doi:10.1111/iej.12501
Strains of Enterococcus faecalis differ in their ability to coexist in biofilms with other root canal bacteria
vez de Paz1, J. R. Davies1, G. Bergenholtz2 & G. Svensa €ter1 L. E. Cha 1
Department of Oral Biology, Faculty of Odontology, Malm€ o University, Malm€ o; and 2Department of Endodontology, The Sahlgrenska Academy at Gothenburg University, Gothenburg, Sweden
Abstract vez de Paz LE, Davies JR, Bergenholtz G, Cha €ter G. Strains of Enterococcus faecalis differ in their Svensa ability to coexist in biofilms with other root canal bacteria. International Endodontic Journal, 48, 916–925, 2015.
Aim To investigate the relationship between protease production and the ability of Enterococcus faecalis strains to coexist in biofilms with other bacteria commonly recovered from infected root canals. Methodology Biofilms with bacteria in mono-, dual- and four-species communities were developed in flow chambers. The organisms used were Lactobacillus salivarius, Streptococcus gordonii and Actinomyces naeslundii and E. faecalis strains, GUL1 and OG1RF. Biovolume and species distribution were examined using 16S rRNA fluorescence in situ hybridization in combination with confocal microscopy and image analysis. The full proteome of the E. faecalis strains was studied using two-dimensional gel electrophoresis. Spots of interest were identified using tandem mass spectroscopy and quantified using Delta 2D software.
Introduction When the enamel/dentine barrier is breached, oral microorganisms gain access to the root canal where they colonize the surfaces of dentine or necrotic tissue, giving rise to aggregates known as microbial
Correspondence: Gunnel Svens€ ater, Department of Oral Biology, Faculty of Odontology, Malm€ o University, Malm€ o S-20506, Sweden (Tel.: +46 406659495; fax +46 406659484; e-mail: [email protected]).
916
International Endodontic Journal, 48, 916–925, 2015
Results All bacteria formed biofilms and an ANOVA analysis revealed that the biofilm biomass increased significantly (P ≤ 0.01) between 6 and 24 h. L. salivarius, S. gordonii and A. naeslundii formed mutualistic biofilm communities, and this pattern was unchanged when E. faecalis GUL1 was included in the consortium. However, with OG1RF, L. salivarius and S. gordonii were outcompeted in a 24-h biofilm. Proteomic analysis revealed that OG1RF secreted higher levels of proteases, GelE (P = 0.02) and SprE (P = 0.002) and a previously unidentified serine protease (P = 0.05), than GUL1. Conclusions Different strains of E. faecalis can interact synergistically or antagonistically with a consortium of root canal bacteria. A possible mechanism underlying this, as well as potential differences in virulence, is production of different levels of proteases, which can cause detachment of neighbouring bacteria and tissue damage. Keywords: antagonism, apical periodontitis, bacterial consortium, multispecies, mutualism. Received 18 September 2014; accepted 5 July 2015
biofilms (Ricucci & Siqueira 2010). In general, bacteria in biofilms have the important and often underappreciated characteristic of being very resistant to a variety of antimicrobial substances, including antibiotics (Gilbert et al. 1997). This property has, in part, been attributed to the presence of a polymeric matrix. Biofilm bacteria are also intrinsically less susceptible to antimicrobial substances due to their slow rate of growth and adoption of a distinct biofilm phenotype, which differs from that of their planktonic counterparts (Costerton et al. 1999).
© 2015 International Endodontic Journal. Published by John Wiley & Sons Ltd
vez de Paz et al. Coexistence of E. faecalis in multispecies biofilms Cha
Root canals of untreated teeth with necrotic pulp tissue support growth of different bacterial consortia. In a study by Sundqvist (1994), the most commonly isolated genera were Fusobacterium and streptococci with Bacteroides spp, Parvimonas spp, Lactobacillus spp, Actinomycetes spp, Veillonella parvula and a series of gram-negative anaerobic rods also present. However, the microbiota of persistent root canal infections, including re-treated cases of teeth with apical periodontitis, differs significantly from that of untreated teeth (Molander et al. 1998, Siqueira et al. 2004). The species diversity is severely restricted, and monocultures, often of E. faecalis, are common (Sundqvist et al. 1998). This finding suggests that root canal treatment causes selection from a multispecies community towards biofilms comprised of one or a few species, often including Enterococcus faecalis. One mechanism behind the strong presence of E. faecalis seems to be its proven resistance to antimicrobial endodontic medicaments (Dahlen et al. 2000, Evans et al. 2002, Portenier et al. 2005), including calcium hydroxide-based sealers (Kayaoglu et al. 2005) and the harsh nutritional conditions that often prevail in canals of root filled teeth (Figdor et al. 2003, Portenier et al. 2005). In biofilm communities such as those found in root canals of infected teeth, the bacteria often exist in close proximity to each other facilitating a range of bacterial interactions. These exchanges can be physical, such as co-adhesion and co-aggregation, or chemical via diffusible signalling molecules and transfer of genetic material. Bacteria in biofilms produce batteries of complementary enzymes, which can act synergistically to effectively exploit complex molecules, thus providing nutrients to the bacteria in the local environment (Marsh 2005). In addition to cooperative interactions, competitive interactions such as the production of bacteriocins and inhibitory metabolic end products can also occur in biofilms. The most studied inhibitory substances include hydrogen peroxide, short chain fatty acids and lantibiotics from Lactobacilli (Hoyo et al. 2009). Thus, the bacterial consortia colonizing the root canal will be determined not only by the species which originally gain entry, but also by synergistic and antagonistic interactions between biofilm bacteria. Clinical studies have shown that E. faecalis is a component of multispecies bacterial communities in infected root canals but may also be found as the sole species, indicating that E. faecalis is capable of outcompeting other members of a polymicrobial infection
© 2015 International Endodontic Journal. Published by John Wiley & Sons Ltd
(Sundqvist et al. 1998). It was hypothesized that different strains of E. faecalis differ in their ability to coexist with other root canal bacteria and that this is related to their protease production in biofilm environments. This may have important implications for differences in virulence between different E. faecalis strains and thus the clinical outcome of apical parodontitis. To explore interactions between root canal bacteria, this study investigated how two strains of E. faecalis (OGRIF – a strain commonly used in endodontic research and GUL1 – a clinical strain isolated from a root canal) can coexist with other bacteria recovered from infected root canals viz. Lactobacillus salivarius, Streptococcus gordonii and Actinomyces naeslundii.
Materials and methods Bacterial strains and culture conditions The bacterial strains, E. faecalis (GUL1), L. salivarius, S. gordonii and A. naeslundii, were isolates from root canals undergoing endodontic treatment presenting with persistent root canal infections (Ch avez de Paz et al. 2003, 2004). E. faecalis (GUL1) was routinely identified by colony morphology and growth on selective tellurite medium. E. faecalis strain OG1RF was originally derived from strain OG1 (formerly designated 2SAR) (Gold et al. 1975). All strains were routinely cultured in blood agar and Bacto Todd–Hewitt broth (TH; Becton Dickinson & Co, Franklin Lakes, NJ, USA) in an atmosphere of 5% CO2 at 37 °C for 24 h.
Biofilm growth Mid-exponential bacterial cultures were created by mixing 0.5 mL of overnight culture with 5 mL fresh TH medium and incubating in an atmosphere of 5% CO2 at 37 °C for 5 h. The optical density at 600 nm was adjusted to 0.6 (approximately 1 9 108 mL 1 cells), and mini flow chambers (l-Slide VI for Live Cell Analysis; Integrated BioDiagnostics, Martinsried, Germany) with hydrophilic uncoated polystyrene surfaces were inoculated with one, two, three or four species according to the outline presented in Fig. 1. Two-, three- and four-species inocula were prepared by mixing 0.5 mL of appropriate combinations of E. faecalis, L. salivarius, S. gordonii and A. naeslundii (equivalent to about 1 9 107 mL 1 cells of each species) (18). Each mini flow chamber was inoculated with 30 lL
International Endodontic Journal, 48, 916–925, 2015
917
vez de Paz et al. Coexistence of E. faecalis in multispecies biofilms Cha
Species
Dual-species biofilms
Three-species biofilm Four-species biofilms
Ef-GUL1
Ef-GUL1
Ef-GUL1 Ls + Ls An
+ An
+ Sg
Ef-OG1RF
+Ls + An +Sg Ls + An + Sg
Ef-OG1RF
Sg Ef-OG1RF + Ls
+ An
+ Sg
+Ls + An +Sg
Figure 1 Outline of the biofilm experiments. The species used were as follows: Enterococcus faecalis strains GUL1 and OG1RF, Lactobacillus salivarius (Ls), Actinomyces naeslundii (An) and Streptococcus gordonii (Sg).
of the bacterial suspension followed by addition of 100 lL of fresh TH to give a final volume of 130 lL per flow chamber. Flow chambers were then incubated under static conditions in an atmosphere of 5% CO2 in air at 37 °C for 6 or 24 h.
Species analysis using 16S rRNA-FISH Bacteria in the biofilms were identified using 16S rRNA-FISH probes (Biomers, Ulm, Germany). E. faecalis was identified using the ENF191 probe (gaaagcgcctttcactcttatgc) labelled with Pacific Blue (kex 410 nm, kem 455 nm), ATTO-488 (kex 504 nm, kem 521 nm) or ATTO-565 (kex 565 nm, kem 590 nm); S. gordonii was identified using the STR405 probe (tagccgtccctttctggt) labelled with ATTO-488 (green); L. salivarius was identified using the LAC722 probe (caccgctacacatgragttccact) labelled with ATTO565 (red); and A. naeslundii was identified using the JF201 probe (gctaccgtcaacccaccc) labelled with Pacific Blue (blue). The 16S rRNA-FISH protocol has been described previously (Ch avez de Paz 2012). Briefly, after 6- or 24-h biofilm cells were fixed with 4% paraformaldehyde in phosphate-buffered saline and permeabilized using a lysozyme solution (10 mg mL 1; Sigma-Aldrich Corporation, St. Louis, MO, USA). The biofilms were then washed with ultrapure water and dehydrated with 50%, 80% and 99% ethanol for 3 min, respectively, after which the flow chambers were inoculated with 30 mL of hybridization buffer [0.9 M NaCl, 20 mM Tris-HCl buffer, pH 7.5, with 0.01% sodium dodecyl sulphate (SDS) and 25% formamide] containing 20 ng mL 1 of the oligonucleotide probes. Probe cocktails containing 20 ng mL 1 of each oligonucleotide were used, and
918
International Endodontic Journal, 48, 916–925, 2015
hybridization was carried out at 47 °C for 3 h in a humid chamber.
Confocal laser scanning microscopy All microscopy was performed using an Eclipse TE2000 inverted confocal scanning laser microscope (Nikon Corporation, Tokyo, Japan). By means of a motorized stage, ten areas within each sample were selected at random and image stack sections composed of ten images, each taken with a variation of 2 lm along the z-position and covering an area of 0.05 mm2 per field of view, were taken. Confocal illumination for the green fluorescence signal was provided by an Ar laser (488 nm laser excitation). For detection of the red fluorescence, a G-HeNe laser was used (543 nm laser excitation), and for the detection of blue fluorescence, a UV laser was used (390 nm laser excitation).
Image analysis and statistical evaluation Confocal laser scanner microscopy (CSLM) images were analysed using software bio-Image_L (Ch avez de Paz 2009). The biovolume represents the overall volume of the biofilm. The results for biofilm growth over time are presented as means from a total of 30 biofilm stacks, that is 10 stacks/flow-chamber channel in three independent experiments. To evaluate the relative proportions of species in dual-species biofilms formed between the two strains of E. faecalis (violet) and L. salivarius (red), A. naeslundii (blue) or S. gordonii (green), the independent subpopulations represented by the different fluorescent colours were evaluated and the percentage of the total biofilm
© 2015 International Endodontic Journal. Published by John Wiley & Sons Ltd
vez de Paz et al. Coexistence of E. faecalis in multispecies biofilms Cha
Two-dimensional gel electrophoresis and protein identification Cell wall proteins and whole cell extracts of GUL1 and OG1RF were prepared as described previously (Davies et al. 2009, Dorkhan et al. 2012). For isolation of extracellular proteins, the culture medium was harvested from batch cultures of cells in exponential growth phase by centrifugation and proteins precipitated overnight at 4 °C using 10% trichloroacetic acid (Sigma-Aldrich Corporation). After centrifugation (16 000 g, 30 min, 4 °C), the pellet was re-suspended in acetone and sonicated for 3 9 10 s. Samples were then re-centrifuged as above, and the resulting pellet dissolved in 500 mL of two-dimensional gel electrophoresis (2DE) rehydration buffer (8 M urea, 2% CHAPS, 10 mM DTT, 2% IPG buffer). The protein concentration was determined using a 2D Quant kit (GE Healthcare Life Sciences, Pittsberg, PA, USA), and a volume corresponding to 50 mg protein (for silver-stained gels) or 150 mg protein (for Coomassie-stained gels) was subjected to isoelectric focusing as described previously (Davies et al. 2009). Cell wall proteins were subjected to gel electrophoresis on 7% SDS-PAGE gels, whereas whole cell extracts and extracellular proteins were separated on 14% SDSPAGE gels. SDS-PAGE was performed at a constant current of 15 mA per gel overnight at 10 °C in a PROTEAN II xi cell (Bio-Rad, Hercules, CA, USA). Rainbow molecular mass standards (GE Healthcare Life Sciences) were run on the acidic side of the IPG strips. Gels were stained with Coomassie brilliant blue or silver according to the relevant protocol from GE Healthcare Life Sciences. All gels were run in triplicate. Silver-stained gels were scanned and the integrated optical intensity (IOD) within the spot boundaries of the major differently expressed proteins determined using DECODON Delta-2D software (Greifswald, Germany). For identification, the differently expressed proteins were excised manually from Coomassie brilliant blue-stained gels and subjected to liquid chromatography–tandem mass spectrometry (LC-MS/MS) as described previously (Davies et al. 2009). Briefly, proteins were reduced with DTT (GE Healthcare Life Sciences; 60 °C, 20 min), alkylated with iodoacetamide
© 2015 International Endodontic Journal. Published by John Wiley & Sons Ltd
(GE Healthcare Life Sciences; 25 °C, 10 min) and then digested with trypsin (37 °C, 8 h). Tryptic peptides were separated and subjected to MS/MS. Peptide peaks were deconvoluted automatically and mass lists in the form of Mascot Generic Files used as the input for Mascot MS/MS Ions searches of the NCBInr database using the Matrix Science web server (www.matrixscience.com).
Statistical analysis Statistical analysis of biofilm formation over 6 or 24 h was performed with the SYSTAT 13 software package using two-way ANOVA with a Tukey post hoc test. IOD values from the 2D gels from the major differentially expressed proteins in strains OG1RF and GUL1 were transferred to Prism 5 (GraphPad, La Jolla, CA, USA). Mean and standard errors were calculated, and the values for each spot compared using Student’s t-test.
Results Ability of the tested organisms to form monospecies biofilms Both strain GUL1 and strain OG1RF adhered to the flow-cell surfaces after 6 h (Fig. 2, white bars), and after 24 h of incubation (Fig. 2, grey bars), the biovolumes had increased significantly (P = 0.015 for
Biofilm biovolume m3 x104
biovolume represented by E. faecalis calculated for 30 biofilm stacks and expressed as the mean SE. 3D reconstructions were produced using the function ‘Structure and Distribution’ in bioImage_L FISH.
8
* *
6
6h 24 h
*
*
4
2
0
GUL1
OG1RF
Ls
An
Sg
Figure 2 Formation of mono-species biofilms in mini flow chambers over time. The bars represent the mean biovolumes of the biofilms formed by Enterococcus faecalis strains GUL1 and OG1RF, as well as Lactobacillus salivarius, Actinomyces naeslundii and Streptococcus gordonii, after 6 or 24 h. The error bars show standard errors calculated from three independent experiments and *P ≤ 0.01.
International Endodontic Journal, 48, 916–925, 2015
919
vez de Paz et al. Coexistence of E. faecalis in multispecies biofilms Cha
P = 0.002, respectively), whereas for A. naeslundii, the 1.4-fold increase seen was not significant (P = 0.25). At 24 h, the greatest biovolume was registered for S. gordonii.
GUL1 and P < 0.001 for OG1RF). Although the biofilms of OG1RF at 24 h contained a greater biovolume than those of GUL1, this difference was not significant (P = 0.28). When the same procedure was undertaken with the clinical isolates of L. salivarius, A. naeslundii and S. gordonii, all showed attachment to the flow-chamber surface at 6 h (Fig. 2, white bars) and after 24 h of incubation, the biovolume increased in all three organisms (Fig. 2, grey bars). For S. gordonii and L. salivarius, the twofold increase in biomass was significant (P = 0.005 and
(a)
(b)
Co-cultivation of the E. faecalis strains with L. salivarius, S. gordonii or A. naeslundii for 24 h resulted in dual-species biofilms (Fig. 3). 16S rRNA-FISH revealed
E. faecalis A. naeslundii
(c)
E. faecalis S. gordonii
OG1RF
GUL1
E. faecalis L. salivarius
Coexistence of E. faecalis strains with other root canal bacteria
E. faecalis (%)
100
+ L. salivarius
100
+ A. naeslundii
100
80
80
80
60
60
60
40
40
40
20
20
20
0
GUL1
OG1RF
0
GUL1
OG1RF
0
+ S. gordonii
GUL1
OG1RF
Figure 3 Relative proportions of species in dual-species biofilms formed between the two strains of Enterococcus faecalis and Lactobacillus salivarius, Actinomyces naeslundii and Streptococcus gordonii. The images show 24-h biofilms stained with species-specific oligonucleotides targeting the 16S rRNA gene. In (a), L. salivarius is visualized in red and E. faecalis is green, and in (b), E. faecalis is visualized in red and A. naeslundii in blue, and in (c), S. gordonii is shown green and E. faecalis in blue. The bar represents 5 lm. The box plots show the mean SE for the percentage of biovolume occupied by E. faecalis strain GUL 1 or OG1RF in each biofilm. All experiments were carried out in triplicate.
920
International Endodontic Journal, 48, 916–925, 2015
© 2015 International Endodontic Journal. Published by John Wiley & Sons Ltd
vez de Paz et al. Coexistence of E. faecalis in multispecies biofilms Cha
the two cell types in each biofilm (E. faecalis and L. salivarius, E. faecalis and A. naeslundii or E. faecalis and S. gordonii) in close proximity to each other. Initially, the two species in each experiment were inoculated in equal numbers, that is the proportion of E. faecalis (GUL1 or OG1RF) in each case was approximately 50%. In dual-species biofilms with L. salivarius, the proportion of GUL 1 after 24 h remained at 50 10% (Fig. 3a). However, when OG1RF was used, the level of this strain in the biofilm increased to 75 5% after 24 h suggesting that this strain outcompeted L. salivarius during biofilm formation. A similar result was seen with S. gordonii where the proportion of GUL 1 was 38 5% after 24 h compared to 66 10% for strain OG1RF (Fig. 3c). The A. naeslundii strain used here appeared to be unaffected by the presence of either the GUL 1 or OG1RF E. faecalis strain which showed levels of 38 5% and 42 5% of the total biofilm biomass, respectively, after 24 h (Fig. 3b). These data thus suggest that E. faecalis (a)
(b) A. naeslundii L. salivarius S. gordonii
strain OG1RF was more competitive at colonizing the surface and establishing a biofilm in the presence of L. salivarius and S. gordonii than GUL1. Thus, the E. faecalis strains tested appear to show different properties; GUL1, exhibiting the capacity to coexist with other species in biofilms and strain OG1RF, which appears to be able to supplant other species.
Behaviour of E. faecalis strains in four-species biofilms Before the introduction of E. faecalis, the ability of A. naeslundii, L. salivarius and S. gordonii to exist in multispecies biofilms was investigated. Equal numbers of cells of each species were inoculated into a mini flow chamber and allowed to form a biofilm community for 24 h. Evaluation of the proportions of each species revealed no significant change from the initial inoculate (Fig. 4a) which suggests a balanced
A. naeslundii L. salivarius S. gordonii E. faecalis (GUL1)
(c)
A. naeslundii L. salivarius S. gordonii E. faecalis (OG1RF)
3% 2%
Figure 4 CSLM micrographs of 24-h biofilm communities stained using 16S rRNA-FISH and pie charts depicting the species distribution in the three- and four-species communities. The three-species consortium (a) shows a homogenous distribution of Lactobacillus salivarius (red fluorescent probe), Streptococcus gordonii (green fluorescent probe) and Actinomyces naeslundii (blue fluorescent probe). The four-species consortium shown in (b), which contains Enterococcus faecalis strain GUL1 (violet), also showed a homogenous distribution of species. In the third consortium (c), which contained E. faecalis strain OG1RF (violet), an increase in E. faecalis at the expense of S. gordonii (green) and L. salivarius (red) was seen. The bar represents 5 lm.
© 2015 International Endodontic Journal. Published by John Wiley & Sons Ltd
International Endodontic Journal, 48, 916–925, 2015
921
vez de Paz et al. Coexistence of E. faecalis in multispecies biofilms Cha
colonization with no one species dominating over another. Four-species bacterial consortia were then established using equal numbers of L. salivarius, A. naeslundii, S. gordonii and E. faecalis strain GUL1 or OG1RF and inoculated into mini flow chambers. In the biofilm with strain GUL1, quantification of the proportions of each species revealed that L. salivarius represented 29%, A. naeslundii (24%), S. gordonii 31% and E. faecalis 16% (Fig. 4b). Image analysis revealed that the four species were homogeneously distributed throughout the biofilm. Thus, the proportions of the species in the biofilm did not differ significantly from those in the original inoculate (25% of each species) showing that the inclusion of E. faecalis GUL1 did not disturb the balanced relationship between the species in the biofilm community. In contrast, when E. faecalis strain OG1RF was introduced into the four-species consortium, the proportions of the different species present in the biofilm after 24 h changed dramatically. The proportion of E. faecalis had increased from around 25% in the original inoculate to 68 3% in the biofilm. In contrast, L. salivarius and S. gordonii were almost completely absent, suggesting that strain OG1RF had outcompeted these two species during biofilm formation. As seen in the dual-species biofilm, A. naeslundii was unaffected by the presence of strain OG1RF and the levels remained similar to those in the inoculate.
Proteome of E. faecalis strains To investigate possible mechanisms underlying the ability of strain OG1RF, but not GUL1, to outcompete other species during biofilm formation, the cellular, cell wall and exo-proteomes of the two strains in the exponential phase of growth were compared using 2DE. Only minor differences were seen between the whole cell extracts and cell wall proteins from the two strains, suggesting that they did not differ to any significant degree in this respect (data not shown). However, the exo-proteomes of the two strains showed three dominant differently expressed spots (Fig. 5) that mass spectrometric analysis revealed to correspond to a gelatinase and a serine protease similar to those identified previously in E. faecalis strain V583. The two spots corresponding to gelatinase (GelE) were significantly more expressed in strain OG1RF than in strain GUL1 (tenfold and 30-fold differences, respectively) whilst the serine protease SprE was six times more abundant (P = 0.002; Table 1). In addition, two spots corresponding to a novel serine protease with no sequence similarity to SprE were identified. These (serine protease a and serine protease b) were also significantly more highly expressed by strain OG1RF than by strain GUL1 [eightfold (P = 0.011) and 13-fold differences (0.046), respectively].
76
(a)
(b)
52
Serine protease (a)
38 Serine protease (b)
GelE(a)
24 SprE GelE (b)
17 12
Figure 5 Silver-stained 2DE gels showing the exo-proteome of Enterococcus faecalis strains. Proteins released from (a) E. faecalis strain OG1RG and (b) E. faecalis strain GUL1 were separated by isoelectric focusing (pI range 4–7) in the first dimension and by SDS-PAGE in 14% gels in the second dimension. The positions of the molecular mass markers are shown. The identities of the most prominent differentially regulated proteins are shown in (a).
922
International Endodontic Journal, 48, 916–925, 2015
© 2015 International Endodontic Journal. Published by John Wiley & Sons Ltd
vez de Paz et al. Coexistence of E. faecalis in multispecies biofilms Cha
Table 1 Identity of major differentially expressed exoproteins from two strains of Enterococcus faecalis
Protein identification
Mean IOD strain OG1RF
Serine protease a Serine protease b GelE 1a GelE 1b SprE
11 315 3451 1992 3026 1241
2027 1106 481 492 124
Mean IOD strain GUL1
P-value (Student’s t-test)
0.011 0.046 0.02 0.004 0.002
1373 258 202 108 217
942 158 0.57 8 87
IOD, integrated optical intensity.
Discussion Dental biofilms consisting of multispecies communities are the cause of dental disease; caries (Becker et al. 2002), periodontitis (Kumar et al. 2005) and endodontic infections (Brito et al. 2007). In multispecies biofilms, bacterial species interact co-operatively or competitively with other members of the community. Examples of mutualistic biofilm communities include those formed between Aggregatibacter actinomycetemcomitans, F. nucleatum and Veillonella spp (Periasamy & Kolenbrander 2009a) as well as between S. gordonii and the periodontal pathogen Porphyromonas gingivalis (Periasamy & Kolenbrander 2009b). In this study, the ability of three root canal bacteria (L. salivarius, A. naeslundii and S. gordonii) to form biofilm communities was investigated. The strains were all recovered from root canals after treatment suggesting that they had survived exposure to the antimicrobial agents used. After 24 h, the relative proportions of the three species were similar to those in the initial inoculate suggesting that they can coexist in biofilms. 16S rRNA-FISH in combination with CLSM showed that S. gordonii and A. naeslundii were in close proximity to each other suggesting some form of cell–cell interaction between these species. Co-adhesion between different species is a common phenomenon amongst early colonizers of the oral cavity and has been demonstrated for A. naeslundii and S. gordonii (Kolenbrander et al. 2010). In addition, these bacteria can participate in mutually beneficial metabolic interactions. For instance, S. gordonii is known to produce H2O2 that at high levels can lead to oxidative stress and cell damage. This effect can be attenuated by the release of catalase from A. naeslundii. Thus, A. naeslundii can rescue S. gordonii cells from the damaging effects of endogenous H2O2 production (Jakubovics et al. 2008). In contrast, L. salivarius showed no obvious cell–cell interaction with
© 2015 International Endodontic Journal. Published by John Wiley & Sons Ltd
S. gordonii or A. naeslundii and no mutualistic interactions have been described for these species. However, from the data presented here, it is evident that L. salivarius can coexist in biofilms with both S. gordonii and A. naeslundii. In common with L. salivarius, A. naeslundii and S. gordonii, E. faecalis is commonly isolated from treated root canals. However, in contrast to L. salivarius, A. naeslundii and S. gordonii, E. faecalis can be found as the only bacteria present (Sundqvist et al. 1998) suggesting that E. faecalis can outcompete other species in the root canal environment. In this study, it was evident that the two strains of E. faecalis differed in their capacity to coexist with L. salivarius, A. naeslundii and S. gordonii. Strain GUL1 was able to establish dual-species biofilms with L. salivarius, A. naeslundii or S. gordonii. In each case, the relative proportions were similar to those in the original suspension suggesting that there was no antagonistic relationship between the two members of the community. Close interspecies cell contact was most evident between GUL1 and A. naeslundii. As E. faecalis is known to produce H2O2 (Huycke et al. 2002), it cannot be excluded that a similar mutualistic relationship to that between A. naeslundii and S. gordonii may exist. In the four-species consortium after growth for 24 h, the biomass of each species was also approximately equal to that in the original inoculate. This illustrates that for GUL 1, mutualistic and/or neutralistic interactions dominate the community behaviour. In contrast, the presence of strain OG1RF in dual-species biofilms with L. salivarius and S. gordonii suppressed growth of the latter species, whereas biofilm growth of A. naeslundii appeared to be unaffected by the presence of OG1RF. This behaviour was confirmed in the four-species consortium where, after 24 h, OG1RF comprised 68% of the total biomass. These data suggest that this strain of E. faecalis acts antagonistically in relation to S. gordonii and L. salivarius, whereas interactions with A. naeslundii were mutualistic. Thus, the ability of E. faecalis to coexist with other potential members of root canal biofilms appears to be related to strain-dependent characteristics. In cases where the strain present in the root canal biofilm has properties similar to OG1RF, the ability of E. faecalis to outcompete other bacteria may result in the presence of only a few bacterial species, or even a monoculture of E. faecalis. In one study of persistent endodontic infections, monocultures of E. faecalis were identified in around 30% of the root canals containing a culturable flora (Sundqvist et al.
International Endodontic Journal, 48, 916–925, 2015
923
vez de Paz et al. Coexistence of E. faecalis in multispecies biofilms Cha
1998), suggesting that the dominance of this species may be associated with total overgrowth of the community. Enterococcus faecalis has been shown to release a large number of proteins, including documented virulence determinants such as cytolysin, gelatinase E (GelE), zinc metalloprotease and serine protease (SprE) (Tendolkar et al. 2003, Shankar et al. 2012). To investigate the extent to which differential expression of virulence factors could contribute to the differences in behaviour seen between OG1RF and GUL 1 in co-culture with L. salivarius and S. gordonii, the cellular, cell wall and exo-proteomes were compared using 2DE. Whilst only minor differences were seen in cellular and cell wall proteins from the two strains, the exo-proteomes showed a number of spots that were significantly more expressed in strain OG1RF than in strain GUL1. These corresponded to the proteins GelE and SprE, as well as a novel serine protease with no sequence similarity to SprE. Previously, it has been shown that different strains of E. faecalis, isolated from cheese or hospital environments, exhibit differences in their expression of these proteases (Pessione et al. 2012). Extracellular proteases can degrade extracellular matrix proteins enabling bacteria to enter host tissues, but their production is also one way in which E. faecalis may affect other members of the root canal consortium, through, for instance detachment and/or damage to neighbouring cells.
Conclusions The data presented lend support to the view that the extent to which E. faecalis is able to co-colonize with other organisms in root canals may depend on its protease production. High producers of proteases would suppress growth of other, but not all, organisms in a biofilm consortium, whereas low producers would be able to coexist well with other species. As proteases can also cause tissue damage and stimulate immune responses, high producers may also represent more virulent E. faecalis strains. Whether or not treatment is affected by this phenotypic property cannot be discerned from these findings, but it can be speculated that strains that can outcompete other root canal bacteria may be more frequently associated with endodontic treatment failures. In the light of this, endodontic treatment should not only be focused on bacterial elimination but also on the pathogenic potential of bacterial strains in root canal biofilms.
924
International Endodontic Journal, 48, 916–925, 2015
Acknowledgements The authors would like to thank Agnethe Henriksson for excellent technical assistance as well as The Aberdeen Proteome Facility, which is funded jointly by the SHEFC, BBSRC and the University of Aberdeen, for help with protein identification using LC-MS/MS. The study was funded by the Knowledge Foundation, Sweden. The authors deny any conflict of interest related to this study.
References Becker MR, Paster BJ, Leys EJ et al. (2002) Molecular species analysis of bacterial species associated with childhood caries. Journal of Clinical Microbiology 40, 1001–9. Brito LC, Teles FR, Franca EC et al. (2007) The use of multiple-displacement amplification and checkerboard DNA-DNA hybridization to examine the microbiota of endodontic infections. Journal of Clinical Microbiology 45, 3039–49. Ch avez de Paz LE (2009) Image analysis software based on color segmentation for characterization of viability and physiological activity of biofilms. Applied Environmental Microbiology 75, 1734–9. Ch avez de Paz LE (2012) Development of a multispes biofilm community by four root canal bacteria. Journal of Endodontics 38, 318–23. Ch avez de Paz LE, Dahlen G, Molander A, M€ oller A, Bergenholtz G (2003) Bacteria recovered from teeth with apical periodontitis after antimicrobial endodontic treatment. International Endodontic Journal 36, 500–8. Ch avez de Paz LE, Molander A, Dahlen G (2004) Gram-positive rods prevailing in teeth with apical periodontitis undergoing root canal treatment. International Endodontic Journal 37, 579–87. Costerton JW, Stewart PS, Greenberg EP (1999) Bacterial biofilms: a common cause of persistent infections. Science 284, 1318–22. Dahlen G, Samuelsson W, Molander A, Reit C (2000) Identification and antimicrobial susceptibility of enterococci isolated from the root canal. Oral Microbiology and Immunology 15, 309–12. Davies JR, Svens€ ater G, Herzberg MC (2009) Identification of novel LPXTG-linked surface proteins from Streptococcus gordonii. Microbiology 155, 1977–88. Dorkhan M, Ch avez de Paz LE, Skep€ o M, Svens€ ater G, Davies JR (2012) Effects of saliva or serum coating on adherence of Streptococcus oralis strains to titanium. Microbiology 158, 390–7. Evans M, Davies JK, Sundqvist G, Figdor D (2002) Mechanisms involved in the resistance of Enterococcus faecalis to calcium hydroxide. International Endodontic Journal 35, 221–8.
© 2015 International Endodontic Journal. Published by John Wiley & Sons Ltd
vez de Paz et al. Coexistence of E. faecalis in multispecies biofilms Cha
Figdor D, Davies JK, Sundqvist G (2003) Starvation survival, growth and recovery of Enterococcus faecalis in human serum. Oral Microbiology and Immunology 18, 234–9. Gilbert P, Das J, Foley I (1997) Biofilm susceptibility to antimicrobials. Advances in Dental Research 11, 160–7. Gold OG, Jordan HV, van Houte J (1975) The prevalence of enterococci in the human mouth and their pathogenicity in animal models. Archives of Oral Biology 20, 473–7. Hoyo K, Nagaoka S, Ohshima T, Maeda N (2009) Bacterial interactions in dental biofilm development. Journal of Dental Research 88, 982–90. Huycke MM, Abrams V, Moore DR (2002) Enterococcus faecalis produces extracellular superoxide and hydrogen peroxide that damages colonic epithelial cell DNA. Carcinogenesis 23, 529–36. Jakubovics NS, Gill SR, Vickerman MM, Kolenbrander PE (2008) Role of hydrogen peroxide in competition and cooperation between Streptococcus gordonii and Actinomyces naeslundii. FEMS Microbiology and Ecology 66, 637–44. Kayaoglu G, Erten H, Alacßam T, Ørstavik D (2005) Short-term antibacterial activity of root canal sealers towards Enterococcus faecalis. International Endodontic Journal 38, 483–8. Kolenbrander PE, Palmer RJ, Periasamy S, Jakubovics NS (2010) Oral multispecies biofilm development and the key role of cell-cell distance. Nature Reviews of Microbiology 8, 471–80. Kumar PS, Griffen AL, Moeschberger ML, Leys EJ (2005) Identification of candidate periodontal pathogens and beneficial species by quantitative 16S clonal analysis. Journal of Clinical Microbiology 43, 3944–55. Marsh PD (2005) Dental plaque: biological significance of a biofilm and community lifestyle. Journal of Clinical Periodontology 32, 7–15. Molander A, Reit C, Dahlen G, Kvist T (1998) Microbiological status of root-filled teeth with apical periodontitis. International Endodontic Journal 31, 1–7. Periasamy S, Kolenbrander PE (2009a) Aggregatibacter actinomycetemcomitans builds mutualistic biofilm communities
© 2015 International Endodontic Journal. Published by John Wiley & Sons Ltd
with Fusobacterium nucleatum and Veillonella species in saliva. Infection and Immunity 77, 3543–51. Periasamy S, Kolenbrander PE (2009b) Mutualistic biofilm communities develop with Porphyromonas gingivalis and initial, early and late colonizers of enamel. Journal of Bacteriology 191, 6804–11. Pessione A, Lambert C, Cocolin L et al. (2012) Different protein expression profiles in cheese and clinical isolates of Enterococcus faecalis revealed by proteomic analysis. Proteomics 12, 431–47. Portenier I, Waltimo T, Orstavik D, Haapasalo M (2005) The susceptibility of starved, stationary phase, and growing cells of Enterococcus faecalis to endodontic medicaments. Journal of Endodontics 31, 380–6. Ricucci D, Siqueira JF (2010) Biofilms and apical periodontitis: study of prevalence and association with clinical and histopathologic findings. Journal of Endodontics 36, 1277– 88. Shankar J, Walker RG, Ward D, Horsburgh MJ (2012) The Enterococcus faecalis exoproteome: identification and temporal regulation by Fsr. PLoS ONE 7, e33450. Siqueira JF Jr, R^ oßcas IN, Rosado AS (2004) Investigation of bacterial communities associated with asymptomatic and symptomatic endodontic infections by denaturing gradient gel electrophoresis fingerprinting approach. Oral Microbiology and Immunology 19, 363–70. Sundqvist G (1994) Taxonomy, ecology and pathogenicity of the root canal flora. Oral Surgery, Oral Medicine, Oral Pathology 78, 522–30. Sundqvist G, Figdor D, Persson S, Sj€ ogren U (1998) Microbiologic analysis of teeth with failed endodontic treatment and the outcome of conservative re-treatment. Oral Surgery, Oral Medicine, Oral Pathology, Oral Radiology and Endodontics 85, 86–93. Tendolkar PM, Naghdayan AS, Shankar N (2003) Pathogenic enterococci: new developments in the 21st century. Cellular and Molecular Life Sciences 60, 2622– 36.
International Endodontic Journal, 48, 916–925, 2015
925
Strains of Enterococcus faecalis differ in their ability to coexist in biofilms with other root canal bacteria
vez de Paz1, J. R. Davies1, G. Bergenholtz2 & G. Svensa €ter1 L. E. Cha 1
Department of Oral Biology, Faculty of Odontology, Malm€ o University, Malm€ o; and 2Department of Endodontology, The Sahlgrenska Academy at Gothenburg University, Gothenburg, Sweden
Abstract vez de Paz LE, Davies JR, Bergenholtz G, Cha €ter G. Strains of Enterococcus faecalis differ in their Svensa ability to coexist in biofilms with other root canal bacteria. International Endodontic Journal, 48, 916–925, 2015.
Aim To investigate the relationship between protease production and the ability of Enterococcus faecalis strains to coexist in biofilms with other bacteria commonly recovered from infected root canals. Methodology Biofilms with bacteria in mono-, dual- and four-species communities were developed in flow chambers. The organisms used were Lactobacillus salivarius, Streptococcus gordonii and Actinomyces naeslundii and E. faecalis strains, GUL1 and OG1RF. Biovolume and species distribution were examined using 16S rRNA fluorescence in situ hybridization in combination with confocal microscopy and image analysis. The full proteome of the E. faecalis strains was studied using two-dimensional gel electrophoresis. Spots of interest were identified using tandem mass spectroscopy and quantified using Delta 2D software.
Introduction When the enamel/dentine barrier is breached, oral microorganisms gain access to the root canal where they colonize the surfaces of dentine or necrotic tissue, giving rise to aggregates known as microbial
Correspondence: Gunnel Svens€ ater, Department of Oral Biology, Faculty of Odontology, Malm€ o University, Malm€ o S-20506, Sweden (Tel.: +46 406659495; fax +46 406659484; e-mail: [email protected]).
916
International Endodontic Journal, 48, 916–925, 2015
Results All bacteria formed biofilms and an ANOVA analysis revealed that the biofilm biomass increased significantly (P ≤ 0.01) between 6 and 24 h. L. salivarius, S. gordonii and A. naeslundii formed mutualistic biofilm communities, and this pattern was unchanged when E. faecalis GUL1 was included in the consortium. However, with OG1RF, L. salivarius and S. gordonii were outcompeted in a 24-h biofilm. Proteomic analysis revealed that OG1RF secreted higher levels of proteases, GelE (P = 0.02) and SprE (P = 0.002) and a previously unidentified serine protease (P = 0.05), than GUL1. Conclusions Different strains of E. faecalis can interact synergistically or antagonistically with a consortium of root canal bacteria. A possible mechanism underlying this, as well as potential differences in virulence, is production of different levels of proteases, which can cause detachment of neighbouring bacteria and tissue damage. Keywords: antagonism, apical periodontitis, bacterial consortium, multispecies, mutualism. Received 18 September 2014; accepted 5 July 2015
biofilms (Ricucci & Siqueira 2010). In general, bacteria in biofilms have the important and often underappreciated characteristic of being very resistant to a variety of antimicrobial substances, including antibiotics (Gilbert et al. 1997). This property has, in part, been attributed to the presence of a polymeric matrix. Biofilm bacteria are also intrinsically less susceptible to antimicrobial substances due to their slow rate of growth and adoption of a distinct biofilm phenotype, which differs from that of their planktonic counterparts (Costerton et al. 1999).
© 2015 International Endodontic Journal. Published by John Wiley & Sons Ltd
vez de Paz et al. Coexistence of E. faecalis in multispecies biofilms Cha
Root canals of untreated teeth with necrotic pulp tissue support growth of different bacterial consortia. In a study by Sundqvist (1994), the most commonly isolated genera were Fusobacterium and streptococci with Bacteroides spp, Parvimonas spp, Lactobacillus spp, Actinomycetes spp, Veillonella parvula and a series of gram-negative anaerobic rods also present. However, the microbiota of persistent root canal infections, including re-treated cases of teeth with apical periodontitis, differs significantly from that of untreated teeth (Molander et al. 1998, Siqueira et al. 2004). The species diversity is severely restricted, and monocultures, often of E. faecalis, are common (Sundqvist et al. 1998). This finding suggests that root canal treatment causes selection from a multispecies community towards biofilms comprised of one or a few species, often including Enterococcus faecalis. One mechanism behind the strong presence of E. faecalis seems to be its proven resistance to antimicrobial endodontic medicaments (Dahlen et al. 2000, Evans et al. 2002, Portenier et al. 2005), including calcium hydroxide-based sealers (Kayaoglu et al. 2005) and the harsh nutritional conditions that often prevail in canals of root filled teeth (Figdor et al. 2003, Portenier et al. 2005). In biofilm communities such as those found in root canals of infected teeth, the bacteria often exist in close proximity to each other facilitating a range of bacterial interactions. These exchanges can be physical, such as co-adhesion and co-aggregation, or chemical via diffusible signalling molecules and transfer of genetic material. Bacteria in biofilms produce batteries of complementary enzymes, which can act synergistically to effectively exploit complex molecules, thus providing nutrients to the bacteria in the local environment (Marsh 2005). In addition to cooperative interactions, competitive interactions such as the production of bacteriocins and inhibitory metabolic end products can also occur in biofilms. The most studied inhibitory substances include hydrogen peroxide, short chain fatty acids and lantibiotics from Lactobacilli (Hoyo et al. 2009). Thus, the bacterial consortia colonizing the root canal will be determined not only by the species which originally gain entry, but also by synergistic and antagonistic interactions between biofilm bacteria. Clinical studies have shown that E. faecalis is a component of multispecies bacterial communities in infected root canals but may also be found as the sole species, indicating that E. faecalis is capable of outcompeting other members of a polymicrobial infection
© 2015 International Endodontic Journal. Published by John Wiley & Sons Ltd
(Sundqvist et al. 1998). It was hypothesized that different strains of E. faecalis differ in their ability to coexist with other root canal bacteria and that this is related to their protease production in biofilm environments. This may have important implications for differences in virulence between different E. faecalis strains and thus the clinical outcome of apical parodontitis. To explore interactions between root canal bacteria, this study investigated how two strains of E. faecalis (OGRIF – a strain commonly used in endodontic research and GUL1 – a clinical strain isolated from a root canal) can coexist with other bacteria recovered from infected root canals viz. Lactobacillus salivarius, Streptococcus gordonii and Actinomyces naeslundii.
Materials and methods Bacterial strains and culture conditions The bacterial strains, E. faecalis (GUL1), L. salivarius, S. gordonii and A. naeslundii, were isolates from root canals undergoing endodontic treatment presenting with persistent root canal infections (Ch avez de Paz et al. 2003, 2004). E. faecalis (GUL1) was routinely identified by colony morphology and growth on selective tellurite medium. E. faecalis strain OG1RF was originally derived from strain OG1 (formerly designated 2SAR) (Gold et al. 1975). All strains were routinely cultured in blood agar and Bacto Todd–Hewitt broth (TH; Becton Dickinson & Co, Franklin Lakes, NJ, USA) in an atmosphere of 5% CO2 at 37 °C for 24 h.
Biofilm growth Mid-exponential bacterial cultures were created by mixing 0.5 mL of overnight culture with 5 mL fresh TH medium and incubating in an atmosphere of 5% CO2 at 37 °C for 5 h. The optical density at 600 nm was adjusted to 0.6 (approximately 1 9 108 mL 1 cells), and mini flow chambers (l-Slide VI for Live Cell Analysis; Integrated BioDiagnostics, Martinsried, Germany) with hydrophilic uncoated polystyrene surfaces were inoculated with one, two, three or four species according to the outline presented in Fig. 1. Two-, three- and four-species inocula were prepared by mixing 0.5 mL of appropriate combinations of E. faecalis, L. salivarius, S. gordonii and A. naeslundii (equivalent to about 1 9 107 mL 1 cells of each species) (18). Each mini flow chamber was inoculated with 30 lL
International Endodontic Journal, 48, 916–925, 2015
917
vez de Paz et al. Coexistence of E. faecalis in multispecies biofilms Cha
Species
Dual-species biofilms
Three-species biofilm Four-species biofilms
Ef-GUL1
Ef-GUL1
Ef-GUL1 Ls + Ls An
+ An
+ Sg
Ef-OG1RF
+Ls + An +Sg Ls + An + Sg
Ef-OG1RF
Sg Ef-OG1RF + Ls
+ An
+ Sg
+Ls + An +Sg
Figure 1 Outline of the biofilm experiments. The species used were as follows: Enterococcus faecalis strains GUL1 and OG1RF, Lactobacillus salivarius (Ls), Actinomyces naeslundii (An) and Streptococcus gordonii (Sg).
of the bacterial suspension followed by addition of 100 lL of fresh TH to give a final volume of 130 lL per flow chamber. Flow chambers were then incubated under static conditions in an atmosphere of 5% CO2 in air at 37 °C for 6 or 24 h.
Species analysis using 16S rRNA-FISH Bacteria in the biofilms were identified using 16S rRNA-FISH probes (Biomers, Ulm, Germany). E. faecalis was identified using the ENF191 probe (gaaagcgcctttcactcttatgc) labelled with Pacific Blue (kex 410 nm, kem 455 nm), ATTO-488 (kex 504 nm, kem 521 nm) or ATTO-565 (kex 565 nm, kem 590 nm); S. gordonii was identified using the STR405 probe (tagccgtccctttctggt) labelled with ATTO-488 (green); L. salivarius was identified using the LAC722 probe (caccgctacacatgragttccact) labelled with ATTO565 (red); and A. naeslundii was identified using the JF201 probe (gctaccgtcaacccaccc) labelled with Pacific Blue (blue). The 16S rRNA-FISH protocol has been described previously (Ch avez de Paz 2012). Briefly, after 6- or 24-h biofilm cells were fixed with 4% paraformaldehyde in phosphate-buffered saline and permeabilized using a lysozyme solution (10 mg mL 1; Sigma-Aldrich Corporation, St. Louis, MO, USA). The biofilms were then washed with ultrapure water and dehydrated with 50%, 80% and 99% ethanol for 3 min, respectively, after which the flow chambers were inoculated with 30 mL of hybridization buffer [0.9 M NaCl, 20 mM Tris-HCl buffer, pH 7.5, with 0.01% sodium dodecyl sulphate (SDS) and 25% formamide] containing 20 ng mL 1 of the oligonucleotide probes. Probe cocktails containing 20 ng mL 1 of each oligonucleotide were used, and
918
International Endodontic Journal, 48, 916–925, 2015
hybridization was carried out at 47 °C for 3 h in a humid chamber.
Confocal laser scanning microscopy All microscopy was performed using an Eclipse TE2000 inverted confocal scanning laser microscope (Nikon Corporation, Tokyo, Japan). By means of a motorized stage, ten areas within each sample were selected at random and image stack sections composed of ten images, each taken with a variation of 2 lm along the z-position and covering an area of 0.05 mm2 per field of view, were taken. Confocal illumination for the green fluorescence signal was provided by an Ar laser (488 nm laser excitation). For detection of the red fluorescence, a G-HeNe laser was used (543 nm laser excitation), and for the detection of blue fluorescence, a UV laser was used (390 nm laser excitation).
Image analysis and statistical evaluation Confocal laser scanner microscopy (CSLM) images were analysed using software bio-Image_L (Ch avez de Paz 2009). The biovolume represents the overall volume of the biofilm. The results for biofilm growth over time are presented as means from a total of 30 biofilm stacks, that is 10 stacks/flow-chamber channel in three independent experiments. To evaluate the relative proportions of species in dual-species biofilms formed between the two strains of E. faecalis (violet) and L. salivarius (red), A. naeslundii (blue) or S. gordonii (green), the independent subpopulations represented by the different fluorescent colours were evaluated and the percentage of the total biofilm
© 2015 International Endodontic Journal. Published by John Wiley & Sons Ltd
vez de Paz et al. Coexistence of E. faecalis in multispecies biofilms Cha
Two-dimensional gel electrophoresis and protein identification Cell wall proteins and whole cell extracts of GUL1 and OG1RF were prepared as described previously (Davies et al. 2009, Dorkhan et al. 2012). For isolation of extracellular proteins, the culture medium was harvested from batch cultures of cells in exponential growth phase by centrifugation and proteins precipitated overnight at 4 °C using 10% trichloroacetic acid (Sigma-Aldrich Corporation). After centrifugation (16 000 g, 30 min, 4 °C), the pellet was re-suspended in acetone and sonicated for 3 9 10 s. Samples were then re-centrifuged as above, and the resulting pellet dissolved in 500 mL of two-dimensional gel electrophoresis (2DE) rehydration buffer (8 M urea, 2% CHAPS, 10 mM DTT, 2% IPG buffer). The protein concentration was determined using a 2D Quant kit (GE Healthcare Life Sciences, Pittsberg, PA, USA), and a volume corresponding to 50 mg protein (for silver-stained gels) or 150 mg protein (for Coomassie-stained gels) was subjected to isoelectric focusing as described previously (Davies et al. 2009). Cell wall proteins were subjected to gel electrophoresis on 7% SDS-PAGE gels, whereas whole cell extracts and extracellular proteins were separated on 14% SDSPAGE gels. SDS-PAGE was performed at a constant current of 15 mA per gel overnight at 10 °C in a PROTEAN II xi cell (Bio-Rad, Hercules, CA, USA). Rainbow molecular mass standards (GE Healthcare Life Sciences) were run on the acidic side of the IPG strips. Gels were stained with Coomassie brilliant blue or silver according to the relevant protocol from GE Healthcare Life Sciences. All gels were run in triplicate. Silver-stained gels were scanned and the integrated optical intensity (IOD) within the spot boundaries of the major differently expressed proteins determined using DECODON Delta-2D software (Greifswald, Germany). For identification, the differently expressed proteins were excised manually from Coomassie brilliant blue-stained gels and subjected to liquid chromatography–tandem mass spectrometry (LC-MS/MS) as described previously (Davies et al. 2009). Briefly, proteins were reduced with DTT (GE Healthcare Life Sciences; 60 °C, 20 min), alkylated with iodoacetamide
© 2015 International Endodontic Journal. Published by John Wiley & Sons Ltd
(GE Healthcare Life Sciences; 25 °C, 10 min) and then digested with trypsin (37 °C, 8 h). Tryptic peptides were separated and subjected to MS/MS. Peptide peaks were deconvoluted automatically and mass lists in the form of Mascot Generic Files used as the input for Mascot MS/MS Ions searches of the NCBInr database using the Matrix Science web server (www.matrixscience.com).
Statistical analysis Statistical analysis of biofilm formation over 6 or 24 h was performed with the SYSTAT 13 software package using two-way ANOVA with a Tukey post hoc test. IOD values from the 2D gels from the major differentially expressed proteins in strains OG1RF and GUL1 were transferred to Prism 5 (GraphPad, La Jolla, CA, USA). Mean and standard errors were calculated, and the values for each spot compared using Student’s t-test.
Results Ability of the tested organisms to form monospecies biofilms Both strain GUL1 and strain OG1RF adhered to the flow-cell surfaces after 6 h (Fig. 2, white bars), and after 24 h of incubation (Fig. 2, grey bars), the biovolumes had increased significantly (P = 0.015 for
Biofilm biovolume m3 x104
biovolume represented by E. faecalis calculated for 30 biofilm stacks and expressed as the mean SE. 3D reconstructions were produced using the function ‘Structure and Distribution’ in bioImage_L FISH.
8
* *
6
6h 24 h
*
*
4
2
0
GUL1
OG1RF
Ls
An
Sg
Figure 2 Formation of mono-species biofilms in mini flow chambers over time. The bars represent the mean biovolumes of the biofilms formed by Enterococcus faecalis strains GUL1 and OG1RF, as well as Lactobacillus salivarius, Actinomyces naeslundii and Streptococcus gordonii, after 6 or 24 h. The error bars show standard errors calculated from three independent experiments and *P ≤ 0.01.
International Endodontic Journal, 48, 916–925, 2015
919
vez de Paz et al. Coexistence of E. faecalis in multispecies biofilms Cha
P = 0.002, respectively), whereas for A. naeslundii, the 1.4-fold increase seen was not significant (P = 0.25). At 24 h, the greatest biovolume was registered for S. gordonii.
GUL1 and P < 0.001 for OG1RF). Although the biofilms of OG1RF at 24 h contained a greater biovolume than those of GUL1, this difference was not significant (P = 0.28). When the same procedure was undertaken with the clinical isolates of L. salivarius, A. naeslundii and S. gordonii, all showed attachment to the flow-chamber surface at 6 h (Fig. 2, white bars) and after 24 h of incubation, the biovolume increased in all three organisms (Fig. 2, grey bars). For S. gordonii and L. salivarius, the twofold increase in biomass was significant (P = 0.005 and
(a)
(b)
Co-cultivation of the E. faecalis strains with L. salivarius, S. gordonii or A. naeslundii for 24 h resulted in dual-species biofilms (Fig. 3). 16S rRNA-FISH revealed
E. faecalis A. naeslundii
(c)
E. faecalis S. gordonii
OG1RF
GUL1
E. faecalis L. salivarius
Coexistence of E. faecalis strains with other root canal bacteria
E. faecalis (%)
100
+ L. salivarius
100
+ A. naeslundii
100
80
80
80
60
60
60
40
40
40
20
20
20
0
GUL1
OG1RF
0
GUL1
OG1RF
0
+ S. gordonii
GUL1
OG1RF
Figure 3 Relative proportions of species in dual-species biofilms formed between the two strains of Enterococcus faecalis and Lactobacillus salivarius, Actinomyces naeslundii and Streptococcus gordonii. The images show 24-h biofilms stained with species-specific oligonucleotides targeting the 16S rRNA gene. In (a), L. salivarius is visualized in red and E. faecalis is green, and in (b), E. faecalis is visualized in red and A. naeslundii in blue, and in (c), S. gordonii is shown green and E. faecalis in blue. The bar represents 5 lm. The box plots show the mean SE for the percentage of biovolume occupied by E. faecalis strain GUL 1 or OG1RF in each biofilm. All experiments were carried out in triplicate.
920
International Endodontic Journal, 48, 916–925, 2015
© 2015 International Endodontic Journal. Published by John Wiley & Sons Ltd
vez de Paz et al. Coexistence of E. faecalis in multispecies biofilms Cha
the two cell types in each biofilm (E. faecalis and L. salivarius, E. faecalis and A. naeslundii or E. faecalis and S. gordonii) in close proximity to each other. Initially, the two species in each experiment were inoculated in equal numbers, that is the proportion of E. faecalis (GUL1 or OG1RF) in each case was approximately 50%. In dual-species biofilms with L. salivarius, the proportion of GUL 1 after 24 h remained at 50 10% (Fig. 3a). However, when OG1RF was used, the level of this strain in the biofilm increased to 75 5% after 24 h suggesting that this strain outcompeted L. salivarius during biofilm formation. A similar result was seen with S. gordonii where the proportion of GUL 1 was 38 5% after 24 h compared to 66 10% for strain OG1RF (Fig. 3c). The A. naeslundii strain used here appeared to be unaffected by the presence of either the GUL 1 or OG1RF E. faecalis strain which showed levels of 38 5% and 42 5% of the total biofilm biomass, respectively, after 24 h (Fig. 3b). These data thus suggest that E. faecalis (a)
(b) A. naeslundii L. salivarius S. gordonii
strain OG1RF was more competitive at colonizing the surface and establishing a biofilm in the presence of L. salivarius and S. gordonii than GUL1. Thus, the E. faecalis strains tested appear to show different properties; GUL1, exhibiting the capacity to coexist with other species in biofilms and strain OG1RF, which appears to be able to supplant other species.
Behaviour of E. faecalis strains in four-species biofilms Before the introduction of E. faecalis, the ability of A. naeslundii, L. salivarius and S. gordonii to exist in multispecies biofilms was investigated. Equal numbers of cells of each species were inoculated into a mini flow chamber and allowed to form a biofilm community for 24 h. Evaluation of the proportions of each species revealed no significant change from the initial inoculate (Fig. 4a) which suggests a balanced
A. naeslundii L. salivarius S. gordonii E. faecalis (GUL1)
(c)
A. naeslundii L. salivarius S. gordonii E. faecalis (OG1RF)
3% 2%
Figure 4 CSLM micrographs of 24-h biofilm communities stained using 16S rRNA-FISH and pie charts depicting the species distribution in the three- and four-species communities. The three-species consortium (a) shows a homogenous distribution of Lactobacillus salivarius (red fluorescent probe), Streptococcus gordonii (green fluorescent probe) and Actinomyces naeslundii (blue fluorescent probe). The four-species consortium shown in (b), which contains Enterococcus faecalis strain GUL1 (violet), also showed a homogenous distribution of species. In the third consortium (c), which contained E. faecalis strain OG1RF (violet), an increase in E. faecalis at the expense of S. gordonii (green) and L. salivarius (red) was seen. The bar represents 5 lm.
© 2015 International Endodontic Journal. Published by John Wiley & Sons Ltd
International Endodontic Journal, 48, 916–925, 2015
921
vez de Paz et al. Coexistence of E. faecalis in multispecies biofilms Cha
colonization with no one species dominating over another. Four-species bacterial consortia were then established using equal numbers of L. salivarius, A. naeslundii, S. gordonii and E. faecalis strain GUL1 or OG1RF and inoculated into mini flow chambers. In the biofilm with strain GUL1, quantification of the proportions of each species revealed that L. salivarius represented 29%, A. naeslundii (24%), S. gordonii 31% and E. faecalis 16% (Fig. 4b). Image analysis revealed that the four species were homogeneously distributed throughout the biofilm. Thus, the proportions of the species in the biofilm did not differ significantly from those in the original inoculate (25% of each species) showing that the inclusion of E. faecalis GUL1 did not disturb the balanced relationship between the species in the biofilm community. In contrast, when E. faecalis strain OG1RF was introduced into the four-species consortium, the proportions of the different species present in the biofilm after 24 h changed dramatically. The proportion of E. faecalis had increased from around 25% in the original inoculate to 68 3% in the biofilm. In contrast, L. salivarius and S. gordonii were almost completely absent, suggesting that strain OG1RF had outcompeted these two species during biofilm formation. As seen in the dual-species biofilm, A. naeslundii was unaffected by the presence of strain OG1RF and the levels remained similar to those in the inoculate.
Proteome of E. faecalis strains To investigate possible mechanisms underlying the ability of strain OG1RF, but not GUL1, to outcompete other species during biofilm formation, the cellular, cell wall and exo-proteomes of the two strains in the exponential phase of growth were compared using 2DE. Only minor differences were seen between the whole cell extracts and cell wall proteins from the two strains, suggesting that they did not differ to any significant degree in this respect (data not shown). However, the exo-proteomes of the two strains showed three dominant differently expressed spots (Fig. 5) that mass spectrometric analysis revealed to correspond to a gelatinase and a serine protease similar to those identified previously in E. faecalis strain V583. The two spots corresponding to gelatinase (GelE) were significantly more expressed in strain OG1RF than in strain GUL1 (tenfold and 30-fold differences, respectively) whilst the serine protease SprE was six times more abundant (P = 0.002; Table 1). In addition, two spots corresponding to a novel serine protease with no sequence similarity to SprE were identified. These (serine protease a and serine protease b) were also significantly more highly expressed by strain OG1RF than by strain GUL1 [eightfold (P = 0.011) and 13-fold differences (0.046), respectively].
76
(a)
(b)
52
Serine protease (a)
38 Serine protease (b)
GelE(a)
24 SprE GelE (b)
17 12
Figure 5 Silver-stained 2DE gels showing the exo-proteome of Enterococcus faecalis strains. Proteins released from (a) E. faecalis strain OG1RG and (b) E. faecalis strain GUL1 were separated by isoelectric focusing (pI range 4–7) in the first dimension and by SDS-PAGE in 14% gels in the second dimension. The positions of the molecular mass markers are shown. The identities of the most prominent differentially regulated proteins are shown in (a).
922
International Endodontic Journal, 48, 916–925, 2015
© 2015 International Endodontic Journal. Published by John Wiley & Sons Ltd
vez de Paz et al. Coexistence of E. faecalis in multispecies biofilms Cha
Table 1 Identity of major differentially expressed exoproteins from two strains of Enterococcus faecalis
Protein identification
Mean IOD strain OG1RF
Serine protease a Serine protease b GelE 1a GelE 1b SprE
11 315 3451 1992 3026 1241
2027 1106 481 492 124
Mean IOD strain GUL1
P-value (Student’s t-test)
0.011 0.046 0.02 0.004 0.002
1373 258 202 108 217
942 158 0.57 8 87
IOD, integrated optical intensity.
Discussion Dental biofilms consisting of multispecies communities are the cause of dental disease; caries (Becker et al. 2002), periodontitis (Kumar et al. 2005) and endodontic infections (Brito et al. 2007). In multispecies biofilms, bacterial species interact co-operatively or competitively with other members of the community. Examples of mutualistic biofilm communities include those formed between Aggregatibacter actinomycetemcomitans, F. nucleatum and Veillonella spp (Periasamy & Kolenbrander 2009a) as well as between S. gordonii and the periodontal pathogen Porphyromonas gingivalis (Periasamy & Kolenbrander 2009b). In this study, the ability of three root canal bacteria (L. salivarius, A. naeslundii and S. gordonii) to form biofilm communities was investigated. The strains were all recovered from root canals after treatment suggesting that they had survived exposure to the antimicrobial agents used. After 24 h, the relative proportions of the three species were similar to those in the initial inoculate suggesting that they can coexist in biofilms. 16S rRNA-FISH in combination with CLSM showed that S. gordonii and A. naeslundii were in close proximity to each other suggesting some form of cell–cell interaction between these species. Co-adhesion between different species is a common phenomenon amongst early colonizers of the oral cavity and has been demonstrated for A. naeslundii and S. gordonii (Kolenbrander et al. 2010). In addition, these bacteria can participate in mutually beneficial metabolic interactions. For instance, S. gordonii is known to produce H2O2 that at high levels can lead to oxidative stress and cell damage. This effect can be attenuated by the release of catalase from A. naeslundii. Thus, A. naeslundii can rescue S. gordonii cells from the damaging effects of endogenous H2O2 production (Jakubovics et al. 2008). In contrast, L. salivarius showed no obvious cell–cell interaction with
© 2015 International Endodontic Journal. Published by John Wiley & Sons Ltd
S. gordonii or A. naeslundii and no mutualistic interactions have been described for these species. However, from the data presented here, it is evident that L. salivarius can coexist in biofilms with both S. gordonii and A. naeslundii. In common with L. salivarius, A. naeslundii and S. gordonii, E. faecalis is commonly isolated from treated root canals. However, in contrast to L. salivarius, A. naeslundii and S. gordonii, E. faecalis can be found as the only bacteria present (Sundqvist et al. 1998) suggesting that E. faecalis can outcompete other species in the root canal environment. In this study, it was evident that the two strains of E. faecalis differed in their capacity to coexist with L. salivarius, A. naeslundii and S. gordonii. Strain GUL1 was able to establish dual-species biofilms with L. salivarius, A. naeslundii or S. gordonii. In each case, the relative proportions were similar to those in the original suspension suggesting that there was no antagonistic relationship between the two members of the community. Close interspecies cell contact was most evident between GUL1 and A. naeslundii. As E. faecalis is known to produce H2O2 (Huycke et al. 2002), it cannot be excluded that a similar mutualistic relationship to that between A. naeslundii and S. gordonii may exist. In the four-species consortium after growth for 24 h, the biomass of each species was also approximately equal to that in the original inoculate. This illustrates that for GUL 1, mutualistic and/or neutralistic interactions dominate the community behaviour. In contrast, the presence of strain OG1RF in dual-species biofilms with L. salivarius and S. gordonii suppressed growth of the latter species, whereas biofilm growth of A. naeslundii appeared to be unaffected by the presence of OG1RF. This behaviour was confirmed in the four-species consortium where, after 24 h, OG1RF comprised 68% of the total biomass. These data suggest that this strain of E. faecalis acts antagonistically in relation to S. gordonii and L. salivarius, whereas interactions with A. naeslundii were mutualistic. Thus, the ability of E. faecalis to coexist with other potential members of root canal biofilms appears to be related to strain-dependent characteristics. In cases where the strain present in the root canal biofilm has properties similar to OG1RF, the ability of E. faecalis to outcompete other bacteria may result in the presence of only a few bacterial species, or even a monoculture of E. faecalis. In one study of persistent endodontic infections, monocultures of E. faecalis were identified in around 30% of the root canals containing a culturable flora (Sundqvist et al.
International Endodontic Journal, 48, 916–925, 2015
923
vez de Paz et al. Coexistence of E. faecalis in multispecies biofilms Cha
1998), suggesting that the dominance of this species may be associated with total overgrowth of the community. Enterococcus faecalis has been shown to release a large number of proteins, including documented virulence determinants such as cytolysin, gelatinase E (GelE), zinc metalloprotease and serine protease (SprE) (Tendolkar et al. 2003, Shankar et al. 2012). To investigate the extent to which differential expression of virulence factors could contribute to the differences in behaviour seen between OG1RF and GUL 1 in co-culture with L. salivarius and S. gordonii, the cellular, cell wall and exo-proteomes were compared using 2DE. Whilst only minor differences were seen in cellular and cell wall proteins from the two strains, the exo-proteomes showed a number of spots that were significantly more expressed in strain OG1RF than in strain GUL1. These corresponded to the proteins GelE and SprE, as well as a novel serine protease with no sequence similarity to SprE. Previously, it has been shown that different strains of E. faecalis, isolated from cheese or hospital environments, exhibit differences in their expression of these proteases (Pessione et al. 2012). Extracellular proteases can degrade extracellular matrix proteins enabling bacteria to enter host tissues, but their production is also one way in which E. faecalis may affect other members of the root canal consortium, through, for instance detachment and/or damage to neighbouring cells.
Conclusions The data presented lend support to the view that the extent to which E. faecalis is able to co-colonize with other organisms in root canals may depend on its protease production. High producers of proteases would suppress growth of other, but not all, organisms in a biofilm consortium, whereas low producers would be able to coexist well with other species. As proteases can also cause tissue damage and stimulate immune responses, high producers may also represent more virulent E. faecalis strains. Whether or not treatment is affected by this phenotypic property cannot be discerned from these findings, but it can be speculated that strains that can outcompete other root canal bacteria may be more frequently associated with endodontic treatment failures. In the light of this, endodontic treatment should not only be focused on bacterial elimination but also on the pathogenic potential of bacterial strains in root canal biofilms.
924
International Endodontic Journal, 48, 916–925, 2015
Acknowledgements The authors would like to thank Agnethe Henriksson for excellent technical assistance as well as The Aberdeen Proteome Facility, which is funded jointly by the SHEFC, BBSRC and the University of Aberdeen, for help with protein identification using LC-MS/MS. The study was funded by the Knowledge Foundation, Sweden. The authors deny any conflict of interest related to this study.
References Becker MR, Paster BJ, Leys EJ et al. (2002) Molecular species analysis of bacterial species associated with childhood caries. Journal of Clinical Microbiology 40, 1001–9. Brito LC, Teles FR, Franca EC et al. (2007) The use of multiple-displacement amplification and checkerboard DNA-DNA hybridization to examine the microbiota of endodontic infections. Journal of Clinical Microbiology 45, 3039–49. Ch avez de Paz LE (2009) Image analysis software based on color segmentation for characterization of viability and physiological activity of biofilms. Applied Environmental Microbiology 75, 1734–9. Ch avez de Paz LE (2012) Development of a multispes biofilm community by four root canal bacteria. Journal of Endodontics 38, 318–23. Ch avez de Paz LE, Dahlen G, Molander A, M€ oller A, Bergenholtz G (2003) Bacteria recovered from teeth with apical periodontitis after antimicrobial endodontic treatment. International Endodontic Journal 36, 500–8. Ch avez de Paz LE, Molander A, Dahlen G (2004) Gram-positive rods prevailing in teeth with apical periodontitis undergoing root canal treatment. International Endodontic Journal 37, 579–87. Costerton JW, Stewart PS, Greenberg EP (1999) Bacterial biofilms: a common cause of persistent infections. Science 284, 1318–22. Dahlen G, Samuelsson W, Molander A, Reit C (2000) Identification and antimicrobial susceptibility of enterococci isolated from the root canal. Oral Microbiology and Immunology 15, 309–12. Davies JR, Svens€ ater G, Herzberg MC (2009) Identification of novel LPXTG-linked surface proteins from Streptococcus gordonii. Microbiology 155, 1977–88. Dorkhan M, Ch avez de Paz LE, Skep€ o M, Svens€ ater G, Davies JR (2012) Effects of saliva or serum coating on adherence of Streptococcus oralis strains to titanium. Microbiology 158, 390–7. Evans M, Davies JK, Sundqvist G, Figdor D (2002) Mechanisms involved in the resistance of Enterococcus faecalis to calcium hydroxide. International Endodontic Journal 35, 221–8.
© 2015 International Endodontic Journal. Published by John Wiley & Sons Ltd
vez de Paz et al. Coexistence of E. faecalis in multispecies biofilms Cha
Figdor D, Davies JK, Sundqvist G (2003) Starvation survival, growth and recovery of Enterococcus faecalis in human serum. Oral Microbiology and Immunology 18, 234–9. Gilbert P, Das J, Foley I (1997) Biofilm susceptibility to antimicrobials. Advances in Dental Research 11, 160–7. Gold OG, Jordan HV, van Houte J (1975) The prevalence of enterococci in the human mouth and their pathogenicity in animal models. Archives of Oral Biology 20, 473–7. Hoyo K, Nagaoka S, Ohshima T, Maeda N (2009) Bacterial interactions in dental biofilm development. Journal of Dental Research 88, 982–90. Huycke MM, Abrams V, Moore DR (2002) Enterococcus faecalis produces extracellular superoxide and hydrogen peroxide that damages colonic epithelial cell DNA. Carcinogenesis 23, 529–36. Jakubovics NS, Gill SR, Vickerman MM, Kolenbrander PE (2008) Role of hydrogen peroxide in competition and cooperation between Streptococcus gordonii and Actinomyces naeslundii. FEMS Microbiology and Ecology 66, 637–44. Kayaoglu G, Erten H, Alacßam T, Ørstavik D (2005) Short-term antibacterial activity of root canal sealers towards Enterococcus faecalis. International Endodontic Journal 38, 483–8. Kolenbrander PE, Palmer RJ, Periasamy S, Jakubovics NS (2010) Oral multispecies biofilm development and the key role of cell-cell distance. Nature Reviews of Microbiology 8, 471–80. Kumar PS, Griffen AL, Moeschberger ML, Leys EJ (2005) Identification of candidate periodontal pathogens and beneficial species by quantitative 16S clonal analysis. Journal of Clinical Microbiology 43, 3944–55. Marsh PD (2005) Dental plaque: biological significance of a biofilm and community lifestyle. Journal of Clinical Periodontology 32, 7–15. Molander A, Reit C, Dahlen G, Kvist T (1998) Microbiological status of root-filled teeth with apical periodontitis. International Endodontic Journal 31, 1–7. Periasamy S, Kolenbrander PE (2009a) Aggregatibacter actinomycetemcomitans builds mutualistic biofilm communities
© 2015 International Endodontic Journal. Published by John Wiley & Sons Ltd
with Fusobacterium nucleatum and Veillonella species in saliva. Infection and Immunity 77, 3543–51. Periasamy S, Kolenbrander PE (2009b) Mutualistic biofilm communities develop with Porphyromonas gingivalis and initial, early and late colonizers of enamel. Journal of Bacteriology 191, 6804–11. Pessione A, Lambert C, Cocolin L et al. (2012) Different protein expression profiles in cheese and clinical isolates of Enterococcus faecalis revealed by proteomic analysis. Proteomics 12, 431–47. Portenier I, Waltimo T, Orstavik D, Haapasalo M (2005) The susceptibility of starved, stationary phase, and growing cells of Enterococcus faecalis to endodontic medicaments. Journal of Endodontics 31, 380–6. Ricucci D, Siqueira JF (2010) Biofilms and apical periodontitis: study of prevalence and association with clinical and histopathologic findings. Journal of Endodontics 36, 1277– 88. Shankar J, Walker RG, Ward D, Horsburgh MJ (2012) The Enterococcus faecalis exoproteome: identification and temporal regulation by Fsr. PLoS ONE 7, e33450. Siqueira JF Jr, R^ oßcas IN, Rosado AS (2004) Investigation of bacterial communities associated with asymptomatic and symptomatic endodontic infections by denaturing gradient gel electrophoresis fingerprinting approach. Oral Microbiology and Immunology 19, 363–70. Sundqvist G (1994) Taxonomy, ecology and pathogenicity of the root canal flora. Oral Surgery, Oral Medicine, Oral Pathology 78, 522–30. Sundqvist G, Figdor D, Persson S, Sj€ ogren U (1998) Microbiologic analysis of teeth with failed endodontic treatment and the outcome of conservative re-treatment. Oral Surgery, Oral Medicine, Oral Pathology, Oral Radiology and Endodontics 85, 86–93. Tendolkar PM, Naghdayan AS, Shankar N (2003) Pathogenic enterococci: new developments in the 21st century. Cellular and Molecular Life Sciences 60, 2622– 36.
International Endodontic Journal, 48, 916–925, 2015
925
