Weakening Effect of Cell Permeabilizers on Gram-Negative Bacteria Causing Biodeterioration

Abstract

Gram-negative bacteria play an important role in the formation and stabilization of biofilm structures on stone surfaces. Therefore, the control of growth of gram-negative bacteria offers a way to diminish biodeterioration of stone materials. The effect of potential permeabilizers on the outer membrane (OM) properties of gram-negative bacteria was investigated and further characterized. In addition, efficacy of the agents in enhancing the activity of a biocide (benzalkonium chloride) was assessed. EDTA, polyethylenimine (PEI), and succimer (meso-2,3-dimercaptosuccinic) were shown to be efficient permeabilizers of the members of Pseudomonas and Stenotrophomonas genera, as indicated by an increase in the uptake of a hydrophobic probe (1-N-phenylnaphthylamine) and sensitization to hydrophobic antibiotics. Visualization of Pseudomonas cells treated with EDTA or PEI by atomic force microscopy revealed damage in the outer membrane structure. PEI especially increased the surface area and bulges of the cells. Topographic images of EDTA-treated cells were compatible with events assigned for the effect of EDTA on outer membranes, i.e., release of lipopolysaccharide and disintegration of OM structure. In addition, the effect of EDTA treatment was visualized in phase-contrast images as large areas with varying hydrophilicity on cell surfaces. In liquid culture tests, EDTA and PEI supplementation enhanced the activity of benzalkonium chloride toward the target strains. Use of permeabilizers in biocide formulations would enable the use of decreased concentrations of the active biocide ingredient, thereby providing environmentally friendlier products.

Stone monuments are subject to the deteriorative and degradative action of the environment and living organisms (12, 27, 45). Growing concern for the preservation of cultural heritage has boosted research on the biological attack on historical buildings (11, 15, 28). Biodeterioration processes result from complex interactions of surface-invading microbes with each other as well as with the surface material. Additionally, environmental factors and physiochemical properties of the surface material in question (e.g., porosity) determine the aggressiveness of the deterioration process and, as a consequence, its result (11). Besides being a potential cause of decay, soiling induced by biological growth results in aesthetical disfiguration of the stone and causes both physical and chemical damage on stone monuments.

Understanding the complex microbial ecosystem of building materials is a prerequisite for controlling the growth of microbial species causing biodegradation. Organisms present on stone monuments include photolithoautotrophs, such as algae and cyanobacteria, chemolithoautotrophic bacteria, mosses, and higher plants (11, 15, 45). The phototrophs algae and cyanobacteria have been considered the primary colonizers of building surfaces, conditioning the surfaces and excreting nutrients and growth factors for heterotrophic microbes (11). A majority of the microbes persist on building surfaces within complex microbial communities and a structured biofilm ecosystem (18), which provides shelter for the microbes. In addition, the endolithic environment, the pore space of rocks, has been reported to be a microhabitat giving protection from intense solar radiation and desiccation as well as providing mineral nutrients, rock moisture, and growth surface (42). Extracellular polysaccharides especially play various roles in the structure and function of different biofilm communities: e.g., excluding and/or influencing the penetration of antimicrobial agents and providing protection against a variety of environmental stresses, such as UV radiation, pH shifts, osmotic shock, and desiccation (10, 13, 35). Besides phototrophic cyanobacteria, many other gram-negative bacterial species, e.g., members of Pseudomonas, Stenotrophomonas, and Sinorhizobium genera, have been isolated from biodeteriorated stone samples (11, 38). Since members of these genera are potential extracellular polysaccharide produces (35), prevention of their growth or adhesion to stone materials would provide means to diminish biofilm formation on stone surfaces.

An additional factor making the prevention of gram-negative bacterial growth extremely difficult is related to the structure of the gram-negative cell envelope (14, 26, 37). The outer membrane (OM) of gram-negative bacteria acts as a permeability barrier that is able to exclude macromolecules and hydrophilic substances, thereby being responsible for the intrinsic resistance of these bacteria to antimicrobial compounds (14, 30, 31). In gram-negative bacteria, the barrier function of the OM is mainly due to the presence and features of lipopolysaccharide (LPS) molecules in the outer leaflet of the membrane, along with various multidrug efflux pumps that also contribute to the resistance of the cells (31, 32, 33). Pseudomonas species especially have been reported to be resistant to many biocides and antimicrobial agents (36, 37). According to Walsh et al. (43), the inner core phosphates of P. aeruginosa appear to play a key role in the intrinsic drug resistance of this bacterium.

Although the OM of gram-negative bacteria protects the cells from many external agents, it is possible to specifically weaken it by various agents, collectively called permeabilizers, which disintegrate the LPS layer and increase the permeability of the OM (40). The classical example is the chelator EDTA, which sequesters divalent cations that contribute to the stability of the OM by providing electrostatic interactions with proteins and LPS (2, 40). Besides EDTA, a number of other permeabilizers are known, some of which act quite differently. Polyethyleneimine (PEI), a cationic polymer, has been recognized as a permeabilizer acting by intercalating into the OM rather than releasing LPS (19, 20). Succimer (meso-2,3-dimercaptosuccinic; DMSA) is an active heavy-metal-chelating agent used, e.g., to treat lead poisoning in humans (16). Bansal-Mutalik and Gaikar (8) reported that sodium bis-(2-ethylhexyl) sulfosuccinate (AOT) was capable of permeabilizing Escherichia coli cells and causing selective enzyme (penicillin acylase) release. Nitrilotriacetic acid (NTA) is a complexing agent of the same general type as EDTA, and it has been reported to be a permeabilizer and to increase the sensitivity of gram-negative bacteria to hydrophobic antibiotics (5, 40).

Mechanistic studies on the action of antimicrobial chemicals advance our understanding of their potential application. The objective of this study was to determine and characterize the effect of selected permeabilizers on the OM of environmental gram-negative bacteria isolated from biodeteriorated surfaces. We especially wanted to study the activity of PEI on the OM of Pseudomonas. To our knowledge, this is the first study using atomic force microscopy (AFM) to study the antimicrobial mechanisms of PEI and EDTA, well-known permeabilizers against Pseudomonas. Furthermore, the effect of permeabilizers on the activity of a biocide, benzalkonium chloride, was studied.

MATERIALS AND METHODS

Chemicals.

HEPES, n-heptadecanoic acid methyl ester, 1-N-phenylnaphthylamine (NPN), polyethylenimine (PEI; mean molecular mass, 70 kDa), meso-2,3-dimercaptosuccinic acid (DMSA), nitrilotriacetic acid (NTA), sodium bis-(2-ethylhexyl) sulfosuccinate (AOT), and benzalkonium chloride (BC) were from Sigma-Aldrich (Steinheim, Germany); EDTA was from Riedel-de-Haen (Seelze, Germany). A stock solution of NPN (0.5 M) was prepared in acetone and diluted to 40 μM into 5 mM HEPES (pH 7.2) for the fluorometric assays. A stock solution of DMSA was prepared in ethanol and of NTA in 1 M NaOH. Solvents were included as controls in the experiments.

Bacterial strains.

The bacterial strains (Table 1) used for the studies were isolated from biodeteriorated mineral materials and deposited at the VTT culture collection as Sinorhizobium morelense (=Ensifer adhaerens) VTT E-022105 (later E2105); Pseudomonas sp. strains VTT E-022106 (E2106), E-022217 (E2217), E-052906 (E2906), and E-052911 (E2911); and Stenotrophomonas nitritireducens E-022107 (E2107). They were identified by partial 16S rRNA gene sequencing according to Saarela et al. (38). The working cultures were stored at −70°C and cultivated on trypticase soy agar (TSA; Oxoid, Basingstoke, United Kingdom) at 25°C. For permeability assays, cells were grown in Luria-Bertani broth (LB) as described by Helander et al. (19). Cultivations were carried out at 25°C with shaking (150 rpm, unless otherwise stated). Growth was monitored by measuring the A₆₃₀ with a Multiskan MCC/340 spectrophotometer (ThermoLabSystems, Helsinki, Finland). Further details of cell treatments are given below under various experimental settings.

TABLE 1.

Target strains used in the study

Strain	VTT code	Identificationb	Isolation site
Sinorhizobium morelense	E-022105	100% S. morelenseT	Marble on fresco, Italy
Pseudomonas sp.a	E-022106	98.3% P. putidaT/“P. gingerii”T/P. fuscovaginaeT/P. aspleniiT	Marble on fresco, Italy
Stenotrophomonas nitritireducens	E-022107	99.8% S. nitritireducens	Marble on fresco, Italy
Pseudomonas sp.a	E-022217	98.9% P. jesseniiT, 98.0% P. putidaT	Marble on fresco, Italy
Pseudomonas sp.a	E-052906	98.8% P. mendocina	Sandstone, Scotland
Pseudomonas sp.a	E-052911	98.9% P. mandeliiT	Sandstone, Scotland

^aProbably a new species.

^bGenBank accession numbers for partial 16S rRNA gene sequences are, in descending order, DQ465005, DQ465006, DQ465007, DQ465008, DQ465009, and DQ465010.

Permeability assays.

Two different methods were utilized to determine permeability properties of the OM: (i) NPN uptake and (ii) susceptibility to hydrophobic antibiotics.

(i) NPN uptake.

NPN uptake by bacterial suspensions was measured using black fluorotiter plates (Catalog no. 9502 867; ThermoLabSystems, Helsinki, Finland) and the automated fluorometer Fluoroskan Ascent FL (ThermoLabSystems) as described earlier (1, 21). Briefly, cells grown to an A₆₃₀ of 0.5 ± 0.02 were deposited by centrifugation at room temperature for 10 min at 1,000 × g and suspended into a half volume of 5 mM HEPES buffer (pH 7.2). Aliquots (100 μl) of this cell suspension were pipetted into fluoroplate wells, which contained NPN (10 μM) and, as test substances, either EDTA (1.0 and 0.1 mM), PEI (10 μg ml⁻¹), DMSA (1 mM), AOT (1 mM), or HEPES buffer (control) to make up a total volume of 200 μl. If desired, MgCl₂ was added to the cell suspension before addition of NPN. Fluorescence was monitored within 3 min from four parallel wells per sample (excitation, 355 nm; half bandwidth, 38 ± 3 nm; emission, 402 nm; half bandwidth, 50 ± 5 nm). Each assay was performed at least three times.

(ii) Antibiotic susceptibility and growth inhibition tests.

The susceptibility of bacterial cultures to hydrophobic antibiotics was tested with the agar diffusion method on Iso-Sensitest agar (Oxoid, Basingstoke, Hampshire, England) with or without PEI supplementation (5 to 250 μg ml⁻¹) using Neo-Sensitab discs (erythromycin, novobiocin, clindamycin, fucidin, and rifampin; Rosco Diagnostica, Taastrup, Denmark). The diameters of inhibition zones were measured after incubation of the plates at 25°C for 24 and 48 h. All determinations were performed with two replicates, and results are presented as mean values. Further susceptibility tests with PEI and combinations of PEI and novobiocin were performed in liquid cultures as described by Helander et al. (19) with an automated turbidometer (Bioscreen C; ThermoLabSystems). Microbiological growth curve data was collected and analyzed with Research Express software (Transcalactic Ltd., Helsinki, Finland).

Results from the permeability assays were analyzed statistically using two-tailed unpaired Student's t tests to determine differences.

Atomic force microscopy (AFM) studies.

Pseudomonas sp. strain E2106 cells grown in LB to an A₆₃₀ of 0.8 ± 0.02 were deposited by centrifugation at room temperature for 10 min at 1,000 × g and washed with 10 mM Tris-HCl buffer (pH 7.2), and the optical density of the suspension was adjusted to an A₆₃₀ of 0.5 ± 0.02 with the same buffer. Cells were harvested by centrifugation (1,000 × g, 10 min at room temperature). For the treatments resuspended into either buffer alone or buffer supplemented with 1 mM EDTA or 10 μg of PEI mg⁻¹, cells were treated at 25°C for 10 min with shaking (150 rpm), harvested by centrifugation (10,000 × g) in an Eppendorf microcentrifuge for 1 min at room temperature, washed with ultrapure water, harvested by centrifugation as described above, and resuspended into sterile ultrapure water. Analysis was done with duplicate cultures.

For AFM analysis, the treated cells were applied on a freshly cleaved mica surface and allowed to dry before imaging (9, 29). To determine the effect of the treatments on the cell membrane, an average of four images on different areas for each sample were imaged. The images were acquired in air under ambient conditions using a NanoScope IIIa Multimode AFM (Digital Instruments, Santa Barbara, CA) equipped with a “J”-scanner. The tapping mode was used with scan rates of 0.5 to 1.2 Hz and as little force as possible, and the ratio of set point amplitude and free amplitude was usually 0.8 to 0.9 with a target amplitude 1 V. Noncontact silicon cantilevers (NSC15/AlBS; μMasch) with the nominal resonance frequency of 350 kHz and a tip radius better than 10 nm were used. The topography and phase-contrast images were captured simultaneously. The phase-contrast image shows the phase difference between the oscillations of the cantilever-driving piezo and the detected oscillations. Nanoscope III 5.12r2 software (Digital Instruments) was used in image processing, which only included flattening in order to remove possible tilt in the image data. The average surface root-mean-square roughness of the treated cells was calculated with Nanoscope III 5.12r2 software from five replicate images with a resolution of 512 pixels.

Microtiter plate assay for biofilm formation (BF assay).

Efficacy of EDTA, PEI, DMSA, and benzalkonium chloride on the biofilm formation of the target strains was assayed on 96-well microtiter plates (Nunclon 167008; Nalge Nunc International) with a protocol modified from Kolari et al. (23). Inoculum for the assay was grown overnight in trypticase soy broth (TSB; Oxoid; 25°C, 150 rpm) and diluted into TSB to obtain a cell density of 10⁵ CFU ml⁻¹. Briefly, each of the microtiter plate wells was filled with a total volume of 250 μl with TSB, test agent (25 μl), and bacterial inoculum (25 μl). The plates were placed on a rotary shaker (120 rpm, 25°C) for 3 days. The wells were emptied, stained with 300 μl of crystal violet (4 g liter⁻¹ in 20% [vol/vol] methanol) for 3 min, washed three times under running tap water to remove planktonic cells, and allowed to dry in air. Stain retained by the biofilm was dissolved in ethanol (330 μl per well, 1 h), and the A₅₉₅ was measured with a Multiskan MCC/340 spectrophotometer. All determinations were performed with three replicates, and results are presented as mean values. Each assay was performed three times.

Testing of the enhancement of in vitro antimicrobial activity.

The capability of permeabilizers (EDTA and PEI) to enhance the activity of benzalkonium chloride was further assayed with an automated turbidometer (Bioscreen; ThermoLabSystems) according to Raaska et al. (34). Briefly, each of the microtiter plate wells was filled with test agents (30 μl) and a dilution of the inoculum (30 μl; initial density, 10⁴ CFU ml⁻¹), and then they were filled with TSB to a total volume of 300 μl. The microtiter plates were incubated at 25°C for 48 h, and the optical density at 600 nm (OD₆₀₀) was measured every 10 min. Microbiological growth curve data were collected and analyzed with Research Express software (Transcalactic Ltd., Helsinki, Finland). The area under the growth curve was used as a measure of growth. All determinations were performed with five replicates, and results are presented as mean values. Each assay was performed three times.

Nucleotide sequence accession numbers.

The sequences determined in the course of this work were deposited in GenBank under accession numbers DQ465005, DQ465006, DQ465007, DQ465008, DQ465009, and DQ465010.

RESULTS

Effect of tested samples on the uptake of NPN.

Environmental isolates were selected for NPN uptake studies in order to reveal possible changes in the OM permeability of gram-negative bacteria causing biodeterioration. NPN fluorescence is associated with the presence of this hydrophobic probe in a glycerophospholipid environment (20), and increased fluorescence values indicate weakening of the OM. The detailed results of the NPN uptake experiments with EDTA, PEI, DMSA, NTA, AOT, and benzalkonium chloride (BC) are presented in Table 2. EDTA (1 mM), PEI (10 μg ml⁻¹), and DMSA (1 mM) brought about a significantly higher NPN uptake than that of control treatments with all other strains except S. morelense E2105. Benzalkonium chloride (0.001%, wt/vol) weakened the outer membrane of the tested microbes, as indicated by a significant increase in the NPN uptake. Addition of 1 mM MgCl₂ into the buffer used in the NPN assay abolished the permeabilizing activities of 0.1 mM EDTA and diminished the permeabilizing activity of PEI. The OM-destabilizing activity of 1 mM EDTA and DMSA was only partially abolished by MgCl₂. DMSA (1 mM) supplementation resulted in pH 4.5 (±0.2) in the test assay, whereas for the other treatments the pH remained at 7.0 (±0.2). NTA at 1 mM concentrations did not significantly increase the NPN uptake of the target strains. AOT (1 mM) increased the NPN uptake of Pseudomonas sp. strain E2906, whereas NPN uptake of other microbes was only slightly affected.

TABLE 2.

NPN uptake induced by EDTA, PEI, DMSA, NTA, AOT, and BC

Strain	MgCl2 (mM) concn	Relative fluorescence (±SD) with the following additive to assay buffera:
		Control	EDTA (1 mM)	EDTA (0.1 mM)	PEI (10 μg ml−1)	DMSA (1 mM)	NTA (1 mM)	AOT (1 mM)	BC (0.001%)
Sinorhizobium morelense
E2105	0	467 ± 72	443 ± 103	415 ± 192	422 ± 16	483 ± 74	327 ± 151	366 ± 70	512 ± 79
E2105	1	256 ± 45	428 ± 78	260 ± 63	257 ± 39	494 ± 67	253 ± 49	333 ± 5*	410 ± 15**
Stenotrophomonas nitritireducens
E2107	0	208 ± 29	303 ± 50*	338 ± 60	293 ± 19*	359 ± 89	287 ± 64	264 ± 44	348 ± 36**
E2107	1	78 ± 12	157 ± 24	76 ± 13	189 ± 12**	324 ± 70*	65 ± 11	199 ± 54	280 ± 69**
Pseudomonas sp.
E2106	0	185 ± 13	354 ± 16***	284 ± 36	324 ± 32**	367 ± 77	265 ± 72	143 ± 9	377 ± 38**
E2106	1	122 ± 11	174 ± 49	130 ± 52	194 ± 40	332 ± 71	150 ± 34	127 ± 12	310 ± 49**
E2217	0	54 ± 29	286 ± 26***	128 ± 20	278 ± 15**	265 ± 50*	65 ± 8	148 ± 2**	265 ± 55***
E2217	1	72 ± 5	56 ± 27	67 ± 6	77 ± 1	211 ± 50**	63 ± 6	122 ± 51*	119 ± 8*
E2906	0	223 ± 53	525 ± 106**	483 ± 76***	390 ± 47**	471 ± 91**	249 ± 58	327 ± 78**	455 ± 21**
E2906	1	87 ± 15	160 ± 47	86 ± 17	269 ± 51*	468 ± 83**	77 ± 14	190 ± 45**	360 ± 9**
E2911	0	143 ± 55	405 ± 73***	352 ± 92***	268 ± 17**	300 ± 42***	120 ± 33	237 ± 100	332 ± 23***
E2911	1	92 ± 13	137 ± 40	100 ± 10	102 ± 40	193 ± 42**	91 ± 11	112 ± 73	143 ± 12

^aThe pH of all treatments was 7.0 ± 0.2, except for the pH of the DMSA treatment, which was 4.5 ± 0.2. *, P < 0.02 compared with the control; **, P < 0.01 compared with the control; ***, P < 0.001 compared with the control.

Antibiotic susceptibility and growth inhibition tests.

A sensitizing effect to hydrophobic antibiotics is one of the indications of OM-permeabilizing action. We tested the susceptibility of the target strains to a set of hydrophobic antibiotics (clindamycin, rifampin, novobiocin, erythromycin, and fucidin) by the agar diffusion method on plates containing different concentrations of PEI. The results are summarized in Table 3, which shows that PEI induced an increased susceptibility of Pseudomonas sp. strain E2106 to erythromycin, novobiocin, and fusidin. The susceptibility of S. nitritireducens E2107 was not, however, significantly enhanced by PEI addition on agar plate tests for these antibiotics (Table 3). PEI supplementation slightly increased the susceptibility of target strains to rifampin. Susceptibility of S. morelense E2105 was only slightly enhanced for novobiocin by PEI supplementation.

TABLE 3.

Effect of PEI on the susceptibility of the target strains to selected antibiotics as determined by the agar diffusion method

Strain and antibiotic	Diam of inhibition zone (mm) with PEI concn (μg ml−1) of:
	0	25	50	250
Sinorhizobium morelense E2105
Clindamycin, 25 μg	21	20	20	NDa
Rifampin, 30 μg	36	40	42	ND
Novobiocin, 100 μg	31	38	38	ND
Erythromycin, 78 μg	25	25	22	ND
Fucidin, 100 μg	20	23	26	ND
Pseudomonas sp. strain E2106
Clindamycin, 25 μg	0	0	0	0
Rifampin, 30 μg	22	24	29	36
Novobiocin, 100 μg	0	12	20	30
Erythromycin, 78 μg	0	0	15	20
Fucidin, 100 μg	0	12	17	18
Stenotrophomonas nitritireducens E2107
Clindamycin, 25 μg	12	17	12	ND
Rifampin, 30 μg	35	41	40	ND
Novobiocin, 100 μg	0	0	0	ND
Erythromycin, 78 μg	21	22	22	ND
Fucidin, 100 μg	17	23	22	ND

^aND, no data (no growth on agar plates supplemented with 250 μg ml⁻¹ of PEI).

The effect of PEI on the growth of Pseudomonas sp. strain E2106 and S. nitritireducens E2107 was also tested using a Bioscreen automated turbidometer (Fig. 1). In the agar diffusion test, a larger concentration of PEI was required (>25 μg ml⁻¹) than in the suspensions (10 μg ml⁻¹) for the sensitization of Pseudomonas sp. strain E2106 cells to novobiocin (Fig. 1a). In suspension experiments, supplementation by 10 μg of PEI ml⁻¹ enhanced the susceptibility of Pseudomonas sp. strain E2106 and S. nitritireducens E2107 cells to novobiocin (already with 10 μg ml⁻¹ of novobiocin). However, the growth of S. nitritireducens E2107 was not fully prevented by the combination of PEI and novobiocin even at higher concentrations tested (Fig. 1b).

FIG. 1.

Enhancement of susceptibility of Pseudomonas sp. strain E2106 (a) and Stenotrophomonas nitritireducens E2107 (b) to novobiocin by PEI. Bacterial growth, expressed as the optical density at 600 nm (OD_600nm), was measured for 48 h. Symbols: ○, control; •, 10 μg of novobiocin ml⁻¹; □, 50 μg of novobiocin ml⁻¹; ▪, 100 μg of novobiocin ml⁻¹; ⧫, 10 μg of PEI ml⁻¹; ▴, 50 μg of PEI ml⁻¹; ⋄, 100 μg of PEI ml⁻¹; Δ, 10 μg of novobiocin and 10 μg of PEI ml⁻¹; _, 10 μg of novobiocin and 50 μg of PEI ml⁻¹; ×, 10 μg of novobiocin and 100 μg of PEI ml⁻¹.

AFM.

Topographic images of the control Pseudomonas sp. strain E2106 cells revealed a compact and smooth surface without notable ruptures or pores on the cell surface (Fig. 2a).

FIG.2.

Surface of control (A)-, 1 mM EDTA (B)-, and 10 μg ml⁻¹ PEI (C)-treated Pseudomonas sp. strain E2106 cells visualized by atomic force microscopy (AFM). Left side, topographic image; right side, phase image.

Phase-contrast images of the control cells revealed that the hydrophilic surface was uniform. Magnification of the topographic images also revealed a uniform OM structure (Fig. 3a). The average surface root-mean-square roughness for the control cells was 2.06 ± 0.45. Figure 2b shows topographic and phase-contrast images of 1 mM EDTA-treated cells. The surfaces of cells visualized in topographic images were rough, and the outer membrane surface appeared damaged, indicating release of LPS and weakening of OM structure. Phase-contrast images revealed large areas with different hydrophilicity/hydrophobicity on the cell surface. The magnification of the topographic images showed extensive disruption of the LPS layer (Fig. 3b). The average surface roughness of EDTA-treated cells was 3.23 ± 0.49 (P < 0.05 compared to the control cells). The release of LPS from the surface of EDTA-treated cells resulted in large and irregularly shaped pits where the cytoplasmic membrane was revealed. The effect of PEI (Fig. 2c) was different from that of EDTA. Treatment of the cells with PEI flocculated the Pseudomonas cells, causing aggregation and adhesion of the cells. In addition, cells were swollen, with increased cell surface area and bulges. Magnification of the topographic image showed smooth OM surfaces with bulges and an increased surface roughness compared to that of control cells (Fig. 3c). Surface roughness of PEI-treated Pseudomonas cells was 7.48 ± 1.46 (P < 0.001 compared to the control cells).

FIG. 3.

Surface of control (A)-, 1 mM EDTA (B)-, and 10 μg ml⁻¹ PEI (C)-treated Pseudomonas sp. strain E2106 cells visualized by atomic force microscopy (AFM) in three-dimensional mode.

Prevention of adhesion and biofilm formation on polystyrene plates.

Figure 4 shows the quantification of biofilm formation by six environmental isolates in the BF assay. The most effective biofilm formers were S. nitritireducens E2107 and Pseudomonas sp. strains E2106, E2906, and E2911. EDTA at 1 mM concentration prevented biofilm formation of the tested strains. Benzalkonium chloride at a concentration of 0.01% significantly prevented biofilm formation of all tested strains compared to that of the control treatment. Also, a lower benzalkonium chloride concentration (0.001%) diminished the biofilm formation of the strains compared to that of control treatments, with Pseudomonas sp. strain E2106 being less affected than the other strains. DMSA at a concentration of 1 mM prevented the biofilm formation of S. morelense E2105 and Pseudomonas sp. strains E2906 and E2911. Supplementation by PEI (10 μg ml⁻¹) did not significantly decrease the biofilm formation compared to that of the control.

FIG. 4.

Efficacy of different permeabilizers (BC, MDSA, EDTA, and PEI) on prevention of biofilm formation on polystyrene plates as determined by crystal violet staining.

Enhancement of in vitro antimicrobial activity.

The capability of selected permeabilizers to increase efficacy of benzalkonium chloride (BC) in suspensions is shown in Table 4. BC alone had a minor growth-inhibitory activity against the Pseudomonas sp. strains E2106 and E2217. Supplementation with PEI (10 μg ml⁻¹) significantly increased (P < 0.001) the activity of BC toward the tested Pseudomonas strains, whereas EDTA (0.1 mM) supplementation did not increase the activity. In the test assay, supplementation by PEI alone diminished the growth of Pseudomonas sp. strain E2106.

TABLE 4.

Effect of permeabilizers on the enhancement of benzalkonium chloride (0.001%) activity in suspensions^a

Tested sample	Area under the growth curve ± SD for:
	Pseudomonas sp. strain E2106	Pseudomonas sp. strain E2217
Control	38 ± 1	41 ± 2
EDTA, 0.1 mM (pH 7)	35 ± 1*	43 ± 1
PEI, 10 μg ml−1 (pH 7)	20 ± 1***	40 ± 2
BC, 0.001% (wt/vol)	35 ± 0**	34 ± 2**
BC + EDTA	33 ± 3	32 ± 3**
BC + PEI	13 ± 5***	21 ± 2***

^aResults are presented as areas under the growth curve (five replicates). *, P < 0.02 compared with the control; **, P < 0.01 compared with the control; ***, P < 0.001 compared with the control.

DISCUSSION

Biocides must traverse the outer cell layer(s) of microbes to reach their target sites, usually present within microbial cells (36). Therefore, permeabilizers can enhance the activity of biocides and other antimicrobial agents, thus enabling application of a reduced amount of biocide. Among the potential permeabilizers examined in this study, EDTA, PEI, and succimer (DMSA) were shown to be efficient permeabilizers for the members of Pseudomonas and Stenotrophomonas genera, as indicated by the increase in the uptake of hydrophobic probe (NPN). This is in agreement with our earlier study where EDTA and DMSA destabilized the outer membrane of Pseudomonas aeruginosa E-97041, as indicated by an increased NPN uptake (3).

Intrinsic and acquired multidrug resistance in gram-negative bacteria is related to the synergy between limited OM permeability and energy-dependent multidrug efflux pumps (32). S. morelense is an opportunistic pathogen and has been reported to be highly resistant to several antibiotics (44). Our environmental S. morelense isolate, E2105, was sensitive to clindamycin, rifampin, novobiocin, erythromycin, and fucidin. Addition of PEI slightly increased the susceptibility of this strain to novobiocin. However, our target strain S. morelense E2105 seemed to have a weak OM structure, since in the NPN uptake assay the uptake values were already high in control cells and no statistically significant difference between various treatments was observed. This weak structure was likely related to the number of stabilizing divalent cations in the OM, since MgCl₂ addition stabilized the control cells but the permeabilizing activity of EDTA was not completely abolished by the MgCl₂ addition.

Pseudomonas species are able to degrade chloride compounds, and therefore they are not very sensitive to quaternary ammonium compounds (10, 17). Loughlin et al. (25) reported that P. aeruginosa cells generated stable resistance to benzalkonium chloride during passage in concentrations beneath the MIC of BC, and this resistance was also later retained in the absence of the disinfectant. In addition, a cross-resistance to the membrane-active antibiotic polymyxin B was also detected. In our studies, in the biofilm formation assay Pseudomonas sp. strain E2106 was the most resistant to benzalkonium chloride of the strains tested. EDTA and PEI enhanced the activity of benzalkonium chloride in suspension experiments toward Pseudomonas. In addition, our studies showed that benzalkonium chloride disintegrated the OM of the target cells, as indicated by an increased NPN uptake. Recently, it was reported that EDTA at high (50 mM) concentration caused rapid dispersion of P. aeruginosa cells from biofilms by chelation of several divalent cations that are required to stabilize the biofilm matrix (7).

Bansal-Mutalik and Gaikar (8) reported that AOT was capable of permeabilizing Escherichia coli cells and causing selective enzyme (penicillin acylase) release. In our study, AOT (1 mM) increased the NPN uptake of Pseudomonas sp. strain E2906, whereas NPN uptake of other microbes was only slightly affected. Nitrilotriacetic acid (NTA) has been reported to increase the sensitivity of gram-negative bacteria to hydrophobic antibiotics (5, 40). In our study, 1 mM NTA weakly destabilized Pseudomonas sp. strain E2106 and S. nitritireducens E2107 cells. However, NTA has been classified as possibly carcinogenic (4), and therefore it is not suitable to be used in biocide formulations intended for environmental applications, although it might have other application areas.

Succimer (DMSA) has been reported to be a potential remover of smear layers in dental applications (41). In our study, succimer was capable of destabilizing the OM of all tested strains. In the NPN uptake assay, MgCl₂ addition only slightly abolished the OM disintegrating activity of DMSA, indicating that OM disintegrating activity of DMSA was only partially related to the removal of stabilizing divalent cations from the OM. Succimer is a hydroxy acid compound, and thereby part of the permeabilizing activity is related to the acidity and structure of the compound. Lactic acid, another hydroxy acid, has been shown to be a potent permeabilizer (1).

Polyethylenimine (PEI) is a weakly basic aliphatic polymer which is polycationic due to the presence of primary, secondary, and tertiary amino groups (6). PEIs are available in different molecular masses and forms, and they are widely utilized as protein and nucleic acid precipitants in process industry (39). Helander et al. (19) demonstrated that PEI is a potent permeabilizer of the OM of pathogenic gram-negative bacteria, as PEI sensitized E. coli, P. aeruginosa, and Salmonella enterica serovar Typhimurium to hydrophobic antibiotics and detergents. Helander et al. (20) also demonstrated that PEI intercalated in the OM and increased the membrane surface area without liberation of LPS-associated cell material from pathogenic gram-negative bacteria. Our study confirms that PEI is also capable of permeabilizing gram-negative environmental strains, representing Pseudomonas and Stenotrophomonas species, since significant NPN uptake and increased sensitivity to hydrophobic antibiotics was observed with these strains.

AFM images displayed massive changes on the OM of Pseudomonas sp. strain E2106 due to PEI treatment. This is not surprising, as Pseudomonas lipopolysaccharides typically are rich in phosphate groups (22, 43), and Pseudomonas cell surface is thus expected to bind polycationic PEI in large amounts. In PEI-treated cells, AFM images visualized the capability of PEI to intercalate in the OM and increase the membrane surface area. This observation is in agreement with the results of Helander et al. (20), who reported the same phenomenon in Salmonella by using transmission electron microscopy. Kotra et al. (24) studied the effect of EDTA on E. coli with AFM, and they reported that release of the LPS from the surface results in large and irregular-shaped pits where the peptidoglycan layer was exposed. Our AFM images from EDTA-treated Pseudomonas cells also revealed patchiness of the damaged OM structure. This nonuniform alteration of the OM by EDTA as revealed by AFM is in accordance with the classical findings that only a certain proportion of LPS can be released by EDTA, indicating the presence of structurally and electrostatically different subpopulations of LPS in the OM. The existence of such structurally distinct LPS populations in spatially separate areas of the OM, as discussed in more detail by Alakomi et al. (2), is further supported by our present findings with AFM.

Alternative and novel biocide formulations are needed to restrict the growth of harmful microbes in sites where traditional biocides or construction alternatives are ineffective or impossible to implement. The application of an effective biocide/permeabilizer combination could aid in the destruction of the microbial biofilms that cause the degradation while allowing the use of reduced concentrations of the biocide. In order to be able to enhance the activity of biocides, e.g., with permeabilizers, knowledge of the mechanism of permeabilizers and factors influencing their activity is essential. However, the efficacy of the formulated biocide products must be further evaluated on stone materials with complex microbial communities and in field trials under various environmental conditions. In addition, compatibility of the formulated products with commercial restoration products such as water repellents and consolidation agents has to be ensured.