Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

An Eye to a Kill: Using Predatory Bacteria to Control Gram-Negative Pathogens Associated with Ocular Infections

  • Robert M. Q. Shanks,

    Affiliation Department of Ophthalmology, Campbell Laboratory of Ophthalmic Microbiology, University of Pittsburgh, Pittsburgh, United States of America

  • Viral R. Davra,

    Affiliation Department of Oral Biology, University of Medicine and Dentistry of New Jersey, Newark, New Jersey, United States of America

  • Eric G. Romanowski,

    Affiliation Department of Ophthalmology, Campbell Laboratory of Ophthalmic Microbiology, University of Pittsburgh, Pittsburgh, United States of America

  • Kimberly M. Brothers,

    Affiliation Department of Ophthalmology, Campbell Laboratory of Ophthalmic Microbiology, University of Pittsburgh, Pittsburgh, United States of America

  • Nicholas A. Stella,

    Affiliation Department of Ophthalmology, Campbell Laboratory of Ophthalmic Microbiology, University of Pittsburgh, Pittsburgh, United States of America

  • Dipti Godboley,

    Affiliation Department of Oral Biology, University of Medicine and Dentistry of New Jersey, Newark, New Jersey, United States of America

  • Daniel E. Kadouri

    Affiliation Department of Oral Biology, University of Medicine and Dentistry of New Jersey, Newark, New Jersey, United States of America

An Eye to a Kill: Using Predatory Bacteria to Control Gram-Negative Pathogens Associated with Ocular Infections

  • Robert M. Q. Shanks, 
  • Viral R. Davra, 
  • Eric G. Romanowski, 
  • Kimberly M. Brothers, 
  • Nicholas A. Stella, 
  • Dipti Godboley, 
  • Daniel E. Kadouri


Ocular infections are a leading cause of vision loss. It has been previously suggested that predatory prokaryotes might be used as live antibiotics to control infections. In this study, Pseudomonas aeruginosa and Serratia marcescens ocular isolates were exposed to the predatory bacteria Micavibrio aeruginosavorus and Bdellovibrio bacteriovorus. All tested S. marcescens isolates were susceptible to predation by B. bacteriovorus strains 109J and HD100. Seven of the 10 P. aeruginosa isolates were susceptible to predation by B. bacteriovorus 109J with 80% being attacked by M. aeruginosavorus. All of the 19 tested isolates were found to be sensitive to at least one predator. To further investigate the effect of the predators on eukaryotic cells, human corneal-limbal epithelial (HCLE) cells were exposed to high concentrations of the predators. Cytotoxicity assays demonstrated that predatory bacteria do not damage ocular surface cells in vitro whereas the P. aeruginosa used as a positive control was highly toxic. Furthermore, no increase in the production of the proinflammatory cytokines IL-8 and TNF-alpha was measured in HCLE cells after exposure to the predators. Finally, injection of high concentration of predatory bacteria into the hemocoel of Galleria mellonella, an established model system used to study microbial pathogenesis, did not result in any measurable negative effect to the host. Our results suggest that predatory bacteria could be considered in the near future as a safe topical bio-control agent to treat ocular infections.


In an era of increasing antibiotic resistance among bacterial pathogens, the search for new antibiotics and novel treatments for infections caused by these organisms is a priority among researchers. One novel treatment is biological therapy using specific bacteriophage for controlling the infecting pathogen [1][3]. Another novel treatment that might hold the potential to treat antibiotic resistant infections are predatory bacteria [4]. Recently, several studies have highlighted the ability of predatory bacteria Bdellovibrio spp. and Micavibrio spp. to prey on Gram-negative pathogens. Among the pathogens which were evaluated were bacteria associated with oral infections [5][7], gastrointestinal infections [8], zoonotic infection [9], pathogens associated with food processing and spoilage [10][12], as well as bacteria linked to systemic infections, burns and wounds [13]. Although the data published so far supports the claim that predatory bacteria could be used to control human pathogens, there is still concern regarding the toxic effects of administering large numbers of Gram-negative bacteria as live antibiotics. Therefore, treatment of local infections where the pathogens are easily accessible to topical or locally injected treatment would be ideal candidates to demonstrate a “proof of concept” that infections can be successfully treated with predatory bacteria.

One such local bacterial infection that is treated by direct administration of antibiotic to the site of infection is keratitis, infection of the cornea. Bacterial keratitis can be caused by both Gram-positive and Gram-negative pathogens. Common Gram-negative pathogens associated with keratitis are Pseudomonas aeruginosa and Serratia marcescens [14][16]. Bacterial keratitis is usually localized to an area of the cornea and is treated with antibiotic solutions delivered topically to the eye.

The first step in demonstrating that predatory bacteria can successfully treat bacterial keratitis caused by Gram-negative bacteria is showing that the predatory bacteria can kill Gram-negative bacteria isolated from keratitis cases. Secondly, we must show that the predatory bacteria are non-toxic and non-inflammatory to human cells. In this study we tested whether predatory bacterial species Bdellovibrio bacteriovorus and Micavibrio aeruginosavorus were able to kill keratitis isolates of P. aeruginosa and S. marcescens, including antibiotic-resistant isolates, in vitro. We also tested whether B. bacteriovorus and M. aeruginosavorus were cytotoxic and inflammatory to human corneal-limbal epithelial cells (HCLE) in vitro. To further test whether these predatory bacteria were pathogenic, we used the Galleria monella pathogenesis model to determine whether B. bacteriovorus and M. aeruginosavorus reduced the viability of the G. monella larvae.

Materials and Methods

Bacterial strains, and growth conditions

The predatory bacteria used in the study were Bdellovibrio bacteriovorus strains HD100, 109J (ATCC 43826) and Micavibrio aeruginosavorus strain ARL-13 [17], [18]. Ten Pseudomonas aeruginosa and nine Serratia marcescens isolates were examined in this study. All clinical isolates were isolated from keratitis patients by Dr. Ritterband at the New York Eye Infirmary and Regis Kowalski at the UPMC Eye Center. Many of these bacteria were fluoroquinolone resistant and previously used in antibiotic efficacy studies [19], [20]. Pseudomonas aeruginosa and Serratia marcescens were grown with aeration and maintained in LB media. Predator stock-lysates were prepared by co-culturing the predators in the presence of host bacteria suspended in diluted nutrient broth (DNB), a 1∶10 dilution of nutrient broth amended with 3 mM MgCl2 and 2 mM CaCl2 [13]. E. coli ZK2686 and P. aeruginosa UCBPP-PA14 were used as host cells for B. bacteriovorus and M. aeruginosavorus, respectively. The co-cultures were incubated on a rotary shaker at 30°C. Fresh predator cultures were prepared as previously described [13], [21], [22], in brief, 2 ml of overnight-grown host cells (∼1×109 CFU/ml) were added to 2 ml of predatory bacteria taken from a stock-lysate and suspended in 20 ml DNB. The co-cultures were incubated for 24 hrs at 30°C to reach ∼1×108 PFU/ml predator’s cells. At this point, the lysates were filtered through a 0.45 µm Millex-HV pore-size filter (Millipore, Billerica, MA) in order to remove any residual host cells (harvested predator).

Predation Experiments

Predation experiments were conducted as previously described [13]. Five ml co-cultures were prepared by adding 0.5 ml of washed host cells to 0.5 ml of freshly harvested predator bacteria in DNB media. The cultures were incubated at 30°C for 48 hrs. The capability of B. bacteriovorus and M. aeruginosavorus to prey was evaluated by the reduction in prey cell viability in the predator co-cultures. Cell viability was measured by dilution plating and CFU enumeration at 24 and 48 hrs. Each co-culture was conducted twice in triplicate.

Cytotoxicity assays

B. bacteriovorus and M. aeruginosavorus were prepared as described above using 5 ml of washed host cells and 5 ml of freshly harvested predator in 50 ml DNB media. The co-cultures were incubated for 24 and 36 hrs for B. bacteriovorus and M. aeruginosavorus, respectively. Thereafter, the lysates were filtered four times through a 0.45-µm Millex-HV pore-size filter to remove any residual host cells. The filtered harvested lysate was washed twice by centrifugation, 15,000 rpm for 30 min, and resuspended in 2 ml of DNB. Aliquots of the predator preparation was removed and plated on agar plates, to confirm that the samples are free from host cells. Samples were also taken to determine predator concentration using standard double-layered agar method [23]. Purifications were conducted on 3 and 4 separate occasions for M. aeruginosavorus and B. bacteriovorus, respectively.

Cytotoxicity assays were conducted as described [24] with some modifications, Human corneal-limbal epithelial (HCLE) [25] cells were cultured in 24-well plates until they were confluent. HCLE cells were grown in Keratinocyte serum-free medium (KSFM) with L-Glutamine, supplemented with 25 µg/ml BPE, 0.2 ng/ml EGF, and 1 mM CaCl2. HCLE cells were seeded without antibiotics to prevent interference of antibiotics in subsequent assays. The plates were incubated in an incubator at 37°C with 5% CO2. The wells were washed 3 times using PBS, pH 7.4 (Sigma-Aldrich, St. Louis, MO) and 450 µl of KSFM media was added to each well. Thereafter, wells were inoculated with 50 µl of each predator prep (∼0.2–1.1×109 PFU/well for B. bacteriovorus strains and ∼2×108 PFU/ well for M. aeruginosavorus) or predator free DNB control for maximum viability. Other controls included 0.25% of triton X-100 to measure total killing and 50 µl of DNB washed P. aeruginosa PA14 (∼2.5×107 CFU/well) as a positive control for bacterial cytotoxicity. Cell cultures were incubated for 4 and 24 hrs. After the incubation, aliquots of medium were removed from each well, centrifuged to remove bacteria, and stored at –20°C for subsequent pro-inflammatory cytokine analysis. The cells were then washed three times with PBS. Alamar Blue viability reagent (Invitrogen) in KSFM containing amikacin (10 µg/ml) was added to each well (500 µl/well) to assess cell viability. Fluorescence was measured after 1.5 hrs of incubation using a Synergy 2 microplate reader (Biotek) at 500/27 nm excitation and 620/40 nm emission wavelength. Experiments were conducted four times using B. bacteriovorus and three times using M. aeruginosavorus. Each experiment was conducted in quadruplicate (4 cell culture wells).

Cytokine analysis

HCLE supernatants were collected at 4 and 24 hrs post bacterial exposure. Four biological samples were used for ELISA and also tested on two different days with a different harvest sample set. IL-8 and TNFα ELISAs were run on the 4 and 24-hour sample sets according to manufacturer’s instructions (for IL-8, R & D Systems®; for TNFα, Thermo Scientific Pierce Biotechnology). Upon completion of the assay, samples were read according to the manufacturer’s instructions on a Synergy 2 plate reader (BioTek). Samples were graphed and statistical analysis was performed using GraphPad Prism 5 using one-way ANOVA with Tukey’s post-hoc test.

Toxicity assay in Galleria mellonella invertebrate infection model

Viability experiments were conducted as described previously [26] with some modifications. Galleria mellonella larvae were obtained from New York worms (New York Worms, Glen Cove, NY). Larvae were in their final instar stages and had equal size and weight (330±30 mg). B. bacteriovorus strains, 109J, HD100, and M. aeruginosavorus ARL-13 were grown and concentrated as descried for the cytotoxicity assays. The predators were suspended in phosphate buffered saline (PBS), pH 7.4 (Sigma-Aldrich, St. Louis, MO). Final predator concentration 2×109 PFU/ml. Five microliters of each sample was injected into the hemocoel of each larva via the last left proleg using a Hamilton 25 µl syringe and 30.5-gauge needle. Prior to use, the syringes were sterilized using bleach. The syringes were cleaned with 70% alcohol and distilled water and the needles were changed between every sample. In addition to the predators, worms were also injected with 5 µl of PBS buffer (negative control) and 5 µl of 8×104 CFU/ml P. aeruginosa PA14 (positive control). After injection, the worms were incubated at 30°C and the numbers of live larvae were scored for 11 days. Larvae were considered dead when they display no movement in response to gentle shaking of the dish or touching with a pipette tip. Six petri dishes containing 5 worms were assigned to each experimental and control groups (30 worms total for each sample).

Statistical Analysis

Graphpad Prism 5 software was used to perform statistical analysis. This analysis consisted of One-way ANOVA with Tukey’s multiple comparison test.


Predation by B. bacteriovorus and M. aeruginosavorus

When exposed to the predators, all of the isolates were found to be susceptible to at least one predator. All S. marcescens isolates were found to be susceptible to predation by both B. bacteriovorus 109J and HD100, with cell reduction ranging from 1.7 log10 to greater than 5 log10, compared to the initial cell concentration and the predator free control (Table 1). 100% of the tested P. aeruginosa isolates were reduced by B. bacteriovorus HD100. However, only 70% of the isolates were reduced by the 109J strain. Eight of the 10 P. aeruginosa isolates were reduced by M. aeruginosavorus with a greater than 2-log10 reduction measured for 87% of the predation positive strains (Table 1). It was previously shown that M. aeruginosavorus ARL-13 is able to use P. aeruginosa as prey; however, it is unable to utilize S. marcescens [13], [22]. Therefore, in this study, M. aeruginosavorus was not tested on S. marcescens.

Table 1. Predation of S. marcescens and P. aeruginosa ocular isolates by predatory bacteria.

Cytotoxicity B. bacteriovorus and M. aeruginosavorus to HCLE cells

As a first step towards judging the suitability of predatory bacteria for ocular infections, we tested whether B. bacteriovorus and M. aeruginosavorus were cytotoxic to HCLE cells in vitro. Bacteria were co-incubated with HCLE cells at an MOI of ∼100 with a known cytotoxic P. aeruginosa strain [27] as a positive control, and >800 for each of the predatory bacteria. Bacteria and HCLEs were co-incubated for 4 and 24 hours, then bacteria were removed and HCLE cells were tested for viability using the fluorescent vital stain alamar blue. Whereas P. aeruginosa was highly cytotoxic at both time points, the predatory bacteria were not significantly different from the mock at either 4 or 24 hrs (p>0.05, ANOVA, with Tukey’s post-test) (Figure 1).

Figure 1. Cytotoxicity to human corneal-limbal epithelial cells in vitro.

Alamar blue vital stain was used to measure cytotoxicity from positive control P. aeruginosa strain PA14 (average MOI  =  111), detergent lysis (LYSIS), medium only negative control (MOCK), and experimental strains B. bacteriovorus strain 109J (average MOI  =  4720), B. bacteriovorus strain HD100 (average MOI  =  1039), and M. aeruginosavorus (Mica, average MOI  =  853). HCLE viability was measured after 4 h (A) and 24 h (B) of exposure. Total independent data points from 4 experiments are shown. Asterisks indicate significant differences (p<0.001, ANOVA with Tukey's post-test). Only PA14 was significantly different than MOCK. Error bars indicate one standard deviation.

Production of pro-inflammatory cytokines following exposure to predatory bacteria

Because the predatory bacteria used in this study are Gram-negative bacteria, we predicted that they may cause adverse inflammatory effects upon exposure to ocular cells. Supernatants of HCLE cells co-incubated for four hrs with B. bacteriovorus and M. aeruginosavorus in the above noted cytotoxicity studies and were analyzed for proinflammatory cytokines IL-8 and TNF-a. These cytokines were chosen because they are expressed by ocular surface cells exposed to bacteria [28], [29]. Whereas the positive control, P. aeruginosa, elicited a strong and significant induction of both cytokines, neither IL-8 nor TNF-a was found to be elevated above the mock negative control in HCLE supernatants co-incubated with any of the three predatory bacteria (Figure 2A–B). The same pattern was observed after co-incubation of the predatory and positive control bacteria at 24 hrs for IL-8; however, TNF-a levels were undetectable (data not shown).

Figure 2. Inflammatory response of human corneal-limbal epithelial cells to predatory bacteria in vitro.

Pro-inflammatory cytokines IL-8 (panel A) and TNF-α (panel B) were measured using ELISA assays. Cell supernatants taken from HCLE cells after 4 hrs of incubation with positive control Pseudomonas aeruginosa strain PA14 (average MOI  =  111), detergent lysis (LYSIS), medium only negative control (DNB), and experimental strains B. bacteriovorus strain 109J (avgerage MOI  =  4720), B. bacteriovorus strain HD100 (average MOI  =  1039), and M. aeruginosavorus (Mica, average MOI  =  853). Total independent data points from 2 experiments are shown. Asterisks indicate significant differences (p<0.001, ANOVA with Tukey's post-test). Only PA14 was significantly different than MOCK. Error bars indicate one standard deviation.

In vivo effect of predatory bacteria on G. mellonella

G. mellonella was recently recognized as a suitable host model system to study microbial pathogenesis and innate immunity [30][34]. The G. mellonella system generally demonstrates a positive correlation between virulence factors found in mammals to those isolated in the insect, emphasizing the ability to utilize the system to bridge between in vitro studies and vertebrates [26]. In order to further evaluate the potential risk of using predatory bacteria G. mellonella worms were exposed to high concentrations of each predator. All of the worms injected with P. aeruginosa PA14 (4×102 CFU/worm) were nonviable 24 hrs post-injection. However, worms injected with PBS, B. bacteriovorus 109J, HD100 (1.1×107 PFU/worm) and M. aeruginosavorus ARL-13 (0.9×107 PFU/worm) were all viable 24 hrs post-injection. 11 days post-injection, the viability of the worms were 96.6%, 100%, 96.6% and 93.3% viability for the control, B. bacteriovorus 109J, HD100 and M. aeruginosavorus ARL-13, respectively. Furthermore, no change in larva pigmentation was observed in the predator-infected worms (Data not shown). In G. mellonella the change in color indicates melanization caused by the host immune response to the microbial challenge [26]. Thus, based on our finding it could be concluded, that unlike other pathogens, predatory bacteria do not provoke an aggressive innate immune response when injected.


In this study, we have demonstrated that M. aeruginosavorus ARL-13 is able to prey on clinical isolates of P. aeruginosa isolated from ocular infections. This finding is in line with earlier reports regarding the host specificity of this predator and its ability to attack P. aeruginosa [13], [22], [35]. Our data also suggest that both B. bacteriovorus 109J and HD100 are capable of using S. marcescens as a host. This finding is in agreement with a study reporting a 3 log10 reduction in cell viability of a non clinical isolate of S. marcescens following predation by B. bacteriovorus 109J [13]. In addition, B. bacteriovorus 109J and HD100 were able to prey on P. aeruginosa. However, the ability of the HD100 strain to attack was broader than that of 109J, preying on 100% and 70% of the isolates, respectively. The narrower ability of B. bacteriovorus 109J to prey on P. aeruginosa is aligned with a recent report in which B. bacteriovorus 109J was able to reduce only 1 out of 4 P. aeruginosa examined strains [5], as wall as earlier findings showing a limited ability of some B. bacteriovorus to prey on P. aeruginosa [8], [36]. The different host- and intra-species strain- specificity demonstrated by Bdellovibrio spp. and Micavibrio spp. is well documented. Furthermore, predation ability was found to be specific to the B. bacteriovorus strain used, with different B. bacteriovorus strains demonstrating unique host specificity [8], [13], [22], [35][38]. As the mechanisms that define the predator’s host specificity are not fully known, we could only speculate on the reason why certain bacteria are recognized as a host while others are not.

A major concern that needs to be addressed when evaluating the potential use of predatory bacteria as topical live-antibiotics are the risks associated with applying Gram-negative microorganisms to human cells. To this end, we have conducted cytotoxicity assays in which HCLE cells were exposed to high concentrations of the predators. Our data indicate that predatory bacteria are significantly less cytotoxic than the control P. aeruginosa. Low cytotoxicity was observed even after an extended exposure period. Although only a small number of studies regarding the safety of using predatory bacteria were conducted, the current data does support the claims that predatory bacteria could be considered safe. In a review article published by Dwider at al [4] the authors cited a study in which Bdellovibrio was injected into mice rabbits and guinea pigs and were found to be non pathogenic to the animals [39]. In a separate study conducted by Lenz and Hespell [40], the investigators attempted to grow the predatory bacteria B. bacteriovorus 109J, Bacteriovorax stolpii UKi2 and Peredibacter starrii A3.12 in the presence of eukaryotic cells. It was concluded that predatory bacteria are unable to grow on hamster kidney cells, mouse liver cells and bovine mammary gland cells. Furthermore, the predators did not grow within rabbit ova, following injection, nor were they able to grow in media containing rabbit ova extracts. Thus, it seems that mammalian cells could not be used by the predators as prey and could not support predator proliferation in the absent of a Gram-negative host. The inability of the predator to grow and establish itself in the intestinal microflora was shown in a study in which Bdellovibrio strain MS7 was fed to Channel catfish, northern leopard frogs, and mice. The predator viability also declined when inoculated into rabbit ileal loops. As Bdellovibrio could not proliferate in vivo it reduces the risk of permanent establishment within the mammalian host, rendering the predator, in the study, as nonpathogenic [41].

In a recent study conducted at the University of Nottingham [9], the in vivo effect of Bdellovibrio in a poultry vertebrate model was examined. It was found that B. bacteriovorus HD100, which was orally administered to chicks, caused no negative health effects on the birds. Furthermore, the authors were not able to recover viable Bdellovibrio from the gut flora, fecal matter or drinking water of the predator-inoculated birds, concluding that the risk of spreading predatory bacteria during treatment is low. Bdellovibrio treatment was also found to improve the well being of the birds colonized with Salmonella Enteritidis in the therapeutic trail.

An additional concern of applying live predatory bacteria is the risk of inadvertently causing inflammation which in could inhibit wound healing and increase the risk of tissue damage. In this study we have demonstrated that exposure to high doses of the predators did not elevate the production of the proinflammatory cytokines IL-8 and TNF-alpha by HCLE cells. Experimental evidence supports that neutrophils attracted by the bacteria-induced inflammation are a major cause of scarring and tissue damage associated with vision loss in keratitis [42]. The low cytotoxic activity of B. bacteriovorus HD100 LPS and its reduced ability to induce TNF-alpha and IL-6, compared to an E. coli control, was previously reported in a study using a human macrophage cell line [43]. The authors attributed their findings to the unique structure of the B. bacteriovorus LPS Lipid A molecule. Unlike Lipid A from many Gram-negative bacteria that contain negatively charged phosphate groups, the B. bacteriovorus Lipid A molecule has α-D-mannose residues which reduced its affinity to LPS receptors thereby lowering inflammation. Our data confirm that B. bacteriovorus HD100 and 109J do not enhance proinflammatory cytokines production. Our data also demonstrate, for the first time, that as reported for B. bacteriovorus o, M. aeruginosavorus also does not enhance inflammation.

Although, our data show that predatory bacteria have little or no adverse effect when applied to human cell cultures, we were interested to evaluate the effect of predatory bacteria in vivo. To address whether predatory bacteria a tolerated by eukaryotic cells in vivo, a G. mellonella microbial pathogenesis model was selected. Although, we did not measure the viability of the predators within the worm over time, we might still conclude that injecting relatively high doses of predatory bacteria do not provoke any measurable toxic or harmful effect to the worm. G. mellonella is recognized as a suitable host model system to study the pathogenesis of both bacteria and yeast and was used to examine pathogenic attributes of many human pathogens including pathogens associated with eye infections [26], [30], [32]. These studies established the use of G. mellonella for a variety of applications such as: examining microbial pathogenicity and lethality, evaluating microbial growth and proliferation, isolating virulence factors, and inspecting putative virulence mechanisms [30], [32]-[34], [44]. Since the innate immune systems of mammals and insects have several features in common [31], G. mellonella could also be used as a model system for studying the host innate immune response to microbial infection as well as identifying microbial virulence factors that mediate the immune response [33], [45].

The potential use of predatory bacteria as a bio-control agent to treat eye infections was first suggested some 40 years ago. In a study conducted in 1972 [46] the “pro-biotic” ability of E. coli and B. bacteriovorus to impact the pathogenesis of Shigella flexneri in animal modules was examined. It was shown that the B. bacteriovorus was able to reduce the severity of keratoconjunctivitis induced by S. flexneri in a rabbit keratoconjunctivitis model. The simultaneous inoculation of Bdellovibrio with S. flexneri was able to prevent the development of the infection. The rate of development of typical keratoconjunctivitis was also decreased when Bdellovibrio was administered within 48 hrs of the initial S. flexneri infection. In a more recent study, the ability of B. bacteriovorus 109J to inhibit growth of and reduce the adherence of Moraxella bovis to Madin-Darby bovine kidney (MDBK) cells, used to mimic bovine keratoconjunctivitis, was confirmed [47]. The ability of Bdellovibrio to survive and prey in human fluids was also demonstrated in an experiment in which B. bacteriovorus 109J was abele to significantly reduce biofilms of Aggregatibacter actinomycetemcomitans in the presence of human saliva that contains many of the same antimicrobial compounds as do tears [5].

In conclusion, our work demonstrates that predatory bacteria have the ability to attack “real-life” Gram-negative human pathogens associated with ocular infection. Furthermore, in vitro studies had revealed that the presence of high concentrations of predatory bacteria don’t appear to be harmful to human cells. These findings, coupled with the ability of predator bacteria to prey in conditions that might be encountered in the eye, emphasize the potential use of applying predator bacteria as a topical agent to treat eye infections caused by pathogens which are resistant to traditional antimicrobials. The efficacy of predatory bacteria to control infection using in vivo models of ocular infection should be the focus of future studies. The long-term goal is to develop a topical predator bacteria product, which might include a single or multispecies predator bacteria mix and could be used alone or in concert with traditional antimicrobial therapies.


Kevin To, Genesis Cano and Tiffany Aston provided technical assistance.

Author Contributions

Conceived and designed the experiments: DEK RMQS. Performed the experiments: VRD EGR KMB DG DEK. Analyzed the data: DEK RMQS NAS. Contributed reagents/materials/analysis tools: RMQS DEK. Wrote the paper: RMQS DEK NAS.


  1. 1. Fukuda K, Ishida W, Uchiyama J, Rashel M, Kato S, et al. (2012) Pseudomonas aeruginosa keratitis in mice: effects of topical bacteriophage KPP12 administration. PLoS One 7: e47742.
  2. 2. Gorski A, Targonska M, Borysowski J, Weber-Dabrowska B (2009) The potential of phage therapy in bacterial infections of the eye. Ophthalmologica 223: 162–165.
  3. 3. Santos TM, Ledbetter EC, Caixeta LS, Bicalho ML, Bicalho RC (2011) Isolation and characterization of two bacteriophages with strong in vitro antimicrobial activity against Pseudomonas aeruginosa isolated from dogs with ocular infections. Am J Vet Res 72: 1079–1086.
  4. 4. Dwidar M, Monnappa AK, Mitchell RJ (2012) The dual probiotic and antibiotic nature of Bdellovibrio bacteriovorus. BMB Rep 45: 71–78.
  5. 5. Dashiff A, Kadouri DE (2011) Predation of oral pathogens by Bdellovibrio bacteriovorus 109J. Mol Oral Microbiol 26: 19–34.
  6. 6. Van Essche M, Quirynen M, Sliepen I, Loozen G, Boon N, et al. (2011) Killing of anaerobic pathogens by predatory bacteria. Mol Oral Microbiol 26: 52–61.
  7. 7. Van Essche M, Quirynen M, Sliepen I, Van Eldere J, Teughels W (2009) Bdellovibrio bacteriovorus attacks Aggregatibacter actinomycetemcomitans. J Dent Res 88: 182–186.
  8. 8. Markelova NY (2010) Predacious bacteria, Bdellovibrio with potential for biocontrol. Int J Hyg Environ Health 213: 428–431.
  9. 9. Atterbury RJ, Hobley L, Till R, Lambert C, Capeness MJ, et al. (2011) Effects of orally administered Bdellovibrio bacteriovorus on the well-being and Salmonella colonization of young chicks. Appl Environ Microbiol 77: 5794–5803.
  10. 10. Lu F, Cai J (2010) The protective effect of Bdellovibrio-and-like organisms (BALO) on tilapia fish fillets against Salmonella enterica ssp. enterica serovar Typhimurium. Lett Appl Microbiol 51: 625–631.
  11. 11. Fratamico PM, Whiting RC (1995) Ability of Bdellovibrio-bacteriovorus 109J to Lyse Gram-Negative Food-Borne Pathogenic and Spoilage Bacteria. J Food Protect 58: 160–164.
  12. 12. Fratamico PM, Whiting RC (1996) Isolation of Bdellovibrios that prey on Escherichia coli O157:H7 and Salmonella species and application for removal of prey from stainless steel surfaces. J Food Safety 16: 161–173.
  13. 13. Dashiff A, Junka RA, Libera M, Kadouri DE (2011) Predation of human pathogens by the predatory bacteria Micavibrio aeruginosavorus and Bdellovibrio bacteriovorus. J Appl Microbiol 110: 431–444.
  14. 14. Das S, Sheorey H, Taylor HR, Vajpayee RB (2007) Association between cultures of contact lens and corneal scraping in contact lens related microbial keratitis. Arch Ophthalmol 125: 1182–1185.
  15. 15. Hume EB, Willcox MD (2004) Emergence of Serratia marcescens as an ocular surface pathogen. Arch Soc Esp Oftalmol 79: 475–477.
  16. 16. Varaprasathan G, Miller K, Lietman T, Whitcher JP, Cevallos V, et al. (2004) Trends in the etiology of infectious corneal ulcers at the F. I. Proctor Foundation. Cornea 23: 360–364.
  17. 17. Rendulic S, Jagtap P, Rosinus A, Eppinger M, Baar C, et al. (2004) A predator unmasked: life cycle of Bdellovibrio bacteriovorus from a genomic perspective. Science 303: 689–692.
  18. 18. Wang Z, Kadouri DE, Wu M (2011) Genomic insights into an obligate epibiotic bacterial predator: Micavibrio aeruginosavorus ARL-13. BMC Genomics 12: 453.
  19. 19. Kowalski RP, Pandya AN, Karenchak LM, Romanowski EG, Husted RC, et al. (2001) An in vitro resistance study of levofloxacin, ciprofloxacin, and ofloxacin using keratitis isolates of Staphylococcus aureus and Pseudomonas aeruginosa.. Ophthalmology 108: 1826–1829.
  20. 20. Rhee MK, Kowalski RP, Romanowski EG, Mah FS, Ritterband DC, et al. (2004) A laboratory evaluation of antibiotic therapy for ciprofloxacin-resistant Pseudomonas aeruginosa.. Am J Ophthalmol 138: 226–230.
  21. 21. Kadouri D, O'Toole GA (2005) Susceptibility of biofilms to Bdellovibrio bacteriovorus attack. Appl Environ Microbiol 71: 4044–4051.
  22. 22. Kadouri D, Venzon NC, O'Toole GA (2007) Vulnerability of pathogenic biofilms to Micavibrio aeruginosavorus. Appl Environ Microbiol 73: 605–614.
  23. 23. Shilo M, Bruff B (1965) Lysis of Gram-negative bacteria by host-independent ectoparasitic Bdellovibrio bacteriovorus isolates. J Gen Microbiol 40: 317–328.
  24. 24. Wingard JB, Romanowski EG, Kowalski RP, Mah FS, Ling Y, et al. (2011) A novel cell-associated protection assay demonstrates the ability of certain antibiotics to protect ocular surface cell lines from subsequent clinical Staphylococcus aureus challenge. Antimicrob Agents Chemother 55: 3788–3794.
  25. 25. Gipson IK, Spurr-Michaud S, Argueso P, Tisdale A, Ng TF, et al. (2003) Mucin gene expression in immortalized human corneal-limbal and conjunctival epithelial cell lines. Invest Ophthalmol Vis Sci 44: 2496–2506.
  26. 26. Jander G, Rahme LG, Ausubel FM (2000) Positive correlation between virulence of Pseudomonas aeruginosa mutants in mice and insects. J Bacteriol 182: 3843–3845.
  27. 27. Ramirez JC, Fleiszig SM, Sullivan AB, Tam C, Borazjani R, et al. (2012) Traversal of multilayered corneal epithelia by cytotoxic Pseudomonas aeruginosa requires the phospholipase domain of exoU. Invest Ophthalmol Vis Sci 53: 448–453.
  28. 28. Willcox MD (2007) Pseudomonas aeruginosa infection and inflammation during contact lens wear: a review. Optom Vis Sci 84: 273–278.
  29. 29. Zhang J, Wu XY, Yu FS (2005) Inflammatory responses of corneal epithelial cells to Pseudomonas aeruginosa infection. Curr Eye Res 30: 527–534.
  30. 30. Garcia-Lara J, Needham AJ, Foster SJ (2005) Invertebrates as animal models for Staphylococcus aureus pathogenesis: a window into host-pathogen interaction. FEMS Immunol Med Microbiol 43: 311–323.
  31. 31. Kavanagh K, Reeves EP (2004) Exploiting the potential of insects for in vivo pathogenicity testing of microbial pathogens. FEMS Microbiol Rev 28: 101–112.
  32. 32. Miyata S, Casey M, Frank DW, Ausubel FM, Drenkard E (2003) Use of the Galleria mellonella caterpillar as a model host to study the role of the type III secretion system in Pseudomonas aeruginosa pathogenesis. Infect Immun 71: 2404–2413.
  33. 33. Peleg AY, Jara S, Monga D, Eliopoulos GM, Moellering RC Jr, et al. (2009) Galleria mellonella as a model system to study Acinetobacter baumannii pathogenesis and therapeutics. Antimicrob Agents Chemother 53: 2605–2609.
  34. 34. Seed KD, Dennis JJ (2008) Development of Galleria mellonella as an alternative infection model for the Burkholderia cepacia complex. Infect Immun 76: 1267–1275.
  35. 35. Lambina VA, Afinogenova AV, Romay Penobad Z, Konovalova SM, Andreev LV (1983) New species of exoparasitic bacteria of the genus Micavibrio infecting gram-positive bacteria. Mikrobiologiia 52: 777–780.
  36. 36. Stolp H, Starr MP (1963) Bdellovibrio Bacteriovorus gen. et sp. n., a Predatory, Ectoparasitic, and Bacteriolytic Microorganism. Antonie Van Leeuwenhoek 29: 217–248.
  37. 37. Jurkevitch E, Minz D, Ramati B, Barel G (2000) Prey range characterization, ribotyping, and diversity of soil and rhizosphere Bdellovibrio spp. isolated on phytopathogenic bacteria. Appl Environ Microbiol 66: 2365–2371.
  38. 38. Rogosky AM, Moak PL, Emmert EA (2006) Differential predation by Bdellovibrio bacteriovorus 109J. Curr Microbiol 52: 81–85.
  39. 39. Verklova ZS (1973) Study of the virulence, toxicity and immunogenicity of different strains of Bdellovibrio bacteriovorus. Gig Sanit 38: 10–13.
  40. 40. Lenz RW, Hespell RB (1978) Attempts to Grow Bdellovibrios Micurgically-Injected into Animal-Cells. Arch Microbio 119: 245–248.
  41. 41. Westergaard JM, Kramer TT (1977) Bdellovibrio and the intestinal flora of vertebrates. Appl Environ Microbiol 34: 506–511.
  42. 42. Hazlett LD (2005) Role of innate and adaptive immunity in the pathogenesis of keratitis. Ocul Immunol Inflamm 13: 133–138.
  43. 43. Schwudke D, Linscheid M, Strauch E, Appel B, Zahringer U, et al. (2003) The obligate predatory Bdellovibrio bacteriovorus possesses a neutral lipid A containing alpha-D-Mannoses that replace phosphate residues: similarities and differences between the lipid As and the lipopolysaccharides of the wild type strain B. bacteriovorus HD100 and its host-independent derivative HI100. J Biol Chem 278: 27502–27512.
  44. 44. Andrejko M, Mizerska-Dudka M, Jakubowicz T (2009) Antibacterial activity in vivo and in vitro in the hemolymph of Galleria mellonella infected with Pseudomonas aeruginosa.. Comp Biochem Physiol B Biochem Mol Biol 152: 118–123.
  45. 45. Dubovskiy IM, Krukova NA, Glupov VV (2008) Phagocytic activity and encapsulation rate of Galleria mellonella larval haemocytes during bacterial infection by Bacillus thuringiensis. J Invertebr Pathol 98: 360–362.
  46. 46. Nakamura M (1972) Alteration of Shigella pathogenicity by other bacteria. Am J Clin Nutr 25: 1441–1451.
  47. 47. Boileau MJ, Clinkenbeard KD, Iandolo JJ (2011) Assessment of Bdellovibrio bacteriovorus 109J killing of Moraxella bovis in an in vitro model of infectious bovine keratoconjunctivitis. Can J Vet Res 75: 285–291.