Anti-Biofilm Activity of a Low Weight Proteinaceous Molecule from the Marine Bacterium Pseudoalteromonas sp. IIIA004 against Marine Bacteria and Human Pathogen Biofilms

Pseudoalteromonas bacteria are known as potential bioactive metabolite producers. Because of the need to obtain natural molecules inhibiting the bacterial biofilms, we investigated the biofilm inhibitory activity of the marine bacterium Pseudoalteromonas sp. IIIA004 against the pioneer surface colonizer Roseovarius sp. VA014. The anti-biofilm activity from the culture supernatant of Pseudoalteromonas sp. IIIA004 (SNIIIA004) was characterized in microtiter plates (static conditions/polystyrene surface) and in flow cell chambers (dynamic conditions/glass surface). The Pseudoalteromonas exoproducts exhibited an inhibition of Roseovarius sp. VA014 biofilm formation as well as a strong biofilm dispersion, without affecting the bacterial growth. Microbial adhesion to solvent assays showed that SNIIIA004 did not change the broad hydrophilic and acid character of the Roseovarius strain surface. Bioassay-guided purification using solid-phase extraction and C18 reverse-phase-high-performance liquid chromatography (RP-HPLC) was performed from SNIIIA004 to isolate the proteinaceous active compound against the biofilm formation. This new anti-biofilm low weight molecule (< 3kDa), named P004, presented a wide spectrum of action on various bacterial biofilms, with 71% of sensitive strains including marine bacteria and human pathogens. Pseudoalteromonas sp. IIIA004 is a promising source of natural anti-biofilm compounds that combine several activities.


Introduction
In the marine environment, submerged surfaces are the subject of active bacterial colonization. Once attached to the substratum, the bacterial communities rapidly form biofilms and secrete extracellular polymeric substances (EPS) which are major components of the biofilm matrix. Rich in polysaccharides, proteins, lipids, DNA, RNA, and water, this matrix protects the microbial cells against stress, antibiotics, host immune system, and insures the stabilization of biofilms [1][2][3][4].
Biofilms are involved in several infectious diseases, both in humans and animals, and are present in a wide range of ecosystems, such as food industries, medical equipment, and natural environments [5][6][7]. In the marine environment, biofilms on submerged surfaces serve as reservoirs for Table 1. Strains used in this study.

Strain
Reference and/or Source Culture Conditions

Preparation of the Pseudoalteromonas sp. IIIA004 Supernatant (SN IIIA004 )
Pseudoalteromonas sp. IIIA004 was grown overnight at 22 • C in Zobell broth supplemented with 30 g L −1 of glucose with shaking for 48 h to optimize the production of antibiofilm compounds. The supernatant, named SN IIIA004 , was harvested by centrifuging the culture (15 min, 7000× g at 4 • C), filter sterilized through 0.22 µm (Millipore PVDF), and stored at −80 • C until use. For some experiments, SN IIIA004 was concentrated 10-fold by lyophilization at a pressure below 450 mTorr at −80 • C (Cryotec freeze-dryer) and named 10X SN IIIA004 .

Anti-Biofilm Assays
Microtiter plate assay (static conditions/polystyrene surface). Bacterial biofilms (marine and non-marine bacteria) were grown in microtiter plates as previously described by Doghri et al. [32]: an overnight bacterial culture was centrifuged 10 min at 7000 g and resuspended in artificial seawater (sea salts Sigma-Aldrich, 35 g L −1 ) for marine strains or in saline solution (NaCl 9 g L −1 ) for non-marine strains, to a final optical density at 600 nm (OD 600 ) of 0.25. A total of 150 µl of the resulting suspensions were then loaded per well of a 96-well microtiter plates (MICROTEST TM 96, Falcon). Artificial seawater or saline solution, without bacteria, served as negative controls. After a bacterial adhesion step of 2 h at 22 • C (marine bacteria) or 37 • C (non-marine bacteria), the wells were gently washed three times with artificial seawater or saline solution, respectively, and 150 µl of Zobell or LB medium, respectively, were added to each well. After incubation at 22 • C or 37 • C for 24 h, the microplates were washed three times. The bacterial biofilms were then stained with a 0.8% w/v crystal violet solution for 20 min and rinsed with ultra-pure water until the wash-liquid was clear. Crystal violet was then eluted from attached cells with 96% ethanol (150 µL per well) and the quantification was carried out by measuring the OD 595nm .
To investigate the effect of SN IIIA004 on bacterial adhesion and biofilm formation, wells were inoculated with biofilm-forming cells resuspended in a solution of 50% v/v SN IIIA004 and 50% v/v artificial seawater. After a 2 h adhesion step, biofilm formation was performed as described above. In this experiment, the culture supernatant was replaced with sterile culture medium in the negative control. Three independent experiments were performed, and for each experiment, the test was repeated in at least three wells per microtiter plate.
Flow cells assay (dynamic conditions/glass surface). Roseovarius sp. VA014 was grown on glass slides in three-channel flow cells (channel dimensions: 1 by 4 by 40 mm) (Technical University of Denmark Systems Biology), as previously described [32]: the flow cells were inoculated with overnight bacterial cultures diluted in artificial seawater to a final OD 600 of 0.1. Bacteria were allowed to attach to the substratum (microscope glass coverslip of 24 × 50 st1, Knittel Glasser) during 2 h at 22 • C without a flow of medium. The channels were then washed to remove non-attached bacteria by applying a flow of artificial seawater for 15 min at a rate of 2 mL h −1 and biofilm growth was performed under a constant flow (2 mL h −1 ) of Zobell broth for 24 h at 22 • C.
To investigate the effect of SN IIIA004 on adhesion and biofilm formation of Roseovarius sp. VA014, several protocols were followed. (i) A solution of 50% v/v SN IIIA004 and 50% v/v artificial seawater was injected without bacteria into the flow cell channels and left for 2 h at 22 • C without flow to coat the glass surface. The channels were then rinsed with artificial seawater before inoculating Roseovarius sp. VA014 bacteria and growing biofilms as described before. (ii) Flow cells were inoculated with Roseovarius sp. VA014 cultures grown for 24 h and resuspended in a solution of 50% v/v SN IIIA004 and 50% v/v artificial seawater to a final OD 600 of 0.1. After the 2-h adhesion step, biofilm formation was performed as initially described. (iii) SN IIIA004 was injected into the channels after the Roseovarius sp. VA014 biofilm maturation step and left for 2 h under static conditions at 22 • C. For each experiment, the culture supernatant was replaced with sterile culture medium in the negative controls.
Microscopic observations were performed by confocal laser scanning microscopy (CLSM) using a TCS-SP2 system (Leica Microsystems, Mannheim, Germany). The biofilms formed were observed by staining the bacteria with 5 µM Syto 61 red for 10 min. The biofilm stacks were analyzed with the COMSTAT software (developed in MATLAB [40]) to estimate the maximal and average thicknesses (µm) and the biovolume (µm 3 µm −2 ) of the biofilm. Each experiment was repeated three times, and three zones of each channel were analyzed per experiment.

Antibacterial Assays
Agar well diffusion assay. The effect of SN IIIA004 on the growth of Roseovarius sp. VA014 bacteria was assayed by adapting the agar well diffusion assay previously described by Sablé et al. [41]. Solid nutrient plates (15 mL) were inoculated with approximately 10 7 cells of the target Roseovarius sp. VA014 strain. Sterile glass rings (4 mm inside diameter) were placed on agar medium and filled with 30 µL of filter-sterilized culture supernatant (SN IIIA004 or 10X SN IIIA004 ). The plates were incubated for 48 h at 22 • C, the optimum growth temperature for the target strain, to allow its growth and the culture supernatant diffusion. The presence of a halo around the glass cylinder indicates an inhibition of bacterial growth if the halo is clear (without cell growth) or eventually a stimulation if the halo is denser than the remaining plate. SN IIIA004 and 10X SN IIIA004 were replaced with Zobell broth and 10X Zobell broth in the respective negative controls.
Liquid antibacterial assay. Target Roseovarius sp. VA014 bacteria grown overnight were resuspended in a solution of 50% v/v SN IIIA004 and 50% v/v artificial seawater at an OD 600nm of 0.25 and incubated for 2 h. An aliquot of the cell suspension was then serially diluted and 100 µL of each dilution were plated, and the colony forming units (CFU) were counted after overnight growth. The remaining undiluted bacterial suspension was centrifuged, resuspended at 20% into fresh medium, and growth was monitored by measuring the absorbance of the cultures at 600 nm.

Microbial Adhesion to Solvent (MATS) Assays
The hydrophobic/hydrophilic and Lewis acid-base characteristics of the Roseovarius sp. VA014 surface were determined using the MATS method described by Bellon-Fontaine et al. [42]. This partitioning method is based on the comparison between microbial cell affinity to couples of solvents. In each pair, one solvent is a monopolar solvent, the other is an apolar solvent, and both must have similar Lifshitz-van der Waals surface tension components. The monopolar solvent can be acidic (electron accepting) or basic (electron donating). The following couples were used: (i) ethyl acetate (electron donating)/decane; (ii) dichloromethane (electron accepting)/tetradecane. All solvents were obtained from Sigma-Aldrich and were of the highest purity grade. Differences between the results obtained with dichloromethane and tetradecane, on the one hand, and between ethyl acetate and decane, on the other hand, indicate the electron donor and the electron acceptor character, respectively, of the bacterial surface. The percentage of cells adhered to tetradecane was used as a measure of cell surface hydrophobicity. Experimentally, a Roseovarius sp. VA014 suspension, containing approximately 10 8 cells mL −1 (OD 400 nm = 0.8), was prepared in artificial seawater. Moreover, 1.5 mL of bacterial suspension was manually mixed for 10 s and vortexed for 120 s with 0.25 mL of the solvent under investigation. The mixture was allowed to stand for 15 min to ensure complete separation of the two phases. The solvent phase was carefully removed and the OD of the aqueous phase was measured at 400 nm. The percentage affinity of bacteria to each solvent was calculated by % Affinity = (1−A/A 0 ) × 100, where A 0 is the OD 400 nm of the bacterial suspension before mixing and A is the OD 400 nm after mixing.

Physico-Chemical Characterization of the Active Compound(s)
To elucidate the biochemical nature of the active compound(s), different treatments were performed on SN IIIA004 . Proteinase K or pronase E were added to SN IIIA004 at a final concentration of 1 mg mL −1 to digest proteins and the reaction mixture was incubated for 1 h at 37 • C. To degrade lipids, lipase acrylic resin form was used at a final concentration of 2 mg mL −1 and the reaction was incubated for 48 h under shaking at 37 • C. DNaseI (100 µg mL −1 ) or RNaseA (25 µg mL −1 ) was added for 12 h at 37 • C to digest the nucleic acids. NaIO 4 was used at a final concentration of 20 mM to hydrolyze polysaccharides by cleaving the C-C bonds and by oxidizing the carbon of vicinal hydroxyl groups [43,44]. After an incubation for 2 h at 37 • C, the excess of NaIO 4 was neutralized with ethylene glycol (1:100) [45] for 2 h at 37 • C. This step was followed by a final overnight dialysis (molecular weight cut-off: 1000 Da, Spectrum Labs.com). To evaluate the heat sensibility, SN IIIA004 was incubated for 1 h at 37 • C, 30 min at 50 • C, 70 • C, or 100 • C. SN IIIA004 was replaced with Zobell broth in the negative controls.
After each of the above treatments, the resulting SN IIIA004 was assayed for anti-biofilm activity by using the microtiter plate assay as described above and its activity was compared with that of the untreated SN IIIA004 .

Biosurfactant Assay
In order to detect the presence of biosurfactant in SN IIIA004 , we used the drop-collapse test as described by Tugrul and Cansunar [46]. Briefly, drops of SN IIIA004 were placed in wells of a microtiter plate coated with sunflower oil. Drops containing biosurfactant would collapse whereas surfactant-free drops would remain stable. Diluted liquid hand washing cream, distilled water and Zobell broth drops were used as controls.

Purification of the SN IIIA004 Anti-Biofilm Compound
A two-step purification protocol was used. In the first step, the crude supernatant was applied to a HyperSep™ C 18 solid phase extraction column (500 mg, 6 mL, Thermo Scientific). Five stepwise elutions were successively performed with 100% Milli-Q water, 20%, 40%, 60% acetonitrile in Milli-Q water and 100% acetonitrile. After acetonitrile evaporation, the different fractions were tested for anti-biofilm activity towards Roseovarius sp. VA014 by using the microtiter plate assay as described above.
The 20% acetonitrile active fraction was concentrated by lyophilization and resuspension in Milli-Q water, filtered, and subjected to reverse-phase-high-performance liquid chromatography (RP-HPLC). This second purification step was performed using a C 18 column (3.5 µm, 150 × 4.6 mm, XSELECT CSH 130, Waters) and conducted with a Waters system (600 Controller, 2996 Photodiode Array detector and 2707 Autosampler). Separation was performed with the following acetonitrile gradient in Milli-Q water: 0 to 20% for 20 min, 20 to 50% for 2 min, 50 to 60% for 2 min, 60 to 90% for 2 min, at a flow rate of 1 mL min −1 . Each eluted fraction, collected according to the chromatographic profile obtained at 215 nm, was concentrated by lyophilization and resuspension in Milli-Q water. For each fraction, the protein concentration was determined by the bicinchoninic acid method (BC protein assay kit, Sigma-Aldrich), using bovine serum albumin as standard and the anti-biofilm activity was tested by using the microtiter plate assay as described above.
The purified active fraction was then subjected to an additional RP-HPLC analysis (the applied gradient is shown on Figure 6b). All the collected fractions containing the pure active molecule were pooled, concentrated by lyophilization, and stored at −80 • C.
To evaluate the molecular weight of the anti-biofilm molecule, the active purified fraction was transferred into a Centricon tube (Millipore) with a 3 kDa Nominal Molecular Weight Limit (NMWL) and centrifuged at 7500× g for 40 min at 4 • C.

Statistical Analyses
All values presented in the "Results" section are the averages of three independent experiments. The standard deviations were calculated using MATLAB software (MathWorks Inc., Natick, MA, USA). For each experiment, at least three technical replicates were performed. In order to analyze differences between a sample and the corresponding control, Student's t tests were performed. Differences were considered significant if p values were <0.05.

Pseudoalteromonas sp. IIIA004 Exoproducts Inhibit Biofilm Formation by Roseovarius sp. VA014
The Pseudoalteromonas sp. IIIA004 marine bacterium was isolated from a complex biofilm closely linked with corrosion products formed on carbon steel structures immersed in a French Atlantic harbor [9]. Here, we found that the culture supernatant of this strain, SN IIIA004 , inhibited the biofilm formation of another bacterial strain sharing the same habitat, Roseovarius sp. VA014 (Figures 1  and 2). When we examined the effect of increasing SN IIIA004 concentrations on biofilm formation in 96-well microtiter plate wells, the Pseudoalteromonas sp. IIIA004 secretome showed a dose-dependent anti-biofilm effect ( Figure 1).
A reduction of about 50% of the Roseovarius sp. VA014 biofilm formation was observed at a 1:2 dilution (50% v/v SN IIIA004 ). This concentration was then used for all subsequent experiments.
Although the polystyrene microtiter plate assay is a simple means of testing bacterial biofilm inhibition, it measures biofilm formation in static cultures, far from a naturally hydrodynamic environment. Therefore, the SN IIIA004 activity was further studied in a flow cell model (on glass surface) that allowed the continuous flow of fresh nutrients into a chamber. The Roseovarius sp. VA014 strain was grown in flow cell in the absence or presence of SN IIIA004 and biofilms were subsequently analyzed using confocal laser scanning microscopy ( Figure 2).
The addition of SN IIIA004 before the Roseovarius sp. VA014 strain (glass-coating with SN IIIA004 , Figure 2b) did not significantly prevent the biofilm formation, which demonstrated that SN IIIA004 is devoid of components able to act on the abiotic surface to reduce Roseovarius sp. VA014 adhesion. On the contrary, when added during the 2 h adhesion step, SN IIIA004 halved the biofilm formation of Roseovarius sp. VA014 (Figure 2c), as observed in static conditions in microtiter plates ( Figure 1). These findings demonstrated that SN IIIA004 also had biofilm inhibitory activity in flow cell under dynamic conditions and that the differences in surface nature (glass or polystyrene) did not influence the anti-biofilm activity. A reduction of about 50% of the Roseovarius sp. VA014 biofilm formation was observed at a 1:2 dilution (50% v/v SNIIIA004). This concentration was then used for all subsequent experiments.
Although the polystyrene microtiter plate assay is a simple means of testing bacterial biofilm The addition of SNIIIA004 before the Roseovarius sp. VA014 strain (glass-coating with SNIIIA004, Figure 2b) did not significantly prevent the biofilm formation, which demonstrated that SNIIIA004 is devoid of components able to act on the abiotic surface to reduce Roseovarius sp. VA014 adhesion. On the contrary, when added during the 2 h adhesion step, SNIIIA004 halved the biofilm formation of Roseovarius sp. VA014 (Figure 2c), as observed in static conditions in microtiter plates ( Figure 1). These findings demonstrated that SNIIIA004 also had biofilm inhibitory activity in flow cell under dynamic conditions and that the differences in surface nature (glass or polystyrene) did not influence the anti-biofilm activity.

SNIIIA004 Disrupts the Established Bacterial Biofilm
While SNIIIA004 exhibited an inhibitory activity against bacterial biofilm formation, it was of interest to explore whether the established biofilms were also sensitive to Pseudoalteromonas sp. IIIA004 exoproducts. This assay was performed by growing mature biofilms of the target strain in flow cell chambers before adding SNIIIA004. The results showed a significant reduction of thicknesses and biovolumes of Roseovarius sp. VA014 biofilms (Figure 2d): about 60% of mature biofilm disappeared. These findings demonstrated that SNIIIA004 contained components able to modify the properties of preformed biofilms and/or to destroy them.

SN IIIA004 Disrupts the Established Bacterial Biofilm
While SN IIIA004 exhibited an inhibitory activity against bacterial biofilm formation, it was of interest to explore whether the established biofilms were also sensitive to Pseudoalteromonas sp. IIIA004 exoproducts. This assay was performed by growing mature biofilms of the target strain in flow cell chambers before adding SN IIIA004 . The results showed a significant reduction of thicknesses and biovolumes of Roseovarius sp. VA014 biofilms (Figure 2d): about 60% of mature biofilm disappeared. These findings demonstrated that SN IIIA004 contained components able to modify the properties of preformed biofilms and/or to destroy them.

SN IIIA004 is Devoid of Bactericidal Activity against Free-Living Cells
Since SN IIIA004 inhibited biofilm formation and disrupted preformed biofilms, we examined whether SN IIIA004 contained an antibacterial substance that could be responsible for these effects. Using the agar well diffusion assay, neither crude SN IIIA004 nor concentrated 10X SN IIIA004 inhibited the growth of Roseovarius sp. VA014, indicating that SN IIIA004 was neither bactericidal nor bacteriostatic toward this target strain. The number of CFU mL −1 was evaluated after incubation of Roseovarius sp. VA014 for 2 h with SN IIIA004 or with Zobell broth (control), and was similar in both cases (6 × 10 8 CFU mL −1 ). Moreover, the number of bacteria was the same for both conditions throughout the experiment (Figure 3).
These results demonstrated that the viability of planktonic bacteria was not affected by the Pseudoalteromonas sp. IIIA004 exoproducts and suggested that SN IIIA004 did not reduce cell viability during the 2 h-adhesion step and the subsequent biofilm formation stages. the agar well diffusion assay, neither crude SNIIIA004 nor concentrated 10X SNIIIA004 inhibited the growth of Roseovarius sp. VA014, indicating that SNIIIA004 was neither bactericidal nor bacteriostatic toward this target strain. The number of CFU mL −1 was evaluated after incubation of Roseovarius sp. VA014 for 2 h with SNIIIA004 or with Zobell broth (control), and was similar in both cases (6.10 8 CFU mL −1 ). Moreover, the number of bacteria was the same for both conditions throughout the experiment (Figure 3). These results demonstrated that the viability of planktonic bacteria was not affected by the Pseudoalteromonas sp. IIIA004 exoproducts and suggested that SNIIIA004 did not reduce cell viability during the 2h-adhesion step and the subsequent biofilm formation stages.

SNIIIA004 Does Not Change the Hydrophilic and Acid Character of the Roseovarius sp. VA014 Cell Surface
Another hypothesis was that SNIIIA004 could modify the properties of bacterial surface and thus affected cell adhesion. We used the Microbial Adhesion To Solvents (MATS) method to determine the hydrophobic/hydrophilic and Lewis acid/base characteristics of Roseovarius sp. VA014 surface, incubated with SNIIIA004 or Zobell broth only (for the control) for 2 h. This partitioning method is based on the comparison of bacterial cell affinity to a monopolar acidic (electron acceptor) or basic (electron donator) solvent and an apolar solvent that have similar Lifshitz-van der Waals surface tension components. The solvent percentage affinity of cells to each solvent is shown in Figure 4.

SN IIIA004 Does Not Change the Hydrophilic and Acid Character of the Roseovarius sp. VA014 Cell Surface
Another hypothesis was that SN IIIA004 could modify the properties of bacterial surface and thus affected cell adhesion. We used the Microbial Adhesion To Solvents (MATS) method to determine the hydrophobic/hydrophilic and Lewis acid/base characteristics of Roseovarius sp. VA014 surface, incubated with SN IIIA004 or Zobell broth only (for the control) for 2 h. This partitioning method is based on the comparison of bacterial cell affinity to a monopolar acidic (electron acceptor) or basic (electron donator) solvent and an apolar solvent that have similar Lifshitz-van der Waals surface tension components. The solvent percentage affinity of cells to each solvent is shown in Figure 4.  Whatever the treatment (Zobell broth or SNIIIA004), the Roseovarius sp. VA014 cell surfaces presented a hydrophilic character by showing a very low affinity for the apolar solvents (decane, tetradecane). Moreover, the percentage affinity for the basic polar solvent (ethyl acetate) was much higher than for dichloromethane, indicating that the cell surface presented a broadly acidic character, whether treated with Zobell broth or with SNIIIA004.

Physico-Chemical Characteristics of SNIIIA004 anti-Biofilm Compounds
To gain information on the biochemical nature of the active compounds, we examined whether Whatever the treatment (Zobell broth or SN IIIA004 ), the Roseovarius sp. VA014 cell surfaces presented a hydrophilic character by showing a very low affinity for the apolar solvents (decane, tetradecane). Moreover, the percentage affinity for the basic polar solvent (ethyl acetate) was much higher than for dichloromethane, indicating that the cell surface presented a broadly acidic character, whether treated with Zobell broth or with SN IIIA004 .

Physico-Chemical Characteristics of SN IIIA004 Anti-Biofilm Compounds
To gain information on the biochemical nature of the active compounds, we examined whether the anti-biofilm molecules present in the Pseudoalteromonas sp. IIIA004 secretome retained their activity after different treatments ( Figure 5). The only treatments that completely canceled the antibiofilm activity of SN IIIA004 were treatments with proteinase K and pronase E. This finding was made both under static ( Figure 5a) and dynamic (data not shown) conditions. The SN IIIA004 anti-biofilm compounds were also heat-sensitive (Figure 5b): the inhibitory effect was significantly affected from 50 • C and clearly decreased when further increasing the temperature.

Isolation and Purification of the Anti-Biofilm Molecule
The proteinaceous molecule responsible of the anti-biofilm activity was purified to homogeneity from the stationary-phase culture supernatant of the Pseudoalteromonas sp. IIIA004 producer. The 20% acetonitrile fraction (50 µg of proteins), obtained by solid-phase extraction on HyperSep C18 cartridges, showed a specifically anti-biofilm activity against Roseovarius sp. VA014. This fraction was concentrated by lyophilization and then subjected to an accurate separation by two successive C18 RP-HPLC. After each separation, only one fraction, corresponding to one OD peak (P2, 10 µg of proteins, Figure 6a and P004, 3 µg of proteins, Figure 6b), presented anti-biofilm activity, showing that no other anti-biofilm molecule was co-purified. From the final C18 RP column, the active molecule was eluted in a single peak at 40 % of acetonitrile. This clearly demonstrated the proteinaceous nature of the active compounds. Moreover, SN IIIA004 drops did not collapse and remained stable on oil-coated surface, showing that SN IIIA004 is devoid of biosurfactant activity.

Isolation and Purification of the Anti-Biofilm Molecule
The proteinaceous molecule responsible of the anti-biofilm activity was purified to homogeneity from the stationary-phase culture supernatant of the Pseudoalteromonas sp. IIIA004 producer. The 20% acetonitrile fraction (50 µg of proteins), obtained by solid-phase extraction on HyperSep C 18 cartridges, showed a specifically anti-biofilm activity against Roseovarius sp. VA014. This fraction was concentrated by lyophilization and then subjected to an accurate separation by two successive C 18 RP-HPLC. After each separation, only one fraction, corresponding to one OD peak (P2, 10 µg of proteins, Figure 6a and P 004 , 3 µg of proteins, Figure 6b), presented anti-biofilm activity, showing that no other anti-biofilm molecule was co-purified. From the final C 18 RP column, the active molecule was eluted in a single peak at 40 % of acetonitrile. The anti-biofilm activity of the HPLC purified eluate treated by proteinase K was also highly affected (70% of loss, Figure 7). This confirmed the proteinaceous nature of the molecule responsible for the anti-biofilm activity. The anti-biofilm activity of the HPLC purified eluate treated by proteinase K was also highly affected (70% of loss, Figure 7). This confirmed the proteinaceous nature of the molecule responsible for the anti-biofilm activity. The anti-biofilm activity of the HPLC purified eluate treated by proteinase K was also highly affected (70% of loss, Figure 7). This confirmed the proteinaceous nature of the molecule responsible for the anti-biofilm activity. Figure 7. Effect of proteinase K treatment on the P004 anti-biofilm compound purified from SNIIIA004 and molecular weight evaluation. Microtiter plate assay was used. P004 was either treated with proteinase K or filtered in a Centricon tube with a 3 kDa Nominal Molecular Weight Limit (NMWL) Figure 7. Effect of proteinase K treatment on the P 004 anti-biofilm compound purified from SN IIIA004 and molecular weight evaluation. Microtiter plate assay was used. P 004 was either treated with proteinase K or filtered in a Centricon tube with a 3 kDa Nominal Molecular Weight Limit (NMWL) (>3 kDa: retentate; <3 kDa: filtrate). Control biofilm: Roseovarius sp. VA014 biofilm treated with Zobell broth instead of SN IIIA004 . Significant differences are indicated by ** (p < 0.01).
The purified eluate was submitted to a centrifugal Centricon filter that retains the components with a molecular weight higher than 3 kDa. The anti-biofilm activity was detected in the filtrate only (Figure 7), clearly showing that the molecular weight of the active component was lower than 3 kDa. This pure molecule was named P 004 .

Spectrum of Action of the P 004 Anti-Biofilm Molecule Purified from SN IIIA004
The microtiter plate assay was used to determine the spectrum of action of the pure compound. P 004 (10 µg) was assayed against 21 bacterial strains able to form stable biofilms in microtiter plates with an OD 595nm > 1 after crystal violet staining. These bacteria include human pathogens, pathogenic marine bacteria such as Flavobacterium, Tenacibaculum, and Vibrio lentus, as well as bacteria potentially involved in biocorrosion or biofouling in the marine environment ( Table 1). The percentages of inhibition of P 004 on monospecies biofilms are presented in Table 2. The most sensitive strains, with inhibition percentages ranging from 62 to 81.5%, were both bacteria widely distributed in the marine environment (Roseovarius sp., Paracoccus sp., and Micrococcus luteus) and well-known human pathogens (Staphylococcus aureus, Yersinia enterocolitica, and Pseudomonas aeruginosa). A second group of bacteria, less sensitive (inhibition percentages from 12 to 45%), included another strain of S. aureus, two Bacillus strains and several marine bacteria belonging to the Flavobacteriaceae family (Zobellia galactanivorans, Tenacibaculum sp., Flavobacterium sp., and Cellulophaga lytica). Some strains were not sensitive to the P 004 biofilm-inhibiting molecule (inhibition percentage < 10%), especially marine bacteria of the Vibrio genus and one Pseudomonas aeruginosa strain.

Discussion
This study reports the isolation and characterization of a new anti-biofilm compound from the Pseudoalteromonas sp. IIIA004 marine bacterium, active against various bacteria. As described in several studies, marine bacteria are a potential source of effective anti-biofilm compounds [23]. Some of them were shown to secrete anti-biofilm molecules active against a wide range of Gram-positive and Gram-negative bacteria including Staphylococcus aureus, Listeria monocytogenes, and Salmonella typhimurium [47]. Here, we examined the in vitro activity of the cell-free supernatant of Pseudoalteromonas sp. IIIA004 against the Roseovarius sp. VA014 marine strain, identified among the pioneer and sustaining surface colonizers particularly on metallic surfaces [9]. This anti-biofilm activity was characterized in microtiter plates (static conditions/polystyrene surface) and in flow cell chambers (dynamic conditions/glass surface). Both techniques highlighted the anti-biofilm potential of this strain. SN IIIA004 was shown to display its activity during two stages of biofilm formation: (i) when the culture supernatant was mixed with Roseovarius sp. VA014 cells during the 2 h adhesion step, biofilm growth was inhibited both on polystyrene and glass surfaces, and (ii) when the culture supernatant was added after the Roseovarius sp. VA014 biofilm formation, the mature biofilm was strongly dispersed. The anti-biofilm effect of Pseudoalteromonas sp. IIIA004 exoproducts is particularly interesting because it combines several activities, as previously demonstrated for the marine bacterial exopolysaccharide EPS273 that exhibited an inhibition of Pseudomonas aeruginosa PAO1 biofilm formation as well as a biofilm dispersion [48]. From 2001, disturbing the multicellular structure of bacterial biofilm was suggested as a promising way to increase antibiotic sensitivity of pathogens in biofilms [49]. Recently, innovative strategies combining antibiotics and anti-biofilm compounds such as polysaccharides [50], synthetic peptides [51], or alginolytic enzymes [15] were proposed in order to increase the susceptibility of microbial biofilms to antibiotics. Interestingly, antibacterial assays with SN IIIA004 demonstrated the lack of bactericidal or bacteriostatic action against free-living Roseovarius sp. VA014 cells. Therefore, the anti-biofilm activity of Pseudoalteromonas sp. IIIA004 is likely to be mediated by mechanisms different from growth inhibition. By contrast, most of the known anti-biofilm molecules are bactericidal or bacteriostatic, such as for instance the AlpP protein secreted by Pseudoalteromonas tunicata [52] or the bioactive compounds of Pseudoalteromonas sp. IBRL PD4.8 [53]. The main characteristics of various anti-biofilm molecules produced by different Pseudoalteromonas strains are summarized in Table 3.
Among the few natural molecules displaying anti-biofilm activity without affecting cell viability, some polysaccharides [47,54] and biosurfactants [55] are well known. We showed that the SN IIIA004 anti-biofilm compound was neither a polysaccharide nor a biosurfactant. We further investigated whether this compound could act by modifying the properties of bacterial cells and/or abiotic surfaces. MATS results clearly showed that SN IIIA004 did not significantly affect the hydrophilic and acid properties of the Roseovarius sp. VA014 cell surface. Moreover, no glass-coating effect was observed with SN IIIA004 compounds. All these findings suggested that the SN IIIA004 anti-biofilm compound did not act on surfaces, whether or not biotic. Other possible mode of action could be considered. Since SN IIIA004 anti-biofilm action is effective during cell adhesion and on mature biofilm without affecting cell multiplication, the anti-biofilm compounds could act as signaling molecules that modulate gene expression of target bacteria, as suggested for several anti-biofilm polysaccharides [47]. For instance, exopolysaccharides released from Lactobacillus acidophilus A4 down-regulated several E. coli genes related to adhesive properties (curli genes) [56]. Moreover, Pseudoalteromonas sp. IIIA004 exoproducts might block lectins or adhesins of fimbriae and pili on the surface of bacteria, which could interfere with the cell-surface and cell-cell adherence, as suggested for the marine bacterial exopolysaccharide A101 [54] or for some microbial branched polysaccharides used as food additives [57]. However, as P 004 does not affect the viability of the tested bacteria, it certainly does not act on membrane permeability, unlike some antibacterial molecules [58].
Among the major kinds of molecules that impair biofilm maturation, there are also the quorum-sensing (QS) inhibitors [7,11,17]. Some enzymes such as N-acyl homoserine lactone (AHL)-lactonases and AHL-acylases degrade signal QS molecules and thus prevent the cell-to-cell communication, which in turn impairs population behavior such as biofilm development [59,60]. Other kind of enzymes can directly target the exopolymeric matrix by degrading its components such as polysaccharides or extracellular DNA. Thus, the use of DNaseI [61], alpha-amylase [62] or Dispersin B [63] has been identified as an efficient means of dispersing biofilms, in vitro and in vivo. Investigation of the potential binding of the Pseudoalteromonas sp. IIIA004 purified exoproducts with adhesins or other matrix compounds would be necessary to better understand the mechanisms underlying their anti-biofilm effects. The anti-biofilm effect of SN IIIA004 during the initial attachment step is likely due to a new molecule. Protease and heat treatments impaired SN IIIA004 ability to inhibit biofilm formation, indicating that the active molecule was of proteinaceous nature. The anti-biofilm peptide P 004 was then purified from Pseudoalteromonas sp. IIIA004 exoproducts and its molecular weight was estimated to less than 3 kDa. To our knowledge, no bacterial molecule with both this molecular weight and a specific anti-biofilm activity (without affecting bacterial viability) has been described yet. On the contrary, several small peptides (natural or chemically synthesized) with dual antimicrobial and anti-biofilm activity have been reported [66,67]. Moreover, the Pseudoalteromonas genus is of great interest to the scientific community because of its prolific metabolite-producing ability [25,68]. Compounds of interest include toxic proteins, polyanionic exopolymers, substituted phenolic and pyrrole-containing alkaloids, cyclic peptides and a range of bromine-substituted compounds with antimicrobial, anti-fouling, algicidal, and various pharmaceutically relevant activities [23,25]. However, among the anti-biofilm proteinaceous compounds synthesized by Pseudoalteromonas strains, only the 190-kDa autotoxic protein (AlpP) produced by P. tunicata D2 and the 23-kDa alginate lyase (AlyP1400) produced by Pseudoalteromonas sp. 1400 were purified and characterized (Table 3) [15,52,69]. Pseudoalteromonas sp. 3J6 and D41 were also identified as producing proteinaceous molecules with a specifically anti-biofilm activity against a wide range of various bacterial biofilms (Table 3) [18,28,29]. However, the anti-biofilm molecule from SN 3J6 was eluted with 50% acetonitrile from a Sep-Pak Plus C 18 cartridge and was recently identified as a 13-kDa protein named alterocin [65], while the SN D41 anti-biofilm molecule was unable to be eluted from the same column either with acetonitrile or with other solvents. These findings support the hypothesis that the anti-biofilm molecules of SN 3J6 and SN D41 are different from the anti-biofilm molecule P 004 eluted with 20% acetonitrile from a C 18 cartridge.
Finally, the new proteinaceous small molecule P 004 presents a wide spectrum of action on various bacteria, with 71% of sensitive strains, including the human pathogens Staphylococcus aureus, Pseudomonas aeruginosa PAO1, and Yersinia enterocolitica (Table 3). Therefore, the potential use of P 004 is not limited to the marine environment to inhibit undesirable bacteria in aquaculture, biofilms involved in biocorrosion or biofouling, but it could also be extended to the medical field.