Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
4,5-dihydroxy-2,3-pentanedione (DPD), a product of the LuxS enzyme in the catabolism of S-ribosylhomocysteine, spontaneously cyclizes to form autoinducer 2 (AI-2). AI-2 is proposed to be a universal signal molecule mediating interspecies communication among bacteria. We show that mutualistic and abundant biofilm growth in flowing saliva of two human oral commensal bacteria, Actinomyces naeslundii T14V and Streptococcus oralis 34, is dependent upon production of AI-2 by S. oralis 34. A luxS mutant of S. oralis 34 was constructed which did not produce AI-2. Unlike wild-type dual-species biofilms, A. naeslundii T14V and an S. oralis 34 luxS mutant did not exhibit mutualism and generated only sparse biofilms which contained a 10-fold lower biomass of each species. Restoration of AI-2 levels by genetic or chemical (synthetic AI-2 in the form of DPD) complementation re-established the mutualistic growth and high biomass characteristic for the wild-type dual-species biofilm. Furthermore, an optimal concentration of DPD was determined, above and below which biofilm formation was suppressed. The optimal concentration was 100-fold lower than the detection limit of the currently accepted AI-2 assay. Thus, AI-2 acts as an interspecies signal and its concentration is critical for mutualism between two species of oral bacteria grown under conditions that are representative of the human oral cavity.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Dental plaque, one of the best characterized natural multispecies biofilms, harbours hundreds of species of bacteria (Kroes et al., 1999; Paster et al., 2001) that live in close association with one another (Listgarten, 1976). Plaque development occurs through sequential colonization of tooth surfaces (Nyvad and Kilian, 1987), and many of the component bacteria are proposed to communicate with one another through small signalling molecules (Kolenbrander et al., 2002). Community development within the biofilm is thought to be facilitated, in part, by specific cell-cell recognition called coaggregation (Cisar et al., 1979; Kolenbrander, 1997) (specific adherence of genetically distinct partner cells which bind to one another to form multicellular networks), and similar cell-cell recognition processes are now being documented in other biofilm systems (Rickard et al., 2003). Oral actinomyces and streptococci coaggregate and are among the first bacteria to colonize a freshly cleaned enamel surface and act as a foundation for other bacterial species to grow upon (Nyvad and Kilian, 1987). Indeed, Diaz et al. (Diaz et al., 2006) have recently demonstrated that greater than 65% of the bacteria present within nascent (< 4 h old) oral biofilms are composed of streptococci and actinomyces. Palmer et al. (Palmer et al., 2001) showed that, when grown together, Streptococcus oralis 34 and Actinomyces naeslundii T14V generate a mutualistic interaction characterized by ‘luxuriant interdigitated growth’ on saliva as the sole nutrient source. While each species attaches to a saliva-conditioned flowcell surface, neither organism exhibits growth as a single-species biofilm. Importantly, the signal required for mutualism between this pair was not identified.
Recently, two structurally similar signalling molecules have been cocrystallized with their respective receptors from the human pathogenic bacterium Salmonella typhimurium and the bioluminescent marine bacterium Vibrio harveyi (Chen et al., 2002; Miller et al., 2004). Both signal molecules have been described as autoinducer 2 (AI-2), an umbrella designation given to a collection of molecules formed from spontaneous rearrangement of 4,5-dihydroxy-2,3-pentanedione (DPD) (Duerre et al., 1971; Semmelhack et al., 2005). In solution, the various forms of AI-2 are in equilibrium (Semmelhack et al., 2005); a given bacterial species could recognize a particular molecule within this AI-2 pool. For example, the two crystallized receptor-AI-2 complexes from S. typhimurium and V. harveyi are distinct in the receptor and in the bound AI-2 derivative (Chen et al., 2002; Miller et al., 2004). The two forms of AI-2 are interconvertible; they can be released from one receptor, enter the AI-2 equilibrium and bind to the other receptor (Miller et al., 2004). Furthermore, Xavier and Bassler (Xavier and Bassler, 2005) demonstrated two-way interconvertible AI-2 communication between Escherichia coli and Vibrio cholerae in batch culture on rich laboratory medium. The data from these three studies support the hypothesized role of AI-2 as an interspecies signal that can interact with structurally and functionally distinct receptors of different bacterial species (Schauder et al., 2001), but a convincing demonstration of interspecies signalling under natural conditions has yet to be provided.
Based on the V. harveyi bioluminescence assay (Surette and Bassler, 1998), AI-2 production has been reported in more than 50 gram-positive and gram-negative bacteria (Federle and Bassler, 2003) including human oral bacteria (Frias et al., 2001; Kolenbrander et al., 2002; Blehert et al., 2003; McNab et al., 2003). Oral streptococci and actinomyces coaggregate in vitro (Cisar et al., 1979; Kolenbrander, 1997) and interact physically in nature (Palmer et al., 2003). We hypothesized that AI-2 could be the interspecies signal required for mutualism between the coaggregating oral bacteria S. oralis 34 and A. naeslundii T14V and, furthermore, that AI-2 levels are critical to the dual-species phenotype of mutualistic interdigitated biofilm growth. We show here that a dual-species biofilm containing an S. oralis 34 luxS mutant and A. naeslundii T14V does not show mutualistic interdigitated growth in saliva, that the genetically complemented luxS mutant does participate in mutualistic growth, and, significantly, that synthetic DPD added at the appropriate concentration restores the mutualistic growth in the luxS mutant/actinomyces biofilm. Thus, a bona fide role for AI-2 as a concentration-dependent, interspecies signal has been identified in a naturally occurring bacterial partnership grown under conditions relevant to nature.
Results and discussion
Identification of luxS and detection of AI-2 in laboratory cultures
By comparative sequence analysis to known luxS genes in other bacterial species, a homologue of luxS was identified in S. oralis 34. A homologue was predicted in A. naeslundii T14V based on the genome sequence of A. naeslundii MG1. AI-2 activity in cell-free supernatants from batch cultures of S. oralis 34 strains and of A. naeslundii T14V was measured (Surette and Bassler, 1998) in a V. harveyi bioluminescence induction assay (Fig. 1). Maximum activity occurred in late exponential phase. S. oralis 34 cell-free culture supernatants exhibited twice the activity of those from A. naeslundii T14V. An S. oralis 34 luxS mutant did not synthesize AI-2, but this activity was restored when the gene was reintroduced on a plasmid. Additionally, in batch culture, the mutant and the complemented strain grew and coaggregated with A. naeslundii T14V like the wild-type S. oralis 34 (data not shown).
Growth of mono- and dual-species biofilms in saliva
As reported by Palmer et al. (2001), wild-type A. naeslundii T14V and S. oralis 34 exhibited minimal mono-species biofilm growth in saliva-fed flowcells (Fig. 2A). Furthermore, deletion of the luxS gene, or genetic complementation via transformation with a plamid-borne luxS, did not alter the minimal ability of S. oralis 34 to grow as mono-species biofilms (Fig. 2A). Quantitative measurement of the mono-species biofilms, over a 22 h period after inoculation, showed that each mono-species biofilm initially grew and then underwent cessation (Fig. 2B). The mono-species biofilm biovolume increased over the first 5 h following inoculation, and was probably due to the cells being harvested from exponential phase batch-cultures. Following this period of biofilm growth in unsupplemented saliva, the biovolume occupied by the mono-species biofilm decreased to a level that was less after 22 h than at either 1 h or 5 h. Thus, the mono-species biofilms never developed and could not sustain a volume greater than the initial inoculum. This absence of sustained mono-species biofilm growth was likely due to the cells dividing at a rate that is less than that needed to prevent being lost from the biofilm.
Dual-species biofilms containing the S. oralis 34 luxS mutant and A. naeslundii T14V were sparse after 22 h (Fig. 3B); similar to mono-species biofilms. However, dual-species biofilms containing A. naeslundii T14V and either the S. oralis 34 wild type or the S. oralis 34 complemented luxS mutant always yielded abundant interdigitated mutualistic growth (Fig. 3A and C). These results suggest that the product of the luxS gene in S. oralis 34 generates AI-2, which is involved in mutualistic growth between these bacteria in dual-species biofilms.
Time-resolved inspection of dual-species biofilm development revealed that each wild type exhibited an increase in total biovolume (Fig. 4A) accompanied by mutualistic interdigitated growth (Fig. 3A). Likewise, each partner in the S. oralis 34 luxS complemented mutant –A. naeslundii T14V pair increased in biovolume (Fig. 4C) and together exhibited mutualistic interdigitated growth (Fig. 3C). In contrast, although the S. oralis 34 luxS mutant –A. naeslundii T14V pair was viable initially, as shown by an increase in biovolume between 1 h and 5 h, this pair could not sustain mutualistic growth (Fig. 4B). By 22 h, biovolume had decreased to a level at or below the initial 1 h biovolume (Fig. 4B), as was exhibited by all mono-species biofilms (Fig. 2B).
Images demonstrating the temporal pattern (1 h, 5 h and 22 h) of biofilm development for each streptococcus-actinomyces pair are shown in Fig. 5. In all pairs, each species independently adheres to the saliva-coated surface and, at 1 h, several adherent cells are in coaggregates (Fig. 5A–C, 1 h). At 5 h, for each pair, coaggregations of interdigitated species are evident, indicating that the cells are viable and interacting (Fig. 5A–C, 5 h). The wild-type pair at 22 h continues to grow, and it exhibits mutualistic interdigitated growth (Fig. 5A, 22 h). The biofilm consisting of the S. oralis 34 luxS mutant –A. naeslundii T14V pair, in contrast, has lost most of its cells at 22 h (Fig. 5B, 22h), indicating an inability to sustain biofilm growth. However, mutualistic interdigitated growth in saliva is restored by complementation of S. oralis 34 luxS mutant with the cloned luxS gene (Fig. 5C, 22h).
Chemical complementation of S. oralis 34 luxS mutant in dual species biofilms
To verify that the luxS complementation result was due to interspecies signalling with AI-2, chemically synthesized DPD was added to dual-species biofilms containing the S. oralis 34 luxS mutant and A. naeslundii T14V. Ten-fold incremental concentrations of synthetic DPD were added to saliva reservoirs that fed developing biofilms. In S. oralis 34 luxS mutant-A. naeslundii T14V biofilms, mutualistic interdigitated growth and biovolume were dependent upon DPD concentration (Fig. 6). DPD at 0.8 nM re-established mutualistic interdigitated growth and resulted in the greatest total biofilm biovolume (approximately 2 × 106 µm3), and this DPD concentration had no effect on single-species biofilms (Fig. 7, cf. Fig. 2B). A 10-fold lower DPD concentration yielded less dual-species biovolume (Fig. 6). The differences are significant but small and suggest that the precise optimal concentration of DPD lies between 0.08 nM and 0.8 nM. DPD at 8.0 nM yielded a reduction in biovolume, and at 800 nM DPD, the biofilm resembled one to which no DPD had been added. A direct 10-fold relationship between biovolume and DPD concentration was not apparent. None of these concentrations (0.08–800 nM DPD) inhibited growth of the streptococci or actinomyces in batch culture. Significantly, in this natural two-species system, the optimal DPD concentration (0.8 nM) is 100-fold lower than the detection limit (Semmelhack et al., 2005) of the AI-2 bioassay (Fig. 8), indicating that only a very low concentration of AI-2 is required for these organisms to conduct AI-2-signalled interspecies communication under natural conditions. In context, attached-biofilm cells that are in intimate contact through cell–cell interactions, such as coaggregation, will be exposed to summed concentrations of AI-2. Thus, in a flowing environment, close proximity may be essential for AI-2 mediated mutualism.
Using the same saliva-based flow system employed in the present report, we have previously shown that a different species, Streptococcus gordonii, upregulated amylase expression in response to a diffusible signal only when intimately associated with Veillonella atypica (Egland et al., 2004). Nearby streptococcal cells not intimately associated with veillonellae did not respond to the signal. These results suggested that the signal was diluted below the effective concentration in the flowing bulk saliva. In our current experiments, the S. oralis 34 luxS mutant-A. naeslundii T14V biofilm did not develop despite the capability of A. naeslundii T14V to produce AI-2. One possible explanation for this result is the presence of inhibitors of AI-2 signalling that effectively compete only when signal concentration is low due to inability to produce AI-2 by the streptococci. A second possible explanation, which we favour, is that in a flowing environment the bioactive threshold concentration for mutualism could not be reached through production of AI-2 by A. naeslundii T14V alone. Thus, a suitable second species that coordinates the optimal production of AI-2 may be required for mutualism of the two species.
A role for luxS in mixed-species accumulations in oral biofilms has been suggested for S. gordonii and Porphyromonas gingivalis (McNab et al., 2003). In a non-growing system, significant biofilm accumulation occurred except when both strains lacked a functional luxS. Another report showed that single-species biofilm formation by an S. typhimurium luxS mutant could be rescued by genetic complementation with luxS under its endogenous promoter (De Keersmaecker et al., 2005). However, addition of synthetic DPD concentrations ranging from 100 nM to 72 µM did not restore biofilm formation to the mutant; perhaps, significantly lower concentrations (pM) could have restored biofilm formation.
In multispecies communities, AI-2 signalling by community members will have reciprocal effects on gene expression by the juxtapositional species. While studies of AI-2 interspecies signalling in natural settings are still in their infancy, Xavier and Bassler (Xavier and Bassler, 2005) clearly showed that AI-2 can mediate two-way communication between E. coli and V. harveyi, when grown in batch coculture; AI-2 produced by either species regulated bioluminescence in V. harveyi and induced lsr expression in E. coli. Considering this result, regulation of gene expression by AI-2 is a likely mechanism operative for mutualism in the oral bacteria studied here. luxS mutants of several oral bacteria are altered in gene expression controlling a variety of metabolic functions including iron transport systems by Actinobacillus actinomycetemcomitans (Fong et al., 2003), haemin acquisition in P. gingivalis (Chung et al., 2001), mutacin I and glucosyltransferase of Streptococcus mutans (Merritt et al., 2005; Yoshida et al., 2005), and glucosyltransferase of S. gordonii (McNab et al., 2003). All of these reports speculate that AI-2 is the signal molecule responsible for interspecies communication. We report here the first proof of this hypothesis using the pure, chemically synthesized DPD. We propose that increased localized concentration of AI-2 in the biofilm mixed-species communities alters gene expression of both species and ultimately results in mutualism.
In nature, bacterial biofilms are typically complex multispecies communities which develop through the adhesion of different species of bacteria to a surface followed by interspecies cell–cell interactions promoting the abundant growth of the community. We have shown here a clear, ecologically relevant role for AI-2 in interspecies signalling; a role that was revealed through experiments that mimic in situ conditions (in biofilms with flowing saliva as the sole nutritional source) and through the use of organisms known to interact in nature. We showed that the mutualistic interdigitated growth characteristic for a pair of human commensal bacteria is dependent on AI-2. Coaggregation is an intrinsic property of oral bacteria (Kolenbrander, 1997), and we propose that coaggregation and AI-2 signalling have synergistic effects in the production of protective commensal communities. In the absence of the concentration of AI-2 required for optimal commensal community growth, opportunistic bacterial pathogens might infiltrate the community more easily. Furthermore, it has been reported that oral periopathogenic bacteria produce higher levels of AI-2 than do commensal actinomyces and streptococci (Frias et al., 2001); once established, oral pathogens might generate levels of signal so high as to be detrimental to optimal growth of commensal communities. Thus, a role for AI-2 in modulating microbial community composition and dynamics is conceivable, and high AI-2 concentrations may be associated with disease-causing communities.
The present work leads to a model for the development of dual species biofilms of S. oralis 34 and A. naeslundii T14V formed under flowing conditions (Fig. 9). The model illustrates the role of AI-2 concentration in mediating mutualism, which is expressed as abundant interdigitated growth. A distinction between global and local AI-2 concentration is necessary. Global concentration refers to a concentration of AI-2 supplied in the saliva reservoir and pumped through the flowcell, whereas local AI-2 concentration within a mixed-species coaggregate is a characteristic dependent upon the site and amount of AI-2 production (cells) and the fluid flow rate. Consequently, if the AI-2 supply source is local, then its production level within the coaggregates must be maintained or washout of AI-2 will occur. Thus, intimate contact, through coaggregation, between AI-2 producing S. oralis 34 and A. naeslundii T14V is required to generate a unique local AI-2 concentration. When wild-type S. oralis 34 is partnered with A. naeslundii T14V (Fig. 9, left panel) the local concentration of AI-2 is higher than when A. naeslundii T14V is partnered with S. oralis 34 luxS mutant (Fig. 9, middle panel). The higher AI-2 concentration is crucial for cells to elicit a response that ultimately leads to mutualism. The global addition of AI-2, in the form of chemically synthesized DPD, complements the AI-2 deficiency between the S. oralis 34 luxS mutant and A. naeslundii T14V (Fig. 9, right panel). Externally added DPD surrounds the developing biofilm, and when this global concentration is optimal (Fig. 9, right panel), it eliminates the restriction for an optimal local concentration (Fig. 9, left panel). The exquisite sensitivity of the dual–species oral biofilm interaction reported here implies that, in the oral cavity, bacterial cell-cell communication might occur at signal molecule concentrations well below those typically investigated or even detected in the laboratory. Furthermore, the signal concentrations may be key to understanding the importance of streptococci and actinomyces in the naissance of dental plaque communities.
Strains and culture conditions
Batch cultures of S. oralis 34 and A. naeslundii T14V were grown anaerobically at 37°C in CAMG (Cisar et al., 1979) or THB broth (Difco, Detroit, MI) supplemented with 1 mM boric acid (THB-B). V. harveyi BB170 and BB152 were grown aerobically in AB medium (Bassler et al., 1994). E. coli was cultivated in Luria–Bertani broth aerobically at 37°C. Where needed, antibiotics were added: erythromycin, 10 µg ml−1 (S. oralis 34) or 300 µg ml−1 (E. coli); kanamycin, 1000 µg ml−1 (S. oralis 34) or 50 µg ml−1 (E. coli).
Analysis of AI-2 production
AI-2 activity was determined in THB-B batch-culture cell-free supernatants using a modification of the bioluminescence assay (Surette and Bassler, 1998). Cells of V. harveyi BB170 were grown at 30°C in batch cultures of autoinducer bioassay (AB) medium according to the method of (Surette and Bassler, 1998). After 14 h growth, V. harveyi BB170 cultures were diluted 1:500 in fresh AB medium. The diluted BB170 culture was stored at −70°C. For analysis of the amount of AI-2 in streptococcal or actinomyces culture supernatants, the supernatants were passed through a 0.2 µm filter unit (Fisher Scientific, Suwanee, GA) and 10 µl of cell-free supernatant was added to 90 µl of thawed diluted BB170 culture in a 96 well plate. Bioluminescence relative to uninoculated medium was calculated as fold induction (Blehert et al., 2003).
Identification and sequence determination of S. oralis 34 luxS
Conserved regions within luxS were identified by alignment of luxS sequences from other Gram-positive bacteria. Primers (5′-GATCACACCATCGTTAAGG-3′ and 5′-CCAAAA GGGGAGCAATCAA-3′) complementary to conserved regions were used to amplify a 206 nucleotide fragment of luxS from S. oralis genomic DNA. The fragment was sequenced and primers were designed (5′-CGAACGAA GACTCAATCCC-3′ and 5′-GTTGCACCAAGCGAATATC-3′) to amplify the whole gene by inverse polymerase chain reaction (PCR). The entire sequence of luxS (483 bp) was determined (GenBank Accession number DQ207739).
luxS mutant construction
An isogenic S. oralis 34 luxS mutant was constructed by inserting the erythromycin resistance gene ermAM (Macrina et al., 1980) into luxS. PCR primers (5′-GGAATTCAGTCAC CTTAGTTCCACTAG-3′ and 5′-CCCAAGCTTGTGGTTGAA CAACAAATTCG-3′) were used to amplify luxS and flanking DNA. The product was cloned into EcoRI and HindIII sites of pDL278 (Blehert et al., 2003), to yield pDL278-So34luxS. PCR was used to amplify ermAM from pVA736 (Blehert et al., 2003). The PCR product was ligated into a unique MscI site located between nucleotides 188–193 of the luxS open reading frame, yielding pDL278-So34luxS::ermAM, and transformed into E. coli XL10 Gold (Stratagene, La Jolla, CA) with selection for erythromycin-resistance. The luxS::ermAM fragment was excised from pDL278-So34luxS::ermAM using EcoRI and HindIII, gel purified, and transformed (Haisman and Jenkinson, 1991) into S. oralis 34 to generate the S. oralis 34 luxS mutant. No genetic system is available to obtain a luxS mutant of A. naeslundii T14V.
Genetic complementation of the luxS mutant
A plasmid containing an undisrupted copy of luxS was constructed by exchanging the erythromycin resistance cassette of pTRKL2 (O’Sullivan and Klaenhammer, 1993) with aphAIII, encoding kanamycin resistance. The amplified S. oralis luxS gene was cloned into the EcoRI site of the modified plasmid and transformed into E. coli XL10 Gold with selection for kanamycin resistance. S. oralis 34 luxS was transformed (Haisman and Jenkinson, 1991) to give S. oralis 34 comp. Complementation was confirmed in three ways: isolation and restriction digestion of the plasmid, PCR amplification of luxS and production of AI-2.
Chemical complementation of S. oralis 34 luxS mutant
The effect of synthetic DPD on the growth of dual-species biofilms was achieved by adding fresh (less than 2 weeks old) DPD (Semmelhack et al., 2005) to 25% saliva, which was subsequently used for cultivation of biofilms.
Cultivation of biofilms
Actinomyces naeslundii T14V, S. oralis 34, the S. oralis 34 luxS mutant and S. oralis 34 comp were each grown to mid-exponential-phase in CAMG medium, washed three times in 25% sterile saliva (Palmer et al., 2001), and adjusted to 5 × 108 cells ml−1 (50 Klett units, 660 nm). Single-species biofilms were grown by inoculation into reusable flowcells (Foster and Kolenbrander, 2004) conditioned for 30 min with 25% saliva flowing at 100 µl min−1. After 20 min of adherence under static conditions, saliva or DPD-supplemented saliva was pumped (100 µl min−1) through the flowcell. Dual-species biofilms were grown by first inoculating the streptococcal strain. Following the adherence period and a subsequent wash (20 min flow with 25% saliva), A. naeslundii T14V cells were injected, and flow was re-initiated after 20 min of static conditions. All flowcells were incubated at 37°C.
Confocal microscopy and image analysis of biofilms
Biofilms were labelled for microscopy by injecting 200 µl of PBS (pH 7.2) supplemented with 1% BSA and 10 µg ml−1 Alexafluor™ (Invitrogen, Carlsbad, Ca.)-conjugated antibodies. Anti-S. oralis 34 antibody (Palmer et al., 2001) was conjugated to Alexa-488 and anti-A. naeslundii T14V antibody (Cisar et al., 1978) was conjugated to Alexa-633 according to the manufacturer’s instructions. A TCS-SP2 confocal microscope (Leica, Exton, PA) with a 40×, 1.25 NA oil immersion lens was used to record confocal image stacks in at least three random locations near the centre of the flowcells, after which biofilm biovolumes were determined by volumetric analyses (IMARIS Ver. 3.3, Bitplane AG, Zurich, Switzerland). Fluorescence intensity thresholds were set manually for red and green pixels, and a voxel size of 1.0 µm3 was used. Three confocal data sets were analysed for each experimental group, and the mean and standard deviation were calculated. All images presented are maximum projections of the entire confocal image stack as produced by the Leica TCS software (Leica, Exton, PA).
We thank B. Swaim, N. Jakubovics and B. Duval for experimental assistance on this project and S. Gill of TIGR for supplying the luxS sequence of A. naeslundii MG1. This research was supported in part by the Intramural Research Program of the NIH, NIDCR.