Isolation of Bdellovibrio sp. from soil samples in Mexico and their potential applications in control of pathogens

Abstract In this study, two strains of Bdellovibrio were isolated from soil samples using the culture‐dependent technique and two members of the family Enterobacteriaceae (Klebsiella sp. and Salmonella sp.) as prey. The Bdellovibrio strains were bacteriolytic, plaque‐forming, and highly motile gram‐negative bacteria. We identified and confirmed the Bdellovibrio strains using microscopy, PCR amplification, and sequencing of the 16S rRNA gene. They were observed to be different strains based on hit locus and prey range analyses. Here, the first report on Bdellovibrio strains isolated from soil in Mexico corroborates earlier report indicating that populations of Bdellovibrio found in soil are heterogeneous thereby the need to identify the various strains.


Introduction
Bdellovibrio and Like Organisms (BALOs) viz. genus Bdellovibrio, Bacteriovorax, Peredibacter, and Halobacteriovorax are a group of obligate predatory bacteria that prey upon gram-negative bacteria for nutrients and reproduction (Snyder et al. 2002;Koval et al. 2015).
Bdellovibrio bacteriovorus, a member of the BALOs represents one of the most studied predatory bacteria. It is a small (0.2-0.5 μm × 0.5-2.5 μm), uniflagellated motile gram-negative bacterium that attacks and hydrolyzes cellular constituents of other gram-negative bacteria, utilizing the derived nutrients for growth and reproduction. The life cycle of B. bacteriovorus involves attachment to a ORIGINAL RESEARCH Isolation of Bdellovibrio sp. from soil samples in Mexico and their potential applications in control of pathogens suitable gram-negative prey, loss of flagellum, followed by penetration into prey periplasmic space. At this stage of intraperiplasmic invasion, the prey cell is altered to form a round structure known as bdelloplast. Digestion of prey cellular constituents then takes place and prior to prey lysis, B. bacteriovorus undergoes septation to produce progeny that are released to carry out further predation (Stolp and Starr 1963;Chatterjee 2009).
Several publications have indicated B. bacteriovorus as a better alternative to treat infections caused by multidrugresistant bacteria (Dashiff et al. 2011;Damron and Barbier 2013;Kadouri et al. 2013). It has also been suggested as a biocontrol agent in aquaculture and animal husbandry as against the conventional antibiotics considering the increasing trend of antibiotic resistance among pathogenic bacteria. Kadouri and O'Toole (2005); Monnappa et al. (2014) reported Bdellovibrio to successfully degrade or inhibit biofilms produced by both gram-positive and -negative bacteria.
B. bacteriovorus is ubiquitous in nature and may be isolated from different sources including plant rhizospheres, freshwater, soil, and gastrointestinal tract of animals (Edao 2000;Jurkevitch et al. 2000;Schwudke et al. 2001;Hobley et al. 2012;Lebba et al. 2013). As mentioned earlier, B. bacteriovorus are known to be obligate intracellular predators of gram-negative bacteria hence it relies solely on the degradation of prey cellular macromolecules and utilization of derived smaller molecular materials for its growth and reproduction. However, Bdellovibrio strains that are capable of growing in the absence of prey as well as on nutrient-rich media mostly referred to as hostindependent Bdellovibrio strains have been isolated using current laboratory protocols (Seidler and Starr 1969;Ferguson et al. 2008). Recently, a B. bacteriovorus strain Tiberius that is capable of growing simultaneously both in the presence and absence of prey was isolated from River Tiber, Rome (Hobley et al. 2012). The molecular derivation of host-independent Bdellovibrio strains have been linked with a mutation in a genetic hit (host interaction) locus coding for proteins that play active role in the attachment and invasion of Bdellovibrio into its prey (Cotter and Thomashow 1992). In addition, about 89% of isolated host-independent Bdellovibrio strains have experienced mutation in this locus, 46% demonstrated mutation involving deletion of 42 bp on the Bd0108 which encodes protein involved in regulation of Type IV pili formation needed for Bdellovibrio attachment and invasion (Sockett 2009;Prehna et al. 2014).
The BALOs, in general, exhibit great phylogenetic diversity and their classification have been dynamic in recent years owing to the distinguishing characteristics that exist between the terrestrial (including the freshwater) and marine groups of BALOs. These two groups have been distributed into two different phylogenetic clades based on differences in features such as G+C content, salt tolerance and prey range (Baer et al. 2000). Consequently, the marine BALOs were separated from the Genus Bdellovibrio and renamed Bacteriovorax. In fact, the nomenclature of BALOs associated with marine habitat was recently changed from Bacteriovorax to Halobacteriovorax (Koval et al. 2015). In the soil habitat, particularly, Bdellovibrio has been described to represent a heterogenous community of predatory bacteria that utilize wide range of gram-negative bacteria as prey (Jurkevitch et al. 2000).
Here, the first report in Mexico on molecular characterization of two Bdellovibrio strains isolated from soil using culture-dependent technique is presented. The two Bdellovibrio strains were found at a neighborhood in the city of Reynosa (a border city to McAllen, Texas in the USA), Tamaulipas State and were characterized using 16S rRNA gene sequencing, hit locus PCR amplification and prey range analyses.

Sample collection
Soil samples were collected at two different sites on a plot of land (26.069678N′, −98.313108W′ and 26.069446N′,−98.312902W′) within the Center for Genomic Biotechnology, National Polytechnic Institute [IPN as in Spanish] located in the city of Reynosa, Mexico. The soil samples were collected with a clean hand trowel after removing about 20-25 mm top of soil and put into a sterile polythene bag. The soil samples were immediately transported to the laboratory for analysis and isolation of Bdellovibrio strains.

Determination of soil pH and electrical conductivity
Soil sample was suspended in 100 mL of deionized water and stirred for 5 min. The suspension was left on the bench overnight (OV) and stirred again afterward. It was further left for 15 min and the liquid portion was transferred into a clean beaker. The pH and electrical conductivity of the soil sample was then measured using HI 991300 TM pH/EC/TDS/Temperature meters (Hanna instruments, Woonsocket, Rhode Island, USA) (Chaudhari et al. 2014).

Host Preparation
The bacterial preys used for preliminary isolation of the Bdellovibrio strains were Klebsiella sp. and Salmonella sp. donated by Biotechnology Institute (Autonomous University of Nuevo Leon [UANL as in Spanish] located in San Nicolás, de los Garza, Nuevo León, Mexico) and Center for Genomic Biotechnology, respectively. The different preys were cultured in Luria Bertani (LB) broth for 24 h and 0.4 mL was mixed with 0.1 mL of filtrate for the double layer agar plating technique. For the liquid culture medium analysis, the bacterial preys were cultured in LB broth for 24 h, harvested by centrifugation, washed and resuspended in 25 mmol/L HEPES buffer (4-[2-hydroxyethyl]-1-piper azineethanesulfonic acid) containing 3 mmol/L CaCl 2. 2H 2 O and 2 mmol/L MgCl 2. 6H 2 O (pH 7.4).

Isolation and determination of lytic activity of Bdellovibrio strains
Soil samples were suspended in 100 mL of HEPES buffer and shaken for 1 h at 200 rpm. The suspension was centrifuged at 1,800g for 5 min. The resulting supernatant was filtered serially using a 0.8 μm and 0.45 μm syringe filter (MF-Millipore TM Membrane, Merck Millipore Ltd, Tullagreen, Carrigtwohill, Co. cork, Ireland). The filtrate was serially diluted and plated on dilute nutrient broth (DNB) agar (0.08% nutrient broth amended with 3 mmol/L MgCl 2 . 6H 2 O, 2 mmol/L CaCl 2 . 2H 2 O, 0.6% agar for top agar, and 1.9% for bottom agar [pH 7.2]) using double-layer agar plating technique (Stolp and Starr 1963;Jurkevitch 2012). The plates were incubated at 30°C and plaque development on the DNB agar was monitored for 7 days. Plaques which emerge on DNB agar between 48 to 72 h and progressively increase in size were taken to be potential Bdellovibrio plaques. The purification of plaques obtained was done by singleplaque isolation technique (Jurkevitch 2012).
For determination of lytic activity, a single pure plaque was cut into 5 mL suspension of bacterial prey cells already washed with HEPES buffer using pipette tip cut with sterile scalpel. The culture was incubated at 30°C and monitored for prey lysis. Microscopic examination was carried out alongside to determine the presence of highly motile Bdellovibrio strains. The clear HEPES buffer culture obtained from the prey lysis was centrifuged three times at 2,800g for 15 min (Schwudke et al. 2001). The resultant supernatant (the lysate) was further used to inoculate large volume of appropriate bacterial prey already washed with HEPES buffer. The initial optical density of the host was read at 600 nm using optizen POP spectrophotometer (Mecasys Co., Ltd, Daejeon, Korea) and the progressive reduction in turbidity with time was monitored to determine the lytic activity of the Bdellovibrio strains. The pure Bdellovibrio strains obtained through separation from prey by successive filtration as described earlier and concentrated at 27,000g as well as lysates were stored in sterile glycerol at 80°C for further study.

Microscopic identification of isolated Bdellovibrio spp using atomic force microscope
The lysate was observed using Olympus U-TVO.35XC-2 light microscope (T2 Tokyo, Japan) for characteristic high motility of Bdellovibrio strains. The atomic force microscopy study (AFM) was done at the Department of Cellular Biology and Genetics (UANL). Briefly, 10 μL of lysate from prey and Bdellovibrio strains cocultured in HEPES buffer was deposited on cleaved mica and allowed to airdry. The bacteria samples were observed using an NT-MDT NTEGRA Prima AFM at room temperature, with a RTESPA probe (Bruker corporation, Beijing China) of spring constant k = 40 N/m in intermittent contact mode. Images of height, deflection, and phase were obtained; 20 × 20, 10 × 10, and 5 × 5 μm 2 image sizes were captured systematically for each sample at three different regions at least. They were analyzed with WSXM software to observe the morphological aspect of the bacteria (Nunez et al. 2003;Horcas et al. 2007).

Amplification of 16S rRNA gene and hit locus by PCR
For the detection of Bdellovibrio strains, a clear lysate was centrifuged three times at 2,800g for 15 min to remove residual prey cells (Schwudke et al. 2001). The final resultant supernatant was centrifuged at 27,000g for 20 min. DNA extraction was done using Wizard ® Genomic DNA purification kit (Madison, Wisconsin, USA) according to manufacturer's instructions. The 16S rRNA gene was amplified using the Ref. Therefore for purification and amplification of considerable length of the 16S rRNA gene, large volume of lysates from prey-predator coculture was centrifuged twice at 1,800g for 15 min at 4°C. The supernatant was serially filtered through 0.80 μm, 0.65 μm, and 0.45 μm syringe filter. At each stage of filtration, microscopic examination to observe the presence of fast moving Bdellovibrio as well as cultivation of the filtrate on LB agar using spread plate technique to ascertain total elimination of prey bacterial cell was done. In addition, double-layer agar plating technique was carried out at each stage of filtration to ascertain the presence of plaque-forming Bdellovibrio strains. Finally, the filtrate obtained was centrifuged at 27,000g for 20 min. The genomic DNA was extracted as previously described above and the 16S rRNA gene was amplified with the Bdello.-fwd. primer; 5′ AGAGT TTGATTCTGGCTCAGA-3′ and Bdello.-rev. primer; 5′-AGGTGATCCAGCCGCAGGTTC-3′ which amplified a 1493 bp fragment of the 16S rRNA gene in an in silico PCR using an online tool available on the website of University of the Basque Country (http://insilico.ehu.es/ PCR/Amplify) with the genomes of different species and strains of Bdellovibrio listed in the online software serving as template. And for the amplification hit locus, the following hit-fwd. primer; TCTAGACAGATGGGATTACTG and hit-rev. primer; GAATTCTGGCATCAACAGC which amplified a 959 bp were used.

Cloning, DNA sequencing, and analysis
The amplified and purified 16S rRNA fragment was ligated into PGEM ® -T vector (Promega ®, Madison, Wisconsin, USA) and transformed into Escherichia coli DH5α competent cells. The cloning procedure was performed according to manufacturer's instructions. Plasmids from positive clones were sequenced by Eurofins MWG Operon© LLC company (www.operon.com; Huntsville, Alabama). The obtained sequences were analyzed using Lasergene program Seqman ® software (DNAstar Inc., Madison, Wisconsin, USA) and subjected to similarity searches against 16S rRNA gene sequences retrieved from ribosomal database project (RDP) and NCBI databases. A neighbor joining (NJ) analysis was carried out using the pairwise distances (Kimura two-parameter model) metric to recover their clustering pattern. Bootstrap values were calculated to test the robustness of interior node support and were obtained by conducting 1000 pseudoreplicates using MEGA© 6.0 software (Tamura et al. 2013).

Prey range analysis
The prey range analysis was carried out using double layer agar plating technique as described above with 36 bacterial isolates including 21 referenced bacterial isolates obtained from the National Collection of Microbial Strains and Cell Cultures at the, Research Center for Advanced Studies (CINVESTAV as in Spanish) of IPN located in México City, México and 15 laboratory bacterial strains obtained from Biotechnology Institute of UANL and Center for Genomic Biotechnology of IPN. The reference bacterial strains were cultured as recommended by the culture collection center while the laboratory strains were cultured in LB medium for 24 h. The bacterial preys used in isolation of the different Bdellovibrio strains (Klebsiella sp. and Salmonella sp.) served as the positive control while bacterial prey without Bdellovibrio and filtrates containing Bdellovibrio without prey served as negative control. The experiment was carried out in duplicates and plaque formation was monitored for at least 7 days.

Statistical analysis
Each experiment and control for determination of lytic activity was carried out in triplicates with Klebsiella sp. and Salmonella sp. preys suspended in HEPES buffer without Bdellovibrio strains serving as control. The transformation of mean values of optical density was done using square root for variance normalization. The transformed mean values of optical density (on y-axis) were plotted against time (on x-axis). Statistical analysis was performed using Excel© for Windows©, 2013 (Microsoft, Redmond, Washington, USA). Student's ttest was used to compare the mean values of the two groups and P < 0.05 was used as indicator of significant difference.

Results and Discussion
Isolation and determination of lytic activity of Bdellovibrio strains Two different strains of Bdellovibrio designated SKB1291214 and SSB218315 were isolated from soil under a banana (Musa paradisiaca L) plant at a neighborhood from the city of Reynosa, Mexico (IPN), using DNB agar with plaque development observed within 2-7 days on Klebsiella sp. and Salmonella sp. preys, respectively. The plaques formed by isolated Bdellovibrio strains were irregular and expanded for many days as previously described by Stolp and Starr (1963) and Jurkevitch (2012). The soil samples were moist dark-brown loamy soil with pH of 7.3 and electrical conductivity of 0.20 mScm −1 suggesting their good agricultural value. The light microscopy examination revealed the isolated Bdellovibrio strains as highly motile, rod (comma) shaped gramnegative bacteria. Further examination of SSB218315 under atomic force microscopy revealed clearly two kinds of bacteria, the small comma-shaped Bdellovibrio (predator) and the bigger long rod-shaped Salmonella (prey). One of the images also showed the attachment of Bdellovibrio to the prey and another one showed the clustering of the Bdellovibrio strains in a fashion that perhaps depict a prey cell that was just lysed (Fig. 1,  panels A, B, C, and D).
Bdellovibrio spp. have been reported to be frequently encountered in soil representing about 80% of all BALOs in soil community (Fulthorpe et al. 2008). However, the ecological role of Bdellovibrio spp. in the different niche where they are encountered is not well understood, perhaps serving as an "ecological balancer" as described by Lebba et al. (2014). This study further supports the use of DNB agar in double agar plating technique as an effective method of isolating Bdellovibrio strains from the soil. In addition, the prey choice can also be one of the determining factors for the successful isolation of Bdellovibrio strains; preferably, gram-negative bacteria that differ from Bdellovibrio strains most especially in size and motility could be considered for successfully isolation of Bdellovibrio strains. This will allow easy differentiation of Bdellovibrio strains from their prey under the microscope.
Bdellovibrio strain SKB1291214 was able to attack and lyse the Klebsiella sp. in liquid medium. This was evident with the reduction in optical density [(1.07-0.26) before statistical square root transformation] within 72 h (Fig. 2). Similarly, Bdellovibrio strains SSB218315 lysed the Salmonella sp. in liquid medium reducing the optical density from 0.98 to 0.15 [values obtained before statistical square root transformation (Fig. 3)]. The statistical analysis of the mean optical density values of the experimental (0.90 ± 0.28 SD) when Klebsiella sp. was infected with Bdellovibrio strain SKB1291214 was significantly different in comparison with the mean optical density values of the control (1.22 ± 0.09 SD) as determined by student's t-test (t = 0.007, P < 0.05). In a similar way, there was significant difference    serving as positive control. The blast search result of the sequenced 16S rRNA gene from RDP showed that the two isolated strains belonged to class δ proteobacteria, family Bdellovibrionaceae, and genus Bdellovibrio. The NJ tree constructed using the 16S rRNA gene (Fig. 4)  The hit locus was successfully amplified in Bdellovibrio strain SSB218315 strains but not in Bdellovibrio SKB1291214 strains (Fig. 5). Sequencing and analysis of the amplified product from Bdellovibrio strain SSB218315 further confirmed it to be hit locus. The blast analysis showed the two study Bdellovibrio strains exhibiting 97% identity with two rhizosphere-derived Bdellovibrio bacteriovorus strains BEP2 and BRP4 reported by Jurkevitch et al. (2000). However, these two strains clustered more closely to Bdellovibrio strain SKB1291214 than SSB218315 on the NJ tree. And interestingly, hit locus was not successfully amplified in this two reported strains (Bdellovibrio bacteriovorus strains BEP2 and BRP4) when PCR technique was used. This is an indication that Bdellovibrio strain SKB1291214 may possibly be a plant rhizosphere-associated strain. Furthermore, since the hit locus has been proposed to be restricted to B. bacteriovorus (Schwudke et al. 2001), electron microscopy may provide more information on Bdellovibrio strain SKB1291214. Perhaps, it may be using different mechanism of action for its predatory activities similar to an epibiotic Bdellovibrio exovorus in which the hit locus has not been amplified as reported by Koval et al. (2013).
In addition, the distant relationship that has been reported to exist between the marine; family Bacteriovoraceae and terrestrial; family Bdellovibrionaceae (including the freshwater) groups of "Bdellovibrio and like organisms" (BALOs) as reported by Baer et al. (2000) and Jurkevitch Figure 4. Phylogenetic tree of partial 16S rRNA gene sequences constructed with neighbor-joining algorithm using the pairwise distances (Kimura two-parameter model) metric to recover their clustering pattern. The numbers at each node represents bootstrap values for 1000 replicates. Thermotoga maritima 16S rRNA gene sequence was used as an external group. et al. (2000) can also be inferred from the NJ tree. These two groups of BALOs were initially grouped together as Bdellovibrionaceae but later separated into two groups based on differences in the characteristics including variations in G+C content, prey preference, and response to salinity.
The 16S rRNA gene sequence data obtained in this study was processed with online software Decipher ® (Wright et al. 2012) to check chimeras and have been submitted to GenBank Databases under accession numbers KT852580.1 and KT807464.1 for Bdellovibrio strain SSB218315 and Bdellovibrio strain SKB1291214, respectively. Details of data submission can be found at GenBank: www.ncbi/nlm.nih.gov.

Prey range analysis
Bdellovibrio strain SKB1291214 was able to form plaque with 13 of 36 (36.11%) bacterial isolates considered for prey range analysis while Bdellovibrio strain SSB218315 was able to prey upon 22 (61.11%) bacterial isolates. None of the Bdellovibrio strains formed plaque on the six grampositive bacteria viz. genera Staphylococcus and Bacillus considered in this study (Table 1). Most of the prey considered belongs to the phylum γ-proteobacteria with two α-proteobacteria (Rhizobium leguminosarum and Agrobacterium tumefaciens) and only one β-proteobacteria (Alcaligenes sp). Bdellovibrio strain SSB218315 preyed upon all the bacteria belonging to family Enterobacteriaceae except Citrobacter freundii. However, Bdellovibrio strain SKB1291214 on the contrary was able to utilize this Citrobacter freundii as prey.
Furthermore, Alcaligens sp. belonging to the phylum β-proteobacteria was preyed upon only by Bdellovibrio strain SSB218315. Stenotrophomonas spp. (family Xanthomonadaceae) and Rhizobium leguminosarum were not preyed upon by the two isolated Bdellovibrio strains.
Bdellovibrio strains have wide prey range but limited to gram-negative bacteria. Here, Bdellovibrio strain SSB218315 was observed to utilize more bacterial prey compared to Bdellovibrio strain SKB1291214. This could be expected because preferential predation has been reported in BALOs (Li et al. 2011). However, the two strains exhibited some preference for the bacterial isolates belonging to the family Enterobacteriaceae and coupled with the fact that Bdellovibrio strains have been isolated from guts of mammals (Schwudke et al. 2001;Lebba et al. 2013), they could be used to stabilize intestinal bacterial flora perhaps as probiotics. Rhizobium leguminosarum lives symbiotically with root of leguminous plants helping in nitrogen fixation which in turn aid plant growth. The inability of the two strains of Bdellovibrio to prey upon Rhizobium leguminosarum suggests the possibility of using the two bacteria synergistically to help plant growth.
One of the setbacks that could limit the application of Bdellovibrio strains is their inability to attack or utilize gram-positive bacteria as prey as equally observed in this study with the inability of the study strains to form plaque on all the gram-positive bacteria considered for the prey range analysis. However, the ability of Bdellovibrio bacteriovorus HD100 to survive in the presence of Staphylococcus aureus using epibiotic mode of attack has been reported (Lebba et al. 2014).
In summary, the two Bdellovibrio strains isolated from soil exhibited the ability to prey upon different types of gram-negative bacteria and this attribute could be considered for future use in the control of pathogenic gramnegative bacteria. The differences observed between the two strains of Bdellovibrio isolated here with respect to amplification of hit locus and prey range further support Jurkevitch et al. (2000) report that population of Bdellovibrio found in the soil is made up of heterogenous groups. Therefore, this suggests the need for further characterization and classification of soil-associated Bdellovibrio for possibility of grouping them into different subgroups (strains). Finally, with the paucity of information on BALOs research in Mexico, this work is expected to pave way for basic line of research in BALOs with ultimate goal of utilizing them for biotechnological applications.