|Home | About | Journals | Submit | Contact Us | Français|
Halobacteriovorax (formerly Bacteriovorax) is a small predatory bacterium found in the marine environment and modulates bacterial pathogens in shellfish. Four strains of Halobacteriovorax originally isolated in Vibrio parahaemolyticus O3:K6 host cells were separated from their prey by an enrichment-filtration-dilution technique for specificity testing in other bacteria. This technique was essential, since 0.45-μm filtration alone was unable to remove infectious Vibrio minicells, as determined by scanning electron microscopy and cultural methods. Purified Halobacteriovorax strains were screened for predation against other V. parahaemolyticus strains and against Vibrio vulnificus, Vibrio alginolyticus, Escherichia coli O157:H7, and Salmonella enterica serovar Typhimurium DT104, all potential threats to seafood safety. They showed high host specificity and were predatory only against strains of V. parahaemolyticus. In addition, strains of Halobacteriovorax that were predatory for E. coli O157:H7 and S. Typhimurium DT104 were isolated from a tidal river at 5 ppt salinity. In a modified plaque assay agar, they killed their respective prey over a broad range of salinities (5 to 30 ppt). Plaques became smaller as the salinity levels rose, suggesting that the lower salinities were optimal for the predators' replication. These species also showed broader host specificity, infectious against each other's original hosts as well as against V. parahaemolyticus strains. In summary, this study characterized strains of Halobacteriovorax which may be considered for use in the development of broad-based biocontrol technologies to enhance the safety of commercially marketed shellfish and other foods.
Bdellovibrio and like organisms (BALO) are Gram-negative, aerobic bacteria which are predatory toward other Gram-negative bacteria and consist of freshwater/terrestrial and marine forms. Bdellovibrio and like organisms are divided into four genera, Bdellovibrio, Bacteriolyticum, Peredibacter, and Bacteriovorax, which was recently renamed Halobacteriovorax (1). These organisms enter into a susceptible host bacterium and reside within the periplasmic space where they utilize the cytoplasmic nutrients of the host to support growth and replication. The replicative form, known as a bdelloplast, elongates within the host and septates into progeny cells as the host is lysed, releasing cells capable of attacking more prey. Halobacteriovorax organisms are the marine forms, which are small, polar flagellated, highly motile, intracellular predators in the class Deltaproteobacteria family Halobacteriovoraceae (1).
Bdellovibrio and like organisms, particularly the nonhalophilic Bdellovibrio bacteriovorus, have been proposed for use in reducing the levels of phytopathogens associated with bacterial blights in soybeans and rice (2, 3) and reducing the levels of Salmonella enterica serovar Typhimurium on freshwater fish fillets (4). Other uses involve the treatment of various infections in animals, including bovine keratoconjunctivitis caused by Moraxella bovis (5), Aeromonas hydrophila in fish (6, 7), Vibrio cholerae and Proteus penneri in shrimp (8, 9), and Salmonella enterica serovar Enteritidis in the chicken gut (10). They have also been applied in the treatment of human disease pathogens, including Pseudomonas aeruginosa from the lungs of cystic fibrosis patients (11), Pseudomonas aeruginosa and Serratia marcescens associated with eye infections (12), and oral pathogens that cause periodontal diseases (13, 14). Bdellovibrio strains were also effective in the treatment of drug-resistant bacteria (15, 16) and in reducing the levels of Escherichia coli and Salmonella spp. from stainless steel surfaces (17).
Less is known about the use of the halophilic Halobacteriovorax strains for the amelioration of diseases in plants and animals or their use for the disinfection of food products or food contact surfaces. Halobacteriovorax has been used to reduce mortalities from vibrios in shrimp larvae (18) and from Aeromonas veronii in fish (19). Richards et al. (20) determined that the low levels of Halobacteriovorax naturally present in seawater were capable of significantly reducing Vibrio parahaemolyticus and Vibrio vulnificus levels in both seawater and oysters. Williams et al. (21) found that Halobacteriovorax has a dose-response relationship in reducing V. parahaemolyticus levels in seawater. In this study, we further characterized Halobacteriovorax strains which may prove useful as biocontrol agents to reduce levels of V. parahaemolyticus, E. coli, and Salmonella spp. in fish and shellfish. Predatory bacteria represent a potential green technology that may prove useful in controlling antibiotic-resistant bacteria in clinical, environmental, and industrial settings.
The size differences between Halobacteriovorax and V. parahaemolyticus are substantial: attack-phase Halobacteriovorax strains are approximately 0.2 to 0.4 μm in width (22) and approximately 0.75 to 1 μm in length (taken from figures of Richards et al. [20, 23]), while V. parahaemolyticus is larger (roughly 0.5 to 0.8 μm in width and 1.4 to 2.4 μm in length). Vibrios are, however, capable of entering a viable but nonculturable state where they shrink in size and become rounded to produce minicells, a state often induced by low temperatures or starvation (24,–27). The presence of minicells prevents the simple separation of host vibrios from the predator, both of which are filterable through a 0.45-μm-pore-size membrane. A method to separate the two is needed to obtain pure cultures of Halobacteriovorax for host specificity assays; otherwise, low levels of original host cells in the Halobacteriovorax culture may lead to contamination of the new host culture and inaccurate assessment of specificities toward alternative host cells.
The purpose of this study was to (i) describe an enrichment-filtration-dilution method to separate Halobacteriovorax strains from their previous host cells, (ii) evaluate the host specificity of four Halobacteriovorax strains against V. parahaemolyticus, V. vulnificus, and Vibrio alginolyticus, the vibrios responsible for the majority of shellfish-related bacterial illnesses and deaths in the United States, as well as E. coli O157:H7 and S. Typhimurium DT104, and (iii) identify the host specificity and salt tolerance of two new strains of Halobacteriovorax, isolated in E. coli O157:H7 and S. Typhimurium DT104.
Four previously described isolates of Halobacteriovorax (G3, OR7, OS1, and S11) were used (23). Strain G3 was isolated from seawater from the Gulf Coast of Alabama, and the others were from seawater from the Delaware Bay. The following strains were also used: V. parahaemolyticus O3:K6 (RIMD2210633), provided by E. Fidelma Boyd at the University of Delaware; V. parahaemolyticus (environmental strain DAL 1094 and clinical strain DIE12 052499), provided by Angelo DePaola, U.S. Food and Drug Administration, as previously described (28); V. vulnificus MLT403 and MLT1003, provided by Mark L. Tamplin while at the University of Florida; S. Typhimurium DT104 H3380, obtained from the Centers for Disease Control and Prevention; three strains of E. coli O157:H7 from the American Type Culture Collection (ATCC), Manassas, VA (strains ATCC 35150, ATCC 43894, and ATCC 43895); and E. coli ATCC 15144, which is a host strain for the BALO Bdellovibrio bacteriovorus described by Abram et al. (29). In addition, V. parahaemolyticus strains AQ 4037, TX 2103, and CPA 7081669 and V. alginolyticus strain 22 were all obtained from the USDA culture collection. All vibrios were enriched in Difco Luria-Bertani (LB) broth (Becton, Dickinson and Company, Sparks, MD) supplemented with 2% NaCl (3% NaCl total), whereas non-Vibrio bacteria were grown in LB broth which contained 1% NaCl. Seawater was obtained from Scotton Landing, on the St. Jones River in Frederica, DE, 6.2 km from the Delaware Bay (20, 23), and was used to screen for Halobacteriovorax strains against E. coli and S. Typhimurium. This location was affected by the tide and contained water that was a mixture of fresh river water and seawater. High-salinity seawater (~30 ppt) for use in the preparation of agar for plaque assays was obtained from the pier at the Cape May-Lewes ferry terminal near the mouth of the Delaware Bay and was autoclaved and then filtered using a 0.22-μm-pore-size filter to remove particulates before being used in making plaque assay agar. Seawater and seawater dilutions (made using distilled water) were used to ensure the salinity of the agar to the desired levels.
Preliminary studies showed that enrichments of all four strains of Halobacteriovorax against V. parahaemolyticus did not totally kill the V. parahaemolyticus host cells in 2- to 3-day-old cultures but that the Halobacteriovorax strains could grow to higher titers than the host vibrios in 2- to 3-day enrichments in seawater. Consequently, we developed an enrichment-filtration-dilution technique to remove host cells and obtain pure cultures of the Halobacteriovorax strains. Two- to 3-day enrichments were filtered through a 0.45-μm-pore-size Millex HV syringe filter (Millipore Corp., Billerica, MA) to remove most of the vibrios, allowing passage of the smaller Halobacteriovorax strains. Filtrates were serially diluted 10-fold in sterile seawater (30 ppt) to 10−7, and V. parahaemolyticus was quantitatively measured by a pour plate method, as previously described (20), using 100 μl of each dilution followed by incubation at 37°C for 24 h to detect the host's presence. Diluted filtrates were also tested for Halobacteriovorax by a double-layer plaque assay using an agar made from BBL polypeptone peptone medium (Pp20 medium; Becton, Dickinson and Company) as described by Richards et al. (20). The higher dilutions containing Halobacteriovorax were Vibrio free and were used in enrichments and plaque assays with other Vibrio and non-Vibrio bacteria to determine the host specificity of all four Halobacteriovorax strains that were originally isolated in V. parahaemolyticus. This method was also employed to separate Halobacteriovorax from E. coli and Salmonella host cells where the absence of host cells in the filtered diluent was confirmed by spread plating 100-μl portions onto LB agar followed by incubation at 37°C.
Cultures of Halobacteriovorax against E. coli and S. Typhimurium were subjected to real-time PCR. Halobacteriovorax-specific primers Bac676F and Bac1442R, as described by Davidov et al. (30), were used to amplify a portion of the 16S rRNA gene of the Bacteriovoraceae (Halobacteriovoraceae) family members. Plaque plugs were picked/stored in sterile seawater and diluted 1:10 in peptone yeast extract (PYE) buffer, and the cells were heated at 99°C for 5 min, which lysed the bacteria to release the DNA. One microliter of each heat-treated strain was subjected to 94°C for 120 s and 45 cycles of PCR at 94°C for 60 s, 56°C for 60 s, and 72°C for 60 s. PCR was performed in 25-μl reaction tubes containing SYBR green Taq ReadyMix (Sigma-Aldrich, St. Louis, MO) to give 10 mM Tris-HCl (pH 8.3), Taq DNA polymerase and SYBR green 1 (amounts proprietary), 50 mM KCl, 3 mM MgCl, 0.2 mM deoxynucleoside triphosphate (dNTPs), 0.2 μM each primer (Integrated DNA Technologies, Skokie, IL), and 1 μl of test sample in a Cepheid Smart Cycler (Cepheid, Sunnyvale, CA).
The sizes of V. parahaemolyticus O3:K6 cells in filtrates of enrichment cultures were determined by scanning electron microscopy (SEM) as follows. V. parahaemolyticus was incubated overnight in LB broth made with 30-ppt salinity seawater at 26°C and 250 rpm. The next day, 100 μl of the enrichment was added to 10 ml of autoclaved and filter-sterilized seawater for further incubation at 26°C and 250 rpm for 6 to 8 h. Five milliliters of the enrichment was filtered through a 0.45-μm-pore-size Millex-HV syringe filter or 3 ml through a 0.22-μm-pore-size Millex-GV syringe filter (Millipore Corp.), 10 μl was applied to coverslips for 30 min, and bacteria were fixed by the addition of 1 μl of 25% glutaraldehyde (Electron Microscopy Sciences, Hatfield, PA) for 1 h. Slides were rinsed with 0.1 M imidazole buffer (pH 7.0) for 30 min, and dehydration was performed using 50, 80, 90, and 100% ethanol gradients for 30 min each. The samples were dried using a Denton DCP-1 critical point dryer (Denton Vacuum, Inc., Cherry Hill, NJ) using liquid carbon dioxide (GTS-Welco Co., Allentown, PA) for approximately 20 min, as previously described (20). The samples were mounted on stubs and sputter gold coated for 30 s (EMS 150R ES; Electron Microscopy Sciences). Samples were observed using an FEI Quanta 200F scanning electron microscope (FEI, Hillsboro, OR) with an accelerating voltage of 10 kV in the high vacuum mode. The sizes of bacteria were measured with FEI XT Docu software. The viability of V. parahaemolyticus O3:K6 minicells after 0.45- and 0.22-μm filtration of 2- to 3-day enrichments was determined on Difco thiosulfate-citrate-bile salts-sucrose agar (TCBS) (Becton, Dickinson and Company) incubated at 37°C for up to 48 h. The newly isolated Halobacteriovorax strains against E. coli O157:H7 and S. Typhimurium DT104 were also imaged by SEM after growth at 26°C and 250 rpm for 48 h in 2 ml of sterile seawater (5 ppt salinity) to which 200 μl of an overnight culture of E. coli O157:H7 or S. Typhimurium DT104 host cells had been added. Samples were glutaraldehyde fixed and processed for SEM as described above.
Halobacteriovorax strains against V. parahaemolyticus, E. coli, and S. Typhimurium were separated from their hosts by the enrichment-filtration-dilution technique and inoculated into alternative hosts to determine their specificities. The specificities were determined by monitoring for plaques on double agar plaque assays in lawns of potential or known host cells. Plaque assays in Vibrio hosts were conducted according to previously described methods (20). To determine the ability of E. coli- and S. Typhimurium-grown Halobacteriovorax strains isolated from low-salinity (5-ppt) seawater to grow at higher salinities, the plaque assay agar was modified by the addition of one-quarter-strength (0.25×) LB broth, Bacto agar (Becton, Dickinson and Company) at 1.5% for the bottom layer of agar and 0.75% for the top overlay, and sterile, natural seawater at 5-, 10-, 20-, or 30-ppt salinity. The addition of the LB broth was required to give denser lawns of E. coli and S. Typhimurium. The optical density at 600 nm (OD600) of the bacterial cultures was also increased to support a denser bacterial lawn, from an OD600 of 0.2 for V. parahaemolyticus (20) to 0.8 to 0.9 for E. coli and S. Typhimurium.
The initial Halobacteriovorax strains used in this study were isolated in V. parahaemolyticus; therefore, it was necessary to remove the host Vibrio to obtain a pure culture of Halobacteriovorax for use as an inoculum in specificity testing in other potential bacterial hosts. Three methods of Halobacteriovorax purification were attempted: filtration through 0.45-μm- or 0.22-μm-pore-size filters, picking Halobacteriovorax from the center of plaques, and an enrichment-filtration-dilution approach. Filtration of enrichments through 0.45-μm-pore-size filters was thought to be a possible means for removing the vibrios from enrichment cultures, since vibrios are generally perceived to be much larger than Halobacteriovorax; however, some V. parahaemolyticus cells were present in the filtrates and appeared as minicells. Minicells are generally considered to be vibrios in their viable but nonculturable state. By SEM, we identified spherical minicells in the filtrates, some of which were <0.22 μm in diameter (Fig. 1). This demonstrates that 0.45-μm filtrates of V. parahaemolyticus cultures are not Vibrio free. A histogram of the size frequencies of apparent spherical minicells is shown in Fig. 1C and indicates the presence of a high proportion of cells capable of passing through 0.45-μm-pore-size membranes.
To test the culturability of V. parahaemolyticus in 0.45- or 0.22-μm filtrates of overnight enrichments, LB broth containing 3% NaCl (total) was inoculated with filtrates and incubated at 37°C and 250 rpm for 24 h. Only vibrios in the 0.45-μm filtrate grew and were confirmed to be V. parahaemolyticus on TCBS agar plates. Although Fig. 1 shows cells smaller than 0.22 μm in diameter, it is suspected that these cells are likely to be too small to contain the double-stranded V. parahaemolyticus chromosome and are thus nonreplicating particles. They are large enough to potentially transport plasmid DNA, but their role in infectivity or the transfer of pathogenicity remains unknown. Bacteria with nominal sizes larger than the pore sizes of filters have also been shown to readily pass through smaller pores (31), and thus filtration alone was insufficient for purifying Halobacteriovorax.
Since filtration was unsuccessful, attempts were made to collect pure Halobacteriovorax by picking agar from the center of large (~5-mm-diameter), clear plaques, where it was assumed that host cells would have succumbed from predation. Our results indicate that this was not the case. Some persistence of host cells remained within the plaques and led to the outgrowth of vibrios in enrichments of the picked materials. We hypothesize that V. parahaemolyticus persists within clear plaques as minicells and that Vibrio minicells are resistant to predation by Halobacteriovorax. Clearly, minicells are smaller than Halobacteriovorax; thus, there is no mechanism for Halobacteriovorax to enter into minicells to replicate. This suggests that minicells may protect Vibrio populations in the environment from total collapse as a survival mechanism during periods when Halobacteriovorax levels are high, thus preserving individual Vibrio species.
A third approach to isolate pure Halobacteriovorax for use in host specificity studies involved an enrichment-filtration-dilution method. This method was successful in separating predators from V. parahaemolyticus O3:K6, E. coli O157:H7, and S. Typhimurium DT104 hosts. The method was successful because the Halobacteriovorax titers were much higher than those of their host cells after enrichment for 48 h followed by 0.45-μm-pore-size filtration. For example, after filtration of enrichments of the four Halobacteriovorax strains which had been isolated in V. parahaemolyticus (G3, OR7, OS1, and S11), Vibrio counts were ≤60 CFU/ml, whereas the titers of Halobacteriovorax in the same filtered enrichments were ≥2.4 × 107 CFU/ml. Thus, the level of Halobacteriovorax was at least 400,000-fold higher than the levels of vibrios in the filtered enrichments. A 1:1,000 dilution of the filtered enrichments allowed host cells to be diluted to extinction, leaving pure cultures of Halobacteriovorax available to perform plaque assays in lawns of alternative host cells.
Table 1 shows the results of the host specificity testing of the four Halobacteriovorax strains (G3, OR7, OS1, and S11) that were originally isolated in V. parahaemolyticus against 13 Vibrio and non-Vibrio bacteria. The Halobacteriovorax strains showed high host species specificity in that they were capable of killing V. parahaemolyticus strains other than O3:K6 but were unable to kill V. vulnificus, V. alginolyticus, several strains of E. coli, or S. Typhimurium. Interestingly, all four Halobacteriovorax isolates exhibited the same specificity profiles, even though G3 and OR7 were phylogenetically defined as members of cluster IX while OS1 and S11 were from cluster XII (23).
In an effort to isolate and identify Halobacteriovorax strains against E. coli and S. Typhimurium and begin to determine their host specificities, low-salinity (5-ppt) water was collected from a tidal river and analyzed by plaque assays on lawns of E. coli O157:H7 or S. Typhimurium DT104. PCR analysis of the picked plaques confirmed that the predators were members of the Halobacteriovoraceae family (1), formerly known as the Bacteriovoraceae family. Scanning electron micrographs are shown for these new isolates grown in 5-ppt-salinity seawater containing E. coli (Fig. 2A) and S. Typhimurium (Fig. 2B) as hosts. Morphologies are typical of those for other Halobacteriovorax strains that are predatory toward V. parahaemolyticus (20, 23). It was determined that Halobacteriovorax against E. coli was also a predator of S. Typhimurium and vice versa. This broader host specificity was in direct contrast with the high species specificity of Halobacteriovorax against V. parahaemolyticus. Specificity testing of the Halobacteriovorax strains against E. coli and S. Typhimurium was performed in three V. parahaemolyticus strains (O3:K6, DIE12 052499, and DAL 1094) in plaque assays conducted at 15-ppt salinity. The results showed that both Halobacteriovorax strains were predatory against all three V. parahaemolyticus strains. The Halobacteriovorax strains isolated against E. coli and S. Typhimurium had a broader host range than the strains originally isolated in V. parahaemolyticus.
Plaque assays were performed with high and low salt concentrations in order to evaluate whether the newly isolated Halobacteriovorax strains from low-salinity (5-ppt) seawater were also capable of replication at higher salinities (10 to 30 ppt) and whether they displayed the same host specificities at elevated salinities. It became obvious that the plaque assay agar used for the Vibrio assays had to be modified, since neither E. coli nor S. Typhimurium grew well on Pp20 agar at higher salinities (20 and 30 ppt). In fact, S. Typhimurium failed to produce bacterial lawns on the plaque assay plates in Pp20 agar containing 30 ppt seawater. Therefore, the current plaque assay procedure for Halobacteriovorax in V. parahaemolyticus (20) was modified to incorporate one-quarter-strength LB broth (added as a powder before autoclaving) with normal-strength Pp20 agar to provide additional nutrients to support E. coli and Salmonella growth at the higher salinities. The results showed that the Halobacteriovorax that was predatory against E. coli was capable of replication at 5-, 10-, 20-, and 30-ppt salinity in E. coli host cells, even though it had been isolated from seawater at 5-ppt salinity. Plaques became smaller as the salinity levels rose, suggesting that the lower salinities were optimal for the predator's growth. Similar results were obtained for the Halobacteriovorax against S. Typhimurium. It readily invaded and replicated in S. Typhimurium at 5-, 10-, and 20-ppt salinity; however, at 30-ppt salinity, plaque formation was observed only after extended incubation (6 days), whereas plaques were visible after only 3 days at the other salinities. Plaques were also smaller on the higher-salinity media. We have shown here that Halobacteriovorax is predatory over a broad range of salinities.
This work suggests that Halobacteriovorax represents a diverse group of predatory bacteria which have various host specificities. Halobacteriovorax isolates against V. parahaemolyticus were highly species specific, predatory only toward V. parahaemolyticus from among the species and strains evaluated. In contrast, Halobacteriovorax isolates against E. coli and S. Typhimurium showed broader species specificity and were able to attack bacteria of other genera, including V. parahaemolyticus. They were also predatory over a broad range of salinities commonly present in the marine environment. We hypothesize, based on the simplicity of isolating Halobacteriovorax strains against E. coli and S. Typhimurium on the first attempt, that many strains of Halobacteriovorax are present in seawater and that they are likely to be predatory toward a broad spectrum of bacteria in the marine environment.
The identification and characterization of Halobacteriovorax against shellfish pathogens are the first steps toward the development of a practical postharvest processing intervention to reduce the levels of these pathogens in shellfish prior to marketing. We are currently evaluating the use of Halobacteriovorax to inactivate V. parahaemolyticus in a modified oyster depuration process with the ultimate goal of designing processing interventions capable of inactivating a host of bacterial pathogens, including V. parahaemolyticus, V. vulnificus, E. coli, and Salmonella spp., in shellfish. The prospect of isolating Halobacteriovorax against an unlimited number of Gram-negative pathogens may permit the development of broad-based biocontrol technologies to enhance the safety of commercially marketed shellfish and other foods.
Support for this study was provided by intramural funding from USDA ARS under CRIS project 8072-42000-065-00D (G.P.R.).
The use of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture (USDA).