Oral Community Interactions of Filifactor alocis In Vitro

Filifactor alocis is a gram positive anaerobe that is emerging as an important periodontal pathogen. In the oral cavity F. alocis colonizes polymicrobial biofilm communities; however, little is known regarding the nature of the interactions between F. alocis and other oral biofilm bacteria. Here we investigate the community interactions of two strains of F. alocis with Streptococcus gordonii, Fusobacterium nucleatum, Porphyromonas gingivalis and Aggregatibacter actinomycetemcomitans, organisms with differing pathogenic potential in the oral cavity. In an in vitro community development model, S. gordonii was antagonistic to the accumulation of F. alocis into a dual species community. In contrast, F. nucleatum and the type strain of F. alocis formed a synergistic partnership. Accumulation of a low passage isolate of F. alocis was also enhanced by F. nucleatum. In three species communities of S. gordonii, F. nucleatum and F. alocis, the antagonistic effects of S. gordonii superseded the synergistic effects of F. nucleatum toward F. alocis. The interaction between A. actinomycetemcomitans and F. alocis was strain specific and A. actinomycetemcomitans could either stimulate F. alocis accumulation or have no effect depending on the strain. P. gingivalis and F. alocis formed heterotypic communities with the amount of P. gingivalis greater than in the absence of F. alocis. However, while P. gingivalis benefited from the relationship, levels of F. alocis in the dual species community were lower compared to F. alocis alone. The inhibitory effect of P. gingivalis toward F. alocis was dependent, at least partially, on the presence of the Mfa1 fimbrial subunit. In addition, AI-2 production by P. gingivalis helped maintain levels of F. alocis. Collectively, these results show that the pattern of F. alocis colonization will be dictated by the spatial composition of microbial microenvironments, and that the organism may preferentially accumulate at sites rich in F. nucleatum.


Introduction
The dental plaque biofilm is comprised of complex communities of microorganisms embedded on tooth surfaces, and is a direct precursor of periodontal disease [1][2][3]. Until fairly recently, a limited number of organisms in the subgingival biofilm, the so called 'red complex', were considered the predominant pathogens in chronic and severe cases of adult periodontitis [4,5]. However, microbiome studies over the last several years have changed our understanding of the multispecies microbial communities that inhabit the oral cavity. The microbial composition of periodontal disease lesions is much more varied than previously recognized and contains high levels of fastidious and as yet-to-be-cultivated taxons [6]. Organisms such as Selenomonas, Synergistes, Desulfobulbus, TM7 and Filifactor alocis have been identified as potential pathogens in a number of independent studies [6][7][8][9][10].
F. alocis is a Gram-positive, slow-growing, obligate anaerobic rod that is found at increased frequency and in higher numbers in periodontal disease sites compared with healthy sites [6,8,9,[11][12][13]. In addition, F. alocis is emerging as an important organism in aggressive periodontitis in children [14], endodontic lesions [15] and pericoronitis [16]. Study of the pathogenic properties of F. alocis is now important to impute a causal association between F. alocis and periodontal disease. In that regard, F. alocis has a number of characteristics consistent with that of a periodontal pathogen.
The organism is resistant to oxidative stress and generally proinflammatory and proapoptotic [17,18]. Furthermore, F. alocis produces several proteases and neutrophil-activating protein A which are upregulated during internalization within epithelial cells [19].
An important early step in the colonization process of periodontal pathogens is the ability to adhere to oral surfaces and accumulate in physiologically compatible heterotypic communities. Schlafer et al. [20] examined the topology of F. alocis within in vivo grown subgingival biofilms from periodontitis patients. F. alocis was frequently present in densely packed groups as a part of concentric bacterial aggregates, and in mushroom-like protuberances on the surface of the biofilm. F. alocis also formed structures resembling test-tube brushes (often observed in dental biofilms [21]). It is likely, therefore, that F. alocis can interact with a variety of oral bacteria and participate in community development. In this study we utilize in vitro models to examine the community forming interactions of F. alocis with common oral organisms of varying degrees of pathogenicity.

Ethics Statement
Saliva collection was approved by the University of Louisville IRB, Protocol # 12.0345 and designated as non-human subjects research as saliva was collected from study principal investigator only.

Saliva Collection
Whole saliva was collected from a healthy volunteer, and dithiothreitol was added to a final concentration of 2.5 mM. Particulate matter was removed by centrifugation at 10 000 g for 10 min. Clarified saliva was diluted to 10% with distilled water, filtered through 0.2 mm pore size nitrocellulose and stored at 280uC. Glass coverslips were reacted with 0.5 ml of 10% saliva (4uC for 16 h) and rinsed with PBS prior to use.

Community Analysis by Confocal Laser Scanning Microscopy (CLSM)
Quantitative and structural analysis of homotypic and heterotypic communities was accomplished by CLSM and subsequent image analysis essentially as previously described [23]. A) Single species. S. gordonii, F. nucleatum, A. actinomycetemcomitans or P. gingivalis cells (2610 8 ) were stained with hexidium iodide (15 mg/ml; Invitrogen, Carlsbad, CA), and F. alocis cells (2610 8 ) were stained with fluorescein isothiocyanate (FITC, 4 mg/ml, Invitrogen). Bacteria were cultured in individual chambers of a Culture Well chambered coverglass system (Grace Bio Laboratories, Bend, OR) in FAB (unless otherwise stated) anaerobically with rocking at 37uC. B) Dual species. S. gordonii, F. nucleatum, A. actinomycetemcomitans or P. gingivalis cells (2610 8 unless otherwise stated) were stained with hexidium iodide were cultured anaerobically in FAB on coverslips overnight with rocking at 37uC. F. alocis cells (5610 7 ) were stained with FITC and reacted with the partner species anaerobically with rocking in FAB at 37uC. C) Three species. S. gordonii stained with hexidium iodide, and F. nucleatum stained with 49,6-diamino-2-phenylindole (DAPI, 1 mg/ml; Invitrogen) were co-cultured on coverslips overnight in FAB anaerobically with rocking at 37uC. After washing, FITC-labelled F. alocis were reacted with the dual species substratum anaerobically with rocking in FAB at 37uC. Coverslips with assembled communities were washed, and quantitative and structural analysis was performed on an Olympus confocal laser scanning microscope (FV1000) with a 660 objective. A series of 0.5-mm-deep optical fluorescent x-y sections (1206120 mm) were collected to create digitally reconstructed 3D images with Volocity software (Perkin Elmer, Waltham, MA).

Statistical Analysis
Community assays were repeated independently four times in triplicate and analysed with a Student's unpaired two-tailed t-test. Pearson's correlation coefficient (PCC) in Volocity was used to ascertain the degree of inter-species colocalization [24,25].

Monospecies Communities
Initially, the structural and quantitative properties of single species communities were determined at 24, 48 and 72 h (Fig. 1). In monospecies accumulations, F. alocis, A. actinomycetemcomitans and P. gingivalis sporadically formed small microcolonies. Community formation by the F. alocis low passage clinical isolate D-62D was sparser compared to the type strain at all time points. F. nucleatum communities developed in unevenly distributed dense clusters. In vivo S. gordonii attaches to the salivary pellicle on enamel surfaces [3], and hence a saliva-coated glass surface was used for S. gordonii community formation. While processing of saliva by centrifugation and filtering can remove mucins and anti-microbial compounds which could influence bacterial growth, S. gordonii developed a markedly thick biofilm, up to 10 mm deep. The biomass of all of the species tested increased over time.

Dual Species Communities
A) F. alocis-S. gordonii. The ability of F. alocis to accumulate on substrata of S. gordonii attached to saliva-coated glass coverslips was investigated. Both F. alocis ATCC 35896 and D-62D strains exhibited sparse accumulation with S. gordonii DL1 ( Fig. 2A). Quantitative measurement of the dual-species communities (Fig. 2B) demonstrated that S. gordonii did not show a significant difference compared to accumulation in single species communities. However, the accumulation of F. alocis strains with S. gordonii showed a dramatic decrease compared to F. alocis alone. At 72 h, the biovolume of strain ATCC 35896 accumulation was reduced 19-fold (p,0.001) by S. gordonii whereas D-62D accumulation was reduced 21-fold (p,0.001). This result suggests that the presence of S. gordonii is strongly inhibitory to F. alocis in F. alocis-S. gordonii heterotypic communities. B) F. alocis-F. nucleatum. Fig. 3A shows that both F. alocis strains accumulated around regions of F. nucleatum abundance. Time-resolved inspection of dual-species biofilm development ( Fig. 3B) revealed that F. alocis strains exhibited an increase in total biovolume: after 48 h in the case of strain ATCC 35896, and after 72 h with strain D-62D. This was accompanied by mutualistic growth of F. nucleatum after 48 h, although synergism was lost at 72 h with strain D-62D. Collectively, these results indicate that F. nucleatum and F. alocis can exhibit a synergistic relationship in the accumulation of dual-species biofilms.
C) F. alocis-A. actinomycetemcomitans. Large aggregations of A. actinomycetemcomitans formed between 48 and 72 h of coculture (Fig. 4A). The biovolume of both F. alocis ATCC 35896 and A. actinomycetemcomitans in heterotypic communities increased following 48 h incubation indicating mutualistic growth. In contrast, co-culture of F. alocis strain D-62D with A. actinomycetemcomitans did not stimulate the accumulation of either species, indicating strain-specific F. alocis interactions with A. actinomycetemcomitans (Fig. 4B). D) F. alocis-P. gingivalis. Heterotypic F. alocis-P. gingivalis communities are shown in Fig. 5A. On substrata of P. gingivalis, both F. alocis strains showed accumulation over a 72 h period; however, the biovolume of F. alocis was reduced with P. gingivalis as compared with F. alocis alone (Fig. 5B). In contrast, P. gingivalis was capable of growth in the presence of F. alocis, reaching greater biovolume at 72 h compared to P. gingivalis alone (Fig. 5B). These results reveal that F. alocis and P. gingivalis can assemble into heterotypic communities; however, while P. gingivalis benefits from this interaction, accumulation of F. alocis is inhibited.

Colocalization Within Communities
To investigate initial physical interactions between bacteria in dual species communities, colocalization analysis with Volocity software was performed, employing Pearson's Correlation Coefficient (PCC) (Fig. 6). F. alocis-S. gordonii heterotypic communities showed a low level of colocalization, reflective of the antagonistic relationship of S. gordonii toward F. alocis. In contrast, F. alocis and F. nucleatum, which exhibit synergy, displayed a higher degree of colocalization in communities. Colocalization between A. actinomycetemcomitans and both strains of F. alocis was low, and hence the mutualistic growth between A. actinomycetemcomitans and F. alocis ATCC 35896 may depend on soluble secreted factors. F. alocis colocalization with P. gingivalis was relatively high, indicating that the two species physically interact before the inhibitory effect of P. gingivalis is manifest.

Interaction Between P. gingivalis and F. alocis
Our data indicated that F. alocis and P. gingivalis physically interact and hence we utilized a panel of P. gingivalis mutants deficient in expression of major surface adhesins to begin to investigate the molecular basis of the interaction. Loss of the major (FimA) fimbriae or the internalin family protein InlJ had no effect on community formation with F. alocis (not shown). In contrast, loss of the minor fimbriae (Mfa1) increased the accumulation of F. alocis with P. gingivalis (Fig. 7A and B). This effect was more pronounced with strain ATCC 35896 than with D-62D. Complementation of the Dmfa1 mutation with the wild type allele in trans reduced heterotypic community, in many instances to levels below those of the wild type (Fig. 7B), presumably the result of elevated expression of Mfa1 from the multicopy plasmid. These results indicate that Mfa1 may have a suppressive role in the development of P. gingivalis-F. alocis communities.
To test for possible chemical communication between P. gingivalis and F. alocis we examined heterotypic community development between F. alocis and a mutant of P. gingivalis with a deletion in luxS, the gene encoding the enzyme responsible for the synthesis of the AI-2 family of signaling molecules. Community biovolume of both F. alocis strains was significantly reduced with P. gingivalis DluxS compared to the parental strain ( Fig. 8A and B), suggestive of a role for AI-2 in the initial interaction between P. gingivalis and F. alocis. Interestingly, levels of P. gingivalis DluxS were also reduced in the dual species communities in comparison to the parental strain. Thus, LuxS appears to be required for maximal accumulation of P. gingivalis with F. alocis, similar to the situation with P. gingivalis and S. gordonii [26]. To further explore a role for AI-2, we compared P. gingivalis DluxS-F. alocis community development in conditioned medium from P. gingivalis parental and DluxS strains. Conditioned medium from the parental, but not the LuxS mutant, strain significantly increased the biovolume of F. alocis ATCC 35896 in a community with P. gingivalis (Fig. 8C and  D). Similar results were obtained with strain D-62D (not shown). Moreover, supplementation of the conditioned medium from the LuxS mutant with DPD, a chemical precursor of AI-2, restored community development to wild type levels for up to 48 h. The effect was lost at 72 h, presumably as a result DPD exhaustion. Quantitative colocalization analysis of P. gingivalis DluxS and F. alocis heterotypic communities showed a decrease in colocalization compared to parental levels (Fig. 6). Collectively, these results show a requirement for interspecies AI-2-dependent signaling for initial association between F. alocis and P. gingivalis.

Comparative Effects of S. gordonii or F. nucleatum on Community Development with F. alocis
In the mixed species biofilms of the oral cavity F. alocis will likely contemporaneously encounter organisms that are synergistic (such as F. nucleatum) or are antagonistic (such as S. gordonii). To assess the relative contributions of S. gordonii and F. nucleatum, we generated a three species community comprised of S. gordonii, F. nucleatum and F. alocis (Fig. 9). Accumulation of F. alocis in this three-species community was minimal, suggesting that the antagonistic effect of S. gordonii supersedes the synergistic effect of F. nucleatum. The nature of the synergistic effect is unknown; however, it may not depend on a reduction in pH by S. gordonii, as F. alocis was capable of monospecies biofilm formation over a pH range of 5-7 (not shown).

Discussion
Dental plaque is a complex multispecies community that develops temporally and spatially through interbacterial binding and communication systems [27,28]. Mitis group streptococci such as S. gordonii rapidly and avidly attached to saliva-coated tooth surfaces, and these organisms then provide an attachment substratum for later colonizers [3]. Moreover, mitis group streptococci influence the pathogenic potential of later colonizers, a property that has led them to be designated as accessory pathogens in the oral cavity [29]. F. nucleatum is abundant in dental plaque and can provide physiological support for other bacteria including P. gingivalis, as well as stabilize interbacterial coadhesion networks [30,31]. Organisms such as P. gingivalis and A. actinomycetemcomitans are associated with periodontal disease, albeit in the context of raising the pathogenic potential of the microbial community as a whole [1,32]. Recent research has implicated F. alocis as an oral pathogen [17,19]; however, the colonization mechanisms of F. alocis have yet to be studied in detail.
In the present study, the community interactions of F. alocis were investigated. S. gordonii had a strongly antagonistic effect on F. alocis, and colocalization and accretion of F. alocis were low in a community with S. gordonii. These results suggest that streptococcal rich regions of plaque will be resistant to colonization by F. alocis. This is in marked contrast to the interaction between S. gordonii and P. gingivalis, in which S. gordonii provides adhesive and metabolic support for P. gingivalis [23,29], and communities of S. gordonii and P. gingivalis are more virulent in mouse alveolar bone loss models than either organism alone [33]. Conversely, arginine deiminase produced by S. cristatus suppresses fimbrial production by P. gingivalis and impedes colonization of the oral cavity [34,35]. Interbacterial interactions in the oral microbial communities would thus appear to exhibit a high degree of species specificity. Furthermore, while F. nucleatum and F. alocis were synergistic in accumulation into dual species communities, the antagonistic influence of S. gordonii predominated in a three species community. The antagonistic effect of S. gordonii would appear, therefore, to extend beyond failure of S. gordonii to provide coadhesive support to F. alocis. Similarly, host responses to S. gordonii-P. gingivalis heterotypic communities can show a bias toward S. gordonii specific responses. Infection of gingival epithelial cells with S. gordonii and P. gingivalis together resulted in S. gordonii modulating the expression of host genes with a broad diversity of physiological functions, and antagonizing the effect of P. gingivalis at the cellular level [36]. Given that oral streptococci can interact with a wide range of bacteria and yeast [29,37], it is likely that their accessory pathogen role has a major influence on community development and oral health status.
The Mfa1 protein is the structural subunit of the minor fimbriae of P. gingivalis. Mfa1 itself can mediate attachment to  the streptococcal SspA/B protein [22,38] and human monocytederived dendritic cells [39]. However, Mfa1 is thought to impede the process of internalization into epithelial cells, and the Dmfa1 mutant invades epithelial cells more efficiently than the parental strain [40]. Similarly, the presence of the Mfa1 protein is detrimental to community formation with F. alocis. Interestingly, initial association between F. alocis and P. gingivalis was not affected by the loss of Mfa1, rather the accumulation into microcolonies was reduced, indicating that Mfa1 may be involved in the transmission of antagonistic signals between the two organisms. This effect was most pronounced with the type strain, suggestive of heterogeneity of F. alocis responses to P. gingivalis signals.
The LuxS enzyme is an AI-2 synthase which is responsible for the production of the AI-2 family of inter-convertible signaling molecules. AI-2 is required for optimal accumulation of P. gingivalis-S. gordonii communities [26], and also controls mixed biofilm formation by various oral streptococcal species [41] and by Actinomyces oris and S. oralis [42]. In the current study LuxS activity was necessary for maximal association between P. gingivalis and F. alocis. The LuxS enzyme is also a component of the activated methyl cycle (AMC) [43] and is responsible for recycling of S-adenosylhomocysteine (SAH) to homocysteine. Disruption of luxS will therefore lead to both a defect in AI-2 mediated signaling and a potential build up of the toxic AMC intermediate, SAH, either of which could affect P. gingivalis-F. alocis interactions. To distinguish between these possibilities, communities comprised of P. gingivalis DluxS and were chemically complemented with either conditioned medium from the P. gingivalis parental strain or with 4,5-dihydroxy-2,3-pentanedione (DPD) an AI-2 precursor. In both cases the wild type phenotype was restored, indicating that the effect of LuxS on P. gingivalis-F. alocis communities relates to its role in AI-2 signaling. An in silico examination of the currently available F. alocis genomic database did not reveal any obvious luxS homologs in F. alocis and thus F. alocis may not produce AI-2 but may be able to sense and respond to the signal, although further studies to resolve this issue are necessary.
F. alocis is one of only a few organisms that is associated with both generalized and localized aggressive periodontitis (LAP). The consensus pathogen in LAP is A. actinomycetemcomitans, and the type strain of F. alocis displayed mutualistic community growth with A. actinomycetemcomitans. This result is consonant with the recent report that the presence of a consortium of A. actinomycetemcomitans, S. parasanguinis, and F. alocis is indicative of future bone loss in LAP [44]. Interestingly, the more recent clinical isolate D-62D did not show this synergy with A. actinomycetemcomitans. This result, along with other differences between the type strain and D-62D reveals heterogeneity within the F. alocis taxon, particularly with regard to potential involvement in LAP. As few F. alocis isolates have been studied, the existence of subgroups with differing properties are yet to be defined. However, in a proteomic study of F. alocis strains, Aruni et al. [19] found more cell wall anchoring proteins in D-62D compared to ATCC 35896, which may have relevance for interactions with A. actinomycetemcomitans. Future studies involving additional F. alocis strains will be necessary to more fully delineate the interspecies coadhesion profile of the organism. Isolates of A. actinomycetemcomitans from the oral cavity also display heterogeneity with respect to levels of leukotoxin and fimbrial production [45,46], and different strains of A. actinomycetemcomitans therefore could also exhibit different patterns of reactivity with F. alocis.

Conclusions
While the dental plaque biofilm develops on all subgingival tooth surfaces in the oral cavity, periodontal disease is more usually localized to specific sites. Thus, spatial variations in the pathogenic potential of the biofilm communities exist. Complex synergistic and antagonistic interactions occur within oral microbial communities and these underlie the success or failure of microbial colonization. The results of this work indicate that the pattern of colonization of F. alocis depends heavily on the antecedent inhabitants of the microbial community. Although dental biofilms can comprise several hundred bacterial species, by practical necessity in vitro studies such as these are limited in the number of organisms and strains that can be investigated, and we recognize that the presence of other bacterial species could modulate the interactions reported herein. Nonetheless, the network of interactions established for F. alocis provides mechanistic insights into the colonization strategies of the organism and form a framework for future studies to define the molecular basis of F. alocis colonization and community formation.

Author Contributions
Conceived and designed the experiments: QW HD SU RL. Performed the experiments: QW CW. Analyzed the data: QW CW. Wrote the paper: QW CW HD SU RL. Figure 8. Role of P. gingivalis LuxS in dual-species community formation with F. alocis. A. P. gingivalis ATCC 33277 (WT), and DluxS (1 6 10 8 , blue, stained with DAPI) were cultured on glass coverslips. F. alocis strains ATCC 35896 and D-62D were stained with FITC (green) and reacted with the P. gingivalis strains for 72 h. B. Biovolume of P. gingivalis or F. alocis in dual species communities at 72 h. Data are representative of four independent replicates. P-value compared with control single species communities was calculated by t-test, and significant differences are p,0.01(**). C. Accumulation of F. alocis ATCC 35896 stained with FITC (green) and cultured in TSB, conditioned medium (CM) from P. gingivalis WT, CM from P. gingivalis DluxS, or CM from P. gingivalis DluxS with 4 mM DPD. D. Biovolume of F. alocis ATCC 35896 cultured in TSB, conditioned medium (CM) from P. gingivalis WT, CM from P. gingivalis DluxS, or CM from P. gingivalis DluxS with 4 mM DPD. Data are representative of four independent replicates. P-value compared with control single species communities was calculated by ttest, and significant differences are at p,0.05 (*) or p,0.01(**). doi:10.1371/journal.pone.0076271.g008 Figure 9. Three-species community formation with F. alocis, S. gordonii and F. nucleatum analyzed by CLSM. A. S. gordonii DL1 (red, stained with hexidium iodide), F. nucleatum (blue, stained with DAPI) were co-cultured on glass coverslips. F. alocis strains ATCC 35896 and D-62D were stained with FITC (green) and reacted with S. gordonii and F. nucleatum for 72 h. B. Biovolume of F. alocis ATCC 35896 and D-62D, S. gordonii DL1 and F. nucleatum ATCC 25586 in three species communities. Data are representative of four independent replicates. Pvalue compared with control single species communities was calculated by t-test, and significant differences are p,0.01(**). doi:10.1371/journal.pone.0076271.g009