Genetic Transformation of an Obligate Anaerobe, P. gingivalis for FMN-Green Fluorescent Protein Expression in Studying Host-Microbe Interaction

The recent introduction of “oxygen-independent” flavin mononucleotide (FMN)-based fluorescent proteins (FbFPs) is of major interest to both eukaryotic and prokaryotic microbial biologists. Accordingly, we demonstrate for the first time that an obligate anaerobe, the successful opportunistic pathogen of the oral cavity, Porphyromonas gingivalis, can be genetically engineered for expression of the non-toxic green FbFP. The resulting transformants are functional for studying dynamic bacterial processes in living host cells. The visualization of the transformed P. gingivalis (PgFbFP) revealed strong fluorescence that reached a maximum emission at 495 nm as determined by fluorescence microscopy and spectrofluorometry. Human primary gingival epithelial cells (GECs) were infected with PgFbFP and the bacterial invasion of host cells was analyzed by a quantitative fluorescence microscopy and antibiotic protection assays. The results showed similar levels of intracellular bacteria for both wild type and PgFbFP strains. In conjunction with organelle specific fluorescent dyes, utilization of the transformed strain provided direct and accurate determination of the live/metabolically active P. gingivalis' trafficking in the GECs over time. Furthermore, the GECs were co-infected with PgFbFP and the ATP-dependent Clp serine protease-deficient mutant (ClpP-) to study the differential fates of the two strains within the same host cells. Quantitative co-localization analyses displayed the intracellular PgFbFP significantly associated with the endoplasmic reticulum network, whereas the majority of ClpP- organisms trafficked into the lysosomes. Hence, we have developed a novel and reliable method to characterize live host cell-microbe interactions and demonstrated the adaptability of FMN-green fluorescent protein for studying persistent host infections induced by obligate anaerobic organisms.


Introduction
The advent of the green-fluorescent protein (GFP) technology generated a myriad of in-vivo and in-vitro cell imaging applications including spatio-temporal analysis and expression of diverse cell signaling molecules, cellular organelles, and the high-throughput functional annotation of genome sequences in various organisms [1,2,3]. Also, the GFP protein and its variants have been successfully used to analyze the prokaryotic organisms, such as bacterial pathogens, and their interactions with various host cells and the cellular machineries [4]. While the GFP-based biosensors revolutionized the imaging techniques, their strict requirement for molecular oxygen as a co-factor for the synthesis of their chromophores significantly limited their applications for anaerobic microorganisms and cellular microenvironments [5].
The recent discovery of oxygen-independent flavin mononucleotide (FMN)-based fluorescent proteins (FbFPs), engineered from the blue-light photoreceptors of Bacillus subtilis and Pseudomonas putida, opened up an exciting avenue of research for real-time imaging of live anaerobic microorganisms and anaerobic inter-and intracellular processes [6]. Drepper at al. showed that the codon optimization of these photoreceptors in the light oxygen voltage domains of E. coli and the ''facultative'' anaerobic bacterium Rhodobacter capsulatus generates significant degree of cyan-green fluorescence expression both in the absence and presence of oxygen. Subsequently, another study illustrated that the FbFPs can be also used as reporter proteins for fungi such as Saccharomyces cerevisiae and the pathogen C. albicans under hypoxic conditions [7]. Recent reports further indicate the versatile applications of these non-toxic anaerobically fluorescent proteins for performing quantitative bacterial real-time assays in addition to examining metabolic activities of marine bacterial species (Roseobacter and Phaeobacter spp.) under various conditions [8].
Porphyromonas gingivalis is a successful opportunistic pathogen of the oral mucosa and prominent member of the oral biofilms. Hostpathogen interactions involving the host-adapted pathogen, P. gingivalis have been studied extensively in various cell types and animal models [9,10,11,12,13,14,15,16]. Nevertheless the precise characterization of the intracellular trafficking and the ultimate fate of P. gingivalis in host cells including its preferential host cell type, gingival epithelial cells, remain incomplete. This has been to a great extent due to the lack of genetic molecular tools for imaging of the live/metabolically active anaerobic microorganisms. We demonstrate here that the obligate anaerobe, P. gingivalis, can be genetically altered for FMN fluorescent protein expression. P. gingivalis-FbFP transformants (PgFbFP) were made by utilizing the FMN-based fluorescent Bs2 protein expression system [6]. The transformants produced bright green-fluorescence and biosynthesis was independent of oxygen. Our functional assays with the PgFbFP demonstrated that this new genetic tool enables the direct examination of metabolically active P. gingivalis infectious processes under oxygen-limited or strictly anaerobic biological systems. The green-fluorescent transformants were used in conjunction with the lysosomal and endoplasmic reticulum (ER) specific fluorescent markers. Our results showed that we can distinctly determine P. gingivalis co-localization with the specific cellular compartments (e.g. lysosomal versus ER) in primary GECs. Since our earlier study indicated that the ATP-dependent Clp serine protease (ClpP) of P. gingivalis is likely to be critical for the organism's optimum adaptation to the intracellular life and survival in oral epithelial cells [17], GECs were infected with PgFbFP and the isogenic ClpPmutant of the wild strain simultaneously. The coinfection of the host cells with the green fluorescent organism and its non-fluorescent mutant (detected by the anti-P. gingivalis-specific polyclonal antibody coupled with blue fluorescent conjugated secondary antibody) provided initial characterization on the cellular machineries differentially involved in the regulation of the P. gingivalis' trafficking.
The non-toxic fluorescent biosensor technology is a significant, developing field and it is of great interest to biologists for the study of a variety of different biological processes. Undoubtedly, the construction of an oxygen-independent fluorescent protein expression in strict anaerobic organisms such as P. gingivalis can provide numerous analysis platforms for studying mechanisms/dynamics of chronic infections in variety of host cells and systems.

Results
Fluorescence Validation Assays for the FbFP Expression in P. gingivalis Transformant P. gingivalis cells were genetically altered via conjugation of the recombinant P. gingivalis-E. coli shuttle vector, pCJO1 (derived from pT-COW) containing FbFP insert from Bs2-evoglow expression plasmid. The resulting transformants, grown anaerobically, were examined for the FbFP expression by epifluorescence microscopy (Fig. 1). Bright green fluorescence levels were visualized in the PgFbFP cells ( Fig.1) while the wild-type P. gingivalis cells did not produce any detectable fluorescence (not shown). DAPI nucleic acid staining was used to confirm the bacterial cells in the samples (Fig. 1). Thus, fluorescence microscopy confirmed a high degree of accumulation of FbFP in the PgFbFP cells which exhibited a large uniform amount of brightness (Fig. 1). The epi-illumination produced slight axial diffraction (fluorescence flare) in some of the cells due to the fluorescent light from above and below the plane of focus.
In parallel experiments, we also measured the fluorescence intensities of PgFbFP and wild-type P. gingivalis using a spectrofluorometer at excitation wavelength, 450 nm (Fig. 2). The analysis of the bacterial samples further confirmed that the transformant bacteria exhibited strong levels of brightness (a maximal fluorescence emission at 495 nm upon excitation with the blue light) similar to the levels previously reported for the FbFP expression in E. coli and C. albicans [7]. There was no fluorescence observed in the wild type (Fig. 2). Hence, these analyses established the adaptability of FbFP expression system in an obligate anaerobe that is a member of the human oral microbiome [18].
Examination of the In-vitro Growth and Host Cell Invasion by P. gingivalis Transformant Prior to performing functional host cell infection assays, we studied the in-vitro growth dynamics of PgFbFP in P. gingivalis' defined TSB liquid media [19]. PgFbFP was inoculated into the anaerobically pre-conditioned TSB media and cultured anaerobically at 37uC for 72 h (late-exponential phase). The results showed that the PgFbFP and the wild type P. gingivalis had similar growth rates indicating that the genetic manipulation of the organism for the FbFP expression did not affect its basic growth/cellular metabolism (Fig. 3).
Next, we wanted to assess whether PgFbFP can efficiently invade and proliferate in the primary GECs as the wild type. The antibiotic protection assays were performed as described before [20]. Infection by both PgFbFP and wild type resulted similar numbers of intracellular bacteria at 24 h post-infection, confirming that the GEC invasion efficiency and intracellular survival level were similar for both strains (data not shown). Subsequently, PgFbFP infection of GECs at 12 and 24 h was visualized by fluorescence microscopy as we described previously [21]. In agreement with the antibiotic protection assay, visualization of the infection of the GECs by the transformed bacteria by fluorescence imaging displayed a high level of cellular infection (Fig. 4). Quantitative analysis of the bacterial fluorescence by NIH ImageJ analysis software indicated that at 24 h infected cells harbored at least 2.5 times more bacteria than at 12 h (data not shown). This was consistent with the ability of wild P. gingivalis to successfully proliferate in GECs over time [22]. Thus, the data further verified that PgFbFP can effectively invade, survive, and replicate in the GECs, which are the central cell type for successful colonization of P. gingivalis in the oral cavity.

Analyses of P. gingivalis Trafficking in the Primary GECs through FbFP Expressing Transformant
P. gingivalis is a host-adapted pathogen that can successfully survive and proliferate in cultures of primary GECs for extended periods of time, and later spread intercellularly [21,23,24]. In line  with this, the infection can modulate and interfere with a variety of GEC responses [25,26,27,28,29,30].
Bacterial intracellular trafficking studies have shown that ER structures are frequently utilized by persistent organisms to promote their intracellular growth due to the nutritionally rich nature of these components [31,32]. P. gingivalis is localized to the perinuclear region of the host cells, however it has not been determined whether the intracellular organism associates with any subcellular compartments in order to further its intracellular life in GECs [14,22].
As we have developed a tool that can unequivocally follow live P. gingivalis inside host cells, GECs were infected with PgFbFP (green) and stained with the ER-specific fluorescent dye (ERTracker-Red) to determine the potential co-localization between ER and the intracellular bacteria ( Fig. 5A and B). Fluorescence microscopy analysis of the samples demonstrated a significant level of overlap between the PgFbFP strain and the ER structures, which are located in the perinuclear compartments typical of the ER (Fig. 5A). The co-localization events were quantified by the JACoP toolbox under NIH ImageJ using Manders' coefficient correlation [33]. This object-based fluorescent intensity analysis revealed ,90% co-localization rate between the P. gingivalis and ER networks in the GECs (Fig. 5B). In contrast, as a control, we infected GECs with a ClpPmutant of P. gingivalis and studied the co-localization events between the mutant bacteria and the ER. The mutant bacterial localization was detected by anti-P. gingivalis-specific polyclonal antibodies coupled with green fluorescent conjugated secondary antibody. As shown in Fig 5A, The ClpPlocalization did not significantly overlap with the ER network (Fig. 5A). Measurement of the co-localization rate by the Manders' coefficient correlation analysis showed a ,34% association, indicating only a partial co-localization (Fig. 5B).
Our previous study and the studies on Listeria spp,. indicated the ClpP expression is essential for the intracellular survival [17,34]. Therefore, P. gingivalis ClpP-infected cells were stained with the lysosome specific fluorescent dye (LysoTracker-Red) and postlabeled with the anti-P. gingivalis antibody coupled with fluorescent conjugated secondary antibody (green) to determine potential subcellular localization of the mutant cells with lysosomes at 24 h infection. (Fig. 6A). The collected images produced an extensive degree of yellow hot-spots (co-localization) and quantitative analysis measured ,66% co-localization between the lysosomal structures and the mutant bacteria (Fig. 6C). The results suggested the majority of mutant bacteria trafficked into the lysosomes and potentially were in the process of lysosomal degradation.
One of the major limitations of antibody based bacterial detection techniques for intracellular trafficking/fate determination studies is that antibody staining cannot distinguish damaged or dead bacteria from live bacteria and/or bacterial antigens from intact/ metabolically active versus degrading microorganisms. Therefore, we developed a novel approach whereby we co-infected GECs with the genetically modified ''green fluorescent P. gingivalis'' and the ClpPsimultaneously to analyze the subcellular compartmentalization of P. gingivalis and its isogenic mutant in the same host cell throughout the infection by triple fluorescent staining. The nonfluorescent mutant (ClpP-) was detected with the anti-P. gingivalis antibodies as described above, and the lysosomes were stained with LysoTracker-Red, as shown in Fig. 6B. Quantitative co-localization analysis of the two P. gingivalis strains utilizing the specific organelle dye within the same host cells further verified the ER network as a prominent subcellular niche for live P. gingivalis infection in GECs, while the majority of ClpPmutants trafficked into the lysosomes at 24 h post-infection ( Fig. 6B and C).

Discussion
Within the last decade, genetically encoded fluorescent molecule-based technologies have become widely used approaches for the study of host cell-parasite interaction [35]. Nevertheless, the absence of oxygen-independent fluorescent protein systems significantly restricted the direct examinations of infectious agents that require limited or strict anaerobic living conditions. P. gingivalis, a gram-negative strict anaerobe and successful colonizer of oral tissues, can survive, replicate, and spread in GECs [18,36]. Since the organism can be cultured in-vitro and is a prominent member of oral biofilms, P. gingivalis is considered a facultative intracellular pathogen. However, it can rather be viewed essentially as an intracellular pathogen if one considers that human epithelial cells are the privileged niche for multiplication, and human oral mucosa as the pathogen's primary reservoir [13,18]. In the absence of fluorescent-based genetic tools only limited details of the P. gingivalis infectious cycle and the interacting host molecules in the oral tissues have emerged.
Several approaches have been used in the past for imaging of anaerobic microorganisms such as P. gingivalis in ex-vivo or in-vitro host models [21,37,38,39]. These techniques frequently involved specific polyclonal antibodies or 16S rDNA probes which cannot distinguish live organisms from dead or damaged/fragmented organisms, or using of non-specific chloromethyl derivative fluorescent dyes which are significantly diluted after a couple of generations.
This study demonstrates for the first time the versatility of the recently reported oxygen-independent and non-toxic FbFPs technology for studying real-time host cell-anaerobic bacteria interaction. The successful transformation of P. gingivalis into a green fluorescent bacterium with a healthy basic metabolism and consistent phenotype for the invasion, multiplication and survival in the GECs were key findings for a variety of future analyses. Moreover, we developed a novel method for an accurate qualitative (visual) and quantitative evaluation of the differentially stained bacteria within the same cells ( Fig. 6B and C). This provided a reliable spatial comparative analysis of two different strains of P. gingivalis and their relationship to a specific organelle (e.g. lysosomes, ER), and further validated the distinct fates of P. gingivalis and its isogenic ClpPmutant in the primary GECs following 24 h infection. These new findings are consistent with the organism's previously described pro-survival phenotype in the GECs [23,24,30]. Also, the preliminary finding of the ClpPmutant trafficking to the lysosomal pathway substantiates the importance of the stress-induced protease, ClpP, for the intracellular survival of P. gingivalis in GECs [17]. This result is also in agreement with the results obtained by Listeria monocytogenes where ClpP serine protease induces the early escape of the pathogen from the phagosomal compartments [34,40].
While this is an initial study in the direct characterization of metabolically active P. gingivalis trafficking in the human gingival cells, it certainly offers more detailed analyses of diverse inter-and intracellular bacterial trafficking and infectious processes induced

Materials and Methods
Bacteria, Growth Conditions, and Construction of PgFbFP P. gingivalis ATCC 33277, ATP-dependent Clp serine proteasedeficient isogenic mutant (ClpP-), and PgFbFP were cultured anaerobically at 37uC in trypticase soy broth (TSB) supplemented with yeast extract (1 mg ml-1), haemin (5 mg ml-1) and menadione (1 mg ml-1). Erythromycin (10 ug ml-1) was added to the media for culture of the mutant strain [17]. The media for PgFbFP strain was supplemented with tetracycline (3 mg ml-1). All bacteria were grown for 24 h, harvested by centrifugation at 6000 g and 4uC for 10 min, washed twice, and resuspended in Dulbecco's Phosphatebuffered saline (Sigma) pH 7.3 before they were reacted with host cells. The number of bacteria was determined using a Klett-Summerson photometer.
Construction of PgFbFP Strain with pCJO1. The evoglow-Bs2plasmid from the Evoglow kit (Evocatal, Germany) was used to amplify a flavin mononucleotide (FMN)-based fluorescent gene to be fused with the P. gingivalis (ATCC 33277) fimA promoter region (Fig S1 and Table S1) [41]. The fusion PCR protocol included an initial step of 95uC for 5 min, denaturation step of 95uC for 1 min, annealing step of 60uC for 1 min, extension step of 72uC for 1 min, and a final elongation step of 72uC for 10 min (Fig S2). The fusion PCR product was then cloned into the Sal1 and SPH1 sites of P. gingivalis-E. coli shuttle vector, pT-COW [42] and transformed into E. coli S17 cells. The recombinant plasmid (named pCJO1) in the E. coli S17 donor cells were subsequently transferred to P. gingivalis through conjugation [28]. Finally, the tetracycline-resistant P. gingivalis clones were grown agar plates and selected for fluorescence analyses.

Generation and Culture of Human Primary Gingival Epithelial Cells
Primary cultures of GECs were generated as described previously [20]. No subject recruitment per se was done. Adult patients were selected at random and anonymously from those Figure 6. Analysis of ClpPmutant association with lysosomes. A. GECs were infected with the ClpP-mutant of P. gingivalis and incubated with the Lyso Tracker (red). The ClpP-mutant strain was detected by green fluorescent antibody. Double staining analysis of the obtained images indicated a large level of co-localization (yellow) between the mutant and the lysosomes. B. Distinct intracellular trafficking of P. gingivalis strains in GECs. GECs were co-infected with PgFbFP (green) and ClpP-mutant simultaneously, and incubated with the Lyso Tracker (red). The samples were reacted with the anti-P. gingivalis antibody coupled with blue fluorescent secondary antibody which labeled both PgFbFP (green) and ClpP-mutant, blue (upper panel). The overlay images of PgFbFP (green) and the anti-P. gingivalis staining (blue) displayed the differentially localized strains in the same cell (yellow for the PgFbFP staining and clean blue for ClpP-mutant). Image analysis showed no significant co-localization (yellow) between PgFbFP (green) and lysosomes (red), whereas the merged images of lysosomes (red) and anti-P.gingivalis staining (blue) produced a significant level of purple (co-localization of ClpP-mutant with lysosomes). C. Quantitative analyses of P. gingivalis co-localization with lysosomes. The co-localization assays depicted in Figure 6A and B were analyzed quantitatively using Mander's coefficient correlation by JACoP/ImageJ analysis software. The results are representative of an average of 50 cells per sample studied from at least two separate experiments performed in duplicate. doi:10.1371/journal.pone.0018499.g006 presenting at the University of Florida Dental Clinics for tooth crown lengthening or impacted third molar extraction. Gingival tissue that would otherwise be discarded was collected after informed written consent by the patient (approved by the Institutional Review Board of University of Florida). No patient information was collected. This study is approved by the Institutional Review Board under the University of Florida human subjects assurance number FWA 00005790. Briefly, healthy gingival tissue was obtained after oral surgery, and surface epithelium was separated by overnight incubation with 0.4% dispase. Cells were cultured as monolayers in serum-free keratinocyte growth medium (KGM) (Lonza) at 37uC in 5% CO 2 . GECs were used for experimentation at 75-80% confluence and cultured at least for 24 h before infection with bacterial cells.

Fluorescence Assays for the Analyses of PgFbFP
FbFP-expressed P. gingivalis strain grown anaerobically at 37uC were washed with PBS, placed on microscopic glass slides and fixed with 10% neutral buffered formalin for 15 min at room temperature. The samples were treated with 4,6-diamidino-2phenylindole (DAPI) (1 mg/ml) (Sigma) to visualize the bacterial nucleic acids. Finally, the PgFbFP samples were washed twice with PBS and analyzed using a Zeiss Axio imager A1 epifluorescence microscope equipped with optical filter sets with excitation, 475/ 40; emission, 500/10 for the green fluorescence, and excitation 360; emission 440/50 nm for the DAPI. The images were collected by a cooled CCD camera (Qimaging). Single exposure images were captured sequentially and saved by Qcapture software as we described previously [26].
Fluorescence emission spectra of the PgFbFP cells were measured with a BioTek Instrument, model Synergy MX Multimode spectrometer (excitation wavelength, 450 nm) and analyzed by Software: Gen5. Fluorescence intensity is expressed in arbitrary units (A.U.).

Growth Measurement of PgFbFP
P. gingivalis ATCC 33277 and the PgFbFP strain were cultured anaerobically at 37uC in the liquid broth media (TSB) as explained above. Fifty microliters of overnight cultures were used to inoculate in 10 ml of TSB medium pre-incubated anaerobically at 37uC. Bacteria were further incubated for 3 h, 6 h, 12 h, 24 h, 48 h, and 72 h. At each incubation time point, OD600 were measured using a BioMate 3 Spectrophotometer (Thermo Electron Corporation).

Epithelial Cell Invasion Assays
Antibiotic Protection Assay. PgFbFP invasion of GECs was determined by the antibiotic protection assay described previously [20]. In brief, bacteria in KGM were incubated with GECs in 6well plates for 60 min at 37uC. After washing with PBS, remaining external bacteria were killed with metronidazole (200 mg ml-1) and gentamicin (300 mg ml-1) and further incubated for 24 h. GECs were washed and lysed with sterile distilled water, and intracellular bacteria were enumerated by culture on blood agar supplemented with haemin and menadione.

Imaging and Quantitation of the PgFbFP Invasion by
Fluorescence Microscopy. GECs were seeded onto 2-well chambered cover-glass slides (Nalge-Nunc International) at a density of 2610 4 cells per well and cultured for 24 h. Cells were infected with PgFbFP at a multiplicity of infection (m.o.i.) of 100 at 37uC in 5% CO 2 incubator. After 12 or 24 h incubation, the slides were washed four times with PBS containing 0.1% Tween 20 to remove the non-adherent bacteria. Cells were fixed in 10% neutral buffered formalin, and rinsed in PBS at room temperature. F-actin was labeled with phalloidin-tetramethylrhodamine B isothiocyanate (TRITC) (Sigma) at 1:100 for 45 min. The samples were mounted in anti-fade mounting medium with DAPI (Vector Laboratories). Samples were visualized using the fluorescence microscope system described above. Acquired images were analyzed for fluorescence intensity with NIH ImageJ analysis software [25].

Fluorescence Imaging of Bacterial Trafficking in GECs
GECs cultivated on the 2-well chambered cover-glass slides were infected with P. gingivalis wild type, ClpPmutant or PgFbFP at an m.o.i of 100 at 37uC for 24 h. The infected live cells were incubated with ERTracker-Red (1 mM/ml) (Invitrogen), a specific fluorescent cell-permeable dye for endoplasmic reticulum labeling or LysoTracker-Red (1 mM/ml) (Invitrogen), a specific fluorescent cell-permeable dye for lysosomes, for 30 min. Cells were fixed with 10% neutral buffered formalin and wild type and Clpinfected cells permeabilized by 0.1% Triton X-100 and reacted with anti-P. gingivalis 33277 rabbit polyclonal antibody (1:1000) followed by Oregon Green 488 goat secondary antibody (1:500) (Invitrogen). For the co-infection studies, GECs were infected with PgFbFP and ClpPmutant strains simultaneously and the live cells were labeled with LysoTracker-Red and then fixed and permeabilized as explained above. The samples were reacted with anti-P. gingivalis 33277 rabbit polyclonal antibody followed by Marina Blue Goat secondary antibody (1:500) (Invitrogen). The samples were mounted in anti-fade mounting medium (Vector Laboratories) and visualized using the fluorescence microscope system described above.

Measuring Co-localization Events
Co-localization analysis was carried out using the JACoP tool under NIH ImageJ software as described previously [33]. Images were pre-processed to correct uneven illumination and to minimize noise and background. The co-localization rates were measured based on Manders' coefficient, which varies from 0 to 1. A coefficient value of zero corresponds to non-overlapping images while a value of 1 reflects 100% co-localization between the images being analyzed. Figure S1 The construction of FbFP expressing P. gingivalis transformant. The codon usage of E. coli FbFP was adapted to P. gingivalis, and the resulting gene was placed under the transcriptional control of the fimA promoter of P. gingivalis.