Busting biofilms: free-living amoebae disrupt preformed methicillin-resistant Staphylococcus aureus (MRSA) and Mycobacterium bovis biofilms

Biofilm-associated infections are difficult to eradicate because of their ability to tolerate antibiotics and evade host immune responses. Amoebae and/or their secreted products may provide alternative strategies to inhibit and disperse biofilms on biotic and abiotic surfaces. We evaluated the potential of five predatory amoebae – Acanthamoeba castellanii, Acanthamoeba lenticulata, Acanthamoeba polyphaga, Vermamoeba vermiformis and Dictyostelium discoideum – and their cell-free secretions to disrupt biofilms formed by methicillin-resistant Staphylococcus aureus (MRSA) and Mycobacterium bovis . The biofilm biomass produced by MRSA and M. bovis was significantly reduced when co-incubated with A. castellanii, A. lenticulata and A. polyphaga, and their corresponding cell-free supernatants (CFS). Acanthamoeba spp. generally produced CFS that mediated biofilm dispersal rather than directly killing the bacteria; however, A. polyphaga CFS demonstrated active killing of MRSA planktonic cells when the bacteria were present at low concentrations. The active component(s) of the A. polyphaga CFS is resistant to freezing, but can be inactivated to differing degrees by mechanical disruption and exposure to heat. D. discoideum and its CFS also reduced preformed M. bovis biofilms, whereas V. vermiformis only decreased M. bovis biofilm biomass when amoebae were added. These results highlight the potential of using select amoebae species or their CFS to disrupt preformed bacterial biofilms.


InTRoduCTIon
A biofilm is a community of bacteria that secrete a matrix consisting of extracellular polymeric substances (EPS), which protects the bacteria adhered to a surface from the environment, host immune responses and other adverse conditions [1,2]. Bacterial pathogens residing in biofilms exhibit increased tolerance to environmental stressors, such as antibiotics (up to 1000-fold as compared with their planktonic counterparts), nutrient deprivation and temperature flux [3][4][5][6]. Biofilms can be especially problematic in hospital settings where patients may be immunocompromised, which increases the risk of infection by commensal microbiota (e.g. Staphylococcus spp.) [7,8]. Bacterial growth in biofilms has contributed to the emergence and spread of antibiotic resistance, which is a public health threat due to the increased prevalence of bacterial pathogens now able to survive current antibiotic treatments, compounded by a decrease in the development of novel antimicrobial drugs [9,10].
Methicillin-resistant S. aureus (MRSA) and Mycobacterium bovis are two important public health threats that form robust biofilms, which likely promotes survival in the environment and during infection of a host [11][12][13]. MRSA has acquired resistance to multiple classes of antibiotics, specifically β-lactams (penicillins and cephalosporins), and is well known for its prevalence in healthcare-acquired infections (HAIs) [14][15][16][17][18]. HAIs that are often associated with biofilm-forming strains of MRSA exhibit a dynamic range of symptoms, which vary from slight skin abrasions to a fatal systemic disease [19,20]. MRSA residing in biofilms are also more resistant to conventional antimicrobial therapies and host immune responses [16,21]. M. bovis is a slow-growing, Gram-positive aerobe found globally in infected soils and mammalian reservoir species [22][23][24][25][26]. The causative agent of bovine tuberculosis, M. bovis, results in tuberculosis-like symptoms in cattle [27,28] and is the aetiological agent of zoonotic tuberculosis in humans [22,29,30]. M. bovis biofilm formation on biotic and abiotic surfaces has also been proposed to contribute to drug tolerance, pathogenicity and environmental persistence [31].
There is an urgent need to find alternative strategies to disrupt biofilm-associated infections, which are a contributing factor in increased patient morbidity and mortality [32]. One such strategy is to trigger biofilm dispersal, resulting in the release of planktonic cells that are generally more susceptible to killing by conventional antibacterial treatments [33]. Free-living amoebae (FLA) graze upon attached Escherichia coli cells and preformed biofilms produced by a variety of bacteria, including Klebsiella pneumoniae, Pseudomonas spp. and Staphylococcus epidermidis [34][35][36]. Predatory FLA gain nutrients by engulfing and degrading bacteria and other protists [37] and are also known to secrete molecules into the environment that impact on bacterial survival [38]. For these reasons, amoebal grazing has been hypothesized to be an important factor in controlling biofilms in aquatic environments [39]. Amoebae are single-cell eukaryotes that live as motile trophozoites in nutrient-rich settings [40]. In the presence of stressors or during periods of nutrient starvation, FLA can encyst and produce a double-walled membrane (e.g. Acanthamoeba and Vermamoeba), thereby becoming extremely resistant to adverse conditions [41,42].
Acanthamoeba are some of the most prevalent amoebae and are widely distributed geographically [43,44]. Members of the genus Acanthamoeba are primarily non-pathogenic but have been reported to opportunistically infect immunocompromised individuals and can also cause amoebic keratitis in contact lens wearers [45,46]. These amoeba are found in soils and aquatic environments, but also in artificial water systems, including sewers and air-conditioning units [47,48]. Vermamoeba vermiformis is another predatory amoeba found in the environment and has been shown to be a potential reservoir for various pathogens, including Legionella pneumophila, Mycobacterium spp. and Stenotrophomonas maltophilia [49][50][51]. Dictyostelium discoideum, another predatory amoeba [52] found globally in soils, is unique in that it can aggregate into a multicellular slug that can transition into fruiting bodies when challenged with unfavourable conditions [53,54].
Based on the predatory nature of amoebae and their potential interactions with biofilm-associated bacteria, we hypothesized that MRSA and M. bovis biofilms may be altered by amoebae and/or secreted factors produced by amoebae. Three amoebae species -Acanthamoeba castellanii, Acanthamoeba lenticulata and Acanthamoeba polyphaga -and their secretory components prepared from cell-free supernatants

Acanthamoeba spp. and CFS disrupt preformed MRSA biofilms
Exposure of preformed MRSA biofilms to starved cultures of A. castellanii, A. lenticulata, or A. polyphaga resulted in a significant reduction in biofilm biomass of 72.3, 61.4 and 62.4 %, respectively, as compared to the media control in static biofilm assays (Fig. 1a). Addition of V. vermiformis and D. discoideum did not decrease MRSA biofilm biomass (Fig. 1b). These results indicate that Acanthamoeba spp. differentially reduce MRSA biofilm as compared to the V. vermiformis and D. discoideum strains tested in this study. MRSA biofilm biomass was further reduced upon co-incubation with CFS obtained from A. castellanii, A. lenticulata and A. polyphaga, which reduced biofilm biomass by 84, 69.9, and 73.6 %, respectively (Fig. 1a). The biofilm-disrupting activity of CFS from A. castellanii, A. lenticulata and A. polyphaga was partially inactivated by boiling. The observed reduction in biofilm biomass from boiled CFS was not as substantial when compared to the reduction observed for co-incubation with amoebae cultures or filtered CFS, suggesting that the biofilmdisrupting activity is only partially heat labile (Fig. 1a). Surprisingly, the addition of D. discoideum or its CFS to preformed MRSA biofilms resulted in a significant increase in biofilm biomass of 10.3 and 19.9 %, respectively. The addition of V. vermiformis or its CFS to the preformed MRSA biofilms did not alter biofilm biomass, although there was a decrease in biofilm biomass (27.5 %) upon the addition of boiled V. vermiformis CFS (Fig. 1b).
GFP-labelled MRSA cultivated as a biofilm were co-incubated with A. castellanii, A. lenticulata, A. polyphaga, V. vermiformis and D. discoideum to visually observe the amoebal interactions with preformed MRSA biofilms. Similar to the biofilm disruption observed in the static biofilm biomass assays, A. castellanii, A. lenticulata and A. polyphaga disrupted the MRSA preformed biofilms. However, the various species of amoebae exhibited different types of behaviours when co-incubated with MRSA biofilms. A. castellanii created circular plaques within the MRSA biofilm ( Fig. 2b), whereas A. lenticulata and A. polyphaga bored channels through the biofilm as they actively migrated (Fig. 2c, d; see also Video S1, available in the online version of this article). This observation supports the hypothesis that amoebae may utilize different mechanisms to migrate within and alter bacterial biofilms. The addition of V. vermiformis or D. discoideum did not produce any observable alterations to the MRSA biofilms. Nearly all the V. vermiformis trophozoites encysted upon co-incubation with preformed MRSA biofilms, while D. discoideum appeared as non-motile round bodies or floating cells in the presence of the biofilms (Fig. 2e, f).

Acanthamoeba spp. and CFS disperse bacteria from preformed MRSA biofilms
We sought to evaluate whether the primary mode of MRSA biofilm mass reduction upon co-incubation with the amoebae or their CFS resulted from the active killing of the bacteria, dispersal from the biofilm, or a combination of the two. Comparison of colony forming units (c.f.u.) dispersing from preformed MRSA biofilms co-incubated with A. castellanii, A. lenticulata, or A. polyphaga resulted in a significant increase in c.f.u. (127.7, 128 and 112.4%) as compared with the biofilms incubated with the media control (Fig. 3). A significant increase in MRSA c.f.u. was also observed when CFS from Acanthamoeba spp. were added to the preformed MRSA biofilms (Fig. 3). The increase in MRSA c.f.u. during co-incubation with the amoebae or their CFS indicates that Acanthamoeba spp. evaluated in these studies are actively dispersing the bacteria from the preformed biofilm. vermiformis and D. discoideum amoebae, CFS and boiled amoebae supernatants after a 24 h incubation period by performing quantitative biofilm biomass assays. These assays consisted of six biological replicates repeated in triplicate. Significance was defined as a calculated P value of less than 0.01 by an unpaired Student's t-test (letters indicate significant differences between treatment populations: a, media control; b, to amoebae; c, CFS). Error bars indicate standard error. Conversely, MRSA was not observed to be dispersing from the biofilm upon the addition of V. vermiformis, D. discoideum, or CFS from these amoebae (Fig. 3). The inability of V. vermiformis and D. discoideum to disperse the MRSA biofilm ( Fig. 3) was consistent with data obtained from MRSA biofilm biomass assays and visualization of the structured biofilm communities (Figs 1 and 2). A significant decrease in MRSA dispersal was measured as compared to the media control when MRSA biofilms were co-incubated with V. vermiformis and D. discoideum (Fig. 3), indicating that these treatments have little effect on disrupting the bacterial biofilm biomass, or may actively reduce planktonic bacterial populations that naturally disperse from the biofilm.

Activity of A. polyphaga CFS on planktonic MRSA
To determine if the Acanthamoeba CFS is responsible for killing planktonic MRSA that dispersed from preformed MRSA biofilms, a planktonic MRSA c.f.u. assay was utilized to evaluate whether A. polyphaga CFS killed MRSA cells growing as planktonic cells as opposed to within a biofilm. Co-incubation of A. polyphaga CFS with MRSA grown to mid-log did not reduce/kill planktonic MRSA as compared with the media controls from 1-4 h post-addition of CFS (Fig. 4). Interestingly, a significant reduction in MRSA was noted when MRSA cells were grown to mid-log, and then diluted 1 : 200 into A. polyphaga-conditioned media and co-incubated, as compared with the media controls. Differences in MRSA population density upon co-incubation with A. polyphaga CFS may account for whether there is an adverse effect of the active component(s) on planktonic MRSA.

Initial characterizations of A. polyphaga biofilmdispersing factor(s)
A. polyphaga CFS was subjected to boiling, mechanical disruption (OmniLyse), or three rounds of freeze/thaw in order to partially characterize the active component(s) responsible for the disruption of preformed biofilms. Inactivation by mechanical disruption may indicate whether extracellular exosomes or vesicles contribute to the active fraction of the supernatant. A significant increase in MRSA biofilm formation was observed when the CFS was either boiled or mechanically disrupted (34.5 and 28.3 %, respectively) as compared with the untreated supernatant (Fig. 5). Boiling and mechanical disruption independently reduced MRSA biofilm biomass (21.6 and 27.8 %, respectively), indicating that the active components of the A. polyphaga supernatant are only partially heat-labile and partially sensitive to mechanical disruption. The anti-biofilm component(s) of the A. polyphaga CFS was unaffected by three freeze/thaw cycles. (Fig. 5).

Acanthamoeba, Vermamoeba and Dictyostelium amoebae and CFS disrupt preformed M. bovis biofilms
Co-incubation of preformed M. bovis biofilms with A. castellanii, A. lenticulata, A. polyphaga, V. vermiformis and D. discoideum led to significant decreases in biofilm biomass of 80.3, 85.9, 82.1, 24.9 and 53.8 %, respectively, in the static biofilm assay (Fig. 6)  Treatments were normalized to a MRSA control biofilm on a log 10 scale. These assays consisted of six biological replicates repeated in triplicate. Significance was defined as a calculated P value of less than 0.01 by an unpaired Student's t-test (letters indicate significant differences between treatment populations: a, media control; b, amoebae). Error bars indicate standard error.
biofilm biomass was observed in the presence of V. vermiformis CFS.
A visible reduction in M. bovis biofilm biomass was also observed with microscopy upon co-incubation with A. castellanii, A. lenticulata, A. polyphaga, V. vermiformis and D. discoideum (Fig. 7a-f). After co-incubation, A. castellanii, A. lenticulata, A. polyphaga and D. discoideum had a large proportion of motile trophozoites (red arrows) actively grazing on the preformed M. bovis biofilm (Fig. 7b-f), whereas V. vermiformis had a large proportion of amoebae encysted (red circles) (Fig. 7b-e). A visible reduction in biofilm biomass by the various amoebae recapitulates the results found in the M. bovis biofilm biomass assays, where a significant decrease in M. bovis biofilm was observed upon co-incubation with all the amoebae species used in this study (Fig. 6).

dISCuSSIon
In the current study, we sought to characterize the interactions between five different species of amoebae (A. castellanii, A. lenticulata, A. polyphaga, V. vermiformis and D. discoideum) with preformed biofilms from two pathogenic bacteria, MRSA and M. bovis. Amoebae appear to possess mechanism(s) to actively disperse bacteria from a biofilm state, although the details of this mechanism have yet to be determined. Our data demonstrate that A. castellanii, A. lenticulata and A. polyphaga and the CFS that they produce significantly disrupted preformed MRSA and M. bovis biofilms. Conversely, V. vermiformis and D. discoideum and the CFS produced by these amoebae had minimal or no ability to disrupt preformed MRSA biofilms, whereas V. vermiformis (amoebae only) and D. discoideum amoebae or D. discoideum CFS significantly reduced M. bovis biofilms. These differential interactions between different species of amoebae and bacteria indicate that there are additional discoveries to be made in this area of research concerning microbe-microbe interactions.
Previous studies have shown that the CFS from A. polyphaga inhibits planktonic MRSA growth after 96 h co-incubation [55] and that MRSA can proliferate within A. polyphaga [56]. However, untreated and heat-treated A. castellanii supernatants killed a significant percentage of planktonic MRSA cells, but did not adversely affect other bacterial species [38]. Bactericidal activity against Xanthomonas oryzae pvs. oryzae and oryzicola by A. lenticulata, A. polyphaga, V. vermiformis and D. discoideum has also recently been described [57]. These studies support our observations that Acanthamoeba spp. secrete substances into the extracellular environment that can affect specific bacterial species, notably planktonic MRSA cells.
The impact of A. polyphaga CFS on planktonic MRSA cells is dependent on the initial concentration and potentially the growth phase of MRSA in the supernatant. When the MRSA population is high (OD 600 0.6), the A. polyphaga CFS does not significantly reduce the planktonic MRSA population. This finding supports our observation that high populations of bacterial cells from a MRSA biofilm can be dispersed by Acanthamoeba spp. without necessarily killing a significant proportion of them in the process. MRSA populations at high cell densities have increased tolerance against antimicrobial factors [21,58], or may be able to neutralize the active killing components secreted by the amoebae.
Studies addressing amoeba-bacteria relationships can be rather complex. Amoebae have the ability to produce antimicrobial compounds that lyse bacteria [59]; amoebae can also engulf bacteria, resulting in either degradation in lysosomes, or the bacteria may begin intracellular replication. After bacteria are phagocytized, the amoebae may thus serve as a potential reservoir [60,61]. Alternatively, bacteria have also been shown to produce compounds that induce amoebae encystment or directly kill the amoebae [55,62,63]. Amoebae and bacterial interactions are especially complex in the context of bacteria residing in biofilms. In the current study, Acanthamoeba spp. actively dispersed cells from both MRSA and M. bovis biofilms. Induced dispersal of bacteria by compounds secreted from amoebae presents the amoebae with a plethora of easily accessible planktonic bacteria to graze upon.
Many of the studies involving the identification and characterization of secreted amoebae compounds have focused on the pathogenicity of amoebae to humans and other hosts.  The pathogenicity of Acanthamoeba spp. can be at least partially attributable to various secreted proteases [64][65][66], an elastase [67] and a pore-forming toxin [68]. These proteases contribute to amoebae pathogenesis and serve as virulence factors [65,69], but their role with regards to altering bacterial growth in a biofilm has yet to be determined. Genomic analysis of A. castellanii identified an alginate lyase gene that has been predicted to potentially contribute to the degradation of bacterial biofilms composed of the exopolysaccharide alginate produced by Pseudomonas spp. [70]. However, the role of this predicted alginate lyase has yet to be characterized. More recently, quantitative proteomic analysis of extracellular vesicles from A. castellanii identified hydrolases and oxidoreductases to be the most abundant protein families, comprising roughly 81 % in these vesicles [71].
In the current study, manipulation of Acanthamoeba spp. CFS demonstrates that the active component(s) responsible for biofilm disruption are partially heat-labile. Additional manipulation of A. polyphaga CFS suggests that the active component(s) are partially inactivated by mechanical disruption (OmniLyse), although the CFS is insensitive to the freeze/thaw process. The observation that activity is lost after mechanical disruption but not after multiple freeze/ thaw cycles leads us to hypothesize that the amoebae may be secreting exosomes or extracellular microvesicles (MVs) into the media that can deliver biofilm-dispersing molecules directly to the biofilm [72]. Exosomes and MVs have been described as a means of transferring materials between cells in response to various physiological stimuli [73], but the observation that amoebae may be secreting them for active disruption of bacterial biofilms represents a potentially new mechanism. Further research will focus on understanding and further characterizing the active components of Acanthamoeba spp. CFS, which may reveal a novel anti-biofilm compound that can be used in combination with current antimicrobial therapies to treat biofilm-associated infections.
Acanthamoeba spp. and V. vermiformis were starved in 1/5 diluted media before performing assays. Amoebae were first dislodged from the flask using a cell scraper and 5 ml of the culture was centrifuged at 1000 g for 10 min at room temperature. The supernatant was removed and the pellet was suspended in 10 ml of 1/5 amoebae-specific media, added to a T25 flask and incubated at 28 °C for 24 h. The amoebae were counted using a haemocytometer and trypan blue to assess viability. Cells were diluted to a final concentration of 1×10 5 amoeba ml −1 . Amoebae supernatants were passed through a 0.22 µm filter to collect CFS. Supernatants were either untreated, boiled at 110 °C for 30 min, frozen at −80 °C and thawed for three repeats, or treated with an OmniLyse kit (ClaremontBio) for 2 min. Media-only controls were processed in the same manner as the supernatant treatments. Each assay utilizing multiple species of amoebae was taken from the same passage number and the amoebae showed similar levels of growth across the different species.

Growth conditions for bacteria
MRSA typed as USA300 strains were used for these studies. S. aureus strain HFH-29568 [74] was used for all static biofilm assays and obtained from BEI resources (catalogue #NR-10314) through the NIH Biodefense and Emerging Infections Research Resources Repository, National Institute of Allergy and Infectious Diseases (NIAID). A GFPexpressing S. aureus strain, AH1726, was used for confocal imaging [75].

Static biofilm assay
Bacterial cultures were grown in a 37 °C incubator until turbid (24 h for MRSA and approximately 2 weeks for M. bovis) as described above. The cultures were diluted to an OD 600 of 0.1 and 100 µl was added to the wells of 96-well flat-bottom plates (Nunc Microwell for MRSA and Corning Biocoat polyd-lysine coated for M. bovis). Plates were sealed in a Ziploc bag and incubated at 37 °C (24 h for MRSA and approximately 6 weeks for M. bovis). Twenty-four hours before the biofilm was processed, 5 ml of each strain of amoebae was starved as described above. After incubation, the media was removed from the 96-well plate and starved amoebae (or CFS treatments) were added in replicates of 6. Plates were incubated for 24 h at 28 °C for Acanthamoeba spp. and V. vermiformis, or room temperature (~21 °C) for D. discoideum. After incubation, the media were removed from the plate and the plate was rinsed in phosphate-buffered saline (PBS) and then stained with 0.05 % crystal violet for 15 min. The crystal violet was then solubilized in 95 % ethanol for 30 min and absorbance (OD 600 ) was measured on either a PerkinElmer EnSpire or a Synergy HT plate reader. These assays were completed in triplicate.

Quantification of MRSA biofilm dispersal
MRSA cultures were grown for 24 h in a shaking incubator at 37 °C, diluted to a final OD 600 of 0.1 in MHB, 100 µl of culture was added to Nunc Microwell 96-well microplates, and then the plate was incubated for 24 h at 37 °C in a sealed Ziploc bag. After the 24 h incubation period, MRSA biofilms were exposed to either starved amoebae or amoebae CFS. The supernatant containing cells dispersing from the biofilm was removed from each well and diluted from 10 −1 to 10 −7 in PBS and plated onto MHB plates. c.f.u. were counted after 24 h incubation at 37 °C. These assays were completed in triplicate.

determination of viability of planktonic MRSA in A. polyphaga (CFS)
MRSA cultures were grown for 24 h in a shaking 37 °C incubator, diluted 1 : 50 in 1/5 PYG and placed in a shaking 37 °C incubator until the culture reached an OD 600 of approximately 0.6 (mid-log). Cultures were added to a 50 ml conical tubes (50 ml total for 1 : 1 dilution and 250 µl total for 1 : 200 dilution) and centrifuged at 4800 g for 20 min, and then the supernatants were removed. The bacterial pellet was suspended in either 1/5 PYG or A. polyphaga CFS, and 3 ml of the culture was added in triplicate to 15 ml plastic tubes. Tubes were placed in a shaking 37 °C incubator and aliquots were plated for c.f.u. in time-increments of 0, 1, 2 and 4 h by dilution from 10 −1 to 10 −7 in PBS and plating on MHB plates. c.f.u. were counted after 24 h incubation at 37 °C. These assays were completed in triplicate.

Confocal imaging
MRSA cultures were grown for 24 h in a shaking 37 °C incubator, the culture was then diluted to an OD 600 0.1 in MHB, 300 µl of the culture was added to the wells of a µ-Slide eight-well glass-bottom IBIDI slide (catalogue #80826), and the slides were incubated at 37 °C in a sealed Ziploc bag. After a 24 h incubation period, the media were removed, and the preformed MRSA biofilms were exposed to starved amoebae or control treatments for 24 h. Confocal imaging was performed using an inverted Olympus FV1000-IX81 confocal imaging system. Time-lapse microscopy was obtained with an imaging speed set to 4 µs/pixel, and an imaging interval set to freerun (image acquisitions repeating at approximately 1.6 s/image). The time-lapse video is displayed at 33 frames s −1 . Images were processed and analysed with Volocity (Perki-nElmer). The utilization of IBIDI slides allowed for multiple biological replicates to be visualized simultaneously and the images shown are representative of results seen across multiple treatment wells and control wells. These assays were completed in triplicate.

Light microscopy
M. bovis cultures were grown for 2 weeks in a shaking 37 °C incubator, diluted to a final OD 600 of 0.1 in modified Sauton's media, 1500 µl was added to the wells of µ-Slide two-well glass-bottom IBIDI slides (catalogue #80286) coated with poly-d-lysine, and incubated at 37 °C in a sealed Ziploc bag. After an incubation period of approximately 6 weeks, the media were removed and preformed M. bovis biofilms were exposed to starved amoebae for 24 h. Imaging was performed using an Olympus IX73 inverted microscope with cellSens software. The images shown are representative of results visualized across treatment and control wells; assays were completed in duplicate.

Statistical analyses
Statistical analyses were conducted using GraphPad Prism 6. Treatments were normalized to bacterial biofilms grown in media controls; media controls consisted of 1/5 amoebaespecific media subjected to the treatment conditions (i.e. 0.22 µm filtration, boiling, freeze/thaw and mechanical disruption) alongside experimental treatments when necessary. The groups were then analysed using an unpaired Student's t-test. Differences were designated as significant when the calculated P-value was less than 0.01. Error bars indicate the standard error. The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of the Army Research Office, DARPA, or the US Government. The US Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation hereon.