For figures, tables and references we refer the reader to the original paper.
Bdellovibrio and like organisms (BALOs) are highly motile proteobacteria that prey on other Gram-negative bacteria. They are obligatory predators, since they need to hijack their prey’s macromolecules as fuel and essential building blocks for multiplication. The most-studied strain of this group is Bdellovibrio bacteriovorus HD100. After attaching to its prey, B. bacteriovorus enters the prey’s periplasmic space and starts to multiply while devouring the cytoplasm of the prey cell. When the predator’s multiplication cycle is completed, it shatters the remains of the prey cell, releasing its progeny into the environment (Sockett and Lambert, 2004). BALOs are ubiquitous in nature and have been isolated from diverse terrestrial and aquatic habitats, from biofilms as well as from animal feces (Schwudke et al., 2001; Davidov et al., 2006).
Periodontitis is an infectious disease primarily associated with Gram-negative periodontopathogens. Therefore, an effect of BALOs on periodontopathogens could be of great clinical relevance. Aggregatibacter actinomycetemcomitans is considered to be a key periodontopathogen and has been associated with aggressive periodontitis (Tonetti and Mombelli, 1999). Periodontitis associated with A. actinomycetemcomitans is difficult to treat reliably by mechanical removal of the subgingival biofilm alone (Mombelli et al., 1994; Takamatsu et al., 1999). Therefore, the treatment is often supplemented with (systemic) antibiotics (van Winkelhoff et al., 1989). Due to the emergence of antibiotic resistance in pathogenic bacteria such as periodontopathogens (van Winkelhoff et al., 2005), alternative treatment strategies need to be developed. This study tested the hypothesis that B. bacteriovorus can prey on and kill A. actinomycetemcomitans under conditions mimicking those in the oral cavity.
A. actinomycetemcomitans strains ATCC 29523, ATCC 43718, ATCC 33384, JP2, and strain 2751 (clinical strain) were grown on blood agar (Blood Agar Base II; Oxoid, Basingstoke, England), supplemented with hemin (5 mg/mL), menadione (1 mg/mL), and 5% sterile horse blood. Broth cultures were grown in Brain Heart Infusion (BHI) broth (Oxoid) at 37°C in 5% CO2.
E. coli ML 35 was provided by E. Jurkevitch (The Hebrew University of Jerusalem, Rehovot, Israel) and was cultured in Luria Bertani broth and on agar medium (Oxoid). Bdellovibrio bacteriovorus HD100 was purchased from DSMZ (Braunschweig, Germany) and cultured on E. coli ML35 prey (Jurkevitch, 2005). To harvest B. bacteriovorus bacteria, we separated B. bacteriovorus cells from the remaining prey by filtration over a 1.2-μm Acrodisk filter (Millipore Filter Corporation, Bedford, MA, USA). After 2 additional filtrations over 0.45-μm Acrodisk filters, the cells were centrifuged (27,000 x g; 20 min). The pellet was subsequently re-suspended in HM medium containing HEPES (25 mM), calcium (3 mM), and magnesium (2 mM) at pH 7.6. The optical density at 600 nm 9 (OD600) was adjusted to 0.100 (≈ 1 x 10 plaque-forming units [PFU] per mL) (Genesys 20, Thermo Electron Corporation, New York, NY, USA). HM medium is the gold standard medium for the cultivation of BALOs (Jurkevitch, 2005). Additionally, cells were re-suspended in Gibbons and Etherden buffered KCl medium (GEB) at pH 7. GEB medium has frequently been used in assays to simulate the ionic composition of saliva (Appelbaum et al., 1979; Gibbons and Etherden, 1982; Peros and Gibbons, 1986; Mombelli, 1999).
B. bacteriovorus Predation of Planktonic Bacteria
Overnight stationary cultures of A. actinomycetemcomitans ATCC 43718 in BHI were centrifuged (7000 x g; 5 min) and washed with HM or GEB medium. Bacteria were then re-suspended to an OD600 of 0.350 in GEB or HM medium (≈ 1 x 109 CFU/mL). Aliquots of 2 mL were transferred to glass tubes. Half of the tubes (test) received 200 μL of Bdellovibrio suspension. The control tubes received 200 μL of a 0.22-μm filtrate of the former Bdellovibrio suspension. The multiplicity of infection (MOI) was 1 PFU predator:1 CFU prey. Pilot investigations determined that this MOI was optimal for study of the dynamics of both predatory as well as prey bacteria in this model. After inoculation, tubes were incubated (37°C) in a rotary shaker at 200 rpm. Samples from test and control tubes were taken at different timepoints for quantitative microbial culturing and for quantitative PCR (qPCR). Samples for qPCR were frozen at −80°C, after which the DNA was extracted with InstaGene matrix (Bio-Rad Life Science Research, Hercules, CA, USA).
Experiments were repeated in GEB medium with A. actinomycetemcomitans strains ATCC 29523, ATCC 33384, and JP2. In these cases, the viability of the pathogen was measured at the start of the experiment and after 18 hrs. Triplicate experiments were performed and repeated on 5 separate days.
B. bacteriovorus Predation of Biofilms
Biofilms of A. actinomycetemcomitans 2751 were developed by inoculation of glass circular slides in a 24-well plate containing BHI. Inoculated slides were incubated overnight at 37°C with 5% CO2. After incubation, the medium was changed, and the biofilm was allowed to grow for one more night. Mature biofilms were washed 3 times with GEB medium. Half of the slides were inoculated with a B. bacteriovorus suspension in GEB medium (test). The other half of the slides were inoculated with a 0.22-μm filtrate of the former suspension (control). All assays were performed at 37°C in a rotary shaker at 100 rpm.
After 0 min, 30 min, and 18 hrs, slides of each series were washed 6 times with 1 mL of GEB medium and fixed for scanning electron microscopy (SEM).
To quantify the destruction of the biofilm, we retrieved 3 glass slides of the test and control sets at different timepoints and washed them 6 times. Thereafter, the biofilm on the slides was fixed by heat and stained with a crystal violet solution for 1 min. Excess crystal violet was washed away under running water. The biofilm was subsequently quantified by measurement of the optical density of the stained biofilm at 550 nm (OD550) with a Genesys 20 photometer.
Scanning Electron Microscopy
The biofilms on glass slides were fixed and dehydrated as previously described (Bergmans et al., 2005). Fixed slides were gold-sprayed and viewed under high voltage (10.0 kV) with a Philips XL30 ESEM-FeG scanning electron microscope (FEI/Philips Electron Optics, Eindhoven, the Netherlands).
Quantitative Real-time PCR (qPCR)
qPCR was performed with the TaqMan methodology recently developed (Van Essche et al., unpublished observations). Probes were designed to anneal to specific regions of the 16S rRNA gene of B. bacteriovorus HD100. This allowed for a sensitive and specific quantification of the Bdellovibrionaceae (Davidov et al., 2006). qPCR was performed on the ABI 7700 Sequence Detection System platform (Applied Biosystems, Foster City, CA, USA). The qPCR reaction mix was composed of 2 x qPCR master mix (Eurogentec, Liège, Belgium), 900 nM of each primer, 50 nM probe, and 3 μL sample DNA. Thermal cycling conditions were: 2 min at 50°C, 10 min at 95°C, followed by 45 repeats of 15 sec at 95°C, and 1 min at 60°C. Data were collected during the annealing phase. In each run, negative controls (no template) were included. As a standard for the qPCR, a fragment of the B. bacteriovorus HD100 16S rRNA gene was cloned into a plasmid with the pGEM-T easy vector system (Promega, Madison, WI, USA). Plasmids were isolated from the clones with the High Pure Plasmid Isolation Kit (Roche Diagnostics GmbH, Mannheim, Germany). The concentration of the plasmid was determined with the GeneQuant RNA/DNA calculator (Amersham Pharmacia Biotech/GE Healthcare, Diegem, Belgium) at a wavelength of 260 nm. A 10-fold dilution series of this plasmid was used in each qPCR run to construct the standard curve.
Statistical analyses were carried out in R for Windows, Version 2.6.1. A linear mixed model was fit to assess difference among strains, media, and time. Residual analysis was carried out, and the data were log-transformed to obtain normally distributed residuals, if necessary. If a factor was found to be significant, multiple comparisons were performed, and the corrections for simultaneous hypotheses were carried out (Bretz et al., 2001).
Predation Kinetics of B. bacteriovorus and Planktonic A. actinomycetemcomitans
In HM medium, B. bacteriovorus decreased A. actinomycetemcomitans viability over time (Fig. 1A⇓). After 8 hrs, the difference in viability between B. bacteriovorus-challenged and control bacteria became statistically significant (p < 0.001) and reached a maximum after 12 hrs (p < 0.001), at which time there was a difference of 2.43 (± 0.13 SEM) log10 CFU/mL between B. bacteriovorus-challenged and control bacteria. During the B. bacteriovorus attack, the number of B. bacteriovorus significantly increased 18.4-fold over the first 12 hrs (p < 0.001).
Kinetics of the lysis of A. actinomycetemcomitans ATCC 43718 by B. bacteriovorus HD100. (A) The results obtained in HM medium. (B) The results obtained in GEB medium. The series with squares represent the number of CFU of A. actinomycetemcomitans per mL sample in the controls that were not exposed to B. bacteriovorus. The series with diamonds represent the number of CFU of A. actinomycetemcomitans per mL sample that were exposed to B. bacteriovorus. The growth of B. bacteriovorus on A. actinomycetemcomitans is represented by the series with triangles and is expressed as genomes per mL sample. Error bars represent the standard error of the mean. N = 5.
In GEB medium, the predation kinetics was similar to the kinetics in HM medium (Fig. 1B⇑), with a significant reduction of A. actinomycetemcomitans viability after 8 hrs of incubation (p < 0.001). The difference in viability between B. bacteriovorus-challenged and control bacteria reached its maximum after 12 hrs (p < 0.001). This 1.85 (± 0.12 SEM) log10 CFU/mL difference, however, was significantly smaller than that obtained in HM medium (p = 0.015).
Similarly, the B. bacteriovorus population increased to a lesser extent (6.28 fold) when compared with the increase in HM medium (p = 0.048).
Susceptibility of different A. actinomycetemcomitans Strains
All 4 tested strains were susceptible to B. bacteriovorus attack (Fig. 2⇓). No significant difference was observed in the reduction of vitality among the strains. Depending on the strain, the A. actinomycetemcomitans viability was reduced, on average, by 1.59–2.38 log10 CFU/mL after 18 hrs of B. bacteriovorus attack in GEB medium.
The reduction of the viability of different A. actinomycetemcomitans strains by B. bacteriovorus HD100. The reduction of viability of the different A. actinomycetemcomitans strains caused by B. bacteriovorus HD100 is expressed in log 10 CFU/mL. To compensate for non-specific decrease of viability of the prey by, e.g., starvation, we present the difference in viability between control (not exposed to predator) and test samples (exposed to predator). The experiment was performed in GEB medium, and samples were taken after 18 hrs of incubation. N = 5.
Predation of A. actinomycetemcomitans Biofilms
The SEM images (Fig. 3⇓) show the breakdown of an A. actinomycetemcomitans 2751 biofilm after a B. bacteriovorus attack. At the beginning of the experiment, the SEM images show large mushroom-like structures of A. actinomycetemcomitans biofilm (Fig. 3A⇓). These structures were well-maintained in GEB medium throughout the 18-hour experiment. However, 30 min after incubation with B. bacteriovorus, many predatory bacteria adhering to the biofilm could be observed (Fig. 3B⇓). After 18 hrs of incubation with the predator, there was a noticeable reduction in the thickness and extent of the biofilm (Figs. 3E, 3F⇓) compared with the control biofilm (Figs. 3C, 3D⇓). The biofilm, however, was not completely cleared. At high magnification, the surface of the control biofilm showed mainly bacterial cells embedded in a small quantity of extracellular material, whereas the biofilm that was attacked by the predator showed few intact prey bacteria remaining embedded in amorphous material.
The effect of predation of B. bacteriovorus HD100 on a biofilm of A. actinomycetemcomitans 2751: SEM images. (A) The biofilm of A. actinomycetemcomitans 2751 at the start of the experiment. (B) The same biofilm 30 min after exposure to B. bacteriovorus HD100. Numerous predatory bacteria can be observed adhering to the biofilm (arrows). (C,D) The control biofilm after 18 hrs of incubation. (E,F) The biofilm after 18 hrs of incubation with the predatory bacterium. The bars represent 1μm, 10 μm, or 20 μm, as noted on the images.
The destruction of the biofilm was quantified by crystal violet staining, followed by measurement of the remaining OD550 (Fig. 4⇓). The difference in OD550 became statistically significant after 12 hrs of incubation (p = 0.029). After 24 hrs, only 26.2% of the OD550 was maintained on the biofilms exposed to the predator, compared with its control (p = 0.0087).
Quantification of the remaining A. actinomycetemcomitans 2751 biofilm after exposure to B. bacteriovorus HD100. The graph represents the OD550 of the crystal-violet-stained biofilm of A. actinomycetemcomitans 2751. The series marked with squares represents the control biofilms not exposed to Bdellovibrio. The series marked with the diamonds represents the OD550 of the biofilms that were exposed to B. bacteriovorus. Error bars represent the standard error of the mean. N = 3.
Several studies have been published describing the possible use of BALOs as biological control agents in environmental as well as medical microbiological settings (Westergaard and Kramer, 1977; Hobley et al., 2006). This study is the first to evaluate the potential use of these micro-organisms for combating A. actinomycetemcomitans in oral infections. Initially, we tested the hypothesis that the type strain B. bacteriovorus (HD100) could attack and kill A. actinomycetemcomitans ATCC 43718.
The interaction between A. actinomycetemcomitans and B. bacteriovorus results in a significant reduction of the pathogenic bacterium, but not in total eradication. This phenomenon is typical of the interaction between B. bacteriovorus and prey bacteria and is not caused by the emergence of stable resistant mutants. The exact nature of this process, however, is not yet fully understood (Shemesh and Jurkevitch, 2004).
In the mouth, pathogenic bacteria mostly reside in biofilms. These biofilm-associated bacteria do not “behave” as planktonic bacteria. For example, the resistance of bacteria to antimicrobial agents is dramatically increased in bio-films (Costerton et al., 1999). Almost without exception, organisms in a biofilm are 1000 to 1500 times more resistant to antibiotics than in their planktonic state. With the increasing interest in the development of improved methods for controlling biofilms, we tested the use of Bdellovibrio for the biological control of A. actinomycetemcomitans biofilms. Similar to the experiments with planktonic A. actinomycetemcomitans, Bdellovibrio was able to attack A. actinomycetemcomitans biofilms successfully. SEM imaging showed that Bdellovibrio was capable of markedly reducing the biofilm biomass, and this reduction occurred within the first 18 hrs of attack. The extent of the damage was visible as a bulk of destroyed biofilm cells, leaving behind what appears to be cell debris, at least at the surface of the remaining biofilm. Analysis of these data, taken together, suggests that BALOs can have a beneficial effect on levels of A. actinomycetemcomitans in the oral cavity. Since most BALOs are obligatory aerobic bacteria, their oxygen requirement might be a limiting factor for their use to prevent or treat infections with anaerobic bacteria such as in periodontitis. However, some BALO strains, such as the halo-philic Bdellovibrios, are able to complete their life cycle under micro-aerophilic conditions. They can survive periods of complete absence of oxygen (Schoeffield et al., 1996). In contrast, many periodontopathogens can be detected in large numbers throughout the oral cavity, and not just in the anaerobic periodontal pocket (Muller et al., 1993). As such, the application of BALOs could help in preventing colonization or translocation of pathogens to different niches in the oral cavity, or in preventing the recolonization of periodontal pockets after periodontal therapy (Quirynen et al., 2001). Since BALOs generally have a wide prey spectrum, it is likely that other Gram-negative periodontopathogens are also susceptible to BALOs. This is particularly interesting in the treatment of periodontitis, which is typically a mixed infection with many Gram-negative pathogens. Currently, we are investigating the effects of BALOs on periodontal pathogens and on population dynamics on in vitro sub-gingival biofilms.
One of the important aspects to be investigated is the effect of BALO treatment on beneficial bacterial populations that are not susceptible to BALOs.
Since BALOs cannot infect mammalian cells (Simpson, 1972), and since no infectious diseases or pathogenic effects have been linked to BALOs, they can be generally regarded as safe. BALOs have been isolated from numerous sites that are in close contact with humans. They have even been isolated from feces of cattle and humans (Edao, 2000). One single observation describes the presence of a Bdellovibrio genomic sequence in the oral cavity (Paster et al., 2002). Therefore, it is unlikely that a harmful effect can be expected from these organisms. It can be hypothesized that BALOs have the potential to be used as a living antibiotic in the treatment of oral infections caused by Gram-negative bacteria.
This study was supported by grants from the Catholic University of Leuven (OT 07/057) and the Research Fund Flanders (G077209N, 1510109N). W. Teughels was supported by the Research Fund of the Catholic University of Leuven (PDM 07/220) and the Research Fund Flanders. We gratefully acknowledge R. De Vos for producing the SEM images.