Host-Derived Pentapeptide Affecting Adhesion, Proliferation, and Local pH in Biofilm Communities Composed of Streptococcus and Actinomyces Species

Bacterial strains. Streptococcus and Actinomyces strains were of defined origin (20, 27, 33) and grown overnight on Columbia-II-agar base plates (Becton Dickinson and Company, Cockeysville, MD), supplemented with 30 ml of a human erythrocyte suspension per liter, at 37°C in an atmosphere with 5% CO₂. Strains intended for hydroxyapatite adhesion tests were metabolically labeled by adding [³⁵S]methionine to the growth medium (33).

Peptides.The peptides RGRPQ, AGAPQ, AARPQ, GGRPQ, RGAPQ, RGRAA, RGRPA, RGRAQ, KAKVN, GQSPQ, and GRPQG were synthesized, purified by using high-pressure liquid chromatography, and characterized by mass spectrometry (Biomolecular Resource Facility, University of Lund, Lund, Sweden). AGRPQ was not possible to synthesize. ProGln (PQ) was from BACHEM AG, Switzerland, and the other peptides were from Sigma Genosys, United Kingdom. Hybrid peptides DzzEEKFLRRIGRFGPRGRPQ (z means l-phosphoserine) and DzzEEKFLRRIGRFGPGQSPQ were synthesized, purified by using high-pressure liquid chromatography, and characterized by mass spectrometry (Mimotopes Pty Ltd., Clayton, Victoria, Australia). The peptides were dissolved in sterile water, the pH was adjusted to 6.8 to 7.0, and aliquots were frozen at −20°C. The RGRPQ peptide for in vivo testing was from Thermo Electron GmbH, Germany.

Proliferation.Bacterial cells were cultured consecutively in a chemically defined medium (44) at 37°C in an atmosphere with 5% CO₂. The second culture, used as an inoculum (1%, 1/100 μl [inoculum/medium]), was grown for 14 h. The inoculum (an A₅₅₀ of 0.03 corresponding to 8 × 10⁷ cells/ml) was added to a minimal growth medium (supplemented with peptide or control) obtained by dilution of the chemically defined medium (with all 20 naturally occurring free amino acids) eight times and filter sterilization in sterile 96-well microtiter plates (Millipore). The wells were incubated at 37°C in a 5% CO₂ atmosphere. Bacterial numbers were recorded every hour by measuring the absorbance at 550 nm (A₅₅₀) using a Spectra MAX 340 spectrometer (Molecular Devices, Sunnyvale, CA). The absorbance values which were used to calculate the generation time, were linearly proportional to the number of bacteria measured by viable counting.

pH increase (ammonia production).An overnight bacterial culture was washed once in 10 ml of sterile water, adjusted to 4.8 × 10⁹ cells/ml with water, and kept on ice for a minimum of 30 min before use (46). The bacterial cell suspension (90 μl) was mixed with test peptide and sterile water to a final volume of 100 μl and a final peptide concentration of 0.5 mM. The test and control suspensions were incubated at 37°C aerobically, and the pH was measured every 30 min for the 330-min test period by using a pH electrode.

Sucrose-induced pH decrease (acid production).To measure bacterial metabolic responses to peptides in the presence of sucrose, bacteria were grown on agar plates, harvested, and washed once in 10 ml of starvation solution [20 mM 2-amino-2-(hydroxymethyl)-1,3-propanediol-HCl, 35 mM NaCl, 62 mM KCl, 2 mM MgSO₄, 1 mM K₂HPO₄ (pH 7.0)]. After centrifugation, the cell pellet was suspended in 10 ml of the same solution, followed by incubation at 37°C for 1 h in an atmosphere with 5% CO₂ for starvation (42). Pelleted bacteria were resuspended in KCl solution (50 mM KCl, 1 mM MgCl₂ · 6H₂O) to 2 × 10⁹ cells/ml and kept on ice until use. Bacterial cells (45 μl) were mixed with the test peptide, and KCl solution was added to give a final volume of 97 μl at a final peptide concentration of 2 mM (S. gordonii SK12) or 5 mM (S. mutans Ingbritt). The pH was measured immediately as the starting pH (0 min). Sucrose solution (3 μl, 1 M) was added to each well, the mixtures and a control cell suspension were incubated at 37°C aerobically, and the pH was measured every 10 min during the initial 30-0min test period and every 30 min during the subsequent 120-min test period by using a pH electrode.

Adhesion and adhesion inhibition.The adhesion of ³⁵S-labeled bacteria to hybrid peptides (19) coated onto hydroxyapatite beads was measured essentially as described previously (5-mg beads, 5 × 10⁸ cells/ml, and 1 h of adhesion) (33). Adhesion to PRP-1 for the adhesion inhibition experiments was performed similarly except for the use of 8-mg beads, 2 × 10⁸ cells/ml, and 30 min of adhesion. Inhibition of adhesion was done by preincubation of cells with peptides for 20 min at room temperature prior to the adhesion assay.

Desorption (reversal of aggregates of bacteria and PRP-1-beads).Aggregation of bacteria by PRP-1-coated latex beads was carried out as described previously (32), except for the use of bacterial suspensions of 6 × 10⁹ cells/ml. Bovine serum albumin-coated latex beads served as a control. Desorption of strain T14V bound to PRP-1-coated beads was measured by adding peptide (5.06 mM) to preformed aggregates of bacteria and PRP-1-beads and visual scoring of the aggregates after 1 min of incubation at scores from 0 to 5 (5, no visible aggregates; 4, small aggregates in unclear solution; 3, moderate aggregates in unclear solution; 2, moderate aggregates in clear solution; 1, large aggregates in clear solution; and 0, clumping of aggregates in clear solution). Aggregation of bacteria by PRP-1-coated beads in the absence or presence of RGRPQ peptide used the same scoring but in the reversed order (e.g., 0 marks no visible aggregates and 5 indicates clumping of aggregates in clear solution; see Fig. 4A).

Identification of RGRPQ released by bacterial proteolysis of the PRP-1 polypeptide. S. gordonii SK12 was cultured on agar plates and incubated in starvation solution at 37°C for 1 h in 5% CO₂. After centrifugation, pelleted cells (2 × 10⁹ cells/ml, 75 μl) were incubated with PRP-1 (5 mg/ml) in M-DIL buffer (0.43% NaCl, 0.042% KCl, 0.1% Na₂HPO₄, 0.1% KH₂PO₄, 1% glycerophosphate disodium salt, 0.024% CaCl₂, 0.01% MgCl₂ · H₂O) for 2 h. After centrifugation, the supernatant (200 μl) was subjected to gel filtration (Superdex-peptide-column HR 10/30; Pharmacia) in 20 mM Tris-HCl-500 mM NaCl (pH 8.0). The flow rate was 0.5 ml/min, and the absorbance was monitored at 214 nm. Purified PRP-1 (50 μg) and synthetic RGRPQ peptide (100 mM, 5 μl) were run in parallel as reference substances.

Mass spectrometry and Edman degradation.Nano-electrospray mass spectrometry was carried out with a Q-Tof instrument (Micromass/Waters) operated as described previously (25) and Edman degradation with a Procise Ht sequencer instrument (Applied Biosystems) operated according to the manufacturer's recommendations. For mass spectrometry, 1 to 2 μl of the peptide fraction (500 μl) was used without prior desalting and mixed with an equal volume of spraying solvent (60% acetonitrile containing 1% acetic acid). For N-terminal sequence analysis, 60 μl of the fraction was applied to a Biobrene precycled glass fiber filter disc.

PRP degradation.PRP degradation was measured essentially as described previously (32). Briefly, PRP-1 (0.2 mg/ml) was coincubated with bacteria (2 × 10⁹ cells/ml) at 37°C for 20 h, followed by native alkaline polyacrylamide gel electrophoresis. The degree of cleavage was scored from 0 to 5 by densitometry (0 = 0 to 10% [negative activity], 1 = 10 to 20% [low activity], 2 = 20 to 40% and 3 = 40 to 60% [moderate activity], 4 = 60 to 80%, and 5 = 80 to 100% [high activity] loss of acidic PRP-1).

In vivo rat model.Sprague-Dawley rats (Møllegard Breeding Laboratories, Ejby, Denmark), 21 days old, were desalivated at day 1 by surgery as described previously (8). The rats were desalivated at day 1 and infected at days 2 and 3 by swabbing the oral cavity after dipping the swab in a bacterial suspension of S. gordonii SK12 and S. mutans NG8 (2 × 10⁹ and 1 × 10⁹cells, respectively) supplemented with peptide (5 mg/ml) or control. Each animal was given orally RGRPQ peptide (4 mg/day) or sterile water four times daily (4 × 100 μl, 10 mg of peptide/ml) from days 4 to 21. At day 9, all rats were reinfected with S. gordonii with peptide (5 mg/ml) or control. The animals received a ground diet supplemented with 56% sucrose and sterile-distilled water. The rats were screened for bacterial growth at days 4, 11, 18, and 21 by using oral swabs, and subsequent dilutions plated on MS (S. gordonii) and MSB (S. mutans) agar plates were used to establish the bacterial numbers (CFU). All animal experiments were approved by the local ethical committee of the Karolinska Institutet at Huddinge University Hospital, Huddinge, Sweden.

Statistics.Data are summarized as means and standard errors of the mean (SEM) or as median values. Unpaired t tests are used for two-group comparisons of peptide measurements referred to in results. For the analyses of the measurements of CFU displayed in Fig. 3, the Wilcoxon rank sum test was used for comparisons between pairs of groups at each time point. For an overall comparison between groups, a derived variable of the total number of CFU across the different time points within each rat is used. All tests are two-sided, and the significance level was set at P < 0.05.

Proteolytic release of RGRPQ from the PRP-1 polypeptide.Release of the RGRPQ peptide from the PRP-1 polypeptide after proteolysis by S. gordonii SK12 was shown in three modes. First, a peptide component eluted in the same position as synthetic RGRPQ upon gel filtration of the proteolysis products (Fig. 1A). Second, the purified peptide revealed the monoisotopic mass 613.46 Da, which corresponds to the singly protonated RGRPQ peptide (Fig. 1B). Third, its sequence, RGRPQ, was demonstrated by Edman degradation (Fig. 1C).

• Open in new tab
• Download powerpoint

FIG. 1.

Identification of a pentapeptide, RGRPQ, derived from proteolysis of PRP-1 by S. gordonii SK12. (A) Isolation of the RGRPQ peptide (marked) by gel filtration of the PRP-1 proteolysis mixture. Synthetic RGRPQ was used as a reference to localize the RGRPQ peptide. (B) Nano-electrospray mass spectrometry of the RGRPQ peptide recovered from gel filtration reveals the expected monoisotopic mass 613.46 Da. (C) Edman degradation of the fraction from gel filtration reveals the sequence RGRPQ.

Stimulation of proliferation of S. gordonii in vitro by RGRPQ and related peptides.We investigated the ability of RGRPQ and Ala substituted peptides to stimulate the growth rate or proliferation of S. gordonii in a minimum growth medium containing free amino acids (1:20 peptide versus free amino acids) (Table 1 and Fig. 2A). A 2.5-fold-increased growth rate occurred for S. gordonii at 150 μM RGRPQ compared to the minimum growth medium (the generation time was reduced from 4.65 to 1.85 h), whereas S. mutans was not affected (data not shown). The growth rate effect by RGRPQ on S. gordonii was dose dependent at peptide concentrations between 25 and 250 μM and optimal at 150 μM.

• Open in new tab
• Download powerpoint

FIG. 2.

Illustration of RGRPQ innate properties and tentative motifs responsible for each property. (A) Effect of RGRPQ derivatives on the growth of S. gordonii SK12 when cultured in a minimum growth medium in the presence or absence (control) of peptide. The bacterial cell numbers were estimated from measuring the A₅₅₀ of the cell culture. The mean absorbance values and the SEM from one experimental run with triplicate determinations are shown. (B) Effect of RGRPQ derivatives on the adhesion (percent binding bacteria out of total added) of A. naeslundii T14V to PRP-1-coated hydroxyapatite beads. (C) RGRPQ innate properties and tentative RGRPQ motifs for each property. Lines indicate each tentative motif with suggested key determinants marked by a broadened boldface line. The proliferation XXXQ motif is marked by a dotted line to emphasize its unique nature compared to the XXXQ adhesion motif. The potential AT hook marks the GRP motif present in some transcriptional factors.

Pentapeptides ending in a Q residue increased the growth rate efficiently, whereas those without a C-terminal Q residue did not (mean A₅₅₀ of 0.060 versus 0.043, P < 0.0001). Ala substitutions (i.e., RARPQ and AARPQ), but not a glycine substitution (i.e., GGRPQ), located in the N-terminal segment reduced the activity, suggesting an extended and C-terminal Q motif (e.g., XXXQ). Moreover, PRP-1-derived peptide sequences with a terminal or internal Q terminus, such as GQSPQ (C terminus) and PQGPPQ (repetitive segment) or GRPQG (internal segment), also stimulated the proliferation of S. gordonii effectively (Table 1).

Stimulation of proliferation of S. gordonii in vivo by the RGRPQ peptide.The ability of the RGRPQ peptide to selectively stimulate oral colonization of S. gordonii in vivo was evaluated by using a rat colonization model (Fig. 3). Rats were coinoculated orally with S. gordonii SK12 and S. mutans NG8 and orally administrated peptide (4 mg/day, n = 14 rats) or placebo (sterile water, n = 14 rats). The RGRPQ peptide, but not placebo, promoted strong selective colonization of S. gordonii in vivo, whereas S. gordonii in the absence of peptide was strongly suppressed under the favorable conditions provided for S. mutans by the desalivation and 56% sucrose diet. The Wilcoxon rank-sum P values between RGRPQ peptide and placebo were P = 0.0001 at day 4 and P = 0.0006 for day 11 and P = 0.0003 for the total counts summed across time points.

• Open in new tab
• Download powerpoint

FIG. 3.

(A and B) Promotion of colonization of S. gordonii SK12 (A), but not of S. mutans NG8 (B), by oral administration of RGRPQ peptide (○) compared to control (sterile water ▪) in a rat model. Rats (n = 14 peptide group, n = 14 control group) were coinfected with S. gordonii SK12 and S. mutans NG8 at the start of the study (day 0). Bacterial numbers (in CFU) are presented as means of the 14 animals in each group, and bars show the SEM. Asterisks indicate statistically significant differences at given time points (***, P < 0.001) based on comparisons between pairs of groups at each time point using the Wilcoxon rank-sum test. Also shown in panel A are the median and the individual S. gordonii CFU values for the RGRPQ-treated rats at the 4- and 11-day time points.

Effects on local pH by the RGRPQ peptide.We next investigated potential mechanisms for the ability of the RGRPQ peptide to block a sucrose-induced pH decrease in oral biofilms in situ (32) (Table 1). When added to S. gordonii SK12 (which exhibits the catabolism of arginine to ammonia), peptides with an N-terminal arginine increased the pH, whereas peptides without this characteristic did not (mean pH 7.28 versus 6.35, P < 0.0001). Moreover, peptides with Ala substitutions located in the C-terminal segment or at the internal Arg residue affected the pH marginally, if at all, suggesting the N-terminal Arg as the responsible motif for ammonia production. In contrast, incubation of peptides with S. mutans Ingbritt (which lacks catabolism of arginine) did not increase the pH (data not shown). Moreover, the RGRPQ peptide blocked pH decrease (an indirect measure of acid production) in both S. mutans Ingbritt and S. gordonii SK12 after challenge with sucrose (Table 1). Other peptide sequences displayed markedly lower blockage of acid production compared to RGRPQ for S. mutans (mean pH 5.1 versus 6.1, P = 0.0145) and for S. gordonii (mean pH 5.6 versus 6.3, P = 0.0387). Thus, the ability of the full-length RGRPQ peptide to block a sucrose-induced pH decrease in oral biofilms in situ may be a function of both ammonia production and the blockage of acid production by viridans streptococci.

PRP degradation and ammonia production are typical properties of commensal viridans streptococci.A panel of oral Streptococcus species of defined arginine catabolism and cariogenic properties (13) was tested for their ability to cleave PRP-1 (and the potential release of RGRPQ) by incubation with PRP-1 and native alkaline gel electrophoresis (Table 2). A high degree and frequency of PRP-1 proteolysis occurred among strains of strictly commensal streptococci (e.g., S. gordonii and S. sanguinis) with arginine catabolism and low acidogenicity, whereas the opposite was observed for highly (e.g., S. mutans) and potentially (e.g., S. mitis) cariogenic streptococci without arginine catabolism but with high acid production (Table 2).

Adhesion inhibition and desorption of A. naeslundii T14V by the RGRPQ peptide.We tested RGRPQ and structural analogues thereof for their ability to inhibit the adhesion of A. naeslundii T14V to PRP-1 coated onto hydroxyapatite beads, and to detach this strain from PRP-1-coated latex beads (Table 1 and Fig. 2B). High-to-moderate inhibitory or desorbing activity occurred with peptides ending in a Q residue, whereas peptides without a Q residue had virtually no activity (mean inhibition of 54% versus 11%, P < 0.0001, mean desorption of 2.6 versus 0.3, P < 0.0001). In addition to Ala substitutions at the C-terminal Q residue, substitution of the P or G residues (RGRAQ and RARPQ/AARPQ, respectively) reduced the inhibitory activity. Moreover, the previously delineated dipeptide PQ lacked inhibitory and desorbing activity (data not shown), and the GGRPQ peptide showed equal activity to RGRPQ. Thus, an XXXQ adhesion motif with a dependence on both glycine and proline (i.e., GXPQ) may be at work.

Actinomyces strains using the same PRP-1 binding sites as S. gordonii are inhibited by RGRPQ.Oral commensal (e.g., S. gordonii and A. naeslundii), as well as pathogenic, bacteria may compete for PRP-1 binding sites. We therefore investigated the ability of the RGRPQ peptide to selectively inhibit the adhesion of a panel of A. naeslundii genospecies 2 (n = 10) and S. gordonii (n = 5) strains (including strain SK12) with avid binding to PRP-1 coated onto hydroxyapatite beads. The RGRPQ peptide blocked the adhesion of half of the A. naeslundii strains (n = 5), but none of the S. gordonii strain (Fig. 4). We next used hybrid peptides terminating in PRP-1 epitopes (RGRPQ and the C-terminal GQSPQ) to delineate potential PRP-1-binding sites for A. naeslundii T14V (inhibitable by RGRPQ), A. naeslundii LY7 (noninhibitable by RGRPQ), and S. gordonii Blackburn (noninhibitable by RGRPQ) (Fig. 4). Whereas the inhibitable A. naeslundii T14V and S. gordonii Blackburn bound avidly to both peptides, A. naeslundii LY7 showed avid binding only to the GQSPQ peptide. Thus, the RGRPQ peptide selectively inhibits Actinomyces strains that use the same PRP-1 binding sites as S. gordonii, which may differ by having a low-affinity binding mode insensitive to free peptides.

• Open in new tab
• Download powerpoint

FIG. 4.

(A) Selective inhibition of the adhesion of certain A. naeslundii genospecies 2 strains by the RGRPQ peptide. Adhesion was determined by measuring the aggregation of the bacteria with PRP-1-coated latex beads in the presence (+RGRPQ) or absence (−RGRPQ) of peptide. Score 5 indicates clumping of aggregates in clear solution, and score 0 indicates no aggregates (see Materials and Methods). (B and C) Adhesion of A. naeslundii and S. gordonii strains to hybrid peptides with a RGRPQ or GQSPQ terminus. Adhesion (percent binding bacteria of total added bacteria) of ³⁵S-labeled A. naeslundii strains T14V and LY7 and S. gordonii Blackburn to a hybrid/RGRPQ peptide (B) or a hybrid/GQSPQ peptide (C) coated onto hydroxyapatite beads. The hybrid peptides are designed with the calcium binding 15-amino-acid N-terminal segment of statherin and a proline residue followed by RGRPQ or GQSPQ, derived from the middle or C-terminal segments of PRP-1, respectively.