• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of jbacterPermissionsJournals.ASM.orgJournalJB ArticleJournal InfoAuthorsReviewers
J Bacteriol. Jul 2007; 189(13): 4850–4859.
Published online Apr 6, 2007. doi:  10.1128/JB.01942-06
PMCID: PMC1913455

Predation by Bdellovibrio bacteriovorus HD100 Requires Type IV Pili[down-pointing small open triangle]


Early electron microscopy and more recent studies in our laboratory of Bdellovibrio bacteriovorus cells indicated the presence of narrow fibers at the nonflagellar pole of this unusual predatory bacterium. Analysis of the B. bacteriovorus HD100 genome showed a complete set of genes potentially encoding type IV pili and an incomplete gene set for Flp pili; therefore, the role of type IV pili in the predatory life cycle of B. bacteriovorus HD100 was investigated. Alignment of the predicted PilA protein with known type IV pilins showed the characteristic conserved N terminus common to type IVa pilins. The pilA gene, encoding the type IV pilus fiber protein, was insertionally inactivated in multiple Bdellovibrio replicate cultures, and the effect upon the expression of other pilus genes was monitored by reverse transcriptase PCR. Interruption of pilA in replicate isolates abolished Bdellovibrio predatory capability in liquid prey cultures and on immobilized yellow fluorescent protein-labeled prey, but the mutants could be cultured prey independently. Expression patterns of pil genes involved in the formation of type IV pili were profiled across the predatory life cycle from attack phase predatory Bdellovibrio throughout the intraperiplasmic bdelloplast stages to prey lysis and in prey-independent growth. Taken together, the data show that type IV pili play a critical role in Bdellovibrio predation.

Bdellovibrio bacteriovorus are small, highly motile, gram-negative deltaproteobacteria that prey upon other gram-negative bacteria such as Escherichia coli, Salmonella, and Proteus species (11, 31). This is accomplished through collision and attachment to prey, followed by invasion and establishment of the predator within the prey periplasm; the prey cell is modified by the predator to form an osmotically stable structure called the bdelloplast. Within the bdelloplast, the predator proceeds to degrade host macromolecules into their constituent monomers and transport them into the growing Bdellovibrio filament, where they are reassembled as required. When the prey cell is exhausted, the filament septates, and progeny Bdellovibrio grow flagella and lyse the prey ghost to escape.

Bdellovibrio organisms also have the ability to survive in the absence of prey species in a growth phase known as the prey- or host-independent (HI) state, and these cells are isolated from prey- or host-dependent (HD) cells (27). HI Bdellovibrio is extremely pleomorphic and can be facultatively predatory or obligately HI; no one mutation has been found to cause this exact phenotype (2), but the phenomenon is well known in Bdellovibrio biology (27). The HI state is a useful tool for culturing and subsequent analysis of mutations that have adverse effects on predation, which would be lethal in the HD state (7). In this study and in previous mutational work (17), we independently derived several lines of HI strains from an isogenic merodiploid HD population containing the target inactivated gene. This allows us to generate a single gene disruption in the chromosome and to study the phenotypic effects that the gene disruption has upon Bdellovibrio predation, without any concern about background variations in HI cell behavior.

The genome sequence of B. bacteriovorus HD100 (22) shows genes encoding a full set of type IV pilus genes dispersed around the chromosome and an incomplete set of genes encoding Flp pili, which are a specific subset of type IVb pili found in diverse bacterial and archaeal species (14). We along with others proposed that Bdellovibrio may use pili as a mechanism of entering the prey cell, possibly via attachment to cell wall through a previously generated pore in the prey outer membrane (22). It can be seen on electron micrographs of invading Bdellovibrio (4) (Fig. 1A and B) that the pore formed in the outer membrane is small and a “tight fit” for the invading Bdellovibrio. Thus, a significant force may be required for prey entry by predator. Attachment of Bdellovibrio to prey cells could not be disrupted by either vortexing or brief sonication (4), indicating a strong interaction between predator and prey. Type IV pili in other bacterial species such as Myxococcus, Neisseria, and Pseudomonas are well characterized and have been shown to be involved in many functions, including host cell adherence and invasion, twitching motility, and fruiting body formation (reviewed in references 5 and 20). Type IV pili have also been demonstrated to have considerable retractile forces of greater than 100 pN (18), which would provide the significant force required to facilitate prey cell entry by Bdellovibrio.

FIG. 1.
Transmission electron micrographs of Bdellovibrio invasion and anterior fibers. (A and B) Wild-type HD100 Bdellovibrio invading E. coli S17-1 prey; prey invasion is accomplished through polar entry of the Bdellovibrio into the prey periplasm. (C) Anterior ...

In this work, we discuss the rationale for and inactivation of the type IVa pilus fiber protein, PilA, in B. bacteriovorus HD100. The resulting mutant strains were incapable of prey entry and therefore predation; as such, they had to be grown HI. Reverse transcriptase-PCR (RT-PCR) analysis of pilA and other type IV pilus-associated genes in the HD100 genome showed logical differential expression at different time points over the Bdellovibrio predatory life cycle and, for some of the type IV genes, in the pilA mutant strain. We therefore conclude that expression of the Bdellovibrio pilA gene, and by inference, type IV pili, is essential to the predatory life style of B. bacteriovorus HD100.


Bdellovibrio culturing on prey and prey independently.

Bacterial strains used in this study are listed in Table Table1.1. Predatory (HD) cultures of B. bacteriovorus HD100 were grown as described previously (16) on E. coli S17-1, containing the pZMR100 kanamycin resistance plasmid as appropriate, in Ca2+-HEPES buffer at 29°C, with shaking at 200 rpm. Prey-independent growth and isolation of HI strains were carried out as described previously (17). Briefly, HI strains were derived several times (see Fig. S10 in the supplemental material) from pilA/pilA::Km prey-dependent merodiploid HD100 strains by filtration through a 0.45-μm-pore-size filter to separate out prey cells; and cells were then plated on rich peptone-yeast extract (PY) medium containing kanamycin to retain selection for the interrupted pilA gene. Subsequent HI growth was axenic, without prey, in PY broth or on PY agar plates as described previously (28). E. coli strains were grown as standard in yeast extract-tryptone broth with appropriate antibiotics at 37°C with shaking at 200 rpm.

Strains used in this study

Insertional inactivation of the pilA gene.

The wild-type pilA gene (accession no. CAE79186, EMBL BX842649.1, and gi 39574009) was amplified from HD100 genomic DNA by PCR using KOD high-fidelity DNA polymerase in buffer 2 for genomic DNA (Novagen). The full-length gene was cloned with 1 kb flanking DNA on either side using the pilA cloning primers (Table (Table2).2). The resulting 2.3-kb fragment was gel purified and digested with BamHI and XbaI using unique sites within the genomic sequence and then ligated into pUC19 digested with the same enzymes (24). A 1.3-kb kanamycin resistance cassette, released from pUC4K (16) by digestion with HincII, was gel purified and blunt ligated into a blunted XcmI unique site 204 bp into the coding sequence of pilA (37). This construct was cloned into conjugative plasmid pSET151 (3) and used to insertionally inactivate (16) the cognate pilA gene in the Bdellovibrio HD100 genome. For the conjugation, 40 ml of donor E. coli S17-1 containing the pSET151 construct pKJE102 was grown to an optical density at 600 nm (OD600) of 0.2 to 0.4, concentrated by centrifugation to 100 μl, and added to 10 ml of overnight attack phase B. bacteriovorus HD100 similarly concentrated onto a nylon filter on a PY plate. After overnight incubation at 29°C, the cells were resuspended from the filter into yeast extract-tryptone medium, and various dilutions were plated onto YPSC (yeast extract, peptone, sodium acetate, calcium chloride,) overlay plates to give single exconjugant plaques for purification. Resulting merodiploid exconjugants were subject to screening for prey-dependent (HD) growth by plaquing on kanamycin-resistant prey E. coli S17-1:pZMR100 (2) and subsequent continual subculturing in standard Ca2+-HEPES and E. coli S17-1:pZMR100 prey liquid lysate cultures in the presence of 50 μg ml−1 kanamycin sulfate (16) The HI growth mode of Bdellovibrio was also employed to allow growth of mutant strains that might be nonpredatory (see details in previous section). HI growth was used in case the interruption of the pilA gene gave a mutation that interfered with normal HD predatory growth. Several lines of cultures were passed through these treatment regimens for both HD and HI growth to guard against any spontaneous mutational events giving rise to alterations in phenotypes (see Fig. S10 in the supplemental material).

PCR and RT-PCR primers used in this study

Transmission electron microscopy.

A total of 15 μl of Bdellovibrio and/or E. coli cells was applied to carbon-formvar-coated electron microscopy grids (Agar Scientific) and allowed to settle for up to 5 min. Excess cells were removed by blotting with Whatman paper, and the grids were stained with 1% uranyl acetate, pH 4, for between 30 s and 1 min. The stain was removed, and the grids were allowed to dry before observation at 100 kV using a JEOL JEM 1010 transmission electron microscope.

RNA isolation and RT-PCR.

For RT-PCR analysis, synchronous predatory cultures were set up as described previously (17) with samples collected at 15, 30, 45, 60, 120, 180, and 240 min postinfection. HI cultures were grown in PY broth, and their OD600 values were matched (to 0.6) as described previously (17). RNA was isolated with modifications published elsewhere (17) on the Promega SV total RNA isolation kit, and RT-PCR was performed using a QIAGEN One-Step RT-PCR kit as described previously (17) with the following conditions: one cycle of 50°C for 30 min and 95°C for 15 min and then 25 cycles of 94°C for 1 min, 48°C for 1 min, and 72°C for 2 min, followed by one cycle of 72°C for 10 min and then a hold at 4°C. Twenty-five cycles of amplification were used so that the PCR did not go to extinction, allowing a semiquantitative view of mRNA levels at a particular time point. Control reactions using the same primers as for RT-PCR to test against the presence of contaminating Bdellovibrio DNA in the RNA samples were carried out using PCR with Taq DNA polymerase (ABgene) under the same PCR conditions. Inclusion of E. coli S17-1 RNA controlled for any cross-reactivity against the Bdellovibrio pil primers (none was seen). RT-PCR was carried out for various pil genes, including pilA, using the primers shown in Table Table22.

Quantitative real-time RT-PCR was carried out on Stratagene MX3005P or MX4000 machines, using the Statagene Full Velocity SYBR Green QRT-PCR kit in one-step reactions. Extensive optimization of primer and template concentrations was carried out to achieve specific amplification of the target gene, and this was confirmed by dissociation curve, agarose gel electrophoresis, and sequencing of the extracted PCR product. For each sample, serial dilutions of template were used to confirm the efficiency of the PCR, and absolute quantification of initial transcript amounts was by comparison to a standard curve using the pure transcript PCR product as a template. Control reactions with no template, with no reverse transcriptase, and with E. coli S17-1 RNA as a template were carried out and gave no significant amplification. Two independent experiments were carried out with a minimum of six replicate experiments on each sample.

Fluorescent assay for predatory capability.

To assess the predatory capability of the pilA::Km mutant, constitutively yellow fluorescent protein (YFP)-expressing Kmr E. coli S17-1:pSB3000 pZMR100 prey grown overnight to an OD600 of 1.5 was challenged, on a solid PY agar surface, with pilAHI::Km mutant and pilAHI/pilA::Km merodiploid strains and also with an HID2 wild-type Bdellovibrio HI strain derived from strain HD100 (experiments were tried on more dilute LB agar but this could not support Bdellovibrio viability during the experiment). Repeat experiments were carried out using independently derived HI pilA::Km mutant strains, paired with their merodiploid parental strains and chosen for their diverse morphologies (see Fig. S1 in the supplemental material). The idea of this approach was that, as the HI phenotype has not been precisely defined by investigators in the Bdellovibrio field (2) and appears to involve phenotypic variation in different HI derivatives (e.g., cell size and shapes vary for individual HI derivatives of any Bdellovibrio strain), we wished to test the effects of pilA interruption in morphologically diverse HI strains that we had derived during our mutant screening. If a common phenotype was seen for all, we were satisfied that this was due to the pilA mutation.

To test predatory properties, the pilAHI strains (merodiploids and mutants) were grown for 3 days in PY medium with 50 μg ml−1 kanamycin at 29°C to an OD600 of 0.8 ± 0.05, and YFP prey were grown on 50 μg ml−1 ampicillin and 50 μg ml−1 kanamycin. A total of 50 μl of prey and 50 μl of Bdellovibrio culture were mixed, spotted onto PY agar plates supplemented with 50 μg ml−1 kanamycin, and incubated at 29°C for 24 h. For the same test using the HID2 strain, cells were grown for 3 days in PY broth only, at 29°C with shaking at 200 rpm, to the same OD as for pilAHI strains, and mixed with YFP prey that had been grown on 50 μg ml−1 ampicillin alone to select for the YFP-expressing plasmid (HID2 has no Kmr gene, and so its growth would be inhibited by the presence of kanamycin). This cell mixture was then spotted onto PY-only agar plates and again incubated for 24 h. At this time, the cells were scraped from the surface of the plate, resuspended in 1 ml of Ca2+-HEPES buffer (17), pelleted by centrifugation for 1 min at 13,000 rpm in a benchtop centrifuge, and resuspended in a final volume of 100 μl. Five-microliter samples were agar mounted and examined under phase-contrast microscopy and YFP optics (excitation, 500 nm) on a Nikon Eclipse E600 epifluorescence microscope; images were taken using a Hamamatsu Orca ER camera and analyzed using IPLab, version 3.6. E. coli uninfected prey and infected bdelloplast numbers were counted (n > 2,500 E. coli cells per experiment).


In this study we aimed to test any predatory role for the type IVa pilus fiber, the pilA gene homologue, from the B. bacteriovorus HD100 genome and to assay the expression of type IV pilus genes found distributed around the chromosome. Type IVa pili are used for twitching motility in Pseudomonas and cell-cell contact leading to social motility of Myxococcus and adherence in Neisseria (reviewed in reference 20). In a previous study Schwudke and coworkers (26) noted varying expression over a predatory cycle of a putative type IVb flp-1 gene homologue in Bdellovibrio (22), which encodes the pilus fiber of Flp pili. In other bacteria these are a specific subset of type IVb pili (14) used for tight adherence to surfaces and are best characterized in Actinobacillus actinomycetemcomitans, the causative agent of juvenile periodontitis (14, 35; reviewed in reference 12).

However, we believe that Flp pili are unlikely to be functional in B. bacteriovorus HD100 as our BLAST analysis of the HD100 genome has shown that, while there is a full set of genes that would encode type IV pili (pil genes), only some of the genes present would encode Flp pili (see Fig. S6 and S7 in the supplemental material). Furthermore, the proteomic approaches employed by Schwudke and coworkers (26) failed to find any Flp-1 protein in cell envelope preparations of B. bacteriovorus HD100 but did find PilA, the type IVa pilus fiber protein.

Bioinformatic analyses revealed a candidate pilus fiber gene in HD100, pilA and its associated full complement of pil genes which, in other bacteria, are sufficient to assemble functional pili. PilA proteins are found in many gram-negative bacteria and show characteristic sequence homologies (8). The HD100 predicted PilA (Bd1290; CAE79186) sequence showed the greatest homology to Myxococcus xanthus PilA; a BLAST search of the HD100 PilA sequence through the NCBI database (25) brought up M. xanthus with an E value of 9e-09, with 28% identity over the whole length of the protein. Figure Figure22 shows an alignment of Bdellovibrio with other type IVa pilins, illustrating the conserved N-terminal and divergent C-terminal regions characteristic of this family of proteins. In M. xanthus, PilA has been shown to be the fiber-forming protein of retractile type IV pili, mediating the cell-cell interactions required for social motility (36). It therefore seemed possible that PilA and the other Pil proteins found in the genome may operate as a retractile apparatus in some aspect of the Bdellovibrio life cycle.

FIG. 2.
Alignment of type IVa pilin proteins (8) showing Bdellovibrio bacteriovorus pilA to have the highly conserved N-terminal sequence homology common among type IVa pilins and phenylalanine that forms the N-terminal residue in the mature protein (indicated ...

Electron microscopy shows pilus-like fibers on the nonflagellar pole.

Pili may, however, have many other adhesive roles in the lifestyle of a bacterium (reviewed in references 8 and 20). Thus, to ascertain whether the type IV pili were expressed in attack phase predatory Bdellovibrio, electron microscopic observations of both HI-grown Bdellovibrio and bacteria freshly liberated from host-dependent predatory growth were carried out (Fig. (Fig.11 and Fig. Fig.3D,3D, frame i). Atomic force microscopy studies of strain 109J by Nunez and colleagues (21) did not show any pilus-like fibers, but their work focused mainly on 109J predation in whole bacterial communities rather than detailed, high-magnification studies of individual cells.

FIG. 3.
(A and B) Predation tests of pilAHI::Km and pilAHI/pilA::Km on immobilized YFP-labeled E. coli S17-1 prey cells on plates of PY medium supplemented with 50 μg ml−1 kanamycin (17). The pilAHI::Km mutant was seen to be nonpredatory, while ...

Our electron microscopy studies confirmed the early work of Shilo (29) and Abram and Davis (1), in which small, polar fibers were shown on Bdellovibrio. Pilus-like fibers were seen in approximately 30% of host-dependent HD100 cells (Fig. (Fig.1C)1C) in uranyl acetate-stained preparations. Agreeing with the findings of Abram and Davis (1), broken fibers were often seen lying in close proximity to the cells, indicating that the shear forces associated with staining seem to affect the proportion of Bdellovibrio organisms seen with intact fibers. Attack phase B. bacteriovorus 109J was also examined under the same conditions and found to have pilus-like fibers at the nonflagellate pole at roughly the same frequency as attack phase HD100 (Fig. (Fig.1D),1D), showing that different strains of Bdellovibrio have the same fibers. The low percentage of wild-type Bdellovibrio cells found with electron microscopically visible pili could be attributed to two things. First, pili are retractile organelles, with PilA subunits being held under the cytoplasmic membrane and polymerized into pilus fibers in a posttranslationally regulated process (15). Thus, many of the cells will contain retracted pili that do not appear on the surface. Probably, pili are only extruded during prey interactions as they would be sheared, if permanently extruded, in such a fast-swimming bacterium as Bdellovibrio; HD100 organisms swim at speeds of up to 160 μm s−1 (17). Second, although pili are strong and withstand immense retractile forces when under tension, they can also break during experimental handling, such as washing and staining for microscopy. Touhami and coworkers (33) note in their paper that “sample washing with water could detach type IV pili from bacteria”; these investigators were also careful to avoid shear forces that would break or deform pili throughout their research. Any stimulus for pilus extrusion in Bdellovibrio is unknown, but we hypothesize that it would be futile to have pili permanently extruded on the surface of such a highly motile bacterium; so possibly contact with a prey cell outer membrane or cell wall could provide that stimulus.

Insertional inactivation of the HD100 pilA gene.

Supplemental Fig. S10 shows the scheme for inactivation of the pilA gene in B. bacteriovorus HD100. Multiple lines of both prey-dependent and -independent pilA/pilA::Km HD100 organisms were subcultured as HI and HD strains (see Fig. S10 in the supplemental material) and screened by PCR and Southern blotting for loss of the wild-type chromosomal copy of the gene and gain of the interrupted pilA::Km form (see Fig. S3 and S4 in the supplemental material). This resulted in nine pilA::Km mutants, all of which were derived from the HI culture stream (24 HI cultures in total) in contrast to the HD culture stream (also 24 cultures), which was always pilA/pilA::Km merodiploid. Continued screening of these prey-dependent (HD) derivatives showed no loss of the wild-type gene, indicating that it was not possible to obtain this mutation under conditions of prey dependency.

All pilAHI::Km strains were unable to grow in predatory liquid cultures, and when spotted onto soft agar overlays of E. coli S17-1 prey, no zones of clearing were seen, unlike those for pilA/pilA::Km merodiploid strains (see Fig. S2 in the supplemental material). Three independent pilAHI::Km isolates with diverse morphologies (see Fig. S1 in the supplemental material) were chosen and assayed for predatory capabilities on YFP-labeled immobilized prey S17-1:pZMR100 E. coli (Table (Table1),1), compared to a parental merodiploid pilA/pilA::Km HI strain for each mutant strain; this was a fair control as each isolate had been subject to the same culture regime. The Bdellovibrio pilAHI/pilA::Km merodiploid strains all gave typically 40 to 50 bdelloplasts per 1,000 E. coli cells. The Bdellovibrio pilAHI::Km mutant was not able to predate the immobilized E. coli host, and no bdelloplasts were seen (Fig. 3A and B). HID2, a wild-type HD100-derived HI strain was assayed as a further control in the same YFP prey assay and found to form similar numbers of bdelloplasts (40 to 50/1,000 E. coli cells) as the parental merodiploid strains (Fig. (Fig.3C;3C; see Fig. S8 in the supplemental material). These results indicated that the loss of pilA results in the inhibition of predatory capability.

Extensive electron microscopy analysis of the pilAHI::Km strains (n > 1,500 cells) revealed no piliated Bdellovibrio organisms in the pilAHI::Km mutant (Fig. 3 D, frames ii and iii) strain compared to the normal, 20 to 30% piliation of pilA+ HI cells seen by examination of the pilAHI/pilA::Km merodiploid strain (Fig. (Fig.3D,3D, frame i).

RT-PCR and qRT-PCR analysis of Bdellovibrio HD100 pilA and RT-PCR analysis of expression of other pil genes in the wild-type predatory cycle.

To further analyze any role for pili in predation, expression of the single Bdellovibrio pilA gene and other pil family genes during the predatory cycle of Bdellovibrio was studied with RT-PCR. RNA was isolated from different time points across the predatory life cycle as described in Materials and Methods. Using 25-cycle semiquantitative RT-PCR (17) (Fig. (Fig.4C),4C), the pilA gene shows constitutively high expression at all time points across the Bdellovibrio life cycle. Attack phase cells have high levels of pilA mRNA, which does not decrease greatly during bdelloplast formation in attack phase cells and after predatory invasion and maturation (Fig. (Fig.4C).4C). The pilA RT-PCR analysis was validated by real-time quantitative RT-PCR (qRT-PCR) carried out on the pilA gene for samples from the attack phase and 30 min postinfection, and this showed that transcript levels of pilA remained approximately constant throughout penetration of the prey cell and establishment of the bdelloplast (Fig. (Fig.4D).4D). Because all predatory cycle RNA preparations were made from cultures that began with the inoculation of identical numbers of predatory Bdellovibrio cells equivalent to those in the attack-phase-only sample, it was possible to compare time points across the life cycle. Any changes in expression observed would be the result of changes in the expression of the Bdellovibrio organisms that had entered prey cells and begun their developmental cycle as any excess attack phase cells would continue to express their genes at the same level as the attack-phase-only control sample. Consistently high transcript levels show abundance, and probably stability, of the pilA mRNA, which would facilitate rapid protein synthesis of a supramolecular fiber structure as necessary in the predatory life cycle. Standard methods of matched RNA/cDNA amounts are not appropriate in Bdellovibrio predation studies as the addition of the prey dilutes the proportion of predatory RNA in total RNA by up to 10-fold, resulting in an apparent, artifactual 10-fold reduction of expression in the Bdellovibrio genes being studied.

FIG. 4.
Comparison of the operon structure of pil genes in M. xanthus and Bdellovibrio. Some cognate genes are highlighted in different colors and shapes to facilitate comparison. (A) Organization of Myxococcus pil genes, taken from Touhami et al. (33). (B) ...

In the B. bacteriovorus HD100 genome, there are three annotated pilQ genes. BLAST analysis showed the best pilQ homologue to be Bd0867 (CAE78812). PilQ is known in other bacteria to form the dodecameric outer membrane pore through which PilA fibers are extruded (6). In terms of gene organization conserved between different bacterial species, Bd0867 seems the more likely candidate since it is in the same operon as other pil genes. Reciprocal BLAST (25) homology searches bring up this gene as the strongest candidate pilQ in the Bdellovibrio genome. As shown by 25-cycle RT-PCR, pilQ shows lower expression in attack phase with transcript levels gradually increasing, peaking at 3 h, and dropping slightly at 4 h. This can be interpreted as the beginning of septation of the growing Bdellovibrio filament within the bdelloplast and the formation of new cell poles with completed pilus basal structures (including PilQ); pilus fibers have been observed to be only polar rather than lateral (Fig. (Fig.1).1). This also means that pilQ expression data would be useful as a marker for late bdelloplast stages and filament septation. BLAST analysis found two good PilT homologues. Bd1510 (CAE79390), named PilT1, has good homologies to other characterized PilT proteins, and the gene organization is conserved with that of M. xanthus (34). PilT in other bacteria is the ATPase needed for retraction of the PilA pilus fiber (20). Bdellovibrio pilT1 shows a steady level of expression throughout the developmental cycle from attack phase through to late bdelloplast.

The other PilT homologue in the B. bacteriovorus genome, Bd3852 (CAE81209), is an even better match than PilT1 to the PilT proteins of other bacterial species. PilT2 is an orphan in the genome but is an extremely good candidate protein. In expression pattern, pilT2 is somewhat unusual, with transcript levels increasing slightly in the early bdelloplast stages, showing the highest level of expression between 45 min and 1 h, dropping slightly at 2 h, reaching its lowest level at 3 h, and then returning to high level at 4 h. As mentioned above, PilT provides the retractile force of the type IV pilus motor, being a hexameric ATPase of the AAA+ family (13) which is held at the base of the type IV pilus under and probably in association with the cytoplasmic membrane.

PilD, the prepilin peptidase that cleaves and methylates the immature PilA fiber protein so that it can be exported and polymerized, is associated with the cytoplasmic membrane of bacteria that utilize type IV pili (reviewed in references 5 and 20). Bdellovibrio pilD shows virtually no expression in attack phase cells, with transcripts starting to appear 45 min into bdelloplast formation (Fig. (Fig.4C).4C). A gradual increase in expression is seen over the time course, peaking at 3 h and decreasing again at 4 h. The expression pattern of pilD is reminiscent of that of pilQ, which is logical, since the peak of expression of both these genes coincides with the beginning of pole formation.

PilG forms part of an ABC-type transporter required for PilA export and functional pilus biogenesis in M. xanthus, with these genes having no known homologues in other type IV pilus-producing species, which indicates a possible restriction to the deltaproteobacteria (36). Bdellovibrio bacteriovorus HD100 has good homologues of all three genes, pilGHI (Bd1291, Bd0860, and Bd0861). Wu et al. conclude that PilGHI may be required for outer membrane localization or export of PilA as mutants do not shed PilA into the surrounding medium and do not produce functional pilus fibers. This could account for the lack of Bdellovibrio genes for minor pilin proteins, which may perform a similar role in other bacteria, such as the Pseudomonas aeruginosa pilE, pilV, pilW, pilX, and fimU genes which have no counterparts that can be found through homology searches of both the Bdellovibrio and Myxococcus genome sequences. The pilE, pilV, pilW, pilX, and fimU genes are required for pilus assembly in P. aeruginosa (reviewed in reference 20) just as the pilGHI genes do in M. xanthus and would be expected to do in Bdellovibrio (9). The pilG gene was chosen for transcriptional assay (Fig. (Fig.4C)4C) as it lies directly downstream of pilA. It must be remembered that the pilA::Km strains had to be grown as HI strains; thus, it was not possible to directly compare the abundance of transcripts between HD and HI strains. This is because the diverse cell lengths and morphologies seen in HI cultures do not allow their comparison to the numbers of uniform short attack phase Bdellovibrio cells. However, examination of transcripts by RT-PCR in the pilAHI::Km strain and in the wild-type control HI strain HID2, allowed a simple nonquantitative examination of the effects of pilA disruption on other pil gene expression.

RT-PCR analysis of pilA and other pil genes in the pilA+ HID2 strain and in the pilAHI::Km strain.

As expected the presence of pilA mRNA is abolished in the pilAHI::Km mutant strain; there was no polar effect on pilG transcription caused by insertion of the kanamycin resistance cartridge into pilA. It seems, therefore, that pilG also has its own promoter even though it seems to lie in an operon with the pilA gene. All other pil genes examined showed no difference in expression levels between the wild-type and pilA mutant RNA samples (which came from cultures matched by OD600 values) except that pilT2 expression was virtually abolished in the pilA knockout strain compared to the pilA+ wild-type HID2.


We conclude that pili are essential for predation by Bdellovibrio as the interruption of the pilus fiber pilA gene rendered independently derived HI strains nonpredatory. Due to the strong homology between the Bdellovibrio HD100 pilus genes and the type IV pilus genes of other bacteria, we suggest that ratcheting or twitching of the pili may be an active and important mechanism for prey entry, although proving the ratcheting hypothesis requires intensive, high-resolution, atomic force microscopic studies on live invading cells, which is beyond the scope of this paper. The extent to which type IV pili may be present on the surface of attack phase Bdellovibrio or extruded when a prey contact is made is difficult to determine by conventional microscopy as it is a feature of type IV pili that pilus subunits are held under the membrane for rapid appropriate assembly (5, 33).

HI-grown pilA+ Bdellovibrio organisms were also found to express pilus-associated genes; but this is not surprising as the pilA+ HI strains studied retain predatory capability. Possibly, the HI growth state mimics intraperiplasmic growth of Bdellovibrio within prey, and thus pili may have a secondary role, post prey-entry, in anchoring the Bdellovibrio in the periplasm, perhaps involving adherence to prey peptidoglycan.

The Bdellovibrio genome contains genes whose products could act as engines of pilus assembly and retraction. These include three annotated pilQ homologues of which one, CAE78812, shows the best homology to other bona fide pilQ genes, such as that in Myxococcus (24% identity at the protein level). In addition, two pilT homologues are seen, both of which have extensive homology (pilT1, CAE79390, shows 24% identity and pilT2, CAE81209, shows 51% identity at the protein level) to pilT of Myxococcus. PilTs are known to provide motive force for type IV pilus retraction (19), and PilQ forms a functional outer membrane pore through which the pilus is extruded or retracted (10). However, both pilT genes and the pilQ gene CAE78812, are again expressed across all HD time points of infection of prey, as determined by RT-PCR (Fig. (Fig.3).3). Thus, the potential to extrude a type IV pilus is retained by Bdellovibrio in all predatory growth stages, possibly indicating roles for the pilus in the bdelloplast after initial prey entry.

Whether type IV pili play important roles anchoring the Bdellovibrio in the developing bdelloplast or providing, by retraction, the immense forces presumably required to squeeze Bdellovibrio organisms through the remarkably small pore generated in the prey cell outer membrane (Fig. 1A and B), we have definitively shown that interruption of the type IV pilus fiber protein-encoding gene, pilA, in Bdellovibrio abrogates the ability of the bacterium to predate entirely. This indicates that type IV pili provide a mechanism that is essential for prey entry or that type IV pili are vital to the productive Bdellovibrio-prey attachments required for intraperiplasmic predatory growth.


We thank Karen Morehouse, Chi Aizawa, and Rob Till for helpful discussions and Mike Capeness, Marilyn Whitworth, and Kath Cheatle for technical assistance. We thank Mitchell Singer and Renate Lux for the gift of the anti-Myxococcus PilA antiserum.

K.J.E. was supported by a BBSRC quota Ph.D. studentship. This work was supported in part by Wellcome grant AL/067712 and by Human Frontier Science Programme Grant RGP57/2005 to R.E.S.


[down-pointing small open triangle]Published ahead of print on 6 April 2007.

Supplemental material for this article is available at http://jb.asm.org/.


1. Abram, D., and B. K. Davis. 1970. Structural properties and features of parasitic Bdellovibrio bacteriovorus. J. Bacteriol. 104:948-965. [PMC free article] [PubMed]
2. Barel, G., and E. Jurkevitch. 2001. Analysis of phenotypic diversity among prey-independent mutants of Bdellovibrio bacteriovorus 109J. Arch. Micorbiol. 176:211-216. [PubMed]
3. Bierman, M., R. Logan, K. O'Brien, E. T. Seno, R. N. Rao, and B. E. Schoner. 1992. Plasmid cloning vectors for the conjugal transfer of DNA from Escherichia coli to Streptomyces spp. Gene 116:43-49. [PubMed]
4. Burnham, J. C., T. Hashimoto, and S. F. Conti. 1968. Electron microscopic observation on the penetration of Bdellovibrio bacteriovorus into gram-negative host cells. J. Bacteriol. 96:1366-1381. [PMC free article] [PubMed]
5. Burrows, L. L. 2005. Weapons of mass retraction. Mol. Microbiol. 57:878-888. [PubMed]
6. Collins, R. F., L. Davidsen, J. P. Derrick, R. C. Ford, and T. Tonjum. 2001. Analysis of the PilQ secretin from Neisseria meningitides by transmission electron microscopy reveals a dodecameric quaternary structure. J. Bacteriol. 183:3825-3832. [PMC free article] [PubMed]
7. Cotter, T. W., and M. F. Thomashow. 1992. Identification of a Bdellovibrio bacteriovorus genetic locus, hit, associated with the host-independent phenotype. J. Bacteriol. 174:6018-6024. [PMC free article] [PubMed]
8. Craig, L., M. E. Pique, and J. A. Tainer. 2004. type IV pilus structure and bacterial pathogenicity. Nat. Rev. Microbiol. 2:363-378. [PubMed]
9. de Bentzmann, S., M. Aurouze, G. Ball, and A. Filloux. 2006. FppA, a novel Pseudomonas aeruginosa prepilin peptidase involved in the assembly of type IVb pili. J. Bacteriol. 188:4851-4860. [PMC free article] [PubMed]
10. Drake, S. L., and M. Koomey. 1995. The product of the pilQ gene is essential for the biogenesis of type IV pill in Neisseria gonorrhoeae. Mol. Microbiol. 18:975-986. [PubMed]
11. Fratamico, P. M., and R. C. Whiting. 1995. Ability of Bdellovibrio bacteriovorus 109J to lyse gram-negative food-borne pathogenic and spoilage bacteria. J. Food Prot. 58:160-164.
12. Henderson, B., S. P. Nair, J. M. Ward, and M. Wilson. 2003. Molecular pathogenicity of the oral opportunistic pathogen Actinobacillus actinomycetemcomitans. Annu. Rev. Microbiol. 57:29-55. [PubMed]
13. Herdendorf, T. J., D. R. McCaslin, and K. T. Forest. 2002. Aquifex aolicus PilT, homologue of a surface motility protein, is a thermostable, oligomeric NTPase. J. Bacteriol. 184:6465-6471. [PMC free article] [PubMed]
14. Kachlany, S. C., P. J. Planet, R. DeSalle, D. H. Fine, D. H. Figurski, and J. B. Kaplan. 2001. flp-1, the first representative of a new pilin gene subfamily, is required for nonspecific adherence of Actinobacillus actinomycetemcomitans. Mol. Microbiol. 40:542-554. [PubMed]
15. Kaiser, D. 2000. How do pili pull? Curr. Biol. 10:R777-R780. [PubMed]
16. Lambert, C., M. C. M. Smith, and R. E. Sockett. 2003. A novel assay to monitor predator-prey interactions for Bdellovibrio bacteriovorus 109J reveals a role for methyl-accepting chemotaxis proteins in predation. Environ. Microbiol. 5:127-132. [PubMed]
17. Lambert, C., K. J. Evans, R. Till, L. Hobley, M. J. Capeness, S. Rendulic, S. C. Schuster, S.-I. Aizawa, and R. E. Sockett. 2005. Characterisation of the flagellar filament and the role of motility in bacterial prey penetration by Bdellovibrio bacteriovorus. Mol. Microbiol. 60:274-286. [PMC free article] [PubMed]
18. Maier, B., L. Potter, S. Magdalene, H. Seifert, and M. P. Sheetz. 2002. Single pilus motor forces exceed 100 pN. Proc. Natl. Acad. Sci. USA 99:16012-16017. [PMC free article] [PubMed]
19. Maier, B., M. Koomey, and M. P. Sheetz. 2004. A force dependent switch reverses type IV pilus retraction. Proc. Natl. Acad. Sci. USA 101:10961-10966. [PMC free article] [PubMed]
20. Mattick, J. S. 2002. Type IV pili and twitching motility. Annu. Rev. Microbiol. 56:289-314. [PubMed]
21. Nunez, M. E., M. O. Martin, L. K. Duong, E. Ly, and E. M. Spain. 2003. Investigations into the life cycle of the bacterial predator Bdellovibrio bacteriovorus 109J at an interface by atomic force microscopy. Biophys. J. 84:3379-3388. [PMC free article] [PubMed]
22. Rendulic, S., P. Jagtap, A. Rosinus, M. Eppinger, C. Baar, L. Christa, H. Keller, C. Lambert, K. J. Evans, A. Goesmann, F. Meyer, R. E. Sockett, and S. C. Schuster. 2004. A predator unmasked: life cycle of Bdellovibrio bacteriovorus from a genomic perspective. Science 303:689-692. [PubMed]
23. Rogers, M., N. Ekaterinaki, E. Nimmo, and D. Sherratt. 1986. Analysis of Tn7 transposition. Mol. Gen. Genet. 205:550-556. [PubMed]
24. Sambrook, J., and D. W. Russell. 2001. Molecular cloning: a laboratory manual, 3rd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
25. Schäffer, A. A., L. Aravind, T. L. Madden, S. Shavirin, J. L. Spouge, Y. I. Wolf, E. V. Koonin, and S. F. Altschul. 2001. Improving the accuracy of PSI-BLAST protein database searches with composition-based statistics and other refinements. Nucleic Acids Res. 29:2994-3005. [PMC free article] [PubMed]
26. Schwudke, D., A. Berhardt, S. Beck, K. Madela, M. W. Linscheid, B. Appel, and E. Strauch. 2005. Transcriptional activity of the host-interaction locus and a putative pilin gene of Bdellovibrio bacteriovorus in the predatory life cycle. Curr. Microbiol. 51:310-316. [PubMed]
27. Seidler, R. J., and M. P. Starr. 1969. Isolation and characterization of prey-independent Bdellovibrios. J. Bacteriol. 100:769-785. [PMC free article] [PubMed]
28. Shilo, M., and B. Bruff. 1965. Lysis of gram-negative bacteria by prey-independent ectoparasitic Bdellovibrio bacteriovorus isolates. J. Gen. Microbiol. 40:317-328. [PubMed]
29. Shilo, M. 1969. Morphological and physical aspects of the interaction of Bdellovibrio with host bacteria. Curr. Topics Microbiol. Immun. 50:174-204. [PubMed]
30. Simon, R., U. Preifer, and A. Puhler. 1983. A broad host range mobilisation system for in vivo genetic engineering: transposon mutagenesis in gram-negative bacteria. Biotechnology 9:184-191.
31. Stolp, H., and M. P. Starr. 1963. Bdellovibrio bacteriovorus gen. et sp. n., a predatory, ectoparasitic, and bacteriolytic microorganism. Antonie Leeuwenhoek J. Microbiol. Serol. 29:217-248. [PubMed]
32. Tomich, M., D. H. Fine, and D. H. Figurski. 2006. The TadV protein of Actinobacillus actinomyectemcomitans is a novel aspartic acid prepilin peptidase required for maturation of the Flp1 pilin and TadE and TadF pseudopilins. J. Bacteriol. 188:6899-6914. [PMC free article] [PubMed]
33. Touhami, A., M. H. Jericho, J. M. Boyd, and T. J. Beveridge. 2006. Nanoscale characterization and determination of adhesion forces of Pseudomonas aeruginosa pili by using atomic force microscopy. J. Bacteriol. 188:370-377. [PMC free article] [PubMed]
34. Wall, D., and D. Kaiser. 1999. Type IV pili and cell motility. Mol. Microbiol. 32:1-10. [PubMed]
35. Wang, Y., and C. Chen. 2005. Mutation analysis of the flp operon in Actinobacillus actinomycetemcomitans. Gene 351:61-71. [PubMed]
36. Wu, S., and D. Kaiser. 1995. Genetic and functional evidence that type IV pili are required for social gliding motility in Myxococus xanthus. Mol. Microbiol. 18:547-558. [PubMed]
37. Yanisch-Perron, C., J. Viera, and J. Messing. 1985. Improved M13 phage cloning vectors and host strains: nucleotide sequences of the M13mp18 and pUC19 vectors. Gene 33:103-119. [PubMed]

Articles from Journal of Bacteriology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...