Proteomic study of plaque fluid in high caries and caries free children

2.1Sampling

According to the third version of the caries diagnosis standard of the Basic Methods of WHO Oral Health Survey, 8 children (CF, dmfs = 0) and CS, dmfs > 10 (3 ∼ 10) were selected respectively, without any oral diseases, systemic health, and no history of medication (including fluorogens) in March, 2020.

The children required no oral hygiene care on the night before sampling, starting from the morning, and 12 h fasting and drinking before sampling. Washing the collection area with sterile saline. Avoid moisture with cotton roll. In CF group, we scraped all the dental plaque of smooth suface of deciduous molar teeth. In CS group, we scraped all the healthy smooth enamel surface [1, 2]. The parents of the child knowingly agreed for the experimental procedure.

2.2Preparation of plaque sample

The plaque was collected in a pre-cooled 0.5 ml centrifuge tube in crushed ice and immediately sent to the laboratory and centrifuged at 4∘C, 15,000 g for 60 minutes to remove the bacteria, insoluble impurities and other substances, and the supernatant liquid obtained was the plaque fluid [3]. The plaque fluid was carefully transferred into the new EP tube, a protease inhibitor (Inhibitor cocktail complete, Roche) was added, and the concentrate was collected with the Millipore (cutoff: 5 KD) ultrafiltration tube at 4∘C, 4500 g. Quantitative protein concentration of the centrifugal fluid was done by Bradford (Bio-Rad, USA), and stored at -80∘C [4, 5].

Figure 1.

Experimental workflow for TMT labeling and analysis: TMT 10-plex labeling was performed for three sets of technical replicates. Each condition was labeled as follows for the three technical replicates: CF (127N, 129N, 130C), HC (128N, 139C, 131). Moreover, equally amount of proteins from all samples were pooled as an internal standard (IS: 126). The labeled fractions were combined and subjected to High-pH Reversed-Phase Fractionation and desalting, followed by separation using liquid chromatography mass spectrometry (LC-MS/MS), and bioinformatics data analysis.

Figure 2.

Glycolysis/Gluconeogenesis.

Figure 3.

Pyruvate metabolism.

2.3Preparation of enzyme solution

In order to reduce the differences between the individuals, the plaque proteins of the children without caries and children with high caries were mixed in equal amounts, and a sample library was established. CF and CS plaque were divided into three parts..Each protein samples were supplemented with lysis buffer (8 mol/Lurea, 40 mmol/L Tris, 65 mmol/L DTT) to the total volume of 100 μl, and mixed with 1 M DTT at 37∘C for 2.5 hours. Subsequently, 10 μl 1 M IAA, was added at room temperature in absence of light and the reaction was continued for 40 minutes. After the above treatment, the protein was completely deformed, the disulfide bond opened up. It was then precipitated with a 5 x volume of precooled acetone (-20∘C) by resting overnight (16 hours) at -20∘C. Then, the mixture was centrifuged at 14,000 g for 40 minutes; the organic solvent was removed using pre-cooled acetone at -20∘C, centrifuged again at 14,000 g for 40 minutes. To remove the salt ions, precooled 70% ethanol (-20∘C) was added and centrifuged at 14,000 g for 40 minutes and then freeze dried. After the enzyme solution was completed, the ultrafiltration membrane was employed with Millipore 10 KD aperture size to collect the filter fluid and freeze dried at -80∘C.

Figure 4.

Tricarboxylic acid cycle.

Figure 5.

Pentose phosphate pathway.

Figure 6.

Fructose mannose metabolism.

2.4Liquid chromatography-mass spectrometry analysis

20 μg of the prepared enzyme solutions were taken and analyzed using the Ettan MDLC liquid chromatography-series mass spectroscopy system (GE Healthcare, Piscataway, NJ, USA). RP trap columns (Zorbax 300 SBC 18, Agilent Technologies, Palo Alto, CA, USA) were employed with an automatic sample feeder. The sample was desalted using the C18 trap column. The sample was separated on the C18 column (Millipore water in A phase and 0.1% FA 84% B finishing water solution, at a gradient rising from 4% to 50% B phase within 2 hours). The separation speed was 2 μl/min. The samples were removed from the column, and subjected directly to electric spray sourceTMLTQTMLinear ion-trap mass spectrometry (Thermo Electron, San Jose, CA, USA). The LTQ mass spectrometry was performed in the automatic gain control (AGC) mode, with the ion source parameters set as: electric spray voltage – 3.2  kV; capillary temperature – 170∘C. Full-scan mass spectrograms were collected in the profile mode, while MS/MS maps were collected in the centroid mode, with 5 centroid mode scans after each profile mode. The analysis was repeated 3 times per sample.

Figure 7.

Fructose mannose metabolism.

2.5Database search and data processing

After obtaining all peptides with quantitative information, It was tested for all p< 0.05, and searched with the SEQUEST program (Bioworks Browser Software suite, Thermo Electron, version 3.1) at the National Center for Biotechnology Information (www.ncbi.nlm.nih.gov) based on the genome information of 24 common oral bacteria in Genome. It was downloaded to get the corresponding protein sequence database (24 common oral bacteria including Streptococcus gordonii str. Challis substr. CH1, Streptococcus mutans UA159, Streptococcus sanguinis SK36, Streptococcus salivarius SK126, Actinomyces odontolyticus ATCC 17982, Lactobacillus acidophilus, Lactobacillus fermentum, Neisseria mucosa ATCC 25996, Neisseria sicca ATCC 29256, Neisseria subflava NJ9703, Veillonella dispar ATCC 17748, Veillonella parvula DSM 2008, Corynebacterium matruchotii, Corynebacterium diphtheriae NCTC 13129, Leptotrichia buccalis DSM 1135, Eikenella corrodens ATCC 23834, Porphyromonas gingivalis, Prevotella melaninogenica ATCC 25845, Capnocytophaga gingivalis ATCC 33624, Capnocytophaga ochracea DSM 7271, Capnocytophaga sputigena ATCC 33612, Fusobacterium nucleatum, Treponema denticola ATCC 35405, Campylobacter rectus RM3267). The polypeptide molecules were identified and their corresponding proteins were relatively quantified. The filter parameters were: when Charge + 1, Xcorr ⩾ 1.9; when Charge + 2, Xcorr ⩾ 2.2; when Charge + 3, Xcorr ⩾ 3.75; where DelCN ⩾ 0.1.

2.6Pathway analysis

The metabolic pathway map was downloaded for the 24 common oral bacteria from the www.kegg.com website to analyze the distribution of the identified proteins in each pathway.

2.7Microbial function and pathway analysis

All identified microbial protein sequences were annotated by BLASTP (version: 2.2.31+) against the UniProtKB/Swiss-Prot database that included 560,118 sequences and the COG database (version: 2014) that included 1,781,653 sequences. Furthermore, InterProScan (version: 5.35–74.0) provided the gene ontology (GO) functional annotations of all identified microbial protein sequences with default parameters. Some in-house Perl scripts and R scripts were used to complete the GO functional classification and statistics analysis. The KEGG orthology (KO) annotation of the proteins were performed using KOBAS (version: 3.0) with E-value≤1e-05 and rank≤5. Moreover, KOBAS was also used to perform a binomial test between the two groups in the KEGG pathway levels. Assignment of the proteins in the KEGG pathway levels was then visualized with Krona Excel Template.

2.8PRM verification

Based on the results of the original label-free based quantitative meta-proteomics analysis, we selected the appropriate target peptides of the candidate proteins and performed targeted shotgun MS to finally determine the peptides of the target proteins with reliable identification information, which was used in the PRM quantification analysis. The peptide information suitable for PRM analysis was imported into the Xcalibur software program for PRM setup. Briefly, 1 μg peptide from each sample was taken for LC-PRM/MS analysis. After sample loading, chromatographic separation was performed using a Thermo Scientific EASY-nLC nano-HPLC system. The following buffer was used: A solution was 0.1% formic acid aqueous solution and solution B was a mixed solution of 0.1% formic acid, acetonitrile, and water (95% of acetonitrile). The column was first equilibrated with 95% A solution. The sample was injected into a Trap column (100 μm × 20 mm, 5 μm-C18, Dr. Maisch GmbH) and subjected to gradient separation through a chromatography column (75 μm × 150 mm, 3 μm-C18, Dr. Maisch GmbH) at a flow rate of 250 nL/min. The liquid phase separation gradient was as follows: 0–25 minutes, linear gradient of B liquid from 5% to 18%; 25–45 minutes, linear gradient of B liquid from 18% to 50%; 45–48 minutes, linear gradient of B liquid from 50% to 95%; and 48–60 minutes, B liquid maintained at 95%. The peptides were separated and subjected to targeted PRM/MS using a Q-Exactive mass spectrometer (Thermo Scientific). The analysis time was 60 min. The parameters were set as follows: detection mode – positive; parent ion scanning range – 350–1500 m/z; capillary voltage – 1.8 kv; isolation width – 1.6 Th; first-order MS resolution – 70,000 @m/z 200; AGC target – 3e6; first-level maximum IT – 250 ms. Peptide secondary MS was performed as follows: for each full scan, target peptides of the precursor m/z were sequentially selected based on the inclusion list for second-order MS (MS2) scan with the parameters as follows: resolution – 35,000@m/z 200; AGC target – 3e6; Level 2 Maximum IT – 120 ms; MS2 Activation Type – HCD; Peptide fragmentation – nitrogen; Isolation window – 2.0 Th; Normalized collision energy – 28 eV. The obtained PRM data of the raw RAW file was analyzed using the Skyline 3.5 software program. Statistical analysis was completed with R and the ‘MetaboAnalystR’ package.

Data availability: All the MS raw files were submitted to the Proteome Xchange Consortium (http://www.proteomexchange.com) via the PRIDE partner repository with the data set identifier.

2.9Statistical analysis

Statistical analysis was performed using the SPSS version 19 software. The experiments were statistically analyzed with the independent sample T test. All statistical analysis were performed at 95% significance level using 2-tailed analysis.

Figure 8.

ABC transport proteins.

Figure 9.

Glycolysis reaction formula.

3.1Label-free 2D-LC-MS/MS Identification

Liquid chromatography-series mass were analyzed in CF, CS group (we repeated the experiments for three times), including 1804 quantitative information peptides, 39 in CF group, and 30 in CS group. The data from liquid chromatography-series mass spectroscopy was analyzed by DecyderMS software, and 603 peptide data sets were obtained, including 202 in CF group. Among them, 33 peptide has a much higher significant differences which was greater than 1.5 times (Table 1), On the other hand, we found 401 peptide in CS group. And there were 199 among them has a greater differences which more than 1.5 times (173 without redundant proteins, Table 2).

3.2Gene ontology function analysis

In total, 9194 (80.16%) identified microbial leading proteins corresponded to at least one GO term using InterProScan and in-house Perl Scripts. The number of proteins was counted at GO level 2 of biological process, cellular component, and molecular function ontology, respectively (Fig. 7). Moreover, the proteins of two groups that corresponded to the GO term of biological process, cellular component, and molecular function ontology were counted. Then, Fisher’s exact test was performed to compare the difference of the protein group number between the two groups. Based on a p< 0.05 level, the GO terms with CS group included 35 in biological process ontology, 2 in cellular component, and 30 in molecular function ontology. The GO terms with CF group included 59 in biological process ontology, 17 in cellular component, and 30 in molecular function ontology (Fig. 8).

The count of proteins corresponding to GO terms of all samples were performed using in-house Perl and R scripts. With a p< 0.05, 80 GO terms (33 in biological process ontology, 5 in cellular component ontology, and 42 in molecular function ontology) were found in atleast12 pair samples with high caries, while only 10 GO terms (8 in biological process ontology and 2 in molecular function ontology) were found in the caries-free specimen. Then, we transformed the p-value with negative log10, and visualized them with heatmaps using R (Fig. 9).

3.3PRM verification

Parallel reaction monitoring (PRM) mass spectrometry was used to verify the target peptides determined by the metaproteomic analysis and a pre-experiment, and the 171 candidate peptides of the target protein were subjected to LC-PRM/MS analysis. In total, 103 candidate proteins were quantified by LC-PRM/MS, including 3 proteins (hinf_c_1_1270, fnuc2539_c_1_361, cgin_c_10_1771) that were determined as the differential expressed protein candidates in the metaprotemics study. The Skyline analysis results of each candidate peptide were shown.

Meanwhile, 17 differential expressed peptide candidates were confirmed by the Mann-Whitney U test with a p< 0.05 cutoff, 15 differential expressed peptide candidates were obtained by paired-samples T test with a p< 0.05 cutoff, and 7 peptides in common (Table 6).

Figure 10.

Glycolysis.

Figure 11.

Gluconeogenesis.

4.1Saccharolysis/sugar lyiogenesis

4.1.1Glycolysis

Saccharolysis (glycolysis) is the first step in the sugar metabolism process of all biological cells. During this process, a molecule of glucose undergoes a ten-step enzyme reaction to form two molecules of pyruvate with the formation of ATP (Figs 8–10). This suggests the presence of active sugar metabolism in the caries-causing plaque, presumably due to the highly expressed enzymes that use intracellular polysaccharide, mainly glycogen as a source of energy, resulting in tooth demineralization [6].

Phosphorylation of the phosphoric fructose (phosphorylation of fructose-6-phosphate) is the third step of glycolysis, further phosphorylation of C on 6-phosphate fructose initially produces 1,6-diphosphate fructose supplied by ATP. The reaction is catalyzed by phosphate fructose kinase (phosphofructokinase, PFK), which is an important speed limiting enzyme in the sugar aerobic oxidation process. Phosphoric transfer of phosphoenol pyruvate is the final reaction in glycolysis that involves the transfer of the high energy phosphorate group from phosphoroll pyruvate to ADP, and is catalyzed by pyruvate kinase, PK, which is another phosphorylation process at the substrate level.

Figure 12.

Tricarboxylic acid.

Table 5

Co expression ABC transporter data set of CS&CF group (P> 0.05)

Num	GI ID	ID details
1	gi|197735467	ABC superfamily ATP binding cassette transport
2	gi|229212097	ABC-type multidrug transport system
3	gi|229212809	ABC-type multidrug transport system
4	gi|229210643	ABC-type oligopeptide transport system
5	gi|42527941	ABC transporter, ATP-binding/permease protein
6	gi|229211946	ABC-type cobalt transport system
7	gi|229211911	ABC-type Fe3+ transport system
8	gi|229211590	ABC-type metal ion transport system
9	gi|42526433	ABC transporter, ATP-binding/permease protein
10	gi|42528282	ABC transporter, ATP-binding/permease protein
11	gi|42525789	ABC transporter, ATP-binding/permease protein
12	gi|42526418	ABC transporter, ATP-binding protein
13	gi|42526693	ABC transporter, ATP-binding protein
14	gi|42525831	ABC transporter ATP-binding protein/peptidase
15	gi|229210567	amino acid/amide ABC transporter membrane prot
16	gi|19704374	branched chain amino acid ABC transporter
17	gi|197736157	dipeptide/oligopeptide/nickel (Ni)2+ ABC supe
18	gi|228274621	excinuclease ABC subunit A
19	gi|42526877	excinuclease ABC subunit B
20	gi|42527978	excinuclease ABC, C subunit
21	gi|42527718	galactoside ABC transporter, ATP-binding protein
22	gi|197736905	iron (Fe)3+ ABC superfamily ATP binding casse
23	gi|42526259	iron compound ABC transporter
24	gi|42526690	iron compound ABC transporter
25	gi|197736077	nickel (Ni)2+ ABC superfamily ATP binding cas
26	gi|197736393	possible nitrate/sulfonate/bicarbonate ABC su

Table 6

Candidate differential expressed peptides were confirmed by Mann-Whitney U test and paired-samples T test in common

Peptide sequence	p-value with Mann-Whitney U test	p-value with paired-sample T test	Fold change
VVEYVEKPVIVYR	4.90E-02	4.63E-02	8.51
YSFSTCYNSER	3.56E-03	4.40E-03	5.21
TAALENAAEGGFNKK	2.19E-02	5.22E-03	4.33
VVVEVLSQGK	2.27E-02	5.87E-03	3.73
LNNCPTSPR	4.35E-02	3.32E-02	3.71
VLDELTALR	4.97E-02	4.18E-02	2.50
SPEEAYEHAK	4.91E-02	2.51E-02	2.14

Figure 13.

A bar chart of protein count at GO level 2 of biological process, cellular component, and molecular function ontology using the ‘ggplot2’ package. The number markers at the bar were protein count of the GO term.

Reversible reactions in glycolysis include 7 steps: glucose phosphate isomerase is involved in the second step; heterogeneous reaction of glucose phosphate (isomerization of glucose-6-phosphate); participation in step 4 1.6-diphosphate pyrolysis reaction (i.e. cleavage of fructose 1, 6 di/bis phosphate from fructose-1, 6-Bisphosphate Aldoase (fructose-1, 6-bisphosphate aldolase); participation in step 6: 3-glyceraldehyde phosphate oxidation reaction (oxidation of glyceraldehydes-3-phosphate); 3-glyceraldehyde phosphate dehydrogenase (glyceraldehyde 3-phosphate dehydrogenase); phosphorate kinase (phosphoglycerate kinase, PGK) participating in high energy phosphate bond transfer of step 1.3-diphosphate glyceric acid. These enzymes showed a high expression in the plaque fluid in both two groups. Studies show that the above catalytic enzymes participate in the sugar heterogeneous reaction at the same time. So, we speculate that both in caries-free or caries causogenic plaque, the glycolysis and sugar heterogeneous reaction can be accurately adjusted, so that the bacteria in the plaque can use the excess sugar in the form of glycogen as energy storage to meet their own energy needs.

The final reversible reaction is step 2-phosphoglyceric acid, catalyzed by enolase, which was highly expressed in the CS group (P< 0.05, ratio = 1.75), and mainly corresponds to the bacterial glycolysis pathway, suggesting that glycololysis plays an extremely important role in carie.

4.2Tricarboxylic acid circulation

The cytoplasm of prokaryotes is the site of tricarboxylic acid circulation, but most enzymes are found in plaque fluid, including isocitrate dehydrogenase (isocitrate dehydrogenase, IDH) (Fig. 12), succinate dehydrogenase, malate dehydrogenase, succinyl-assisted A synthase, and α-ketovaltarate dehydrogenase complex in the CS group are unknown.

The IDH superfamily is ancient and large, widespread in the three boundaries of life (archaea, bacteria and eukaryotes). IDH catalyzes isocitric acid to α-ketopenta in tricarboxylic acid (TCA) cycle, bringing NAD+ or the NADP+ to NADH or a NADPH. It not only plays an important role in energy metabolism, amino acids and vitamin synthesis, but also plays a key regulatory role in the TCA circulation and the carbon flux distribution of acid bypass [11, 12, 13, 14].

4.3ABC transporter protein

This study identified more than 50 corresponding adenosine triphosphate binding box transporters (ATP-binding cassette transporter, ABC transporters) in children who were caries-free or with high caries. Associated with the transport of iron, nickel, cobalt ions, iron compounds, metal ions, oligopeptides, dipeptides, amino acids, branched chain amino acids, lactosidase, it is seen that ABC transporters are involved in various physiological functions of bacteria. It is worth noting that we identified 13 highly-expressed ABC transporters in the CS group, 6 proteins expressed more than 1.5 times, while only 4 ABC transporters were highly expressed in the CF group and all were less than 1.5 times, suggesting that material transformation in the process of caries-causing transformation of bacteria. Among them, ABC transporter of the group CS was expressed 2.6 times high, which may be the potential target molecule for caries activity evaluation.

4.4Molecular partner

This experiment identified 34 molecular companion peptide segments, belonging to the Hsp60 (GroEL) family, Hsp70 (DnaK) family, Hsp90 (Http pG) family, and Hspl00 (CIp) family. GroEL’s E. coli is a homologous oligomer complex that plays an important role in the correct folding and assembly of newborn proteins and the recovery of degenerative proteins under thermal or chemical adversity. Although it has been determined that GroEL is located in the cytoplasm, the surface of some pathogens can express GroEL, and this is generally associated with the role the molecular partners play during adhesion [15]. GroEL expression was also detected in plaque fluid. It is speculated that the Hsp molecular companion, as the main antigen of most pathogens, may cause GroEL rearrangement on the bacterial cell membrane in the process of bacterial infection or under stress stimulation. Unlike the first two HSP, we found that HtpG was significantly upregulated in the CS group, but relevant studies showed that HSP90 synthesis speed and synthesis volume after stress were not significantly different than the above HSP, so its mechanism of cell protection needs to be further clarified.

4.5Phosphorylation modification

In this experiment, the presence of two-component systems and Phosphotransferase system (PTS) was detected and consistently increased in expression in group CS, suggesting active protein phosphorylation modification and intercellular signaling in pathogenic biofilms.

The PTS system usually consists of five proteins including the enzyme I, enzyme (including three subunits of A, B and C) and phosphoolenol pyruvate (phosphoenolpyruvate, HPr). High expression of PTS system IIA, IIC in the CS group was also observed in this trial, indicating that the glucose intake within the flora was active during the occurrence of caries and development.

4.6Bacterial phages

At present, specific phages for bacteria such as Actinomyces, Actinobacillus Actinomycetecomitans, Actinomyces viscosus, Enterococcus faecalis (Actinomyces Actinobacillus actinomycetecomitans, Actinomyces viscosus, Enterococcus faecalis have been isolated in plaque and saliva respectively, In this experiment, eight phage-related proteins were detected in two groups, two of them (phages and microstructural proteins) were highly expressed in the high caries group. Although less data on this finding is available, some scholars have speculated on the potential prospect of bacteriophages in the caries prevention and treatment process [16, 17, 18, 19]. Taking a phage as a plaque control method may be a new area.

4.7Membrane protein

The outer membrane is the contact surface of the bacteria and the external environment, and its main components are lipid, lipoprotein and outer membrane protein. Outer membrane protein is the main component of the outer membrane, which plays an important role in material transport, information identification, cell adsorption, and outer membrane protein and secreted protein are also the first choice protein of vaccine antigens. A variety of high-expression of bacterial outer membrane protein, lipoprotein, and apolipoprotein were detected in this experiment, but its function in the occurrence and development of caries needs yet to be studied specifically.

4.8Other key proteins

After PRM verification, we found some key proteins which may play an important role during the development of dental caries in children (Fig. 13), which are discussed below.

Our experimental results show that although PFK and PK are highly expressed in both CF, CS groups, it exceeds 1.5 times in the CS group, indicating that PEK and PK play an important role in the occurrence and development of caries, especially in bacterial caries, and also shows that the glycolysis process can provide energy for the life activity of bacterial cells and maintain the bacterial physiological function in mature plaque.

In addition, this experiment also found that glycan phosphate isomerase was highly expressed in the plaque fluid in children with high caries, indicating that this enzyme may play an important role in the plaque caries-causing process by promoting the effective energy generation of the plaque bacteria.

ADH is a zinc-containing metallicase widespread in human and animal liver, plant and microbial cells, with a broad substrate specificity to convert pyruvate produced by glycolysis into acetaldehyde and NAD and NAD+, thereby generating the energy needed for glycolysis. In this experiment, we found that ADH was highly expressed in the high caries group. It remains to be further studied, whether it also plays a role in the caries-induced transformation of plaque and the information exchange between bacteria.

Hydrohydrogenase (hydrogenase) is an important class of biological enzymes present in the microorganisms that catalyze the oxidation of hydrogen or hydrogen production from reduced protons. In our experimental results, the presence of iron hydrogenase and ferroxygen reduction protein was detected in both groups and significantly highly expressed in the high caries group, presuming that this may be one of the mechanisms where acid-producing and acid-resistant bacteria survive in the acid-induced plaque.

In conclusion, in the present study, we obtained 1804 peptides with quantitative information, including 395 in CF group, 30 in CS group, 1735 peptides in both the groups. The DeCyderTMMS software conducted further statistics and analysis, and obtained 603 data sets of different peptide expression. The function of 391 peptides was unknown in this experimental dataset, and 47 peptides were highly expressed in the high caries group. Their specific function and their relationship with the caries are still uncertain. Nevertheless, macroproteomics, secretory proteomics and bioinformatics analysis has still provided a very good platform for our research, so that we could conduct a good comprehensive analysis of the proteins known to-date. Further research will be conducive to clarify the cause, and looking for disease-related biomarkers, which will be one of our future research directions.