Ethics statement

This study was carried out in accordance with the guidelines of, and after approval by, the Ethics Committee at Estácio de Sá University, Rio de Janeiro, Brazil, and written informed consent was obtained from the patients.

Subjects and Sample Collection

Samples were taken from patients who had been referred for root canal treatment or emergency treatment to the Department of Endodontics, Estácio de Sa University. Only teeth from adult patients (ages ranging from 20 to 39 years) with carious lesions, necrotic pulps, and radiographic evidence of apical periodontitis were included in this study. Samples were obtained from the root canals of 12 teeth with asymptomatic apical periodontitis and from aspiration of pus from two cases of acute apical abscess, which showed localized swelling, fever, lymphadenopathy, and malaise. No apparent communication from the abscess to the oral cavity or the skin surface was observed. Selected teeth showed no significant gingival recession and were free of periodontal pockets deeper than 4 mm.

Samples from the root canals of teeth with asymptomatic apical periodontitis were taken as follows. After the tooth crown was cleansed with pumice, a rubber dam was placed and the tooth and the surrounding field were decontaminated by a protocol using 3% hydrogen peroxide followed by 2.5% sodium hypochlorite (NaOCl) solution. Complete access preparations were made using sterile burs without water spray. The operative field, including the pulp chamber, was again swabbed with 2.5% NaOCl, which was then inactivated with sterile 5% sodium thiosulfate. If the root canal was dry, a small amount of 10 mM Tris–HCl (pH 8.0) was placed in the canal. A K-type file no. 15 was introduced up to approximately 1 mm short of the root apex, based on radiographs, and used to gently file the canal walls. Afterwards, the fluid in the canal was aspirated using a sterile disposable syringe and transferred to a cryotube containing protease inhibitor phenylmethanesulfonylfluoride (PMSF) and immediately frozen at −80°C. This root canal sample was called S1. Samples from acute apical abscesses were taken by aspiration of the purulent exudate from the swollen mucosa over each abscess. The overlying mucosa was disinfected with 2% chlorhexidine (CHX), and a sterile syringe was used to aspirate pus, which was immediately injected into cryotubes containing PMSF and frozen.

Root canal samples from teeth with asymptomatic apical periodontitis were also taken after chemomechanical procedures. Canals were instrumented at the same appointment in all cases by using BioRaCe instruments (FKG Dentaire, La Chaux-de-Fonds, Switzerland) with the working length (WL) established 1 mm short of the radiographic apex. Master apical files ranged from BR5 (40/.04) to BR7 (60/.04), depending on both the root anatomy and the initial diameter of the root canal. Patency of the apical foramen was confirmed with a K-type file no. 20 throughout the procedures. Irrigation was performed with either 2.5% NaOCl (6 teeth) or 2% CHX (6 teeth), using disposable syringes and NaviTip needles (Ultradent, South Jordan, UT, USA) inserted up to 4 mm short of the WL. Post-instrumentation (S2) samples were taken from the root canals as described for S1 samples.

Polymerase chain reaction for bacterial presence

Ten microliters from clinical samples were subjected to DNA extraction by using the QIAamp DNA Mini Kit (Qiagen, Valencia, CA, USA), following the protocol recommended by the manufacturer. Presence/absence of detectable bacteria in clinical samples was determined by using end-point PCR with universal 16S rRNA gene primers 8f (5′- AGA GTT TGA TYM TGG C - 3′) and 519r (5′- GTR TTA CCG CGG CTG CTG - 3′). Positive and negative controls were included in each batch of samples analyzed. Positive controls consisted of DNA extracted from Enterococcus faecalis (ATCC 29212). Negative controls consisted of sterile ultrapure water instead of sample. All reactions were run in triplicate. PCR reactions were performed in 50 µL of reaction mixture containing 1 µM of each primer, 5 µL of 10× PCR buffer (Fermentas, Burlington, ON, Canada), 3.8 mM MgCl₂, 2.5 U of Taq DNA polymerase (Fermentas) and 0.2 mM of each deoxyribonucleoside triphosphate (Invitrogen Life Technologies, Carlsbad, CA, USA). PCR amplifications were performed in a DNA thermocycler (Mastercycler Personal, Eppendorff, Hamburg, Germany). Cycling conditions consisted of initial denaturation step at 95°C/2 min, followed by 36 cycles at 95°C/30 s, 55°C/1 min, and 72°C/1 min, and final extension at 72°C/10 min. PCR products were subjected to electrophoresis in a 1.5% agarose gel–Tris-borate-EDTA buffer. The gel was stained with GelRed (Biotium, Hayward, CA, USA) and visualized under ultraviolet illumination.

In-solution trypsin digestion

With the purpose to identify the exoproteome, samples were centrifuged at 9.000 g for 20 min, the supernatant was collected and the pellet discarded. Because pilot tests showed that individual root canal samples had very low detectable levels of proteins, samples were pooled for metaproteome analyses: one pool for 6 S1 samples and another pool for the 6 respective S2 samples from canals treated with NaOCl as the irrigant; the same was done for canals irrigated with CHX. Twenty microliters of each sample was used in each pool. Samples from acute apical abscesses were analyzed individually. Before enzymatic digestion with trypsin, samples were reduced with 10 mM dithiothreitol for 1 h at 56 °C and alkylated with 40 mM iodoacetamide for 30 min at room temperature in the dark. Fifty micrograms of proteins from clinical samples, previously dosed by the Folin-Lowry method [17], were digested with trypsin (Promega, Madison, WI, USA) (1∶50, w/w) overnight at 37°C and the peptides were desalted by Zip Tip® (Millipore, Billerica, MA, USA). Samples were vacuum-dried and reconstituted with 0.1% formic acid.

Nanoflow liquid chromatography coupled with LTQ Velos Orbitrap

An aliquot containing 4.5 µL of each pool of root canal samples and the individual samples from abscesses were loaded on LTQ Velos Orbitrap mass spectrometer (Thermo Fisher Scientific, Waltham, MA) connected to nanoflow LC (nLC-MS/MS) by an EASY-nLC system (Proxeon Biosystems, West Palm Beach, FL, USA) through a Proxeon nanoelectrospray ion source. Peptides were separated by a 2%–90% acetonitrile (ACN) gradient in 0.1% formic acid using an analytical column PicoFrit Column (20 cm × ID75 µm, 5 µm particle size, New Objective, Woburn, MA), at a flow of 300 nL/min over 45 min. The nanoelectrospray voltage was set to 1.7 kV and the source temperature was 275°C. All instrument methods for the LTQ-Orbitrap Velos were set up in the data-dependent acquisition mode. The full scan MS spectra (m/z 300–1,600) were acquired in the Orbitrap analyzer after accumulation to a target value of 1e6. Resolution in the Orbitrap was set to r = 60,000 and the 20 most intense peptide ions with charge states ≥2 were sequentially isolated to a target value of 5,000 and fragmented in the linear ion trap by low-energy collision induced dissociation (CID) as dissociation or fragmentation method, normalized collision energy of 35%. The signal threshold for triggering an MS/MS event was set to 500 counts. Dynamic exclusion was enabled with an exclusion size list of 500, exclusion duration of 60 s, and repeat count of 1. An activation q = 0.25 and activation time of 10 ms were used. Samples were analyzed in duplicate.

Peak lists (msf) were generated from the raw data files using Proteome Discoverer version 1.3 (Thermo Fisher Scientific) with Sequest search engine and searched against Human (200,740 sequences and 86,640,852 residues) and Bacterial (377,577 sequences and 139,313,674 residues) protein database downloaded from UniProt (http://www.uniprot.org) in September 2012, with carbamidomethylation as fixed modification, oxidation of methionine as variable modification, one trypsin missed cleavage and a tolerance of 10 ppm for precursor and 1 Da for fragment ions. All datasets were processed using ScaffoldQ+v.3.3.1 software with false discovery rate less that 1% as cut-off. Gene Ontology annotation for functions category (biological or molecular) were obtained using UniProt tools and database.

LC-QTOF analysis

Samples were vacuum-dried and reconstituted in 20 µL of 0.1% formic acid. A total of 4 µL of reconstituted peptide mixture was injected onto an LC-MS system consisting of a 1260 series liquid chromatograph, HPLC-Chip Cube MS interface, and 6530 QTOF mass spectrometer (all Agilent Technologies, Santa Clara, CA). The system was equipped with an HPLC-Chip (Agilent Technologies) that incorporated a 360 nL enrichment column and a 150 mm ×75 µm reverse phase column was packed with Polaris-C18, 3 µm particles. Three analytical replicates were analyzed by mass spectrometry. For each mass spectrometry experiment, peptides were loaded onto the enrichment column with solvent A (water with 0.1% formic acid). A two-step gradient generated at a flow rate 0.3 µL/min was used for peptide elution. This included a linear gradient from 3% solvent B (acetonitrile with 0.1% formic acid) to 40% B over 40 min followed by a sharp increase to 90% B within 5 min. The total running time, including column reconditioning, was 65 min. The column effluent in all cases was directly analyzed by the 6530 QTOF mass spectrometer that was interfaced with an HPLC-Chip Cube nanospray source. The latter was operated at a capillary voltage of 1950 V with a capillary current of 0.085 µA in extended dynamic range (2 GHz) mode. The MS data were acquired in the positive ionization mode using Agilent MassHunter Workstation QTOF B.04.00. During the course of data acquisition, the fragmentor voltage, skimmer voltage, and octopole RF were set to 150, 65, and 750 V, respectively. Auto-MS/MS was performed using scan speed varied based on precursor abundance option with a maximum cycle time of 6.8 s. In each cycle, MS spectra were acquired at 8 Hz (eight spectra/s) (m/z 295–1700), and the twenty most abundant ions (with charge states 2+, 3+, and >3+) exceeding 1000 counts were selected for MS/MS (m/z 50–1700). A medium isolation (4 m/z) window was used for precursor isolation. A collision energy table (m/z 300–1500, z1 [CE 9–69], z2 [CE 9–69], z3 [CE 9–54], z>3 [CE 9–54] was used for fragmentation with beam-typefragmentation or high energy collision dissociation (HCD) in 6530 QTOF. Reference mass correction was activated using a reference mass of 1221.9906. Precursors were set in an exclusion list for 0.15 min after one MS/MS spectra.

LC-QTOF data was searched against the UniProt micro-organism and human database, using the Agilent Spectrum Mill Server software (Rev B.04.00.127). Data were extracted and peak lists were created with the Spectrum Mill Data Extractor program with the following attribute: scans with the same precursor ±0.03 m/z were merged within a time frame of ±60 s. The UniProt microorganism and human database was searched for tryptic/non-tryptic peptides with a mass tolerance of 10 ppm for the precursor ions and a tolerance of 50 ppm for the fragment ions. Two missed cleavages, fixed Modification (carbamidomethylation) and variable modifications [oxidized methionine (M)] were allowed. Spectrum Mill autovalidation was performed at peptide and protein level (1% FDR).

Protein classification

Identified bacterial proteins were categorized according to biological function as follows: transcription and translation, cell division/peptidoglycan synthesis, chemotaxis, DNA process, energy metabolism, fatty acid metabolism, methanogenesis, membrane, nucleotide metabolism, antibiotic resistance, adhesion, pathogenesis/virulence, proteolysis, stress response, transport, vitamin biosynthesis, and other/unknown. Human proteins were classified as follows: cellular process and metabolism, immune system, circulatory system, extracellular connective matrix, and other/unknown. Keratins and other presumed contaminants were removed from the entire dataset.

Microbial proteins involved in pathogenicity

Biofilm formation is an important virulence trait of many bacterial pathogens [18]. It has been shown that apical periodontitis is a disease caused by biofilms colonizing the root canal system [6]. In the present study, three proteins involved in bacterial adhesion were detected: glycosyltransferase 1, tight adherence protein G, and coagulation factor 5/8 type domain protein. The enzyme glycosyltransferase 1 was found in initial root canal samples. This enzyme is involved with production of extracellular polysaccharides that are important for bacterial adhesion to surfaces and biofilm formation [19], [20].

Eight proteins with proteolytic activity were found in this study, most of them associated with abscesses. They included a collagenase, a metalloprotease, a serine protease, an extracellular protease, and endopetidases. These enzymes may play an important role in several ecological and pathogenic effects, including tissue invasion, acquisition of nutrients from proteins, destruction of the connective tissue matrix, and inactivation of host defense molecules [21].

Streptopain (SPE B), a protein of streptococcal origin, was found in one pool of S1 samples. This is a cysteine protease-like exotoxin with several biological effects, including cleavage of fibronectin, vitronectin, and interleukin 1-β precursor and participation in the apoptosis of monocytes and epithelial cells [22]. Therefore, this exotoxin can participate in bacterial invasion of connective tissues, induction of inflammation, reduction of phagocytic activity, and inhibition of tissue repair [22]. Another protein involved in tissue invasion, a putative invasin, was detected in an abscess sample. Tissue invasion can be regarded as a crucial initial step for the pathogenicity of several bacteria [23].

Microbial proteins involved in resistance/survival processes

Several stress proteins were detected in the samples tested. Cells respond to environmental stress by inducing or accelerating the synthesis of specific proteins known as stress proteins, including heat-shock proteins (HSPs), which act as molecular chaperones in the assembly and folding of proteins, as well as proteases when damaged proteins need to be degraded. In addition to being found intracellularly, HSPs can also be located on the cell surface and on outer membrane vesicles released by bacteria [24]. The chaperone DnaK (HSP70) was found in one abscess sample and in S2 samples. Another chaperone, (GroEL or HSP60), was found in S1 samples from asymptomatic root canal infections. HslU, which is another HSP, was also found in this study and has been shown to provide an essential ATPase activity [25]. These findings may indicate a stress response to host defenses and treatment, in an attempt to survive and/or recover from damage. These stress proteins may also act as virulence factors via the following mechanisms: cytotoxicity, adhesion, modulation of host cell activities, and induction of synthesis of pro-inflammatory cytokines [24], [26].

Others stress response proteins detected in this study include alkyl hydroperoxide reductase, whose expression has been shown to be up-regulated in some anaerobes upon oxidative stress [27]; a periplasmic sensor signal transduction histidine kinase, which senses specific environmental stimuli [28]; a tetratricopeptide TPR_2 repeat protein, which mediates protein-protein interactions and the assembly of multiprotein complexes, and may be involved in protein folding [29], [30].

Catalase and quinone oxidoreductase are involved in the response to oxidative stress [31], [32] and were found in post-treatment samples. Catalase also plays an important role in the reactivation of bacteria present in a viable but noncultivable state [33].

A toxin component of the putative toxin-antitoxin system was detected in an abscess sample. This system may mediate the general stress response, and participate in the regulation of biofilm and persister cell formation, which is an important mechanism of reduced susceptibility of biofilms to antibiotics [34].

Proteins involved in DNA repair are also of interest because they may participate in the response to stress conditions induced by treatment or host defenses. Some of the DNA repair-related proteins detected in this study were A/G-specific adenine glycosylase, DNA ligase, ATP-dependent helicase, and UvrABC system proteins. Of the 12 proteins related to DNA repair, 8 were found in abscesses; in these cases, there may be a strong oxidative stress caused by phagocytes during the combat to infection [35].

Two enzymes that confer resistance to antibiotics were detected: beta-lactamase (resistance to penicillins and cephalosporins) and TetR (resistance to tetracyclines). Previous studies have detected the genes for these antibiotic resistance enzymes directly in root canals samples [36] or in isolates from endodontic infections [37], but in situ expression of the gene products have not been previously determined. In a previous study using a metaproteomic approach, Nandakumar et al. [16] also identified enzymes related to resistance to tetracyclines and beta-lactams. These findings confirm that endodontic bacteria may produce antibiotic resistance enzymes in the root canal environment and this may have some important implications, since antibiotics are used for some acute conditions [38] or have even been proposed as topical irrigants or intracanal medication during root canal treatment [39], [40].

Other proteins of interest

Another noteworthy finding was the identification of the enzyme methyl coenzyme M reductase of archaeal origin. Methanogenic archaea have been detected in endodontic infections by DNA-based molecular methods, including by an assay directed to mcrA, which encodes for the methyl coenzyme M reductase [41]. The presence of products of this microorganism in endodontic infections indicates metabolic activity in a complex network of microbial interactions.

CRISPR consists of identical repeated DNA sequences interspaced by highly variable sequences (spacers). CRISPR-associated (cas) genes encode conserved proteins that together with CRISPRs compose the CRISPR/Cas system, which is present in many prokaryotes and confer protection against invasion by phages, plasmids, and transposons [42]. A previous study reported the occurrence of CRISPR-cas in several endodontic isolates of E. faecalis and suggested a role for this system in modulation of interactions in the endodontic polymicrobial biofilm [43]. In the present study, CRISPR-associated proteins were identified in samples taken before and after treatment, confirming that this prokaryote defense system is present in members of endodontic infections.

Host defense proteins

Human proteins were identified in all samples, being more frequent in abscess samples and in S1 samples from asymptomatic teeth. The fact that most proteins identified are intracellular and are involved in cellular and metabolic processes can be possibly explained by rupture of cells during the inflammatory process in response to infection, especially in abscesses. Another possible source of intracellular proteins is the remnants of necrotic pulp tissue in the canal. However, it must be recognized that cellular rupture may also have occurred during sample taking or processing steps. In addition, 24 proteins related to the innate or adaptive immune system were found. They include defensins 1 and 3 and myeloperoxidases, which are produced by polymorphonuclear neutrophils. These proteins are not only involved with defense against microorganisms but also in tissue destruction [44], [45]. Other immunity-related proteins of interest include components of the complement system, imunoglobulins, protease inhibitors, receptors involved in the regulation of T cells, such as CTLA4 (Cytotoxic T-lymphocyte protein 4), and T-LAK cell-originated protein kinase (Lymphokine-activated killer T-cell-originated protein kinase), which is involved in the activation of lymphoid cells. Databases for human proteins are much more complete and accurate, reducing the number of proteins classified as unknown.

Abscess versus asymptomatic cases

Metaproteomic analyses revealed a larger number of proteins in the abscess samples when compared to samples from asymptomatic infection, in spite of the fact that samples from asymptomatic teeth came from a larger number of cases. Studies have shown that the number of bacterial cells and species in abscesses are greater than in asymptomatic cases [38], which may help explain our findings. Also, it is expected that bacterial cells engaged in an acute infection are in active processing of protein expression machinery. However, one should not discard the possibility that in abscesses, samples were taken from pus, in which an acute host response is established and a high killing rate of host cells and bacteria is expected. This would generate more proteins, including cytoplasmic proteins.

Post-treatment samples

Analysis of the total number of proteins detected after treatment revealed a decrease in the CHX group but an increase in the NaOCl group. The purpose of analyzing post-instrumentation samples was two-fold: first, to evaluate the effects of treatment on the metaproteome; second, to evaluate if the proteins expressed after treatment procedures would differ from those present before treatment, considering the possibility that microorganisms resisting to treatment may produce a different pattern of proteins. The reduction in the CHX group was completely expected, but the increased in NaOCl group may be related to the tissue dissolving ability of this substance and its high destructive effects on cells, releasing cytoplasmic proteins to the environment. This is coherent with the types of proteins found in S2 samples from the NaOCl group. In the group of CHX, one stress response protein was found, and this may be related to a response to treatment. However, a more distinct pattern was not evident. This may have been due to one of the following reasons: limitation of the sampling approach; few bacterial cells remaining after treatment; and the short time elapsed between treatment and sampling, not allowing sufficient time for bacteria to change their pattern of protein expression in order to adapt themselves to the altered environment. If the latter is true, further studies should be performed to analyze the metaproteome associated with bacterial persistence after treatment, and for this, researchers should wait a little longer to collect samples after intracanal procedures.

Technical and analytical considerations

Actually, in the present study, efforts were expended towards evaluation of the exoproteome. However, many cytoplasmic proteins were detected. This may be due to cells that recently died and released their protein content in the environment (especially in abscesses and in samples taken after treatment), which are still considered as part of the exoproteome if these proteins are stable and remain in abundance [46]. However, presence of cytoplasmic proteins may also have been a result of sample processing that resulted in cell lysis.

Several proteins identified in this study were not linked to endodontic bacteria, which may be related to databases with incomplete information for oral bacteria [47]. This also may help explain why a large number of proteins remained unidentified and were classified as having unknown function. It is salient to point out that the majority of the bacteria found in the endodontic microbiome, especially the as-yet-uncultivated portion, have not had their genomes sequenced. The same is true for bacteria in other human and environmental sites. Although the extent of protein sequence conservation is largely unknown, the possibility exists that most of the peptides attributed to nonoral bacteria actually belong to phylogenetically similar oral relatives of those species, which remain to be sequenced [48]. It is also important to note that proteins from uncultivated or not-yet-sequenced cultivated species that are abundant in the community may potentially pass unnoticed as there is no representative sequence in the database.

The number of proteins identified may have also been affected by the principle of parsimony adopted by the softwares used in this study, which reduce the occurrence of redundant sequences [49]. Moreover, the utilization of rigid filters, such as the 1% False Discovery Rate, and filtering of data generated from Proteome Discoverer by Scaffold may have contributed to a limited but more reliable protein identification [50], [51]. Another limitation of this study is that, unlike DNA, proteins cannot be amplified, therefore less abundant proteins may be undetected [52].

In contrast to bacterial isolates cultivated in the laboratory, human clinical samples, especially from a small environment like the root canal, provide a reduced amount of biomass available for proteomic analysis. Therefore, for root canal samples to be properly analyzed they had to be pooled. In addition, many proteins were found only by one of the methods used. This may be related to the different aliquots used in each identification, which may influence detection of proteins found in low abundance. The peptide mixture generated by the shotgun proteomic strategy is so complex to permit that the mass spectrometer acquires the mass spectra of all peptides in a single LC-MS/MS run. Consequently, these experiments must be frequently repeated so as to increase the number of peptides obtained by mass spectra [53]. This approach was used in the present study.