Functional Advantages of Porphyromonas gingivalis Vesicles

Porphyromonas gingivalis is a keystone pathogen of periodontitis. Outer membrane vesicles (OMVs) have been considered as both offense and defense components of this bacterium. Previous studies indicated that like their originating cells, P. gingivalis vesicles, are able to invade oral epithelial cells and gingival fibroblasts, in order to promote aggregation of some specific oral bacteria and to induce host immune responses. In the present study, we investigated the invasive efficiency of P. gingivalis OMVs and compared results with that of the originating cells. Results revealed that 70–90% of human primary oral epithelial cells, gingival fibroblasts, and human umbilical vein endothelial cells carried vesicles from P. gingivalis 33277 after being exposed to the vesicles for 1 h, while 20–50% of the host cells had internalized P. gingivalis cells. We also detected vesicle-associated DNA and RNA and a vesicle-mediated horizontal gene transfer in P. gingivalis strains, which represents a novel mechanism for gene transfer between P. gingivalis strains. Moreover, purified vesicles of P. gingivalis appear to have a negative impact on biofilm formation and the maintenance of Streptococcus gordonii. Our results suggest that vesicles are likely the best offence weapon of P. gingivalis for bacterial survival in the oral cavity and for induction of periodontitis.


Introduction
Bacterial vesicles are produced generally by Gram-negative bacteria through the blebbing and pinching-off of the bacterial outer membrane [1].Bacterial vesicles have been found in culture media, samples of dental plaque and even the open ocean [2][3][4][5].Using nanoparticle tracking analysis, Biller et al. recently demonstrated that vesicles in the culture medium of Porchlorococcus were as abundant as the number of cells in early growth phase, while they were 10 times as numerous as the number of cells found in the exponential and stationary phases [2].As they carry many of the virulence factors found in bacterial outer membranes and the periplasm, vesicles are known to be an efficient vehicle involved in the pathogenicity of Gram-negative bacteria.Studies have shown that vesicles are involved in adherence, biofilm formation, invasion, host cell damage, and modulation of host immune responses [6].They have distinct functional advantages over whole bacterial cells.For example, vesicles that are enriched in bacterial virulence factors and signal molecules are protected from dilution and proteolytic degradation and are able to travel to distant targets.Thus Pseudomonas aeruginosa packages 2-heptyl-3-hydroxy-4-quinolone into vesicles that transport this molecule within a population of cells to communicate and coordinate social activities of the bacteria [7].
P. gingivalis, in particular, is a gram-negative bacterium associated with chronic periodontitis.Vesiculation in P. gingivalis was first reported in the 1980s [8].Earlier studies demonstrated that the vesicles serve as a vehicle for toxins, proteolytic enzymes, and adhesins, based on observations that vesicles were able to: 1) degrade collagen, Azocoll, and N-alpha-benzoyl-DLarginine p-nitroanilide, and 2) promote bacterial adherence between two non-coaggregating bacterial species such as Capnocytophaga ochracea and Eubacterium saburreum [3].Previous studies revealed that P. gingivalis vesicles were also able to enhance the attachment and invasion of Tannerella forsythia to epithelial cells.They were also able to mediate coaggregation between Treponema denticola and Lachnoanaerobaculum saburreum and between myceliumtype Candida albicans and Staphylococcus aureus [9][10][11], processes which did not occur in the absence of P. gingivalis vesicles or in the presence of heat-treated vesicles.Recent studies using proteomic tools have validated that P. gingivalis vesicles may act as intermediaries that carry a wide variety of virulence factors provided by their parent cells.Most virulence factors displayed on the outer membrane of P. gingivalis were found in the vesicles, including FimA, Mfa1, HagA, and gingipains [12,13].More interestingly, some well-known virulence factors were enriched in vesicles when compared to levels found in P. gingivalis cells.We recently reported an approximately 3-5 fold increase in gingipain levels in vesicles compared to the levels observed in surface extracts of the originating cells [14].Among the functional advantages of P. gingivalis vesicles, was the finding that they were able to induce immune responses.Using a mouse model, Nakao et al. demonstrated that vesicles effectively elicited P. gingivalis-specific serum IgG and IgA as well as salivary IgA five weeks after an initial intranasal immunization, whereas whole cells did not [15].It was suggested that the increased antigenicity found in the vesicles might result from the more concentrated immune-dominant determinants on the vesicles compared to P. gingivalis' cell surfaces.
Although P. gingivalis vesicles have been intensively studied recently, the virulence features of the vesicles with respect to periodontitis, especially when compared to P. gingivalis cells, are not completely understood.In order to further define the functional advantages conferred by P. gingivalis vesicles, we determined the invasive efficiencies of both P. gingivalis cells and vesicles and role of P. gingivalis vesicles in biofilm formation.Moreover we now show that a previously unrecognized property of P. gingivalis vesicles, that of horizontal gene transfer.These findings may represent an opportunity to utilize characteristics of P. gingivalis vesicles to develop strategies to reduce bacterial virulence.

Materials and Methods
Bacterial strains and vesicle preparation P. gingivalis strains are listed in Table 1, They were grown from frozen stocks in TSB (trypticase soy broth) or on TSB blood agar plates supplemented with yeast extract (1mg/ml), hemin (5 μg /ml) and menadione (1μg/mL), and incubated at 37°C in an anaerobic chamber (85% N2, 10% H2, 5% CO2).Streptococcus strains were grown in Trypticase Peptone Broth (TPB) supplemented with 0.5% glucose at 37°C under aerobic conditions.P. gingivalis vesicles were prepared as previously described [16].Briefly, P. gingivalis was grown to the late exponential phase and growth media were collected by centrifugation at 10,000 × g for 15 min at 4°C and filtered through a 0.22-μm-pore-size filters (CellTreat) to remove residual bacteria.Vesicles were collected by ultracentrifugation at 126,000 x g for 2 h at 4°C and resuspended in phosphate-buffered saline (PBS) containing 10% glycerol.Since quantifying vesicles by their protein or lipid content in weight represents the most common way to normalize data [1], for these studies proteins were extracted from vesicles using a BugBuster Protein Extraction Reagent (Novagen) and protein concentrations were determined with a Bio-Rad Protein Assay Kit (Bio-Rad).

Treatment of host cells with P. gingivalis vesicles.
Human gingival fibroblasts (HGF) and human oral keratinocytes (HOK) were purchased from ScienCell Research Laboratories and human umbilical vein endothelial cells (HUVEC) from ATCC.These host cells were cultured in specific media, respectively according to the manufacturer's instructions.Prior to treatment, HGFs, HOKs, and HUVECs (2 × 10 4 ) were seeded and grown overnight in poly-L-lysine coated 35mm glass bottom dishes (MatTek Corporation, Ashland, MA), and then exposed to P. gingivalis 33277 (2 × 10 6 ) or its vesicles (100 ng) for 0 or 1 h.After removal of unbound vesicles, HGFs were cultured for another 3 or 24 h, at 37°C, 5% CO2, using fresh media.The cytotoxicity of vesicle treatment was evaluated with a Pierce LDH Cytotoxicity Assay Kit (Thermo Scientific).There was no cytotoxicity detected under our experimental conditions.

Confocal microscopy
HOKs and HGFs were fixed with 3.8% formaldehyde in a sodium phosphate buffer at room temperature for 10 min after treatment, permeabilized with 0.1% Triton X-100 for 10 min, and blocked with 5% bovine serum albumin in PBS for 1 h.HOKs and HGFs were then immunostained with pan-specific antibodies of P. gingivalis 33277 and followed by goat anti-rabbit IgG conjugated to Alex Fluor 546 (Invitrogen).Nuclei were stained with 4 0 , 6-diamino-2-phenylindole (DAPI, 1 μg/mL; Invitrogen).Confocal images were acquired using a Nikon A1R confocal microscope.

Vesicle-associated DNA and RNA purification
To eliminate any surface-associated DNA and free DNA in the suspension, vesicles (5 μg) purified from P. gingivalis 33277 cultures were resuspended in PBS and treated with or without 2U of DNase I (Invitrogen) in a 30 μl final reaction volume, and incubated at 37°C for 30 min according to the manufacturer's instructions.Any DNase I was then inactivated at 75°C for 15 min, and the protected DNA inside of vesicles was released by treatment of vesicles at 100°C for 10 min.The samples were then centrifuged at 16,000 × g for 5 min.The supernatants containing the DNA were collected for PCR analysis.For vesicle-associated RNA purification, vesicle samples (25 μg) were treated with 5 μl of GES lysis buffer (50 mM guanidinium thiocyanate, 1 mM EDTA, and 0.005% (w/v) sarkosyl, final concentration) in a 100 μL final reaction volume [2].RNA was purified from the reaction mixture using an Aurum Total RNA Mini Kit (Bio-Rad) following the manufacturer's protocol.

PCR and real-time PCR analyses
The extracted DNAs were amplified by PCR in a 50 μL final reaction volume containing 1 μL of the supernatant of heat treated vesicles, the primers listed in S1 Table, and 0.2 μL of Platinum Taq DNA polymerase (Invitrogen).The amplifications were performed in a thermal cycler (Bio-Rad) at 94°C for 45 s, 56°C for 30 s, and 72°C for 30 s for a total 30 cycles, followed by 10 min of elongation at 72°C.PCR products were separated by agarose gel electrophoresis, stained with ethidium bromide, and visualized by UV transillumination with an EpiChemi III Darkroom (UVP).To detect RNA in P. gingivalis vesicles, RNAs were reverse-transcribed using iScript reverse Transcription Supermix (Bio-Rad), and real-time RT-PCR was performed by using an SsoAdvanced SYBR Green Supermix on a C1000 Touch Thermal Cycler (Bio-Rad) according to the manufacturer's instructions.RNA levels in vesicle samples were determined relative to the non-reverse transcribed calibrator sample by using the comparative cycle threshold (ΔC T ) method.ΔC T was calculated by subtracting the average C T value of the test sample from the average C T value of the calibrator sample, and was then used to calculate the ratio between the two by assuming a 100% amplification efficiency.

Horizontal DNA transfer with vesicles
Gene transfer between P. gingivalis cells and vesicles was examined using vesicles isolated from a mutant strain generated from P. gingivalis 49417 by introducing a 2.1-kb ermF-ermAM cassette into the fimA gene.Since the erm resistant gene was detectable in these vesicles, we used these vesicles as donors in a DNA transfer assay.In brief, P. gingivalis 33277 was anaerobically cultured in TSB at 37°for 6 h to reach an optical density at 600 nm (OD 600 ) of 0.5.Cells (10 8 ) were pelleted, resuspended, and mixed with the vesicles (2.5 μg) in 1 ml of fresh TSB.The suspension was incubated for 24 h, and recipients were selected by plating the bacterial suspension on TSB blood plates supplemented with erythromycin (5 μg/mL).After culturing for 5-7 days, the colonies were counted and collected for future PCR analysis.To confirm that the erm resistant gene was indeed incorporated into the fimA gene, P. gingivalis chromosomal DNA was isolated from the recipients using an Easy-DNA Purification Kit (Invitrogen).PCR was performed using DNA isolated from the recipients as a template, while the primers corresponded to sequences of the fimA gene.

Biofilm formation and dispersal of Streptococcus gordonii
Attachment of S. gordonii DL1 to human saliva coated surfaces was quantified by the method of O'Toole and Kolter [17].Human whole saliva was collected from periodontally healthy donors.The saliva collection protocol was approved by the IRB of Meharry Medical College and all participants gave informed written consent.After centrifugation and filter sterilization (pore size, 0.22 μm) to remove cellular debris, saliva was incubated in a 96-well polystyrene plate (Corning Incorporated) for 1 h.The plate was then washed with phosphate-buffered saline (PBS, 100 mM NaH2PO4, 150 mM NaCl).S. gordonii DL1 was grown in TPB to mid log phase (OD600 = 1.0) and collected by centrifugation.The bacterial cells were resuspended in fresh TPB and added to each well (5×10 7 /well) of a 96-well plate that was pre-coated with human whole saliva with or without P. gingivalis vesicles, and then incubated at 37°C in an anaerobic chamber for 24 h.To determine the role of P. gingivalis vesicles in biofilm dispersal of S. gordonii, the vesicles were added to the wells including an S. gordonii biofilm in ¼ TPB (1 vol phosphate buffered saline TBP and 3 vol PBS) and incubated for another 24 h before biofilm quantitation measures.The bacterial biofilms were stained with 1% crystal violet for 15 min.After washing with PBS three times, the biofilms were destained with 200μL of 95% ethanol for 30 sec; 125μL of absolute ethanol from each well was transferred to a fresh well for reading at 570 nm in a Benchmark Plus Microplate Spectrophotometer (Bio-Rad).

Statistical analyses
A student's t-test was used to determine statistical significance of the differences in the invasive activities of P. gingivalis cells and vesiclesm and on the effect of the vesicles on S. gordonii biofilms.A P<0.05 was considered significant.Values are shown ± SD unless stated otherwise.

Comparison of the invasive efficiency between P. gingivalis vesicles and their originating cells
Our recent study showed that P. gingivalis OMVs were able to efficiently invade HOKs [14].
We further compared the invasive activities of P. gingivalis cells and vesicles.To expose HOKs, HGFs, and HUVECs to P. gingivalis 33277 cells or their vesicles, host cells (2 × 10 4 ) were cultured with P. gingivalis cells (2 × 10 6 ) or vesicles (100 ng, equivalent to vesicles released by 2 × 10 4 of P. gingivalis in late log-phase) [18] for 1 h.After removal of the unbound vesicles, cells were cultured for another 3 or 24 h with fresh media.The invasive efficiency of P. gingivalis and its derived vesicles were examined under a confocal microscope.After exposure to P. gingivalis cells or vesicles, respectively, both the bacterial cells and the vesicles were found in HOKs, HGFs, and HUVECs but with significantly different efficiencies (Fig 1A ).By counting 10 random areas (1.34 × 1.34 mm), we found that 82 ± 5.4, 69 ± 6.5, and 80 ± 13.4% of the HOKs, HGFs, and HUVECs engulfed vesicles 3 h post exposure, respectively, and up to 85 ± 7.1, 93 ± 2.9, and 87 ± 6.9% of the HOKs, HGFs, and HUVECs contained the vesicles after a 24 h exposure (Fig 1B , 1C, and 1D).Far more vesicles were internalized in individual HOKs or HGFs at 24 h than at 3 h, as shown by the greater intensity of red fluorescence.We observed a significantly lower invasive efficacy of P. gingivalis cells compared to vesicles.Two to three fold less the host cells carried intercellular bacteria 3 h post exposure compared to the number of the cells with intercellular vesicles.Moreover, none of the host cells were found with engulfed bacteria or vesicles when the bacteria or vesicles were immediately removed after exposure (0 h values).These data provide strong evidence that P. gingivalis vesicles possess a higher invasive activity than their parent cells.

Detection of DNAs and RNAs in P. gingivalis vesicles
Nucleic acids have been found as components in outer membrane vesicles of some bacteria [2,19].To test if P. gingivalis vesicles carry any DNA fragments, we first pre-treated purified vesicles with or without DNase, and any residual protected DNA was released from vesicles by heat treatment.The DNA was amplified using PCR with16 primer pairs specifically corresponding to P. gingivalis genes, and included some well-known virulence genes.As shown in Fig 2A, all DNA fragments tested were amplified, although at different efficiencies.DNA fragments of genes encoding the major subunit of long fimbriae (fimA) and superoxide dismutase (sod) were detected at relatively higher levels, suggesting there may be a preferential internal packaging of DNAs by P. gingivalis vesicles.To further examine if some of the DNA fragments were large enough to encode a gene, we amplified and detected a 1.2 kilobase of the fimA gene including its promoter and terminator regions (fimA-FL, Fig 2A), indicating that some DNA in vesicles is large enough to encode a virulence factor.When comparing levels of PCR products using DNA templates from vesicles with or without DNase pre-treatment, we found that level of DNA amplification with certain specific primer pairs was decreased after DNase treatment (Fig 2B).This suggests that cytoplasmic DNAs may not only be packaged into vesicles, but also bind to the surfaces of the vesicles.The latter would likely be degraded when it voyages through oral microbial biofilms and periodontal tissues.To determine if RNAs are also incorporated into P. gingivalis vesicles, RNAs were isolated from purified vesicles of P. gingivalis 33277 and were reverse-transcribed. qPCR was performed with the same set of primers used for DNA detection.RNA sequences from the open reading frames of 16 genes (S1 Table ) were all recovered using RT-PCR (Fig 3).Not surprisingly, there was a very high rate of detection of 16S rRNA, followed by a number of the mRNAs of wellknown virulence factors such as mfa1, sod, and fimA.

DNA transfer through vesicles in P. gingivalis
Vesicle-mediated horizontal gene transfer has been reported in some Gram-negative bacteria including Acinetobacter baumannii and E. coli [20,21].To test if DNA detected in P. gingivalis 49417 (type III fimA strain) vesicles may be transferred to P. gingivalis strain 33277 (type I fimA strain), we isolated vesicles from a fimA mutant derived from P. gingivalis 49417.This mutant carries a 2.1-kb ermF-ermAM cassette encoding an erythromycin resistant gene incorporated into the fimA gene [22].PCR analysis confirmed the presence of the ermF-ermAM cassette in P. gingivalis 49417 vesicles (data not shown), therefore, this antibiotic resistant gene was used as a marker to select any recipients.After incubating with purified vesicles from the fimA mutant for 24 h, P. gingivalis 33277 gained full resistance to erythromycin was detected with efficiency at 1.9 × 10 -7 .(Fig 4A ).A lower efficiency was found when bacteria were incubated with the vesicles pretreated with DNase.Moreover, no erythromycin resistant colony was found on the plates inoculated with equivalent amount of P. gingivalis cells or vesicles alone.It should be pointed out that the efficiency is only based on the chance of P. gingivalis 33277 taking up the erm gene from the vesicles.Since multiple genes were packaged into the vesicles, genes other than the erm gene may also be transferred into P. gingivalis 33277.In addition, a difference was found when comparing the efficiencies of gene transfer when vesicles with or without DNase treatment were used as donors, suggesting that DNA fragments exist within vesicles and on the surface and both of them may be transferred into P. gingivalis cells.
After gene transfer into recipients, a gene must be integrated into the chromosomal DNA in order to persist within the cells.It is presumed that the erm gene in the vesicles of the fimA mutant is flanked with fimA sequences at both ends, and that homologous DNA recombination occurs between vesicle donor DNA and the chromosomal DNA of the recipient.To determine if the erm gene incorporated into the fimA gene of the recipient through an allelic-exchange, we analyzed the insertion sites of the erm gene using PCR with genomic DNA isolated from ten P. gingivalis recipients as templates, and primers corresponding to the promoter and termination regions of the fimA gene as probes.As shown in Fig 4B, the erm gene found in all recipients was incorporated into the fimA gene, since size of PCR products is about 3.3 kb that includes full sizes of the fimA gene and the erm gene.

Effect of P. gingivalis vesicles in S. gordonii biofilms
Co-aggregation between P. gingivalis and S. gordonii is well-known example of heterotypic biofilm formation.Previous studies revealed involvement of multiple sets of adhesins in the interaction of these two bacteria.The first set of adhesins includes P. gingivalis long fimbriae (FimA) and GAPDH present on the surface of streptococci [23]; the second set involves P. gingivalis short fimbriae (Mfa1) and streptococcal SspA/B (antigen I/II) adhesins [24,25].As both FimA and Mfa1 are major proteins associated with P. gingivalis 33277 vesicles [14], we Vesicles carrying an erythromycin resistant gene were derived from a fimA mutant and used as donors.P. gingivalis 33277 was the recipients.(A) The P. gingivalis 33277 receiving the erm gene was selected on TSB blood plates supplemented with erythromycin.Each bar represents the number of colonies detected on the plates, after P. gingivalis 33277 was incubated with vesicles without DNase pre-treatment, vesicles with DNase pre-treatment, and P. gingivalis 33277 without vesicles respectively, or vesicles alone without P. gingivalis 33277.An asterisk indicates a significant difference between colonies detected in P. gingivalis 33277 incubated with the vesicles without DNase treatment vs. those found in other groups (n = 3, p <0.05; t test).(B) The erm gene was detected by PCR with the fimA primers as a probe, and PCR products were visualized on agarose gels.The PCR product of the fimA gene is shown at 1.2 kb, and the PCR product of the chimeric gene (fimA + erm) is at 3.3 kb.doi:10.1371/journal.pone.0123448.g004therefore examined the role of P. gingivalis vesicles in biofilm formation of S. gordonii.A biofilm formation assay was first performed by inoculating vesicles derived from P. gingivalis strains with S. gordonii DL1into the 96 well plates covered with human saliva for 48 hr.In the presence of vesicles from wild type 33277, there was a 60% decrease in the biofilm formation of DL1 compared to DL1 biofilm formation in the absence of vesicles.While, there was a 20-30% decrease in biofilm formation of DL1 formation in the presence of vesicles from wild type strain W83 (lacks fimA), the fimA mutant FAE, or the mfa1 mutant MFAE (Fig 5A).However, there was no significant difference detected when comparing DL1 biofilm formation with or without the addition of vesicles from the gingipain-null mutant (rgpA -, rgpB -and kgp -) strain KDP128 [26].
The role of vesicles in biofilm disposal was then tested by the addition of vesicles after DL1 biofilms were formed.Similar results were obtained.Vesicles from wild type strain 33277 presented the highest ability to release and disperse DL1 cells from biofilms compared to W83, the fimA mutant.Addition of KDP128 or MFAE vesicles had no effect on DL1 biofilms (Fig 5B).These findings demonstrate that gingipains and fimbrial proteins are involved in inhibition of biofilm formation of S. gordonii.Presumably, fimbrial protein-mediated attachment of vesicles establishes an intimate relationship to S. gordonii, which in turn, facilitates the degradation actions of gingipains.

Discussion
An intriguing feature of P. gingivalis is its ability to invade a variety of host cells.A comparison of the invasive capabilities of P. gingivalis cells and its vesicles revealed a significant higher invasive capability of vesicles.The invasive ability of P. gingivalis cells has been well documented [27,28].P. gingivalis invasion is initiated by attachment to host cells, which is mediated by the interaction between the major fimbriae and surface receptors on host cells, for example primarily integrin α5β1 found on epithelial cells [29].However, the mechanism (s) responsible for invasion of P. gingivalis vesicles is not well understood.There is existing evidence that a different mechanism(s) is likely involved in vesicle invasion, per se.Thus far fimbriae have not been observed on bacterial vesicles, but our previous studies showed that P. gingivalis vesicles consist of several well-known adhesins, such as fimbrial proteins and gingipains [14].In addition, vesicles derived from the fimA mutant did not show any effect on their invasive capabilities compared to found in vesicles from the wild type strain, although the mutant still had a significantly decreased invasion capability compared to its parent cells [30].Moreover, internalization of P. gingivalis cells has been reported to be involved in secretion of the serine phosphatase SerB, which can be found in host cells.SerB activates the actin-depolymerizing molecule cofilin by dephosphorylating the Ser 3 residue [31], resulting in a transient and localized disruption of actin structures that allow entry of the organism into the cytoplasm.Since SerB was not found in P. gingivalis vesicles [12,14], other unknown mechanism(s) may be responsible for the rearrangement of actin in this setting.
Horizontal gene transfer is known as one of the driving forces for the evolution of bacteria; it also represents an efficient step allowing bacteria to acquire new functions.Besides three well-known mechanisms of gene transfer (transformation, transduction, and conjugation), a fourth mechanism for horizontal gene transfer has been identified and involves bacterial vesicles.For example, vesicle-mediated gene transfer was found from the food borne pathogen Escherichia coli O157:H7 to E. coli JM109 and Salmonella enterica [32].In a series of studies on horizontal gene transfer, Tribble et al. reported that DNA transfer occurs in P. gingivalis mainly through a transformation-like process, while a conjugation-like mechanism may only play a minor role [33][34][35].Here, we demonstrated that vesicle-associated DNA was transferred between two strains of P. gingivalis, although the DNA transfer efficiency of vesicles pre-treated with DNase I was lower than that of vesicles in the absence of DNase I pre-treatment.Since native plasmids and bacteriophages have not found in P. gingivalis, it is not known how the chromosomal DNA is packaged into the vesicles.Berleman and Auer proposed three possible routes of incorporation of DNA into bacterial vesicles [36]: 1) cell-free vesicles passively take up extracellular DNA; 2) DNA is packaged into vesicles during the process of cell death; and 3) in some bacteria, DNA is taken up during bacterial replication through a budding process.However, there is no evidence that any of these mechanisms is used for DNA incorporation into P. gingivalis vesicles.Another important question is how DNA is released from vesicles into P. gingivalis' cytoplasm.Previous studies of Pseudomonas aeruginosa demonstrated that vesicles from this bacterium were fused to E. coli, which facilitated the attachment of these E. coli to P. aeruginosa parent cells [37,38].Although there remains a lack of direct evidence for this process, we speculate that vesicles from the fimA mutant may merge with the outer membrane of P. gingivalis 33277 and release vesicular contents into the bacterial cells.
Finding DNA within P. gingivalis vesicles make them good candidates for induction of atherosclerosis.Atherosclerotic cardiovascular disease, a chronic inflammatory disease of the blood vessels, is one of the most common causes of morbidity and mortality world-wide [39].Dysfunction of endothelial cells induced by hypertension, reactive oxygen species, or modified low-density lipoprotein cholesterol (LDL) express more adhesive molecules and promoted adhesion and infiltration by immune system cells [40].Bacteria-mediated inflammation is also known to be, at least in part, responsible for the initiation of atherosclerosis.One of the best model pathogens in pathogenesis of atherosclerosis is P. gingivalis [41].Involvement of P. gingivalis in atherosclerosis is supported by observations from epidemiological, clinical, immunological and molecular studies [42][43][44][45][46].Some of the most convincing evidence of its role in atherosclerosis is the finding of P. gingivalis in cardiovascular samples including atheromantous plaques using PCR and in situ hybridization.A total of 21 studies from 2000 to 2012 found P. gingivalis DNA in 0-88.85% of cardiovascular samples, with a mean of 58% [41].Studies that attempted to detect live P. gingivalis in atheromantous plaques were unsuccessful.However, a study by Kozarov et al. showed that both P. gingivalis and A. actinomycetemcomitans were detected within human coronary artery endothelial cells using immune-staining after cells were incubated with a homogenate of carotid atherosclerotic plaque sample from one patient [47].Another report also discovered P. gingivalis colonies in one of seven homogenates of atherectomy specimens after a pre-incubation of the individual specimens with human monocytic cells, but not when directly incubated with homogenates [48].It remains unknown how monocytic cells facilitated isolation of viable P. gingivalis.Based on our observations, including the efficient invasive activities and detection of vesicle-associated DNA, it is conceivable that vesicles are a major contributor to the development of atherosclerosis and that previous detection of DNA and proteins of P. gingivalis in atherosclerotic samples may most likely originate from P. gingivalis vesicles, although it cannot be ruled out that P. gingivalis cells may also disseminate to the arterial wall.
Dental plaque is a multi-species microbial biofilm.Interactions between/among different species are established by the specific recognition between adhesin and receptor.Most of the well-known adhesive molecules are found in P. gingivalis vesicles such as FimA and Mfa1.Therefore, vesicles, once released, likely act as representatives of P. gingivalis in order to communicate with other oral bacteria.We found that P. gingivalis vesicles functioned as an inhibitor and disperser of S. gordonii biofilms, which mainly involved gingipains, since a gingipainnull P. gingivalis mutant completely lost its inhibitory activity.P. gingivalis gingipains are multifunction proteins containing a proteolytic domain and adhesive domains, therefore, they may act as scissors and/or a glue in dental plaques.A previous in vitro study showed that lysgingipain (Kgp) is involved in the degradation of synthetic peptides derived from the SspB protein of S. gordonii [49].Beside an interaction with Mfa1 of P. gingivalis, S. gordonii SspB also play a role in the bacterial aggregation with Candida albicans through its interaction with the cell wall protein, Als3 [50].It is likely that scissors and/or the glue function of P. gingivalis vesicles may create a favorable environment by eliminating competitors within the oral microbial community.

Fig 1 .
Fig 1. Invasive activity of P. gingivalis vesicles into HOKs, HGFs, and HUVECs.Infected cells presented by differential interference contrast (DIC) images.P. gingivalis cells and vesicles were stained with primary anti-33277 serum and a secondary antibody conjugated with Alex Fluor 546 (red); nuclei were stained with DAPI (blue) (A).Scale bars, 20 μm.The numbers of HOKs (B), HGFs (C), and HUVECs (D) with intercellular P. gingivalis and vesicles were determined by counting ten random areas.Each bar represents the percentage of HOKs, HGFs, and HUVECs with intercellular bacteria or vesicles.The SEs are indicated (n = 3).An asterisk indicates the statistical significance of invasive rates between P. gingivalis 33277 cells and their derived vesicles (P <0.05; t test).doi:10.1371/journal.pone.0123448.g001

Fig 2 .
Fig 2. Detection of DNA fragments in P. gingivalis vesicles.(A) PCR amplification was performed with DNA released from P. gingivalis 33277 vesicles.PCR products were visualized on agarose gels.(B) Semiquantitation of PCR products was conducted with ImageJ software.Each bar represents the intensity of the PCR product band from amplification of DNA in vesicles treated with DNase I relative to that in vesicles without DNase I treatment.Means and SDs are indicated (n = 3).doi:10.1371/journal.pone.0123448.g002

Fig 3 .
Fig 3. Detection of RNAs in P. gingivalis vesicles.RNA associated with vesicles from P. gingivalis 33277 was identified with a reverse transcription and qPCR analyses.Each bar represents the fold increase of DNA fragment levels detected after reverse transcription and qPCR analyses normalized by those detected in the vesicles without a reverse transcription.Values represent means ± SD of three replicates.doi:10.1371/journal.pone.0123448.g003

Fig 4 .
Fig 4. Horizontal gene transfer in P. gingivalis strains.Vesicles carrying an erythromycin resistant gene were derived from a fimA mutant and used as donors.P. gingivalis 33277 was the recipients.(A) The P. gingivalis 33277 receiving the erm gene was selected on TSB blood plates supplemented with erythromycin.Each bar represents the number of colonies detected on the plates, after P. gingivalis 33277 was incubated with vesicles without DNase pre-treatment, vesicles with DNase pre-treatment, and P. gingivalis 33277 without vesicles respectively, or vesicles alone without P. gingivalis 33277.An asterisk indicates a significant difference between colonies detected in P. gingivalis 33277 incubated with the vesicles without DNase treatment vs. those found in other groups (n = 3, p <0.05; t test).(B) The erm gene was detected by PCR with the fimA primers as a probe, and PCR products were visualized on agarose gels.The PCR product of the fimA gene is shown at 1.2 kb, and the PCR product of the chimeric gene (fimA + erm) is at 3.3 kb.

Fig 5 .
Fig 5.Inhibition of S. gordonii biofilm formation by P. gingivalis vesicles.S. gordonii DL1 was introduced to saliva-coated wells for establishment of S. gordonii biofilms.Vesicles derived from wild type strains 33277 or W83, the fimA mutant (FAE), the mfa1 mutant (MFAE) or a gingipain-null mutant (KDP128), were either added into the wells at the beginning (A) or after S. gordonii biofilms were already formed (B).The S. gordonii biofilm remaining on the surfaces of the wells was determined, and the relative binding ability of each strain was represented and compared to that of wild type strain 33277 = 1 unit.The results presented are the averages of four independent experiments.Error bars represent SDs.An asterisk indicates there was a statistically significant difference in the biofilm formation between S. gordonii in the presence or absence of P. gingivalis vesicles (P <0.05; t test).doi:10.1371/journal.pone.0123448.g005

Table 1 .
Bacterial strains used in this study.