Autoinducer 2 Is Required for Biofilm Growth of Aggregatibacter (Actinobacillus) actinomycetemcomitans

Abstract

Autoinducer 2 (AI-2) is required for the growth of Aggregatibacter (Actinobacillus) actinomycetemcomitans in culture under conditions of iron limitation. However, in vivo this organism thrives in a complex multispecies biofilm that forms in the human oral cavity. In this report, we show that adherent growth of A. actinomycetemcomitans on a saliva-coated surface, but not planktonic growth under iron-replete conditions, is defective in a LuxS-deficient background. Biofilm growth of the luxS mutant exhibited lower total biomass and lower biofilm depth than those for the wild-type strain. Normal biofilm growth of the luxS mutant was restored genetically by introduction of a functional copy of luxS and biochemically by addition of partially purified AI-2. Furthermore, introduction of S-adenosylhomocysteine hydrolase, which restores the metabolism of S-adenosylmethionine in the absence of LuxS, into A. actinomycetemcomitans did not complement the luxS mutation unless AI-2 was added in trans. This suggests that AI-2 itself is required for biofilm growth by A. actinomycetemcomitans. A biofilm growth deficiency similar to that of the LuxS-deficient strain was also observed when a gene encoding the AI-2-interacting protein RbsB or LsrB was inactivated. Biofilm formation by A. actinomycetemcomitans was virtually eliminated upon inactivation of both rbsB and lsrB. In addition, biofilm growth by wild-type A. actinomycetemcomitans was reduced in the presence of ribose, which competes with AI-2 for binding to RbsB. These results suggest that RbsB and LsrB function as AI-2 receptors in A. actinomycetemcomitans and that the development of A. actinomycetemcomitans biofilms requires AI-2.

Quorum sensing is a mechanism that allows a bacterial population to monitor and respond to its cell density through the action of secreted soluble signal molecules, or autoinducers. The chemical structure of quorum-sensing signals is diverse and includes acylated homoserine lactones (19, 36), small peptides (10), quinolones (37), or furan derivatives (54). As the microbial population expands and the concentration of an autoinducer increases accordingly in the local environment, a critical threshold is reached that triggers changes in gene expression. This process may allow organisms to synchronize the expression of genes that may be required or beneficial for community survival (1, 16, 52).

Autoinducer 2 (AI-2) is a furan-like quorum-sensing signal that was initially identified in Vibrio harveyi (3), and its synthesis requires the luxS gene (39). The enzyme encoded by luxS is involved in the metabolism of S-adenosylmethionine (SAM) (53) and cleaves S-ribosylhomocysteine to produce homocysteine and 4,5-dihydroxy-2,3-pentanedione (DPD) (53), which undergoes intramolecular rearrangement to produce AI-2. Two structural forms of AI-2 have been identified to date: Salmonella enterica serovar Typhimurium produces 2R,4S-2,3,3,4-methyltetrahydroxytetrahydrofuran (R-THMF), whereas V. harveyi produces the borate diester form of S-THMF (7, 31).

The luxS gene is highly conserved in a wide range of gram-positive and gram-negative bacteria, and many, if not all, of these organisms produce an AI-2-like signal that is capable of inducing the expression of the lux operon of Vibrio harveyi. As a result, AI-2 has been suggested to represent a widely distributed quorum-sensing signal that is recognized by many organisms (39). Inactivation of luxS has been reported to influence a wide variety of cellular processes, including type III secretion (43), cell motility (26, 42), the development of biofilms (4, 8, 17, 18, 29, 30, 56), the expression of virulence factors (12, 22, 28), and iron uptake (11). The signal transduction mechanism that governs the cellular response to AI-2 has been most thoroughly characterized in Vibrio spp. In the genus Vibrio, the detection of AI-2 requires LuxP, a periplasmic AI-2 receptor (7) that associates with the LuxQ sensor kinase-phosphatase (32) and initiates a phosphotransfer cascade involving LuxU (14) and the response regulator LuxO (13). LuxO controls the expression of multiple small regulatory RNAs (quorum-regulatory RNAs [Qrr]) that, in conjunction with the chaperone Hfq, influence the expression of LuxR (24, 25), the master regulator of the lux operon (25). Recently, an additional two-component system, encoded by varSA, in Vibrio cholerae has been shown to converge on the quorum-sensing circuit by regulating the expression of the CsrBCD small RNAs, which control the expression of Qrr via CsrA (24). However, many organisms that possess luxS lack the dedicated two-component response circuit that exists in Vibrio spp. This raises the possibility that AI-2 may not function as a quorum-sensing signal in all organisms (9, 53).

Our studies have focused on the role and function of AI-2 in the oral pathogen Aggregatibacter actinomycetemcomitans, an organism that is associated with aggressive forms of periodontitis and other systemic infections (5, 33, 41, 57). This organism produces an AI-2-like signal that induces V. harveyi bioluminescence (12) and regulates the growth of A. actinomycetemcomitans under iron limitation conditions by controlling the expression of a variety of iron storage and uptake genes (11). However, A. actinomycetemcomitans thrives in vivo as part of a complex multispecies biofilm that forms in the human oral cavity (23, 51). In this report, we show that the adherent growth of A. actinomycetemcomitans on a saliva-coated surface is defective in a LuxS-deficient background. The luxS mutation was complemented genetically by the introduction of a functional luxS gene and biochemically by the addition of partially purified AI-2. Introduction of S-adenosylhomocysteine hydrolase to A. actinomycetemcomitans, which restores the metabolism of SAM in the absence of LuxS, did not complement the luxS mutation unless AI-2 was added in trans. The biofilm growth defect was also observed in RbsB- and LsrB-deficient backgrounds, as well as with wild-type A. actinomycetemcomitans in the presence of ribose, which competes with AI-2 for binding to RbsB (21). Together, these results suggest that AI-2 itself is required for A. actinomycetemcomitans biofilm growth and that RbsB and LsrB function as AI-2 receptors in A. actinomycetemcomitans.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

A. actinomycetemcomitans strains were cultured under microaerophilic conditions at 37°C in brain heart infusion (BHI; Difco) medium supplemented with 40 mg of NaHCO₃ per liter. The LuxS-deficient (12) and RbsB-deficient (21) strains of A. actinomycetemcomitans were grown as described above but in medium supplemented with 25 μg per ml kanamycin. An A. actinomycetemcomitans lsrB mutant (40) was grown in BHI containing 25 μg per ml spectinomycin, and the A. actinomycetemcomitans lsrB rbsB double mutant (40) was cultured in BHI with 25 μg per ml kanamycin and 25 μg per ml spectinomycin. Growth of the mutant strains in broth culture was compared to that of wild-type A. actinomycetemcomitans JP2 by diluting overnight cultures of each 1:20 into fresh medium without antibiotics and monitoring the optical density at 600 nm over 24 h. The specific growth rate for each strain during exponential growth was determined for two independent cultures by calculating the change in cell count versus time using a standard curve relating A. actinomycetemcomitans cell density and turbidity at 600 nm. Escherichia coli strains were grown in Luria-Bertani (LB) medium (1% tryptone, 0.5% yeast extract, 0.5% NaCl) with aeration at 37°C. E. coli strains containing plasmid pGEM-T, pYGK, or pYGS were cultured as described above in LB supplemented with 100 μg per ml ampicillin, 25 μg per ml kanamycin, or 25 μg per ml streptomycin, respectively. V. harveyi BB170 (sensor 1⁻ sensor 2⁺) was a gift from B. Bassler (Princeton University) and was grown overnight in AB medium (2) with aeration at 30°C. AB medium consists of 0.3 M NaCl, 50 mM MgSO₄, 0.2% Casamino Acids, 10 mM potassium phosphate (pH 7.0), 1 mM l-arginine, 2% glycerol, 1 μg per ml thiamine, and 10 ng per ml riboflavin.

Biofilm growth and analysis.

A. actinomycetemcomitans biofilms were grown on a saliva-coated coverglass in a polycarbonate flow chamber (model FC81; Biosurface Technologies Corp., Bozeman, MT). Chamber dimensions are 50.5 mm by 12.7 mm by 2.54 mm) at 25°C. Saliva was collected from a single donor, filter sterilized (pore size, 0.22 μm), and incubated with the coverglass (60 mm by 24 mm) for 20 min at 37°C. The saliva-coated coverglass was then washed with phosphate-buffered saline (PBS) (100 mM NaH₂PO₄, 150 mM NaCl) for 10 min at a flow rate of 60 ml per h using a peristaltic pump. The flow rate for all subsequent steps of biofilm growth was 5.8 ml per h, with one exception (see below). Multiple biofilms were generated simultaneously using a multichannel peristaltic pump (ManoStat Sarah cassette; Fisher Scientific Co.). The flow cells were inoculated for 1 h with an overnight culture of A. actinomycetemcomitans that was suspended in PBS to an optical density of 0.5 at 600 nm and then washed with PBS for 30 min. Bound cells were fed with BHI medium for as long as 60 h. The resulting biofilm was stained with 0.2 mg/ml fluorescein isothiocyanate (Sigma, St. Louis, MO) in the dark for 1 h and washed with PBS for 2 h. Biofilms were visualized by using an Olympus Fluoview FV500 confocal scanning laser microscope (Olympus, Pittsburgh, PA) under ×600 magnification using an argon laser. Confocal images were captured from 10 to 20 randomly chosen frames from each flow chamber. Biofilm depth was determined by performing z-plane scans from 0 to 100 μm above the glass surface, and total biomass was determined by integrating fluorescence intensity across the z-stack using the Fluoview FV500 software provided with the confocal microscope. Biofilm depth and total biomass were also determined using the COMSTAT image-processing software package (20). For COMSTAT analysis, confocal images were converted to grayscale with PhotoShop CS and threshold values were set using the LOOKTIF function of COMSTAT. Biomass data were analyzed using a pairwise t test (Graphpad Software, Inc.) and expressed as the mean ± standard deviation calculated from all of the frames obtained for a given biofilm. For some experiments, the BHI medium was supplemented with partially purified AI-2 obtained from A. actinomycetemcomitans. The amount of AI-2 used for these experiments was normalized based on the response of the reporter strain V. harveyi BB170. The partially purified AI-2 sample (see below) was suspended in 1 ml of PBS, and serial dilutions were assayed for induction of the bioluminescence of V. harveyi BB170 as described below. The highest dilution that induced a 400-fold increase in bioluminescence after 6 h was chosen. In a typical experiment, filter-sterilized, partially purified AI-2 was diluted 1:1,000 in the BHI growth medium. In other experiments, the BHI growth medium was supplemented with ribose (0 to 100 mM). Biofilms that were grown under high shear force were fed growth medium at a flow rate of 35 ml per h.

Partial purification of A. actinomycetemcomitans AI-2.

A fraction enriched for A. actinomycetemcomitans AI-2 was produced essentially as described by Sperandio et al. (45). Briefly, cells from an overnight culture of A. actinomycetemcomitans were diluted 1:20 into fresh medium and cultured at 37°C to mid-exponential phase (optical density, 0.3). Cells were harvested by centrifugation, and AI-2 was obtained from a 7.2-ml aliquot of the conditioned medium. The culture supernatant was filtered through a 0.22-μm-pore-size filter and subsequently through a Centricon YM-3 3-kDa exclusion filter (Millipore). The filtrate was then lyophilized, suspended in 1 ml of cold 5 mM sodium phosphate buffer, pH 6.2, and chromatographed on a C₁₈ Sep-Pak reverse-phase column (Waters Co.) according to the manufacturer's instructions. AI-2 activity in the column fractions was followed by monitoring the induction of bioluminescence of Vibrio harveyi BB170 (sensor 1⁻ sensor 2⁺). Active fractions were lyophilized and stored at 4°C.

Determination of V. harveyi bioluminescence.

AI-2-mediated induction of V. harveyi bioluminescence was determined as follows. An aliquot of an overnight V. harveyi BB170 culture was diluted 1:25,000 into fresh sterile AB medium, and 90 μl of the diluted cells was added to each well of a microtiter plate. Positive-control wells then received 10 μl of V. harveyi AI-2, and negative-control wells received 10 μl sterile AB medium. AI-2 from V. harveyi was obtained from conditioned medium of overnight V. harveyi BB170 cultures grown in AB medium. After the cells were harvested by centrifugation, the conditioned medium was filtered through a 0.22-μm-pore-size filter and used immediately. Experimental wells received 10 μl of the partially purified A. actinomycetemcomitans AI-2 that was diluted 1:25 with sterile AB medium. The microtiter plates were then shaken at 500 rpm and 30°C, and bioluminescence was measured at hourly intervals using a Wallac Victor3 multilabel counter (Perkin-Elmer).

Cloning of sahH and introduction into A. actinomycetemcomitans.

The sahH gene with its promoter was amplified from Pseudomonas aeruginosa PAO1 genomic DNA with primers H5 (5′-GGTACCCGCTATAATCGCCCGCTCAG-3′) and H3 (5′-GGATCCTGGTTGTAGTGATCGGCGAG-3′) as described by Walters et al. (49). The underlined sequences in the primers above represent KpnI and BamHI restriction sites to facilitate cloning into the pYGS shuttle vector (see below). Amplification was performed using the following PCR program: 95°C for 5 min, followed by 30 cycles of 95°C for 1 min, 60°C for 2 min, and 72°C for 3 min. The resulting PCR product was ligated with pGEM-T Easy (Promega) and transformed into E. coli DH5α to produce strain E. coli pGEMT-sahH. Recombinant clones were confirmed by EcoRI digestion to release the 1,680-bp sahH insert. Plasmid pGEMT-sahH was then digested with KpnI and BamHI, and the released fragment was cloned into the E. coli-A. actinomycetemcomitans shuttle vector pYGS. pYGS is a derivative of the pYGK shuttle vector (6) that contains a streptomycin resistance marker replacing the kanamycin resistance gene of pYGK. The resulting plasmid was introduced into the A. actinomycetemcomitans luxS mutant by electroporation (47), and streptomycin-resistant clones were selected. A plasmid isolated from streptomycin-resistant clones was digested with KpnI and BamHI to confirm the presence of the 1,680-bp sahH gene. The resulting strain was designated A. actinomycetemcomitans luxS-psahH.

Real-time PCR.

Overnight cultures of the A. actinomycetemcomitans luxS mutant and A. actinomycetemcomitans luxS-psahH were diluted 1:20 into fresh BHI medium and incubated at 37°C until the mid-exponential-growth phase (optical density, 0.3). Cells were harvested by centrifugation, and total RNA was isolated using TRIzol (Invitrogen) according to the manufacturer's instructions. Total RNA was digested with RNase-free DNase until no product was observed in the absence of reverse transcriptase by using the sahH primers given below. To confirm sahH expression in A. actinomycetemcomitans luxS-psahH, first-strand cDNA was prepared by using Ready-To-Go You-Prime first-strand beads (Amersham Biosciences) as described by the manufacturer, and the resulting cDNA was amplified with primers HRT5 (5′-ATGAGCGCTGTCATGACGCCC-3′) and HRT3 (5′-CGGCATACTTGCGGCGCAGG-3′). Each reaction mixture contained 7.5 pmol of the appropriate antisense primer and 50 ng of bacterial RNA. After completion of the cDNA synthesis, serial dilutions of cDNA were amplified by addition of 7.5 pmol of the sense primer and 0.5× SYBR green dye (Roche Applied Science). Amplifications were carried out using the Smart Cycler system (Cepheid) with the following profile: denaturation at 95°C for 15 s, annealing at 55°C for 15 s, and elongation at 72°C for 30 s for 45 cycles, followed by extension at 72°C for 1 min. Control reaction mixtures that did not contain reverse transcriptase were included in all experiments. All reactions were normalized against real-time PCRs using primers specific for the A. actinomycetemcomitans 5S RNA gene. The threshold cycle for each reaction was determined from a second derivative plot of total fluorescence as a function of the cycle number by using the software package supplied with the Smart Cycler system. All reactions were carried out on at least two independently obtained RNA preparations with consistent results.

Complementation of the A. actinomycetemcomitans lsrB mutant.

The lsrB gene was amplified from A. actinomycetemcomitans genomic DNA using primers lsrB5 (5′-GCGGGATCCGAAAACGCGTGTAAAA-3′) and lsrB3 (5′-GCGTCTAGATGAGAAATTGTAGTTGTCG-3′). Amplification was carried out using the following profile: 95°C for 5 min, followed by 30 cycles of 95°C for 1 min, 60°C for 2 min, and 72°C for 3 min. The resulting amplicon was ligated with pGEM-T Easy (Promega). After transformation of E. coli DH5α, clones were confirmed by plasmid isolation and the release of the appropriate lsrB fragment after digestion with BamHI and XbaI (underlined sites in the PCR primers above). The released fragment was then ligated into pYGK-ltxP that was cleaved with the same restriction enzymes. PYGK-ltxP contains the promoter of the A. actinomycetemcomitans leukotoxin operon introduced into the KpnI and BamHI sites of the shuttle vector pYGK (6) and positioned to drive lsrB expression. The resulting plasmid was confirmed by restriction digestion and then introduced into the A. actinomycetemcomitans lsrB mutant by electroporation. The biofilm growth of the resulting strain was determined as described above.

RESULTS

luxS and AI-2 are necessary for A. actinomycetemcomitans biofilm growth.

Our previous studies showed that inactivation of A. actinomycetemcomitans luxS reduced planktonic growth of A. actinomycetemcomitans under iron-limiting conditions and influenced the expression of several operons encoding iron storage and transport proteins (11). Since A. actinomycetemcomitans is present as part of a mixed microbial community in the oral cavity and iron acquisition is essential for survival in the host, we next sought to determine if luxS regulates the adherent growth of A. actinomycetemcomitans in biofilms. To accomplish this, flow cells containing a saliva-coated coverglass as the substrate were inoculated with bacterial cells obtained from a broth culture of A. actinomycetemcomitans JP2 or the luxS mutant, and adherent cells were cultured for as long as 60 h at 25°C in sterile BHI medium at a flow rate of 5.8 ml per h. After incubation for 30 h, both strains formed visible biofilms, but the total biomass was greater for strain JP2 (data not shown). After 60 h, A. actinomycetemcomitans JP2 formed a highly structured biofilm that extended up to 50 μm off the saliva-coated surface (see representative images in Fig. 1A). In contrast, the A. actinomycetemcomitans luxS mutant exhibited reduced total biomass and attained a depth of only <20 μm. Supplementation of the growth medium with partially purified AI-2 obtained from the wild-type strain restored the biofilm growth of the luxS mutant to wild-type levels, as measured both by total biomass and by the average depth of the biofilm (Fig. 1A, bottom right). Integration of fluorescence intensity across the representative compiled z-stack image shown in Fig. 1A confirmed that the total fluorescence and z-axis depth of the biofilm were restored to near-wild-type levels upon addition of AI-2 (Fig. 1B). Similar results were obtained when the luxS mutant was genetically complemented with a plasmid-borne copy of luxS. A summary of the data obtained from 40 to 60 random microscopic fields from at least three independent biofilm cultures of the wild-type, mutant, and complemented strains is shown in Table 1. Although total biomass and biofilm depth were similar to those for the wild-type strain, the architecture and surface topology of the complemented biofilm differed. Since these biofilms were supplemented with partially purified AI-2 and no direct biochemical assay is available to determine the AI-2 concentration, it is possible that the levels of AI-2 in these experiments exceeded that produced endogenously by A. actinomycetemcomitans. Increasing the shear force under which biofilms formed by increasing the flow rate of the culture medium resulted in a reduction in the overall accumulation of both wild-type and luxS mutant biofilms, but the luxS mutant still exhibited significantly reduced biomass and biofilm depth relative to those for the wild type (Table 1). These results indicate that luxS is important for A. actinomycetemcomitans biofilm growth and suggest that biofilm growth requires the presence of AI-2.

FIG. 1.

AI-2 influences the biofilm growth of A. actinomycetemcomitans. (A) (Top) Representative confocal images of biofilms formed by wild-type (strain JP2) (left) and LuxS-deficient (luxS mutant) (right) A. actinomycetemcomitans. (Bottom) Biofilm growth of the luxS mutant complemented with a plasmid-borne copy of luxS (luxS-pluxS) (left) or in the presence of growth medium supplemented with partially purified A. actinomycetemcomitans AI-2 (luxS mutant + AI-2) (right). The upper images are in the x/y plane, and the lower images show the corresponding x/z reconstruction from a Z-stack from 0 to 100 μm in 1-μm increments above the saliva-coated coverglass. Scale bar, 100 μm. (B) Graphic representation of biofilm biomass and depth expressed as integrated fluorescence of wild-type (•) or LuxS-deficient A. actinomycetemcomitans grown either in medium alone (▪) or in medium supplemented with partially purified AI-2 (▴). Total fluorescence in each Z-stack image was measured using the Fluoview software provided with the Olympus Fluoview 500 laser scanning confocal microscope and was plotted (as relative fluorescence units) against distance above the saliva-coated coverglass (z axis depth).

TABLE 1.

Biofilm formation of wild-type and LuxS-deficient A. actinomycetemcomitans^a

Strain	Exogenous AI-2	Flow rate (ml/h)	Mean biofilm depth (μm)	Biomass (μm3/μm2)
JP2 (wild type)	−	6	44.4 ± 12.3	0.274 ± 0.105
luxS mutant	−	6	19.9 ± 8.5*	0.028 ± 0.006**
luxS mutant	+	6	42.8 ± 13.6	0.250 ± 0.072
luxS-pluxS	−	6	45.2 ± 11.8	0.220 ± 0.051
JP2	−	35	14.3 ± 4.1	0.112 ± 0.027
luxS mutant	−	35	5.5 ± 2.4*	0.015 ± 0.004*

^a*, P < 0.01; **, P < 0.001.

S-Adenosylhomocysteine hydrolase does not complement luxS mutation in A. actinomycetemcomitans.

Inactivation of luxS potentially affects the metabolism of SAM, since LuxS catalyzes the conversion of S-ribosylhomocysteine to homocysteine and DPD, the precursor of AI-2 (54). However, in organisms that do not possess luxS, S-adenosylhomocysteine is converted directly to homocysteine by S-adenosylhomocysteine hydrolase (SahH). To confirm that the deficiency in A. actinomycetemcomitans biofilm growth arises from the absence of AI-2 rather than from an effect of luxS mutation on SAM metabolism, sahH was cloned from Pseudomonas aeruginosa and introduced into the A. actinomycetemcomitans luxS mutant. As shown in Table 2, real-time PCR confirmed the presence of sahH transcripts in the transformed strain but not in the luxS mutant. Biofilms formed by the transformed organism exhibited a twofold increase in total biomass (P < 0.01), as shown in Table 2. However, biofilms formed by the transformed organism in the presence of exogenous AI-2 exhibited total biomass similar to that of the wild-type strain. These results suggest that reconstitution of SAM metabolism in the absence of LuxS fails to fully complement the luxS mutation, and together with the complementation experiments for which results are shown in Fig. 1, they suggest that AI-2 itself is necessary for optimal biofilm growth by A. actinomycetemcomitans.

TABLE 2.

Complementation of the luxS mutation by SahH

Strain	Phenotype	Cycle threshold	AI-2	Biomass (μm3/μm2)a
luxS mutant	SahH−	28.84	−	0.025 ± 0.007
luxS-psahH	SahH+	12.02	−	0.051 ± 0.01*
luxS-psahH	SahH+	NAb	+	0.193 ± 0.072**

^a*, P < 0.01; **, P < 0.001.

^bNA, not applicable.

Inactivation of the putative AI-2 receptors of A. actinomycetemcomitans reduces biofilm growth.

Our previous work suggested that the LsrB protein of A. actinomycetemcomitans interacts with AI-2 and may function as a receptor for AI-2 (40). To determine if LsrB plays a role in biofilm growth, an A. actinomycetemcomitans lsrB mutant (40) was inoculated and grown in flow cells as previously described. The representative biofilm images in Fig. 2A and the total biofilm biomass (Fig. 2B) show that inactivation of lsrB resulted in reduced biofilm growth, which was restored by genetic complementation with a plasmid-borne copy of the lsrB gene. Interestingly, the lsrB mutant exhibited reduced biomass relative to that of the wild-type strain but maintained significant colony depth (compare the x/z frames in Fig. 2A). The total biomass of the biofilm formed by the lsr mutant was also significantly reduced relative to that for the wild type (Fig. 2B). Our previous studies showed that the RbsB protein of A. actinomycetemcomitans also interacts with AI-2 (21). Therefore, we examined biofilm formation by a strain with knockouts in both rbsB and lsrB (the A. actinomycetemcomitans rbsB lsrB double mutant). As shown in Fig. 2A, cells that were deficient in both rbsB and lsrB were ineffective at establishing biofilms and exhibited sparse, low-lying microcolonies with little structure. The accumulated biomass of these biofilms was significantly reduced (P < 0.001) compared to that for the lsr mutant (Fig. 2B), suggesting that both LsrB and RbsB may function as receptors for AI-2 in A. actinomycetemcomitans.

FIG. 2.

Inactivation of lsrB and rbsB results in reduced biofilm growth. (A) Representative confocal images of 60-h biofilms formed by wild-type A. actinomycetemcomitans (JP2), the lsrB mutant, and the lsrB rbsB double mutant. Biofilm growth of the lsrB mutant was also complemented with a plasmid-borne copy of lsrB. The top panels represent images in the x/y plane, and the lower panels show the corresponding x/z reconstruction from a Z-stack from 0 to 100 μm in 1-μm increments above the saliva-coated coverglass. Scale bar, 100 μm.(B) The biomass of each biofilm was determined using COMSTAT image-processing software as described in Materials and Methods.

Ribose inhibits A. actinomycetemcomitans biofilm growth.

Inactivation of lsrB and rbsB completely inhibited biofilm growth (Fig. 2B) and also significantly reduced the rate of AI-2 depletion from solution by intact cells (40). We also previously showed that ribose and AI-2 compete for the same binding site of RbsB (21), suggesting that ribose may block AI-2 interaction with RbsB and LsrB and interfere with AI-2-dependent phenotypes. To determine if ribose inhibits A. actinomycetemcomitans biofilm growth, biofilms of strain JP2 were generated by feeding adherent cells with BHI containing 50 mM or 100 mM ribose. These concentrations of ribose had previously been shown to effectively compete with AI-2 for binding to RbsB (21). As shown in the representative images in Fig. 3, biofilm formation was significantly reduced in the presence of ribose in a dose-dependent manner. As shown in Table 3, biofilm biomass was significantly reduced in the presence of ribose, but the average biofilm depth was significantly decreased only in the presence of 100 mM ribose. Although the concentrations of ribose required for inhibition are high, these results demonstrate proof of principle that the AI-2-dependent formation of biofilms by A. actinomycetemcomitans can be reduced by targeting LsrB and/or RbsB. Ribose itself had no significant effect on the planktonic growth of A. actinomycetemcomitans: exponential-phase doubling times were 76.2 ± 5.7 min for strain JP2 (wild type) without ribose and 81.9 ± 6.9 min with 50 mM ribose. Indeed, both of the mutant strains of A. actinomycetemcomitans exhibited similar exponential-phase doubling times when cultured in BHI broth (80.7 ± 7.3 min for the lsr mutant and 77.7 ± 3.1 min for the lsr rbs double mutant), suggesting that the inactivation of quorum-sensing genes appears to specifically influence adherent growth of A. actinomycetemcomitans.

FIG. 3.

Ribose inhibits the development of A. actinomycetemcomitans biofilms. A. actinomycetemcomitans JP2 biofilms were grown in the presence of 50 mM or 100 mM ribose and were analyzed by laser scanning confocal microscopy. The top panels show representative images in the x/y plane; the lower panels show the corresponding x/z reconstruction from a Z-stack from 0 to 100 μm in 1-μm increments above the saliva-coated coverglass. Scale bar, 100 μm.

TABLE 3.

Inhibition of A. actinomycetemcomitans biofilm growth by ribose^a

Ribose concn (mM)	Mean biofilm depth (mm) ± SD	Biomass (μm3/μm2)
0	43.4 ± 7.4	0.229 ± 0.049
50	37.1 ± 5.5	0.130 ± 0.092**
100	18.0* ± 6.3	0.041 ± 0.006**

^a*, P < 0.02; **, P < 0.01.

DISCUSSION

The oral biofilm is a complex microbial community composed of as many as 700 bacterial species, but relatively little is known about the mechanisms that allow these organisms to detect and respond to the changing environment of the oral cavity. Quorum-sensing circuits that detect and respond to acylhomoserine lactone and/or AI-2 have been reported to regulate the biofilm growth of numerous bacterial species (4, 8, 17, 18, 29, 30, 56). However, many oral species do not have the capacity to produce or respond to acylhomoserine lactone signals (15, 51) but do possess luxS and secrete AI-2. Our results suggest that AI-2 has little effect on the planktonic growth of A. actinomycetemcomitans, since the doubling times in culture of strains deficient in LuxS or the putative AI-2 receptors (RbsB and LsrB) were similar to that of wild-type A. actinomycetemcomitans. However, each of these mutant strains exhibited significantly reduced adherent growth in flow cell cultures, suggesting that AI-2 is clearly important for biofilm growth of A. actinomycetemcomitans. These results are consistent with other recent studies of oral bacteria showing that biofilm growth of Streptococcus mutans (50), Porphyromonas gingivalis, and Streptococcus gordonii (29) and the mutualistic growth of Streptococcus oralis and Actinomyces naeslundii (38) are dependent on luxS and AI-2. However, it should be noted that the A. actinomycetemcomitans strains used in this study do not express the tad fimbriae, which are found in clinical isolates. These structures confer nonspecific adherence to biotic and abiotic surfaces. Nonspecific adherence mediated by tad fimbriae may precede and facilitate more specific interactions between bacterial adhesins and salivary or epithelial cell ligands in the oral cavity. Our results clearly show that A. actinomycetemcomitans adheres to a saliva-coated surface and develops biofilms even under high shear forces and thus may be representative of events that occur after the initial nonspecific adherence of A. actinomycetemcomitans to tissue or mineral surfaces.

The role of LuxS in the activated methyl cycle (53, 54) raises the possibility that disruption of SAM metabolism may contribute to the phenotypes observed in the mutant strains. However, several lines of evidence suggest that this is not the case with the A. actinomycetemcomitans luxS mutant. First, biofilm growth was restored by introducing a plasmid-borne copy of luxS and also by the addition in trans of partially purified AI-2. In addition, “deaf” strains (i.e., the AI-2 receptor mutants) that were capable of producing AI-2 exhibited a biofilm phenotype similar to that of the luxS mutant strain. Finally, transformation of the luxS mutant with a plasmid-borne copy of sahH, encoding S-adenosylhomocysteine hydrolase, did not restore biofilm growth unless AI-2 was added to the growth medium. Together, these results indicate that biofilm growth of A. actinomycetemcomitans requires AI-2 itself and is not a phenotype arising from the disruption of SAM metabolism.

A. actinomycetemcomitans appears to respond differently to AI-2 than Vibrio spp., since it lacks the dedicated sensor kinase/phosphatase LuxQ, which initiates signal transduction in Vibrio. However, A. actinomycetemcomitans expresses two proteins (RbsB and LsrB) related to the AI-2 receptor of Vibrio, and our previous studies suggest that these two putative periplasmic proteins function as receptors for AI-2 (21, 40). The rbsB and lsrB genes are both present in operons encoding putative ABC transporters, and indeed, the A. actinomycetemcomitans lsr operon is closely related to Salmonella serovar Typhimurium lsrACDBFGE, which functions to actively import AI-2 (48, 55). Inactivation of either rbsB or lsrB significantly reduced the rate at which intact bacteria depleted AI-2 from their environment and reduced biofilm growth of A. actinomycetemcomitans. This suggests that LsrB and RbsB may be periplasmic receptors that facilitate the internalization of AI-2.

Mechanistically, how AI-2 uptake contributes to the development of biofilms remains to be determined. Gonzalez-Barrios et al. (17) showed that Lsr-dependent importation of AI-2 regulates motility and biofilm development by E. coli and that the AI-2-dependent two-component system QseBC (46) may be involved in this process. Interestingly, A. actinomycetemcomitans also possesses QseBC and the quorum-sensing response regulator QseA (44), yet it is a nonmotile organism. Our preliminary results also suggest that QseBC are regulated by AI-2 in A. actinomycetemcomitans, and we are currently investigating their role in biofilm development (D. R. Demuth, unpublished data). However, the ultimate fate of internalized AI-2 in A. actinomycetemcomitans and E. coli remains to be determined.

In contrast, Xavier et al. (55) showed that AI-2 is phosphorylated by LsrK and is subsequently degraded by LsrG to form 2-phosphoglycolate in S. enterica. Glycolate can be degraded by two independent pathways, one requiring glyoxylate carboligase and tartronate semialdehyde reductase to form glycerate, which feeds into glycolysis (http://biocyc.org), and one requiring malate synthase G to produce malate, a substrate of the tricarboxylic acid cycle. Thus, AI-2 appears to be utilized as a carbon source by S. enterica, and Xavier at al. (55) suggest that S. enterica may antagonize intercellular signaling in microbial communities by scavenging and destroying AI-2. However, if AI-2 functions solely as a nutrient, it is not clear why it is secreted and then actively internalized rather than being directly metabolized after synthesis by LuxS. This raises the intriguing possibility that AI-2 performs another function(s) that may influence biofilm development and that requires it to be secreted. For example, boron or borate is a micronutrient that is required for sugar transport, flower retention, and pollen formation in plants (27), but little is known about boron requirements and homeostasis in microbes. Borate is present in the human diet and has recently been shown to be important for the growth and proliferation of human cells (35). In addition, a specific Na⁺-coupled borate transporter, NaBC1, has been identified in humans (34) and is expressed at high levels in salivary glands. Given that borate forms very stable diesters with cis-diols on furanoid rings (27), AI-2 represents an ideal microbial scavenger for borate. Thus, it may be possible to reconcile the apparent signaling and metabolic roles of AI-2 if boron acquisition is important for microbial growth and biofilm development. We are currently testing these hypotheses by investigating biofilm growth by our A. actinomycetemcomitans mutants in the presence of exogenous glycolate and/or borate.

It is evident that A. actinomycetemcomitans does not utilize AI-2 as a quorum-sensing signal in the paradigm of the Vibrio spp., but clearly there is substantial evidence for AI-2-dependent phenotypes that are not readily explained if AI-2 is simply a minor metabolite that feeds into general metabolism. Determining the extra- and intracellular physiologic roles of AI-2 in A. actinomycetemcomitans may clarify the potential signaling function of AI-2 and its relationship to biofilm growth.