Possible Release of an ArgGlyArgProGln Pentapeptide with Innate Immunity Properties from Acidic Proline-Rich Proteins by Proteolytic Activity in Commensal Streptococcus and Actinomyces Species

Abstract

This study suggests degradation of salivary acidic proline-rich proteins (PRPs) into potential innate-immunity-like peptides by oral Streptococcus and Actinomyces species. PRP degradation paralleled cleavage of Pro-containing substrates. PRP degradation by S. gordonii strain SK12 instantly released a Pyr₁-Pro₁₀₄Pro₁₀₅ and a Gly₁₁₁-Pro₁₄₉Gln₁₅₀ peptide together with a presumed Arg₁₀₆Gly₁₀₇Arg₁₀₈Pro₁₀₉Gln₁₁₀ pentapeptide. The synthetic Arg₁₀₆Gly₁₀₇Arg₁₀₈Pro₁₀₉Gln₁₁₀ peptide desorbed bound bacteria and counteracted sucrose-induced decrease of dental plaque pH in vitro.

The acidic proline-rich proteins (PRPs), encoded by the PRH1 and PRH2 loci on chromosome 12p13.2 (4), are major saliva proteins (15). As polymorphic and multifunctional proteins (4, 15, 20), they are potential determinants of host susceptibility to dental caries (23, 24).

Acidic PRPs adsorb to hydroxyapatite surfaces, regulate calcium phosphate and hydroxyapatite crystal equilibrium (15), attach commensal Actinomyces and Streptococcus species to teeth (13, 21), and inactivate ingested plant polyphenols (tannins) (5). While the proline-poor N-terminal 30-residue domain confers hydroxyapatite and calcium binding (15), the proline-rich middle/C-terminal domain binds tannins via proline-rich repeats (5) and bacteria via the ProGln terminus (13, 21).

Acidic PRPs consists of large allelic (e.g., PRP-1 and PIF-s) and small posttranslational (e.g., PRP-3 and PIF-f) variants (4). The small acidic PRPs resulting from proteolytic cleavage at Arg₁₀₆-Gly₁₀₇ display poor bacterial adhesion activities but high affinities for hydroxyapatite surfaces (15).

After secretion, the acidic PRPs are rapidly enriched on tooth surfaces and degraded into potential innate-immunity peptides by dental plaque proteolysis (22). Both gram-negative and gram-positive bacteria display complex profiles of glycosidases and proteases, but little is known about turnover of acidic PRPs by commensal and early-colonizing Streptococcus and Actinomyces species (8). In this study, we used mass spectrometry of peptide mixtures to suggest turnover of acidic PRPs into innate-immunity-like peptides by commensal Streptococcus and Actinomyces species.

PRP-1 and PRP-3 were purified from parotid saliva of three subjects homozygous for PRP-1 and PIF-s by DEAE-Sephacel column chromatography (15 by 1.6 cm; Pharmacia, Uppsala, Sweden) using a linear gradient of 25 to 1,000 mmol of NaCl/liter in 50 mmol of Tris-HCl/liter (pH 8.0). The acidic PRP fractions were then concentrated (Centriprep 10 concentrator; Amicon Inc., Beverly, Mass.) and separated by gel filtration (HiLoad 26/60 Superdex S-200 prep-grade column; Pharmacia) in Tris-HCl (20 mmol/liter)–NaCl (500 mmol/liter), pH 8.0. The resolved PRP-1 and PRP-3 fractions were dialyzed against Tris-HCl buffer and subjected to a Macroprep High Q column (15 by 1.6 cm; Bio-Rad, Hercules, Calif.) using a linear gradient of 25 to 1,000 mmol of NaCl/liter in 50 mmol of Tris-HCl/liter, pH 8.0. The purified proteins were dialyzed against water, lyophilized, and stored at −20°C.

Streptococcus and Actinomyces strains (14, 17) were grown at 37°C for 18 h in 5 ml of Trypticase soy bean glucose limiting broth (1.7% peptone, 0.3% soy peptone, 0.15% yeast extract, 12.5 mmol of glucose/liter, and 12.5 mmol of NH₄HCO₃/liter in NaH₂PO₄-K₂HPO₄ buffer [1 mol/liter], pH 7.3) in an atmosphere with 5% CO₂. Pelleted (17,000 × g for 5 min) cells were washed twice 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) and resuspended in M-DIL buffer at a concentration of 2 × 10⁹ cells/ml. Bacterial cells and cell-free supernatants (obtained after pelleting of bacteria) were kept on ice prior to degradation experiments.

PRP degradation was assayed by mixing equal volumes (300 μl) of protein (0.6 mg/ml) and bacteria (2 × 10⁹ cells/ml), both dissolved in M-DIL buffer, followed by incubation at 37°C for various times (15 min, 4 h, 20 h, and 1 week). After pelleting (17,000 × g for 10 min) of bacterial cells, the supernatants were aliquoted, lyophilized, and stored at −80°C prior to native alkaline polyacrylamide gel electrophoresis (PAGE) and densitometric analyses as previously described except for the use of Tris-glycine 7.5% Ready gels (Bio-Rad) (7).

The substrate specificity of PRP degradation was measured essentially as described elsewhere (11). Briefly, 75 μl of bacterial suspension (2 × 10⁹ cells/ml in M-DIL buffer) was diluted with 75 μl of 0.2 mol of Tris-HCl (pH 7.5)/liter, followed by addition of 20 μl (5 mM in dimethyl sulfoxide) of each substrate: H-Arg-Pro-pNA, H-Lys(Abz)-Pro-Pro-pNA, H-Pro-pNA, H-Glu-Ala-Leu-Phe-Gln-pNA, Z-Gly-Pro-pNA, and B-Arg-pNA (Bachem, Dubendorf, Switzerland). After incubation at 37°C for 20 h, the extent of cleavage was estimated by monitoring the absorbance at 414 nm.

PRP-derived peptide structures were established using a hybrid quadrupole time-of-flight mass spectrometer (Micromass, Manchester, United Kingdom) with a Z-configured nanospray source and gold-coated spraying needles (Protana, Odense, Denmark). Detection was all times in the positive ion mode.

The adhesion-blocking activity of synthetic ArgGlyArgProGln (Biomolecular Resource Facility, University of Lund, Lund, Sweden) was measured by mixing equal volumes (15 μl) of suspensions of bacteria (10⁸ cells/ml) and PRP-1-coated latex beads (16) on a glass plate for 2 min in the presence and absence of pentapeptide. Aggregation was scored visually, and in some experiments the pentapeptide was added to already established bacterium–PRP-1-latex bead aggregates.

The ability of the pentapeptide to counteract the sucrose-induced decrease of dental plaque pH was measured using dental plaque from one healthy donor who had refrained from eating and oral hygiene for 12 h. Sampled plaque was washed twice, suspended (16 mg/ml) in sterile distilled water, and added (90 μl) to microtiter wells, followed by addition of (i) 10 μl of distilled water (control); (ii) 5 μl of sucrose (7.0 mM) plus 5 μl of distilled water; (iii) 5 μl of sucrose (7.0 mM) plus 5 μl of ArgGlyArgProGln (100 mM); and (iv) 5 μl of ArgGlyArgProGln (100 mM) and 5 μl of distilled water. The microtiter wells were incubated at 37°C for 1 h and continuously monitored using a pH electrode.

PRP degradation by commensal Streptococcus and Actinomyces species.

Native alkaline PAGE of acidic PRPs (PRP-1 and PRP-3) after incubation with washed bacterial cells revealed PRP degradation by certain species and in the following order (Table 1; Fig. 1a): S. gordonii, S. sanguis, and A. odontolyticus ≫ S. anginosus > S. mitis, and S. oralis. In contrast, S. mutans, S. sobrinus, A. naeslundii genospecies 1 and 2, and A. viscosus lacked PRP degradation activity.

TABLE 1.

Degradation of acidic PRPs by Streptococcus and Actinomyces species

Species	Strain(s)	PRP degradationa
		PRP-1	PRP-3	Typeb
S. gordonii	SK12, SK184	5	1	I
	SK120	5	0	I
	SK33	0	0
S. sanguis	SK85	4	0	I
	SK112	2	1	II
	SK37	2	1	II
	SK162	2	1	III
S. mitis	SK304	1	0	II
	SK305	0	0
	SK96, SK142	0	0
S. anginosus	SK215	3	1	III
	SK52	1	1	II
	SK63	1	0	II
	SK218	0	0
S. oralis	SK143	1	0	IV
	SK2, SK92	0	0
	C 104	0	0
S. mutans	Ingbritt, JBP, NCTC 10449	0	0
S. sobrinus	SL-1	0	0
A. odontolyticus	T-5-G	5	5	IV
	T-1-K	2	1	IV
	T-23-N, T-3-G	1*	1	IV
	T-21-N, T-22-N, PK 984	0	0
A. naeslundii
Genospecies 1	ATCC 12104, P-3-N, P-5-N, P-11-N, B-2-G, PK 947	0	0
Genospecies 2	T14V, P-1-N, P-7-N, P-1-K, P-1-G, B-7-N, P-2-N	0	0
A. viscosus	ATCC 19246	0	0

^aPRP degradation was assayed by coincubation of PRP-1 or PRP-3 and bacteria for 15 min, 4 h, and 20 h followed by native alkaline PAGE (Fig. 1a). The degree of cleavage at 4 h was scored from 0 to 5 by densitometry as follows: 0, 0 to 10%; 1, 10 to 20%; 2, 20 to 40%; 3, 40 to 6%;, 4, 60 <80%; and 5, 80 to 100% loss of acidic PRPs. Score 1 labeled with an asterisk denotes the formation of degradation products after 20 h of incubation; 95% confidence intervals of densitometry runs of control PRP-1 corresponded to 8% of mean. 

^bThe strains were classified into four degradation patterns (types I to IV) based on peptide banding patterns (Fig. 1a). The type I pattern displayed a peptide migrating just below PRP-3; the type II pattern displayed a peptide migrating just below PRP-3 and another between PRP-1 and PRP-3; the type III pattern displayed two peptides migrating just below and above PRP-3, one between PRP-1 and PRP-3, and a fourth just below PRP-1; the type IV pattern displayed a peptide migrating identically to PRP-3 and another between PRP-1 and PRP-3, although closer to PRP-1 compared to patterns II and III. 

FIG. 1.

Peptide structures generated by degradation of acidic PRPs by S. gordonii SK12. (a) Native alkaline PAGE of degradation of PRP-1 by S. gordonii SK12. Shown are control PRP-1 and PRP-3 (lanes 1 and 8, respectively) and PRP-1 (lanes 2 to 4) and PRP-3 (lanes 5 to 7) after incubation with bacterial cells for various times. (b) Gel filtration chromatograms (Superose 12; Pharmacia) of peptides from degradation of PRP-1 by S. gordonii SK12 for various times. Numbering of peaks refers to mass spectrometric identification of the corresponding peptide structures (c): 1, PRP-1; 2 and 3, 15 min of degradation; 4, 20 h of degradation; 5, 1 week of degradation. Arrows denote the retention times of pure PRP-1 and PRP-3. (c) Peptide structures identified by mass spectrometry of peptide peaks from gel filtration (b). Masses were calculated from the multiply charged ions (m/z) in the mass spectrum (Fig. 2). The numbering of peptides as A to G corresponds to the signals in the mass spectrum (Fig. 2). Pyr indicates a pyroglutamic acid. Mass numbers with asterisks are average masses from deconvoluted mass spectra, while unlabeled mass numbers are consistent with the monoisotopic mass of the peptide. (d) Simplified map of generated peptides in comparison to PRP-1. The potential cleavage sites at peptide bonds formed at Pro or Gln residues are given. The ArgGlyArgProGln pentapeptide presumed to be instantly released is expanded, ProGln termini are marked by black circles, and the posttranslational cyclization of the N-terminal Glu residue to a pyroglutamic acid residue is marked Pyr. The verification by mass spectrometry of phosphorylation of Ser at positions 8 and 22 is marked.

Native alkaline PAGE distinguished four qualitatively distinct degradation patterns (patterns I to IV) with N-terminal peptides migrating close to PRP-3 but otherwise deviating banding patterns (Fig. 1a; Table 1). A strain representative of each of the four degradation patterns cleaved Pro-containing substrates (S. gordonii SK12, S. mitis SK304, S. anginosus SK215, and A. odontolyticus T-5-G, respectively) with similar, narrow (types I to III) or deviating, broad (type IV) substrate specificities the S. gordonii strain SK33 devoid of PRP degradation did not cleave Pro-containing substrates). The broad substrate specificity of A. odontolyticus (type IV) paralleled its high cleavage rate of both PRP-1 and PRP-3 compared to streptococci (types I to III) with a low PRP-3 cleavage rate.

PRP degradation by cell-free supernatants occurred also. Preparative gel electrophoresis of the cell-free supernatant of S. gordonii strain SK12 revealed PRP-degrading activity by at least two different protein fractions. There was no direct relationship between PRP degradation and adhesion to acidic PRPs or soluble immunoglobulin A1 protease activity.

Time dependency and structural features of PRP degradation.

The peptides obtained by degradation of PRP-1 by S. gordonii SK12 for various times were separated by gel filtration and analyzed by mass spectrometry (Fig. 1b to d and 2). The peptide peaks obtained after 15 min of incubation contained an N-terminal 105-residue peptide, Pyr₁-Pro₁₀₄Pro₁₀₅ (peak 2), and a C-terminal 40-residue peptide, Gly₁₁₁-Pro₁₄₉Gln₁₅₀ (peak 3). The additional peptide peak (peak 4) appearing after 20 h of incubation contained a series of 15- to 47-residue peptides: Pro₉₆-Pro₁₀₉Gln₁₁₀, Gly₁₁₁-Pro₁₃₀Gln₁₃₁, Gly₁₁₁-Pro₁₃₄Pro₁₃₅, Gly₁₁₁-Pro₁₃₅Gln₁₃₆, Gly₁₁₁-Pro₁₄₀Gln₁₄₁, Gly₁₁₁-Pro₁₄₉Gln₁₅₀, and Pro₁₀₄-Pro₁₄₉Gln₁₅₀. The peptide peak appearing after 1 week of incubation (peak 5) contained oligopeptides and amino acids, as identified by peptide gel filtration.

Innate-immunity-like properties of synthetic Arg₁₀₆Gly₁₀₇Arg₁₀₈Pro₁₀₉Gln₁₁₀.

The Arg₁₀₆Gly₁₀₇Arg₁₀₈Pro₁₀₉Gln₁₁₀ pentapeptide presumed to be instantly released by PRP degradation counteracted sucrose-induced decrease of dental plaque pH in vitro (Fig. 3a). The pentapeptide alone increased dental plaque pH. In addition, the pentapeptide desorbed bound cells and blocked adhesion of Actinomyces strain T14V, while strain LY7 with another PRP binding specificity was unaffected (Fig. 3b).

FIG. 3.

Ability of synthetic Arg₁₀₆Gly₁₀₇Arg₁₀₈Pro₁₀₉Gln₁₁₀ to counteract the sucrose-induced decrease in dental plaque pH (a) and to desorb attached Actinomyces cells (b) in vitro. (a) Plaque pH following addition of distilled water (Plaque), 0.35 mM (0.12 mg/ml) sucrose (Plaque+sucrose), and 0.35 mM sucrose plus 5 mM (3.0 mg/ml) ArgGlyArgProGln (plaque+sucrose+ArgGlyArgProGln) to dental plaque suspended in microtiter wells. (b) Reversal of aggregates of A. naeslundii strain T14V and PRP-1-coated latex beads (but not of aggregates induced by strain LY7 with another PRP binding specificity) by 5 mM (3.0 mg/ml) of ArgGlyArgProGln (black bars). The formation of aggregates in the absence of pentapeptide is also shown (white bars).

This study suggests turnover of acidic PRPs into innate-immunity-like peptides by proteolytic activity in commensal Streptococcus and Actinomyces species. Like degradation of casein by lactococci (19), degradation of acidic PRPs is time dependent and releases oligopeptides instantly and after prolonged degradation. The instant release of an Arg₁₀₆Gly₁₀₇Arg₁₀₈Pro₁₀₉Gln₁₁₀ pentapeptide seems likely from the appearance of N-terminal 105-residue Pyr₁-Pro₁₀₄Pro₁₀₅ and C-terminal 40-residue Gly₁₁₁-Pro₁₄₉Gln₁₅₀ peptides after short-time (15-min) incubation of PRP-1 with S. gordonii SK12. Both endoprotease specificity per se and intrinsic properties of acidic PRPs, such as the extended structure of proline-rich stretches (9), may explain the preferentially cleavage in the middle/C-terminal proline-rich region by degradation types I to III; PRP-3 was resistant to cleavage, PRP-3-like peptides were generated, and fragments from the region spanning residues 96 to 140 were structurally identified. A prolyl endoprotease, alone or in combination with additional enzyme activities, is indicated by PRP degradation and cleavage of Pro-containing substrates paralleling each other. Furthermore, the peptide structures released by strain SK12 indicated cleavage at peptide bonds formed at Pro or Gln residues: Pro₉₅-Pro₉₆, Arg₁₀₃-Pro₁₀₄, Pro₁₀₅-Arg₁₀₆, Gln₁₁₀-Gly₁₁₁, Gln₁₃₁-Gly₁₃₂, Pro₁₃₅-Gln₁₃₆, Gln₁₃₆-Gly₁₃₇, and Gln₁₄₁-Gly₁₄₂.

The synthetic Arg₁₀₆Gly₁₀₇Arg₁₀₈Pro₁₀₉Gln₁₁₀ pentapeptide possessed innate-immunity-like properties. (i) It blocked the sucrose-induced decrease in dental plaque pH in vitro. Like sialin, an Arg-containing tetrapeptide in saliva (10), the pentapeptide containing two Arg residues increased the dental plaque pH in the absence of sucrose, suggesting acid neutralizsation by Arg catabolism to ammonia (arginine deiminase pathway [12]) as one underlying mechanism. Inhibition of lactate production could be another mechanism. (ii) It desorbed attached A. naeslundii strains and blocked adhesion to PRP-1, depending on PRP binding specificity (21). The desorbing activity suggests a high-affinity binding epitope composed of the ProGln binding site and neighboring amino acids. Notably, PRP degradation generated multiple peptides with ProGln termini that could affect commensal and potentially pathogenic bacteria binding to acidic PRPs and statherin (13, 21).

Commensal streptococci possessing Arg catabolism displayed PRP degradation (S. sanguis, S. gordonii, S. anginosus, and S. mitis), while cariogenic S. mutans and S. sobrinus lacking Arg catabolism did not (17). Consequently, PRP degradation could be a characteristic of streptococcal biotypes with tooth-protective properties. Notably, Arg seems to affect the ecological relationship between S. sanguis and S. mutans (25), and caries-susceptible and caries-resistant subjects seems to differ in ability to increase the dental plaque pH after acidification (1).

Innate-immunity-like properties of acidic PRPs are in line with the association between allelic variation in acidic PRPs, saliva-mediated adhesion, and caries susceptibility; the Db variant, with a 21-amino-acid internal tandem repeat, is associated with altered saliva adhesion of bacteria and caries susceptibility (23, 24). Interestingly, the pentapeptide contains the ArgGlyArgPro motif (AT hook) present in transcriptional factors (3), and innate immunity peptides have been found to affect transcription (6). Both intracellular effects by salivary histatins (18) and bactericidal activity of a poly-l-proline II conformation of the proline-rich tandem repeat sequences of salivary MG2 (2) have been demonstrated. Nevertheless, future studies supporting the concept of PRP-derived innate immunity peptides should demonstrate their presence in oral biofilms, biological functions, and nature of underlying bacteria and endoproteases.

FIG. 2.

Mass spectrum of a peptide mixture from degradation of PRP-1 for 20 h by S. gordonii SK12 (cf. peak 4 in Fig. 1b). All ions (m/z) are labeled with a letter indicating the identity of the peptide structure (Fig. 1c). Superscripts indicate the charge state of the peptide. The insert shows an expansion of the m/z scale for the quadruply charged ion 4F at m/z 983.50, corresponding to a mass of 3,930 Da (Fig. 1c).