2.1

Ethics statement

This study was ethically approved by Peking University Biomedical Ethics Committee (issuing number: PKUSSIRB-2013060) and followed the principles of the Declaration of Helsinki. Parents of the pediatric participants have all signed the informed consent form before the start of the study.

2.2

Subjects

According to previous studies, at least 10 subjects in each group (ECC and control) were needed. From the epidemiological literatures in China, we have found that the incidence of primary caries in 3–4 year-old children was about 43.6% (60% in a two-year study ). As we set the lost rate to 20% for each time point, at least 72 individuals were required to meet our need as a rough estimate. Hence, eighty-two preschool children aged 3 years old in one kindergarten in Beijing, China, who were found to be caries-free on examination (total number of decayed, missing and filled tooth surfaces = 0), were initially recruited at baseline in May 2014 and September 2014. Then, they were followed up every 6 months approximately and this would last for 1 year. The oral health status of each individual was determined by one professional dentist. The diagnostic criteria were according to the WHO Oral Health Surveys Basic Methods (5th edition, 2013). 36 children were excluded at the other two time points (details shown in S Table 1). Finally, 15 children who both meet diagnosis of ECC and the inclusion criteria at the 6- or 12-month time points were classified into the experimental group (Group C). The other 31 remained caries-free, and 15 of them (who were age- and sex-matched) were selected to be the healthy group (Group H). All study subjects were systemically healthy. Those with influenza or an upper respiratory tract infection and those who had received antibiotic therapy within 1 month of any sampling time were excluded from the study. Those who were absent at any time point were excluded. S Table 2 shows detailed information and caries status of the subjects.

2.3

Saliva collection and processing

All individuals were instructed to rinse orally with water after breakfast and then rest for 10 min before saliva collection at 8:30 a.m. Stimulated whole saliva samples were collected for 5 min. The 1.5 mL saliva samples were immediately placed on ice. Insoluble material, cells and debris were removed by centrifugation at 10,000 × g for 10 min at 4 °C. The supernatants were collected, and 1 mM ethylene diamine tetraacetic acid (Sigma, St. Louis, MO, USA) and 1 mM phenylmethyl sulfonyl fluoride (Sigma) were added to inhibit protease activity. Protein concentrations were measured by the Lowry method and the ELx808 Protein Assay (BioTek, Hercules, CA, USA). Supernatants were kept at −80 °C separately until further analysis.

2.4

WCX fractionation and MALDI-TOF MS

All saliva samples were fractionated using WCX magnetic beads (MB) (Bioyong Tech). Samples were purified and isolated by the following steps: (1) The beads were turned upside down to mix them well before use; (2) 20 μL of beads, 150 μL of MB-WCX binding solution, and 10 μL of salivary sample were gently mixed and incubated for 5 min at room temperature; (3) the tubes were placed on the MB separation device (Bioyong Tech) for 1 min, and the beads adsorbed onto the tube wall; (4) the supernatant was then removed, and the beads were washed by mixing thoroughly with 150 μL of MB washing solution. Two minutes later, the tubes were placed on the separation device for 1 min; (5) step 4 was repeated and the supernatants removed; (6) 10 μL of MB elution solution were added and mixed thoroughly with the beads, then the tubes were put on the separation device for 2 min; (7) the clear supernatant was transferred into a fresh tube, and the peptides were analyzed directly on a ClinTOF instrument (Bioyong Tech) or stored at −20 °C and analyzed within 24 h.

The matrix solution was 8 mg/mL CHCA in 50% acetonitrile/0.1% TFA/49.9% deionized water. First, 1 μL of purified peptide solution was spotted onto a MALDI-TOF MS target from ClinTOF. After drying at room temperature, 1 μL of matrix solution was spotted to cover the sample, which was dried again before analysis. MALDI-TOF MS was performed using a ClinTOF instrument. Calibration of the MALDI-TOF MS was done by a three-peptide mixture (monoisotopic molecular weights of 1533.8582, 2465.1989, and 5730.6087 Da; Sigma product numbers P2613, A8346, and I6279, respectively) before analysis. Profile spectra were acquired from an average of 400 laser shots per sample. The mass range 1000–10,000 Da was collected. Each sample of saliva was analyzed three times, and the mean values of the intensities and masses of each peak were used for analysis.

2.5

Data processing and statistical analysis

All the spectra obtained from the saliva samples were analyzed using BioExplorer (Bioyong Tech) to obtain the mean relative peak intensities; then baselines were subtracted, spectra normalized using total ion current, and peak m / z values and intensities determined in the mass range 1000–10,000 Da. A signal-to-noise (S/N) ratio >6 was required. To align the spectra, a mass shift of no more than 0.1% was determined. The peak area was used for quantitative standardization. The k-nearest neighbor (KNN) algorithm in this software suite was used to establish the best pattern of differentiation model for the development of ECC. Analysis of variance (ANOVA) or Kruskal-Wallis test was used to identify differences in protein levels of saliva samples in the two groups among the three time points. The data were analyzed using the BioExplorer statistical package. p < 0.05 were considered to indicate statistical significance.

2.6

Identification of potential peptide biomarkers by nano-LC/ESI–MS/MS

First, the supernatants were processed by WCX as described above, and then the impurities like beads were removed by centrifugation at 10,000 × g for 20 min at 4 °C. Before LC–MS/MS, the supernatants were purified by a 0.22 μm-level filter. Each sample (20 μL) was separated by nano-HPLC system EASY-nLC1000 (Proxeon Biosystems, now Thermo Fisher Scientific). The flow rate of the equilibrated BEH nano ACquity column 100 × 100 mm was 400 nL/min.

Tandem mass spectrometry (MS/MS) was then performed using Q-Exactive mass spectrometer (Thermo Fisher Scientific). The analyzing time was set as 120 min, and the voltage of ion source was 3.5 kV. Fragmentation of peptides was performed using higher-energy collision dissociation (HCD). Mass-to-charge ratio ( m / z ) of the peptide fragments was collected by gathering 20 fragmented fingerprint after each single full scan. Survey scans were acquired at a resolution of 70,000 at m / z 200. Resolution for HCD spectra was set to 17,500 at m / z 200. After the MS/MS figure was input to the PD software (Proteome Discoverer 1.4, Thermo Fisher Scientific), an initial screening was carried out. The parameters used were shown below: range of molecular weight of parent ion was 350–6000 Da, the minimum number of peaks in MS/MS figures was set as 10, and the thresholds of S/N ratio was 1.5. After initial screening, the mass spectrum was searched by Mascot software (version: 2.3.2) based on fixed modification (Cysteine carbamidomethylation) and variable modification (Oxidation of methionine and N-acetylation of protein). Then we used database NCBInr 20120419 (17893860 sequences; 6141683785 residues) and Percolator method (false discovery rate [FDR] ≤ 1%; accuracy of peptide tol: 20 ppm) for quantitative analysis of the spectrum.

2.1

Ethics statement

This study was ethically approved by Peking University Biomedical Ethics Committee (issuing number: PKUSSIRB-2013060) and followed the principles of the Declaration of Helsinki. Parents of the pediatric participants have all signed the informed consent form before the start of the study.

2.2

Subjects

According to previous studies, at least 10 subjects in each group (ECC and control) were needed. From the epidemiological literatures in China, we have found that the incidence of primary caries in 3–4 year-old children was about 43.6% (60% in a two-year study ). As we set the lost rate to 20% for each time point, at least 72 individuals were required to meet our need as a rough estimate. Hence, eighty-two preschool children aged 3 years old in one kindergarten in Beijing, China, who were found to be caries-free on examination (total number of decayed, missing and filled tooth surfaces = 0), were initially recruited at baseline in May 2014 and September 2014. Then, they were followed up every 6 months approximately and this would last for 1 year. The oral health status of each individual was determined by one professional dentist. The diagnostic criteria were according to the WHO Oral Health Surveys Basic Methods (5th edition, 2013). 36 children were excluded at the other two time points (details shown in S Table 1). Finally, 15 children who both meet diagnosis of ECC and the inclusion criteria at the 6- or 12-month time points were classified into the experimental group (Group C). The other 31 remained caries-free, and 15 of them (who were age- and sex-matched) were selected to be the healthy group (Group H). All study subjects were systemically healthy. Those with influenza or an upper respiratory tract infection and those who had received antibiotic therapy within 1 month of any sampling time were excluded from the study. Those who were absent at any time point were excluded. S Table 2 shows detailed information and caries status of the subjects.

2.3

Saliva collection and processing

All individuals were instructed to rinse orally with water after breakfast and then rest for 10 min before saliva collection at 8:30 a.m. Stimulated whole saliva samples were collected for 5 min. The 1.5 mL saliva samples were immediately placed on ice. Insoluble material, cells and debris were removed by centrifugation at 10,000 × g for 10 min at 4 °C. The supernatants were collected, and 1 mM ethylene diamine tetraacetic acid (Sigma, St. Louis, MO, USA) and 1 mM phenylmethyl sulfonyl fluoride (Sigma) were added to inhibit protease activity. Protein concentrations were measured by the Lowry method and the ELx808 Protein Assay (BioTek, Hercules, CA, USA). Supernatants were kept at −80 °C separately until further analysis.

2.4

WCX fractionation and MALDI-TOF MS

All saliva samples were fractionated using WCX magnetic beads (MB) (Bioyong Tech). Samples were purified and isolated by the following steps: (1) The beads were turned upside down to mix them well before use; (2) 20 μL of beads, 150 μL of MB-WCX binding solution, and 10 μL of salivary sample were gently mixed and incubated for 5 min at room temperature; (3) the tubes were placed on the MB separation device (Bioyong Tech) for 1 min, and the beads adsorbed onto the tube wall; (4) the supernatant was then removed, and the beads were washed by mixing thoroughly with 150 μL of MB washing solution. Two minutes later, the tubes were placed on the separation device for 1 min; (5) step 4 was repeated and the supernatants removed; (6) 10 μL of MB elution solution were added and mixed thoroughly with the beads, then the tubes were put on the separation device for 2 min; (7) the clear supernatant was transferred into a fresh tube, and the peptides were analyzed directly on a ClinTOF instrument (Bioyong Tech) or stored at −20 °C and analyzed within 24 h.

The matrix solution was 8 mg/mL CHCA in 50% acetonitrile/0.1% TFA/49.9% deionized water. First, 1 μL of purified peptide solution was spotted onto a MALDI-TOF MS target from ClinTOF. After drying at room temperature, 1 μL of matrix solution was spotted to cover the sample, which was dried again before analysis. MALDI-TOF MS was performed using a ClinTOF instrument. Calibration of the MALDI-TOF MS was done by a three-peptide mixture (monoisotopic molecular weights of 1533.8582, 2465.1989, and 5730.6087 Da; Sigma product numbers P2613, A8346, and I6279, respectively) before analysis. Profile spectra were acquired from an average of 400 laser shots per sample. The mass range 1000–10,000 Da was collected. Each sample of saliva was analyzed three times, and the mean values of the intensities and masses of each peak were used for analysis.

2.5

Data processing and statistical analysis

All the spectra obtained from the saliva samples were analyzed using BioExplorer (Bioyong Tech) to obtain the mean relative peak intensities; then baselines were subtracted, spectra normalized using total ion current, and peak m / z values and intensities determined in the mass range 1000–10,000 Da. A signal-to-noise (S/N) ratio >6 was required. To align the spectra, a mass shift of no more than 0.1% was determined. The peak area was used for quantitative standardization. The k-nearest neighbor (KNN) algorithm in this software suite was used to establish the best pattern of differentiation model for the development of ECC. Analysis of variance (ANOVA) or Kruskal-Wallis test was used to identify differences in protein levels of saliva samples in the two groups among the three time points. The data were analyzed using the BioExplorer statistical package. p < 0.05 were considered to indicate statistical significance.

2.6

Identification of potential peptide biomarkers by nano-LC/ESI–MS/MS

First, the supernatants were processed by WCX as described above, and then the impurities like beads were removed by centrifugation at 10,000 × g for 20 min at 4 °C. Before LC–MS/MS, the supernatants were purified by a 0.22 μm-level filter. Each sample (20 μL) was separated by nano-HPLC system EASY-nLC1000 (Proxeon Biosystems, now Thermo Fisher Scientific). The flow rate of the equilibrated BEH nano ACquity column 100 × 100 mm was 400 nL/min.

Tandem mass spectrometry (MS/MS) was then performed using Q-Exactive mass spectrometer (Thermo Fisher Scientific). The analyzing time was set as 120 min, and the voltage of ion source was 3.5 kV. Fragmentation of peptides was performed using higher-energy collision dissociation (HCD). Mass-to-charge ratio ( m / z ) of the peptide fragments was collected by gathering 20 fragmented fingerprint after each single full scan. Survey scans were acquired at a resolution of 70,000 at m / z 200. Resolution for HCD spectra was set to 17,500 at m / z 200. After the MS/MS figure was input to the PD software (Proteome Discoverer 1.4, Thermo Fisher Scientific), an initial screening was carried out. The parameters used were shown below: range of molecular weight of parent ion was 350–6000 Da, the minimum number of peaks in MS/MS figures was set as 10, and the thresholds of S/N ratio was 1.5. After initial screening, the mass spectrum was searched by Mascot software (version: 2.3.2) based on fixed modification (Cysteine carbamidomethylation) and variable modification (Oxidation of methionine and N-acetylation of protein). Then we used database NCBInr 20120419 (17893860 sequences; 6141683785 residues) and Percolator method (false discovery rate [FDR] ≤ 1%; accuracy of peptide tol: 20 ppm) for quantitative analysis of the spectrum.