Interaction of adrenomedullin and calcitonin gene-related peptide with the periodontal pathogen Porphyromonas gingivalis

Authors


  • Editor: Alex van Belkum

Correspondence: R.P. Allaker, Oral Microbiology, Institute of Cell & Molecular Science, Newark Street, London E1 2AT, UK. Tel.:+44 20 7882 2388; fax:+44 20 7882 2191; e-mail: r.p.allaker@qmul.ac.uk

Abstract

The nature of the interaction between Porphyromonas gingivalis and the multifunctional peptides adrenomedullin and calcitonin gene-related peptide (CGRP) was investigated. Growth of P. gingivalis was not inhibited in the presence of either of these peptides [minimal inhibitory concentration (MIC)>250 μg mL−1]. The ability of the arginine- and lysine-specific proteases from P. gingivalis to breakdown these peptides was investigated. Adrenomedullin and CGRP were incubated with culture supernatants from wild-type and protease gene knockout strains. No significant effect on antimicrobial activity against the indicator organism Escherichia coli BUE55 was found (MIC=6.25 μg mL−1 in all cases). The role of anionic components on the surface of P. gingivalis, which may alter binding of these cationic peptides, was also investigated in relation to adrenomedullin. Growth of gene knockout strains lacking surface polysaccharide and capsule components was not inhibited (MIC>250 μg mL−1). It is suggested that a lack of sensitivity to adrenomedullin and CGRP may enable P. gingivalis to persist in the oral cavity and cause disease.

Introduction

The oral epithelium constitutes a protective interface between external and internal environments, and provides the first line of defence against infection by pathogenic microorganisms. The production of antimicrobial peptides by epithelial cells is now considered to be a key mechanism in such protection (Schroder, 1999). Studies at mRNA level suggest that gene expression for these peptides occurs in a tissue/organ-specific manner, which possibly relates to their antimicrobial spectrum and conditions of expression. Some epithelial antimicrobial peptides are constitutively expressed, while others are inducible, either by the presence of microorganisms or by endogenous cytokines (Schroder, 1999).

Adrenomedullin is a 52-amino-acid multifunctional peptide (Hinson et al., 2000) produced by a wide variety of tissues and cells. A degree of homology with calcitonin gene-related peptide (CGRP) is shown and thus adrenomedullin is included in the calcitonin, CGRP and amylin peptide family (Kitamura et al., 1993). The molecule has a single intramolecular disulphide bond between residues 16 and 21, along with an amidated tyrosine at the carboxyl terminus (Fig. 1). Adrenomedullin exhibits a net positive charge (+6) through an abundance of arginine and lysine residues. In common with other antimicrobial peptides, including human β-defensin-2, the amphipathic structure characterized by spatially separated hydrophobic and charged regions permits bacterial membrane intercalation (Table 1).

Figure 1.

 Adrenomedullin structure showing properties of amino acid residues.

Table 1.   Comparison of the structural and functional properties of adrenomedullin and human β-defensin-2
PropertyAdrenomedullinβ-Defensin-2
ExpressionConstitutiveInducible
ChargeCationic +6Cationic +7
Amphipathic designYesYes
Size52 amino acids41 amino acids
Structural motifsone disulphide bridgethree disulphide bridges
Cytokine stimulationYesYes
Antibacterial activityYesYes
Antifungal activityNoYes

Previous work has demonstrated that adrenomedullin and CGRP have antimicrobial activity against a number of members of the normal human microbial community (Zihni et al., 1999; Allaker & Kapas, 2003). Further support for an antimicrobial role for adrenomedullin has also been provided by a number of other studies. Adrenomedullin protein and mRNA levels are increased when epithelial cells are exposed to whole bacteria and culture supernatants (Kapas et al., 2001). Investigation of adrenomedullin expression in the intestinal epithelium of cows suffering from bovine paratuberculosis, a chronic inflammatory disease caused by infection with the pathogen Mycobacterium paratuberculosis, has also revealed a higher level of expression of adrenomedullin mRNA in infected cattle (Allaker & Kapas, 2003).

The opportunistic pathogen Porphyromonas gingivalis is a proteolytic, gram-negative, anaerobic bacterium, and is a frequent component in the subgingival dental plaque of patients with chronic adult periodontitis. The major extracellular arginine-specific (arg-gingipains; RgpA and B) and lysine-specific (lys-gingipain; KgpA) cysteine proteases of P. gingivalis are considered to be potential virulence determinants (Curtis et al., 1999). Proteolytic activities of P. gingivalis are known to contribute to nutrient acquisition, tissue destruction and deregulation of the inflammatory response and host defences. It is also possible that production of Arg-X and Lys-X proteases enables P. gingivalis to evade the action of antimicrobial peptides, including adrenomedullin and CGRP. The major cell surface macromolecules of P. gingivalis, capsular polysaccharide (or K antigen) and lipopolysaccharide, are also considered to be of importance in the pathogenesis of periodontitis (Lamont & Jenkinson, 1998). It is well known that macromolecules on the surface of pathogenic bacteria form a defensive barrier against the host's immune system. The overall charge of these molecules may act to alter binding of cationic antimicrobial peptides (Devine, 2003).

The aim of this study was to examine the nature of the interaction between the periodontal pathogen P. gingivalis and the multifunctional peptides adrenomedullin and CGRP. A possible role of the extracellular arginine/lysine proteases and cell surface components (capsule and anionic polysaccharide) in this interaction was further investigated using wild-type and isogenic mutant strains.

Materials and methods

Organisms

Wild-type P. gingivalis strains W83 and W50 (K1 serotype) were used; these have been extensively characterized (Lamont & Jenkinson, 1998). Gene knockout P. gingivalis strains from the parent W50: K1A (kgp mutant/lacking lysine-x-specific activity), E8 (rgpA rgpB double mutant/lacking arginine-x-specific activity), polysaccharide and capsular mutants (Table 2) were kindly supplied by Prof. M.A. Curtis (Molecular Pathogenesis Group, Queen Mary, University of London). Escherichia coli BUE55 was used as the indicator organism to assess antimicrobial activity in P. gingivalis culture supernatant and peptide mixtures. This strain was originally isolated because of its increased sensitivity to polymyxin B and has been used in other studies of antimicrobial peptides (Moore et al., 1996).

Table 2. Porphyromonas gingivalis polysaccharide and capsular gene knockout strains
StrainGenotype*Notes
P. gingivalis W50 (K1 serotype)Wild-type 
porR region
 P. gingivalis porRpg1138 (porR)Reduced protease activity, loss of cell-surface anionic polysaccharide, inability to pigment
 P. gingivalis Beg6Δpg1135-pg1141 (including porR)Deletion of complete porR locus. Defects as above
Capsule region
 P. gingivalis GPAΔpg0116-pg0120Region encodes 3′-end of capsule locus. Capsule loss
 P. gingivalis GPBpg0117Gene encodes carbohydrate translocase of capsule locus
 P. gingivalis GPCΔpg0109-pg0118Complete deletion of capsule locus
 P. gingivaliswbpBpg2119Defect led to reduced protease activity and distribution, loss of cell-surface anionic polysaccharide, slow rate of pigmentation
 P. gingivalis 381 (K-, nonencapsulated)Wild-type 

Peptides

Synthetic adrenomedullin and adrenomedullin fragments (residues 1–12, 1–21, 13–52, 16–21, 16–52, 22–52, 26–52, and 34–52) and α-CGRP (ACDTATCVTHRLAGLLSRSGGVVKNNFVPTNVGSKAF) were obtained from Phoenix Pharmaceuticals (Karlsruhe, Germany). Peptides and fragments were made up at the appropriate concentrations in phosphate-buffered saline and stored at −20°C until use.

Adrenomedullin fragment positions in the adrenomedullin molecule are shown as follows: YRQSMNNFQGLR12S13FGC16RFGTC21T22VQKL26AHQIYQFT34DKDKDNVAPRSKISPQGY52

Determination of peptide activity

Growth inhibition assays

The broth microdilution assay used to determine growth inhibition (Allaker et al., 2006) was based upon the method used by Devine (Devine et al., 1999) with P. gingivalis and cationic peptides. The ability of adrenomedullin, CGRP and adrenomedullin fragments to inhibit the growth of wild-type and isogenic mutant strains of P. gingivalis was tested. With P. gingivalis strains all steps were carried out in an anaerobic work station (80% N2 v/v; 10% CO2 v/v; 10% H2 v/v). Serial 1 : 2 dilutions of peptides and fragments were made in prereduced brain heart infusion broth (BHI; Oxoid CM225) with haemin (1 mg L−1) and menadione (0.5 mg L−1). Wells were inoculated with a 1 : 10 dilution of an overnight bacterial culture (final concentration in inoculum of c. 5 × 105 CFU mL−1). After anaerobic incubation at 37°C for 24 h, growth was monitored spectrophotometrically at OD540 nm using a microplate reader (Anthos HT III). Positive controls included wells with E. coli BUE55, and negative controls without either peptide or bacteria.

The ability of P. gingivalis (wild-type and protease gene knockout) culture supernatant and peptide (adrenomedullin and CGRP) mixtures (see below) to inhibit the growth of E. coli BUE55 was also determined in the broth microdilution assay. Serial 1 : 2 dilutions of peptides were made in isosensitest broth (ISB; Oxoid CM473). Wells were inoculated with a 1 : 10 dilution of an overnight bacterial culture (final concentration in inoculum of c. 5 × 105 CFU mL−1). After incubation at 37°C for 24 h in air with 5% CO2, minimal inhibitory concentrations (MIC) were recorded as the lowest concentration of peptide inhibiting growth as measured at OD540 nm. Controls with P. gingivalis culture supernatant, without peptide or bacteria were included.

Porphyromonas gingivalis W50 was exposed to adrenomedullin (100 μg mL−1) for 24 h at 37°C under anaerobic conditions as described above. The culture was then centrifuged (10 000 g for 5 min) and the supernatant filter-sterilized. This was then tested for activity against E. coli BUE55 in the growth inhibition assay as described above.

Double-layer diffusion activity assay

Antimicrobial activity of wild-type P. gingivalis culture supernatant and adrenomedullin mixtures (see below) was also assessed by determining the ability to kill E. coli BUE55 in a sensitive double-layer agarose diffusion assay. This was based upon the method used by Devine (Devine et al., 1999) with P. gingivalis and cationic peptides. Escherichia coli was grown in ISB to exponential phase, and the suspension adjusted to an OD540 nm corresponding to c. 2 × 106 CFU mL−1. After inoculation of molten (held at 50°C) medium with 100 μL E. coli suspension, plates were poured containing 12 mL half-strength ISB plus 1% (w/v) low-electroendosmotic agarose (Sigma) and 0.02% (v/v) Tween 20 (Sigma). Peptide preparations (2 μL) were added to 1.5-mm-diameter wells (cut with sterile Pasteur pipettes). Plates were then incubated at 37°C for 3 h in air with 5% CO2. Finally, 12 mL of molten (50°C) double-strength ISB containing 1% (w/v) agarose was overlaid onto the plates. Plates were then incubated at 37°C for 24 h in air with 5% CO2.

Interaction of peptides with bacterial culture supernatants

Porphyromonas gingivalis wild-type (W50 and W83) and protease gene knockout (E8 and K1A) strains were grown in BHI broth with haemin and menadione under identical conditions to achieve maximal Arg-X and Lys-X active protease in wild-type strains as measured using chromogenic substrates (Devine et al., 1999). Cultures at equivalent cell densities were centrifuged (10 000 g for 5 min) and equal volumes of filter-sterilized culture supernatants and peptide solution (100 μg mL−1) were mixed and incubated at 37°C for 24 h under reduced conditions. Antimicrobial activity of these combinations was then tested by broth microdilution and double-layer diffusion assays as described above.

Results

Growth inhibition of wild-type P. gingivalis by adrenomedullin and CGRP

No significant growth inhibition of wild-type P. gingivalis W50 and W83, as determined using the broth microdilution assay (Fig. 2; n=7), with either adrenomedullin or CGRP was demonstrated. No effect of increasing concentration over the range 0.24–250 μg mL−1 was observed. Positive controls demonstrated marked inhibition of E. coli BUE55 in the presence of BHI broth with haemin and menadione (MIC=6.25 μg mL−1).

Figure 2.

 Growth inhibition tests with Porphyromonas gingivalis W50 and W83, as measured by OD in a broth microdilution assay, with decreasing concentrations (μg mL−1) of adrenomedullin and CGRP (mean±SE; n=7).

Growth inhibition of gene knockout strains of P. gingivalis by adrenomedullin

In a similar manner to wild-type strains, no significant growth inhibition of P. gingivalis protease (E8), polysaccharide (porR and Beg6) and capsular (GPB, GPC and wbpB) gene knockout strains, and wild type 381 (K-, nonencapsulated) strains were observed in the presence of adrenomedullin (12 different concentrations of peptide; range 0.24–250 μg mL−1; n=3). At the maximum concentration tested (250 μg mL−1) significant growth inhibition (17% reduction; P<0.01, Student's t-test) of the capsule-deficient mutant GPA was observed. Growth yields (OD540 nm; mean±SEM) for P. gingivalis strains in the presence of the maximum concentration of adrenomedullin (250 μg mL−1) ranged from a minimum of 0.559±0.011 with the gene knockout Beg6 strain to a maximum of 1.087±0.025 with the wild-type W50 strain. The respective growth yields in the absence of adrenomedullin were 0.583±0.009 and 1.114±0.011. Corresponding time zero growth yields (OD540 nm) were <0.05 for all peptide and P. gingivalis combinations.

Growth inhibition of wild-type P. gingivalis by adrenomedullin fragments

Apart from a small reduction with the carboxy-terminal fragment (34–52) all fragments were unable to reduce the growth of P. gingivalis W50. At the maximum concentration tested (25 μg mL−1) growth inhibition (4% reduction; P<0.05, Student's t-test) of P. gingivalis W50 with fragment 34–52 was detected.

Interaction of adrenomedullin with wild-type and protease knockout P. gingivalis cultures and culture supernatants

No difference in the ability to inhibit the growth of E. coli BUE55 was observed between adrenomedullin incubated with culture supernatants from P. gingivalis wild-type strains (W50 and W83) and those derived from protease gene knockout strains (E8 and K1A) (Fig. 3). No significant difference was observed between the arginine (no Arg-X activity) and lysine (no Lys-X activity) mutants. A similar result was obtained with CGRP (Fig. 4). In all cases, E. coli MIC values were 6.25 μg mL−1. In the absence of P. gingivalis culture supernatant the growth of E. coli BUE55 in ISB broth was inhibited by adrenomedullin and CGRP (Fig. 5). In this case, the lowest concentration of both peptides able to prevent bacterial growth was 6.25 μg mL−1.

Figure 3.

 Growth inhibition of Escherichia coli BUE55, as measured by OD in a broth microdilution assay, with decreasing concentrations (μg mL−1) of adrenomedullin incubated with culture supernatants from Porphyromonas gingivalis W50, W83, E8 (arginine protease gene knockout) or K1A (lysine protease gene knockout) (mean±SE; n=5).

Figure 4.

 Growth inhibition of Escherichia coli BUE55, as measured by OD in a broth microdilution assay, with decreasing concentrations (μg mL−1) of CGRP incubated with culture supernatants from Porphyromonas gingivalis W50, W83, E8 (arginine protease gene knockout) or K1A (lysine protease gene knockout) (mean±SE; n=5).

Figure 5.

 Growth inhibition of Escherichia coli BUE55, as measured by OD in a broth microdilution assay, with decreasing concentrations (μg mL−1) of adrenomedullin and CGRP (mean±SE; n=9).

No significant difference (Student's t-test) between adrenomedullin alone and adrenomedullin incubated with culture supernatants from P. gingivalis W50 or W83 was shown using the sensitive double-layer diffusion assay (n=5) against the indicator organism E. coli BUE55. The mean diameter of growth inhibition zones (±SEM) was 11.96±0.14, 11.7±0.20 and 11.67±0.20 mm, respectively.

Supernatant from a P. gingivalis W50 culture exposed to adrenomedullin (100 μg mL−1) for 24 h and then tested for activity against E. coli BUE55 in the broth microdilution assay demonstrated an MIC value of 6.25 μg mL−1, i.e. no decrease in adrenomedullin activity following exposure to P. gingivalis.

Discussion

This study demonstrated that P. gingivalis is able to resist the antimicrobial action of the peptides adrenomedullin and CGRP. This organism has been shown to be able to resist the action of a range of antimicrobial peptides from nonhuman sources through the cleavage of these peptides by its cysteine proteases (Devine et al., 1999). However, P. gingivalis W50 has been shown to be sensitive to the silk moth cationic peptide cecropin B and yet able to inactivate the peptide through the action of its proteases. The lack of protection by proteases in this case may be explained by the slow activity of the enzymes in comparison with the action of the peptide (Devine et al., 1999). Recently, P. gingivalis has been shown to be able to resist the antimicrobial action of a number of human antimicrobial peptides, including cathelicidins and β-defensins, using both killing and growth inhibition assays, respectively (Guthmiller et al., 2001; Joly et al., 2004). The mechanism of resistance in these cases has yet to be determined. Digestion of the adrenomedullin and CGRP peptides by extracellular proteases does not appear to be a significant resistance mechanism in P. gingivalis, as demonstrated with culture supernatant and peptide combinations in this study. This was further supported by the finding that no decrease in the ability of adrenomedullin to inhibit the growth of E. coli was observed after exposure of adrenomedullin to cultures of P. gingivalis. HPLC-MS analysis has also provided additional evidence that exposure of adrenomedullin to P. gingivalis proteolytic enzymes does not result in the production of inactive breakdown products (R.P. Allaker, unpublished observations).

Resistance to antimicrobial β-defensins and other peptides may be due to altered outer membrane proteins or lipopolysaccharide structures. For example, Treponema denticola, which lacks a traditional lipopolysaccharide, is naturally resistant to human β-defensin 2 (Brissette & Lukehart, 2002). Lipopolysaccharide variation between strains of the same species could partly explain the variable susceptibility pattern observed. Studies of lipopolysaccharide and antimicrobial peptides have highlighted the importance of the lipid A and core oligosaccharide in resistance, but in some bacteria (e.g. Prevotella and Porphyromonas spp.) the O-polysaccharide may be significant (Kirikae et al., 1999). Porphyromonas gingivalis possesses a highly unusual O-polysaccharide; 60% of the α-rhamnose residues in the repeating unit are phosphorylated due to the addition of phosphoethanolamine (Paramonov et al., 2001). As a result the increased negative charge of the lipopolysaccharide may increase binding of the cationic adrenomedullin molecule; this may then contribute to resistance by the prevention of access of adrenomedullin to the core oligosaccharide and lipid A. In addition, the binding to the O-polysaccharide could influence the likelihood of inactivation of adrenomedullin by membrane-bound proteases. However, this hypothesis was not supported by the finding that a gene knockout strain of P. gingivalis (Beg6) lacking in surface O-polysaccharide was resistant to growth inhibition with adrenomedullin. Strains of the obligate anaerobic species Prevotella intermedia, Prevotella nigrescens, Prevotella endodontalis and Prevotella denticola have also been shown to be resistant to concentrations of adrenomedullin up to 250 μg mL−1 (R.P. Allaker, unpublished observations). However, as with P. gingivalis certain strains of Prevotella nigrescens have been shown to be sensitive to adrenomedullin. Further studies to examine the binding affinity of these peptides to the lipopolysaccharide or its components from P. gingivalis and Prevotella spp. may help to explain the differences in sensitivity. In the current study, growth of the P. gingivalis GPA mutant, which does not possess a capsule, was slightly inhibited. However, growth of the nonencapsulated P. gingivalis 381 was not inhibited. It therefore does appear that the presence of a capsule is not significant. The precise role of binding of antimicrobial peptides to surface anionic molecules, in terms of a decrease or enhancement of subsequent activity, remains to be determined.

The ability of human β-defensins 2 (HBD-2) and 3 (HBD-3) to inhibit the growth of oral microorganisms has demonstrated strain-selective activity (Joly et al., 2004). Obligate anaerobic species, including P. gingivalis and other opportunistic oral pathogens, were shown to include resistant strains. For example, the nonencapsulated P. gingivalis ATCC 33277 was shown to be sensitive to HBD-2 and HBD-3 whereas W50 (K1 capsule serotype) was resistant (MIC>250 μg mL−1). In a previous study (Allaker & Kapas, 2003) we have shown P. gingivalis ATCC 33277 to be sensitive to adrenomedullin (MIC <0.1 μg mL−1), although no sensitivity of this strain to the shorter CGRP peptide was demonstrated (Zihni et al., 1999). It should be noted that the kinetics of P. gingivalis growth during a challenge with antimicrobial peptides can be variable with respect to growth parameters and the method of measurement used (Shelburne et al., 2005). Indeed, with P. gingivalis ATCC 49417 a radial diffusion assay demonstrated this strain to be resistant (MIC>250 μg mL−1) (Joly et al., 2004) whereas using a microtitre method and a DNA binding dye to assess growth inhibition this strain was recorded as sensitive (MIC of 10 μg mL−1) (Shelburne et al., 2005).

Specific physicochemical properties of peptides may render them unable to inhibit the growth of given microorganisms (Yeaman & Yount, 2003). Internal disulphide bonds between residues 16 and 21 and the occurrence of multimeric structures may be responsible in the case of adrenomedullin and P. gingivalis. Adrenomedullin fragments were therefore tested against P. gingivalis. A significant structure – function relationship has previously been found between the antimicrobial effect of adrenomedullin against E. coli and the carboxyl-terminal portion of the peptide, in which carboxy-terminal fragments were shown to be up to 250-fold more active than the parent molecule (Allaker et al., 2006). However, apart from a small reduction in growth with the carboxy-terminal fragment (residues 34–52), these were shown to be inactive against P. gingivalis.

It is generally agreed that the adrenomedullin concentration required to inhibit bacterial growth is in excess of the levels measured in plasma (Hinson et al., 2000). However, it is not inconceivable under certain circumstances, such as localized inflammatory conditions, that elevated levels may be effective. In terms of interaction with other immune components, the effect on the activity, or lack of activity, of adrenomedullin on P. gingivalis has yet to be established. However, it is known that normal human serum (45%, v/v) has no effect on the ability of culture supernatants from P. gingivalis to inactivate the antimicrobial peptide cecropin B. This peptide was also shown to be antimicrobial in the presence of 5–45% human serum (Devine et al., 1999). In periodontitis, a chronic inflammatory disease of the supporting structures of the teeth, adrenomedullin concentrations found in gingival crevicular fluid have been measured in the range 1–2 μg mL−1 (Lundy et al., 2006). This lends support to the proposed role for adrenomedullin in the protection against periodontal pathogens, including P. gingivalis. However, the prevalence, persistence and possible in vivo significance of the resistance demonstrated by this bacterium to adrenomedullin, CGRP and other antimicrobial peptides expressed within the oral cavity remains to be fully evaluated.

Acknowledgements

This work was funded by St Bartholomew's and the Royal London Charitable Foundation.

Ancillary