact experiment filtrates were determined by dynamic light scattering within 30 min of filtration using the Malvern Instruments Nano-Series Zeta Sizer (as for zeta potential measurements; see above) at 25°C. Dynamic light scattering was also performed for 12-h washed P. aeruginosa cells (no TiO₂) treated with sodium azide (10 min) to arrest motility, because cell swimming interfered with Brownian motion-based measurement of hydrodynamic diameter (not shown).

Liquid tensiometry.

Since P. aeruginosa can produce surfactants during growth (34), which could contribute to dispersion, surface tension was measured as an indicator for biosurfactant production in culture supernatant. Samples were dispensed into 10 ml acid (H₂SO₄ plus NoChromix; Godax Laboratories, Inc., Cabin John, MD)-washed glass beakers for measurements at room temperature using a K10 digital liquid tensiometer (Krüss, Hamburg, Germany) and a small-volume plate.

Transmission electron microscopy.

Primary particle size and morphology for industrial and laboratory-synthesized TiO₂ nanoparticles were determined from transmission electron micrographs (see the supplemental material).

Environmental scanning electron microscopy.

For experiments in LB broth, aliquots (400 μl) were prepared for microscopy by centrifugation (10,000 × g, 10 min), washing (3 times, 400 μl nanopure H₂O), and then resuspension in 100 μl nanopure H₂O. Washing removed residual LB salts, which convolute images, but did not significantly alter agglomerate size or morphology (see the supplemental material). Images were acquired using an FEI Co. XL30 FEG environmental scanning electron microscope (Philips Electron Optics, Eindoven, Netherlands) (see the supplemental material). For each treatment, 15 random images were taken (×2,500 magnification; 5 images for each triplicate sample). For experiments in salt marsh water, aliquots were not washed prior to imaging and 5 random images were taken for each treatment (×625 magnification).

Quantitative analysis of microscopy images.

Size distributions for nanoparticle agglomerates in LB media were determined from environmental scanning electron microscope images. Equivalent diameters (X) for nanoparticle agglomerates (classified as small where 0 < X < 6 μm, medium where 6 ≤ X < 12 μm, and large where 12 ≤ X < 18 μm) were measured for each image using NIS-Elements Basic Research software (Nikon Instruments Inc., Melville, NY) (see the supplemental material). Statistical analysis of size distribution data was performed by Tukey's honestly significant difference test.

Cryogenic scanning electron microscopy.

Specimens from growth experiments were washed (once with nanopure H₂O) prior to liquid nitrogen cryogenic fixation. Samples were transferred to a Polaron cryogenic stage (Quorum Technologies, Guelph, Ontario, Canada) for etching (−70°C, 3 min), followed by gold sputter coating (3 to 5 nm). Imaging was performed using the FEI XL30 instrument as described above, at a 20-kV accelerating voltage and 10.3-mm working distance.

RESULTS

Nanoparticle and cell characteristics.

Industrial TiO₂ is highly heterogeneous; thus, a size range was determined rather than a mean diameter. Equivalent diameters of industrial TiO₂ nanoparticles (from TEM images) ranged from 6.4 nm to 73.8 nm, with 75% having an equivalent diameter between 15 and 60 nm. The primary particle size for laboratory-synthesized TiO₂ nanoparticles was 16.0 ± 1.5 nm. The specific surface area of industrial TiO₂ (54 m²/g) was measured by the Brunauer-Emmett-Teller (BET) technique using a Tristar 3000 (Micrometrics, Norcross, GA). The specific surface area for laboratory-synthesized TiO₂ (94 m²/g) was calculated assuming spherical particles, uniform diameter (16 nm), and an estimated density of 3.97 g/cm³ based on the weighted anatase-to-rutile ratio. The zeta potentials for P. aeruginosa and industrial TiO₂ nanoparticles in LB media were −9.1 mV and −17.9 mV, respectively.

Nanoparticle agglomerate shifts in growth experiments.

TiO₂ nanoparticles were highly agglomerated in LB media both before and immediately after inoculation with P. aeruginosa (Fig. 1; see the supplemental material). Quantitative analysis of environmental scanning electron microscope images showed no effect of inoculation on the maximum size of TiO₂ agglomerates (Table 1). The sizes of washed large agglomerates appeared consistent with sizes of agglomerates in unwashed LB media (see the supplemental material), suggesting that the washing steps did not significantly alter agglomerate size or morphology.

FIG. 1.

Environmental scanning electron microscope images of industrial TiO₂ agglomerates in uninoculated (A) and inoculated (B) LB broth after 12 h. At higher magnification, individual cells are either circumscribed with small nanoparticle clusters (C) or are sorbed onto nanoparticle agglomerates (D). Scale bars, 10 μm (A and B), 1 μm (C), and 2 μm (D).

TABLE 1.

Maximum TiO₂agglomerate sizes for cell growth, cell contact, and supernatant contact experiments

Expt	Exposure (h)	Maximum agglomerate sizea (μm) for:
		Industrial TiO2	Laboratory-synthesized TiO2
Sterile LB media	0	12.9 ± 0.8	13.4 ± 1.3
	12	16.3 ± 0.8	9.8 ± 0.8
Growth	0	10.8 ± 1.0	13.8 ± 1.5
	12	4.9 ± 0.3	5.5 ± 0.4
Cell contact	0	4.9 ± 0.6	NDb
Supernatant contact	0	8.7 ± 0.8	ND

^aSize (maximum equivalent diameter) was measured from environmental scanning electron microscope images (n = 15 images per experiment).

^bND, not determined.

After 12 h under abiotic conditions, the maximum size of industrial TiO₂ agglomerates increased while the maximum size of laboratory-synthesized TiO₂ agglomerates decreased (Table 1). In cell growth treatments, industrial TiO₂ agglomerates appeared smaller after 12 h than abiotic controls and cells and nanoparticles appeared to be highly colocalized (Fig. 1B). Dispersion of large agglomerates after 12 h of bacterial growth was further evidenced by a 3-fold increase in the frequency of small agglomerates, which could be remaining abiotic agglomerates or cells encrusted by nanoparticles, while the frequency of large and mid-sized agglomerates was significantly decreased (Fig. 2). The maximum size of industrial TiO₂ agglomerates in 12-h cultures was less than one-half of the initial size (Table 1). At higher magnification, industrial TiO₂ nanoparticle clusters were observed at individual cell surfaces (Fig. 1C). In several micrographs, bacteria appeared to be associated with, and detaching from, large TiO₂ agglomerates (Fig. 1D). Similar results were found for laboratory-synthesized TiO₂ nanoparticles (see the supplemental material).

FIG. 2.

Frequency of three size ranges (small [S], medium [M], and large [L]) of industrial TiO₂ nanoparticle agglomerates in culture media with and without P. aeruginosa cells. (A and B) Experiments with uninoculated media (black bars) and inoculated media (gray bars) at 0 h (A) and 12 h (B). (C) Contact experiments. Striped bars, cell-free supernatant; white bars, exponential-phase cells added to fresh LB medium. Like letters indicate values that are not significantly different (α = 0.05) for each treatment within each size range. Bars to the left of the dashed line are related to the left y axis; bars to the right are related to the right y axis. ESEM, environmental scanning electron microscope.

Dispersion of large industrial TiO₂ agglomerates in P. aeruginosa cell growth experiments was also observed in cryogenic scanning electron microscope images. Cryogenic scanning electron microscopy is used for visualizing nanoparticle agglomerate morphology (36), microbe-mineral interactions (4, 16), and soil colloids (28) and was used here to visually confirm selected environmental scanning electron microscopy results in light of concerns about the effect of moderate drying and deformation, which can occur during extended imaging (17). In 12-h uninoculated samples, very large (>20-μm) highly branched TiO₂ agglomerates were observed. After 12 h of cell growth, large agglomerates did not exist and bacterial cells appeared encrusted by TiO₂ (Fig. 3 C and E). Large agglomerates were also apparent at lower magnification in cryogenic scanning electron microscope images of 12-h abiotic samples, and the specimen surface was heterogeneous and textured. In contrast, the surfaces from 12-h cell growth treatments appeared smooth and homogeneous and no large agglomerates were observed (see the supplemental material). Thus, environmental and cryogenic scanning electron microscope images, in combination with quantitative data from image analysis, demonstrate that initially large agglomerates were extensively dispersed in P. aeruginosa growth experiments.

FIG. 3.

Cryogenic scanning electron microscope images of industrial TiO₂ nanoparticles (A), P. aeruginosa (B and D), and P. aeruginosa with industrial TiO₂ (C and E). White arrows indicate bare cells in treatments not containing TiO₂ (B and D) or cells encrusted by nanoparticles (C and E).

Contact experiments.

To test if bacterial growth was required for agglomerate dispersion, industrial TiO₂ was brought into contact with either washed cells or culture supernatant, in which cells had been grown for 12 h and subsequently removed by centrifugation. In culture supernatant, industrial TiO₂ agglomerates appeared smaller than agglomerates in fresh LB media (see the supplemental material) and the maximum agglomerate size was 8.7 ± 0.8 μm, which was significantly smaller than the maximum agglomerate size in fresh LB media (Table 1). Quantitative image analysis revealed that, in comparison to the abiotic controls (0 h), there were fewer large and medium agglomerates and more small agglomerates in cell-free culture supernatant (Fig. 2C). However, any differences in aqueous chemistries between the fresh LB broth (Fig. 2A) and the cell-free spent LB broth (Fig. 2C) only minimally changed the overall agglomerate size distributions. Accumulation of biosurfactants did not contribute to dispersion, as the liquid surface tensions of cell-free culture supernatant and fresh LB media were statistically equivalent (54.1 ± 2.1 mN/m and 52.3 ± 2.8 mN/m, respectively).

Environmental scanning electron microscope images from cell contact experiments revealed extensive dispersion of preformed industrial TiO₂ nanoparticle agglomerates, which closely resembled the level of dispersion in 12-h cell growth experiments (see the supplemental material). Medium and large agglomerates were not observed in contact experiments, and average maximum agglomerate sizes for the cell contact and cell growth experiments were statistically equivalent (4.9 ± 0.6 μm and 4.9 ± 0.3 μm, respectively) (Table 1). Agglomerate size distributions for the cell contact and 12-h cell growth experiments showed a similar shift from medium and large to small agglomerates and cell-nanoparticle conglomerates (Fig. 2). Together with scanning electron micrographs, in which cells appear to peel away initially agglomerated nanoparticles (Fig. 1), these data suggest that bacterial contact with TiO₂ agglomerates and sequestration of nanoparticle clusters onto cell surfaces (i.e., biosorption) are the dominant processes leading to dispersion.

In order to determine if bacterially mediated dispersion occurs under environmental conditions, contact experiments were also performed in water collected from a salt marsh. When suspended in marsh water, TiO₂ nanoparticles formed large agglomerates that were similar in size and morphology to those formed in LB medium (see the supplemental material). After cell contact, large agglomerates were entirely dispersed and only a few small agglomerates existed. At higher magnification, bacteria were seen to be associated with the remaining agglomerates or separated from agglomerates and encrusted by TiO₂ nanoparticles (see the supplemental material).

Agglomerate size fractionation and dynamic light scattering.

In abiotic samples (0 h) prepared in LB medium, 81.3% of TiO₂ by mass was retained on the filter membrane, while 24% was retained in cell contact treatments. Dynamic light scattering could not be performed on the entire (unfiltered) samples due to the presence of medium and large (>6-μm) agglomerates, which interfere with the analyses. In the abiotic sample filtrate, 51.5% of TiO₂ (by volume) was agglomerated to larger than 3 μm (Fig. 4) and 5.1% of TiO₂ was in the form of very small (0.2- to 0.7-μm) agglomerates. This contrasted with cell contact experiment filtrate, where 0.7% of TiO₂ was agglomerated to larger than 3 μm. The small-size (0.2- to 0.7-μm) populations in the abiotic filtrate were not observed in cell contact filtrate (Fig. 4), suggesting that dispersed TiO₂ nanoparticles were removed from suspension by bacterial biosorption. The average particle size in cell contact experiments (1.9 ± 0.1 μm) was larger than the average size of P. aeruginosa without TiO₂ (1.4 ± 0.0 μm) (Fig. 4) due to TiO₂ adsorption onto cells, as observed in scanning electron microscope images (Fig. 1C and D and 3C and E).

FIG. 4.

Dynamic light scattering-based particle size distributions for filtered (<5-μm) suspensions of industrial TiO₂ in uninoculated LB media (TiO₂), sodium azide-immobilized P. aeruginosa cells in the absence of TiO₂ (P. aeruginosa), and P. aeruginosa cultured with industrial TiO₂ (TiO₂ + P. aeruginosa). The y axis indicates the percentage of the total particle volume associated with a specific particle size. Data points represent the average volumetric percentages from triplicate samples (error bars represent standard errors).

DISCUSSION

TiO₂ nanoparticle agglomerates are held together by weak electrostatic or van der Waals forces (38) in natural waters. The degree and stability of agglomeration depend highly on environmental chemistry, and abiotic factors affecting agglomeration are well studied. For example, divergence from the nanoparticle isoelectric point (the pH at which the net particle surface charge is neutral) leads to interparticle electrostatic repulsion, thereby decreasing agglomeration (7, 8). Organic acids (e.g., humic and fulvic acids) can similarly decrease agglomeration due to the coating of nanoparticles, leading to steric stabilization (6). Increasing ionic strength and addition of divalent cations facilitate nanoparticle agglomeration by compressing the electrical double layer at nanoparticle surfaces, thereby reducing repulsion (8). Yet, while effects of abiotic chemistry on agglomeration are well established, the potential effects of biotic components of aqueous environments, including bacteria, are not understood.

This study shows that an environmental strain of P. aeruginosa promotes dispersion of initially agglomerated industrial and laboratory-synthesized TiO₂ nanoparticles. In cell growth experiments, both industrial and laboratory-synthesized TiO₂ nanoparticle agglomerates were largely dispersed after 12 h of growth (Fig. 1; see the supplemental material). Dispersion was defined as an increase in the frequency of small agglomerates (Fig. 2), which could be abiotic or bacterium-nanoparticle conglomerates, produced as a consequence of nanoparticles being removed from medium and large agglomerates upon biosorption to bacterial cells (Fig. 1 and 3; see the supplemental material). Subsequent experiments (with industrial TiO₂ only) revealed that changes in culture medium chemistry, either from bacterial metabolism of medium nutrients or production of biosurfactants, contributed little to the observed level of dispersion. In addition to image-based methods (environmental and cryogenic scanning electron microscopy), filtration and dynamic light scattering were used to quantify the phenomenon.

The extinction of small (0.2- to 0.7-μm) abiotic nanoparticle agglomerates in the presence of P. aeruginosa (Fig. 4) suggests that biosorption sequesters dispersed nanoparticles, a result similar to that reported by Limbach et al. (21), where initially dispersed CeO₂ became associated with activated sludge bacteria in wastewater. This phenomenon may be more relevant when considering lower concentrations of TiO₂ or dilute aqueous chemistries that are not expected to facilitate extensive agglomeration. In this study, the high electrolyte concentration in LB growth media (5 g liter⁻¹ NaCl) promoted initial abiotic TiO₂ nanoparticle agglomeration and thus dispersion by bacteria was observable. Amino acids in LB media (1.4 to 19.1 mM) (33) also likely contributed to extensive TiO₂ agglomeration, perhaps analogous to the agglomeration of initially dispersed ZnS nanoparticles in the presence of cysteine (26). Yet dispersion of initially agglomerated TiO₂ nanoparticles was independent of cell growth and solely due to cells contacting agglomerates.

In another study, positively charged CeO₂ nanoparticles were sorbed to the negatively charged surface of E. coli (37). Here, both cell and industrial TiO₂ surfaces are negatively charged in LB media (−9.05 mV and −17.88 mV, respectively), and thus averaged electrostatic interactions between cells and nanoparticles would not account for their association. Cationic molecules within LB media could create cross-linkages between negatively charged cells and nanoparticles, perhaps analogous to aggregation of positively charged gold nanoparticles by citrate bridging (30). Alternatively, as generally described by Jucker et al. (14), adhesion could be mediated by bacterial surface polymers. Adsorption to TiO₂ has been observed for several bacterial cell surface polymers, including pyoverdine, a membrane-associated siderophore produced abundantly by P. aeruginosa (25) that binds TiO₂ via catechol bridges and can serve as a binding site for various metal ions (2). Cell surface lipopolysaccharides isolated from bacteria formed hydrogen bonds with TiO₂ (12), and negatively charged higher-order structures of bacterial lipopolysaccharide, i.e., micelles, sorbed onto negatively charged TiO₂ (13). Li and Logan (20) reported lipopolysaccharide length-dependent adhesion of E. coli to solid surfaces: strains with relatively long lipopolysaccharide polymers adhered more extensively to TiO₂ than strains with truncated lipopolysaccharide chains. Lipopolysaccharides were also involved in biosorption of gold (5), silver (35), and CeO₂ nanoparticles (37) to E. coli, suggesting that lipopolysaccharides play a general role in nanoparticle adhesion to cell surfaces. Extracellular polymeric substances increased adhesion of a Pseudomonas species to TiO₂ (20), though the mechanism was not reported. Sorption of TiO₂ nanoparticles with P. aeruginosa in this study may similarly be due to interactions with bacterial surface polymers, but further research in this area is needed.

In conclusion, this work presents new evidence that the preferential adsorption of TiO₂ nanoparticles to bacterial cell surfaces can disperse nanoparticles that are agglomerated prior to their interaction with bacteria. Bacterially mediated dispersion is not limited to LB medium, as dispersion was also observed in salt marsh water samples. These findings may therefore be transferable to natural waters, particularly those which promote extensive nanoparticle agglomeration and high bacterial density such as salt- and nutrient-rich environments (e.g., estuaries and salt marshes). Bacterial involvement in nanoparticle dispersion could be a generally important process for nanoparticle mobility in the environment, but extrapolation to other conditions is predicated upon nanoparticles agglomerating abiotically and then associating in nanoparticle-bacterium combinations.