Differential microbiome response to sucrose in caries-free and caries-affected individuals

Abstract

This study investigated the early salivary microbiome response to a sucrose challenge in caries-free and caries-active individuals. Eighteen healthy adults aged 24–40 years were recruited, including eight caries-free and ten caries-active participants. Saliva samples were collected at baseline and at 30 min and 2 h after a 20% sucrose rinse. Microbial diversity and composition were assessed using 16 S rRNA gene sequencing to compare dynamic changes between the two groups. The salivary microbiome of caries-active individuals showed a greater and more rapid increase in amplicon sequence variants (ASVs) and more complex microbial interaction networks following sucrose exposure. After 2 h, Actinomyces graevenitzii increased considerably in both groups, while Schaalia odontolytica was enriched only in caries-active individuals, suggesting a microbial signature associated with caries susceptibility. Although global diversity metrics (Shannon and Chao1 indices) did not change significantly after rinsing, caries-active individuals exhibited more pronounced ecological shifts, indicating heightened responsiveness to sucrose. These findings highlight the potential of the salivary microbiome as an early indicator of caries risk and underscore the importance of microbial dynamics in caries pathogenesis. The study provides new ecological insights into how dietary sugars influence oral microbiota and may inform microbiome-based strategies for caries prevention and early diagnosis.

Introduction

Dental caries is one of the most prevalent and consequential oral diseases worldwide, affecting approximately 2.8 billion individuals globally (Peres et al. 2019). It is a chronic infectious condition marked by the progressive demineralization of dental hard tissues (Pitts et al. 2021). Despite substantial public health efforts over the past three decades, the global prevalence of dental caries has declined by only 4% (Peres et al. 2019). This persistently high prevalence underscores the fact that dental caries remains a major public health concern. Understanding the etiological microbiology of caries is therefore essential for developing effective preventive strategies. According to contemporary etiological theory, caries arises from dysbiosis of the oral microbiome, characterized by an overgrowth of cariogenic bacteria in response to risk factors such as inadequate oral hygiene and frequent saccharide intake (Pitts et al. 2017; Zhang et al. 2022). Species strongly associated with caries risk include Streptococcus mutans, Lactobacillus spp., and Scardovia wiggsiae, which are acidogenic and aciduric organisms capable of thriving in low-pH environments (Nyvad and Takahashi 2020; Tanner et al. 2018). In contrast, health-associated taxa such as Streptococcus sanguinis, Streptococcus gordonii, and Actinomyces naeslundii help maintain ecological balance (Spatafora et al. 2024).

Saliva serves as the primary reservoir of oral microorganisms and is the main source of bacteria for dental plaque formation. Notably, saliva has been recognized as a valuable diagnostic medium for assessing caries risk (Belibasakis et al. 2019; Grier et al. 2021). Numerous studies have reported significant associations between oral microbial profiles and caries development. High-throughput 16 S rRNA gene sequencing has revealed that caries-free individuals often exhibit greater microbial diversity, including higher representation of health-associated commensals, compared to caries-active individuals, who show reduced diversity and enrichment of acidogenic taxa (Kim et al. 2018; Jiang et al. 2016; Xu et al. 2018). Shifts in the salivary microbiome may, therefore, serve as predictors of increased caries risk. A better understanding of how salivary microbiota responds to sugar challenges may offer novel insights into oral microbial pathogenicity.

Dietary sugar, particularly sucrose, plays a pivotal role in the transition from a symbiotic to a dysbiotic oral microbiome (Echeverria et al. 2022). Sucrose is considered more cariogenic than other sugars. Early animal studies showed that sucrose consumption led to more severe carious lesions than other sugars at comparable concentrations (Birkhed et al. 1981; Horton et al. 1985). Similarly, Du et al. (2020) demonstrated that sucrose disrupts the balance between acidogenic and alkali-producing bacteria, promoting a caries-favoring microbial environment. Other investigations have shown that sucrose significantly reduces initial salivary bacterial diversity and increases the relative abundance of the Streptococcus genus (Onyango et al. 2020; Lundtorp-Olsen et al. 2021). However, many of these studies either targeted specific microbial species or relied primarily on in vitro models, which may not accurately reflect the in situ dynamics of the oral microbiome.

To date, limited knowledge exists regarding the immediate response of oral microbiota to sucrose challenge in individuals with and without caries. Therefore, the present study aimed to compare the early salivary microbiome response to sucrose exposure between caries-free and caries-active subjects under in situ conditions, and to identify potential microbial biomarkers indicative of caries susceptibility.

Materials and methods

Ethics statement

This study was approved by the Ethics Committee of Tianjin Stomatological Hospital (Tianjin, China; approval number: PH2020-B-013) and was conducted in accordance with the Declaration of Helsinki (General Assembly of the World Medical Association 2014). All participants were fully informed of the study’s purpose and provided written informed consent. All procedures were performed following approved institutional and ethical guidelines. To minimize contamination, saliva was collected using sterile tubes, and gloves and masks were worn during handling. Samples were immediately transferred to ice and processed under aseptic conditions.

Subject recruitment

Eighteen volunteers aged 24–40 years were recruited at Tianjin Stomatological Hospital, including 8 caries-free (CF) and 10 caries-active (CA) individuals. All individuals were approached during routine dental visits to the outpatient clinic. Of those invited, 4 declined participation. All participants were residents of Tianjin, China, where the fluoride concentration in drinking water is 0.85 mg/L. Medical history and oral examinations were completed prior to sample collection. Eligible participants were systemically healthy, had no oral disease other than documented caries (DMFT index: decayed, missing, or filled teeth due to caries), and did not wear any dental appliances or dentures. Periodontal pocket depths were ≤ 3 mm and negative for bleeding on probing at four sites (mesiobuccal, distobuccal, mesiolingual, and distolingual) on all teeth, measured using a periodontal probe after saliva collection.

General exclusion criteria included: (a) severe systemic illness, (b) salivary gland or oral mucosal diseases, (c) diabetes, (d) recent use of antibiotics, topical antimicrobials, mouthwash, or dental treatment within the past 30 days, (e) eating disorders, food intolerances, or allergies, and (f) pregnancy or breastfeeding. Inclusion criteria for the CA group required a DMFT score of ≥ 4. Dental caries activity was confirmed clinically by the presence of active cavitated lesions; no radiographs were taken. For the CF group, participants were examined clinically and confirmed to be free of carious lesions or restorations. Age-, race-, and gender-matched caries-free individuals without restorations served as controls. The difference in numbers between groups (8 CF vs. 10 CA) resulted from the availability of eligible volunteers during recruitment. The mean number of teeth present was also recorded to account for differences in dentition.

Given the exploratory nature of this study and the absence of prior data on sucrose-induced temporal changes in the salivary microbiome, a formal sample size calculation was not feasible. Instead, the sample size was based on the availability of eligible participants, consistent with prior pilot microbiome studies (Belstrøm et al., 2016; Zaura et al. 2009; Jiang et al. 2016).

Assessment of dental caries and plaque accumulation

Dental caries was assessed using the Decayed, Missing, and Filled Teeth (DMFT) index according to WHO (2013) criteria. All permanent teeth, except third molars, were examined with a mouth mirror and explorer, and each tooth was recorded as decayed, missing due to caries, or filled. The sum of these values represented the DMFT score for each participant. Oral hygiene status was evaluated using the Plaque Index (Siln4) by examining six index teeth (16, 12, 24, 36, 32, and 44) across four surfaces (mesial, distal, buccal, and lingual). Plaque wess and Löe, 196as scored on a 0–3 scale (0 = no plaque, 1 = thin film, 2 = moderate accumulation, 3 = abundant plaque), and the mean score per participant was calculated. For both indices, results were expressed as mean ± SD at the group level.

Sample collection and dental examination

Sample collection and dental examinations were performed at Tianjin Stomatological Hospital by an experienced dentist. Participants were instructed to abstain from toothbrushing after the previous night and to avoid food or drink for at least one hour before sampling. At baseline, ≥ 2 mL of unstimulated saliva was collected via passive drooling into sterile tubes. Unstimulated saliva was chosen rather than stimulated saliva or plaque, as it better reflects whole-mouth microbial status and is less variable with chewing or mechanical stimulation (Dawes and Wong 2019; Fábián et al. 2012). All collections were performed in the morning to minimize circadian variation in salivary microbiota. Participants refrained from eating, drinking, or oral hygiene procedures for at least 2 h prior to collection. Participants were then instructed to rinse once with a 20% sucrose solution for 1 min (single rinse, not repeated every 10 min; clarified for reproducibility). The use of 20% sucrose was selected to mimic the high sugar concentration found in sweetened beverages and confectionery, which are commonly consumed in the population (Bowen and Koo 2011; Zero 2004). Additional saliva samples were collected at 30 min and 2 h post-rinse (de Sousa et al. 2020). The groups were labeled as follows: CF.B (caries-free, baseline), CF.30 M (30 min), CF.2 H (2 h), CA.B (caries-active, baseline), CA.30 M (30 min), and CA.2 H (2 h). All samples were immediately chilled on ice and stored at − 80 °C until further processing. From each sample, 1 mL of saliva was used for DNA extraction.

DNA extraction and 16 S rRNA gene sequencing

Genomic DNA was extracted using the cetyltrimethylammonium bromide (CTAB) method, modified with lysozyme treatment. DNA concentration and purity were measured using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific). DNA was diluted to 1 ng/µL using sterile water. The V3–V4 region of the 16 S rRNA gene was amplified using primers 515 F (GTGCCAGCMGCCGCGGTAA) and 806R (GGACTACHVGGGTWTCTAAT) with unique barcodes. Each PCR reaction included 15 µL Phusion^® High-Fidelity PCR Master Mix (New England Biolabs), 0.2 µM of each primer, and 10 ng of DNA. The thermal cycling protocol was: 98 °C for 1 min; 30 cycles of 98 °C for 10 s, 50 °C for 30 s, 72 °C for 30 s; and a final extension at 72 °C for 5 min. PCR products were visualized via 2% agarose gel electrophoresis, mixed in equal volumes, and purified using the Qiagen Gel Extraction Kit (Qiagen, Germany).

Sequence processing

Sequencing libraries were constructed using the NEBNext^® Ultra™ II DNA Library Prep Kit (Cat No. E7645) and quality-checked on a Qubit 2.0 Fluorometer (Thermo Scientific) and an Agilent Bioanalyzer 2100. Paired-end sequencing (2 × 250 bp) was performed on the Illumina NovaSeq platform. Reads were demultiplexed, trimmed, and merged using FLASH (v1.2.11) (Magoč and Salzberg 2011). Quality filtering was done with fastp (v0.20.0) to obtain clean tags. Chimeras were identified and removed using Vsearch (v2.15.0) against the SILVA database (https://www.arb-silva.de/) (Haas et al. 2011). Denoising was performed using DADA2 within QIIME2 (v2020.6) to generate amplicon sequence variants (ASVs). ASVs with < 5 reads were filtered out. Taxonomic classification was based on the SILVA 138.1 database, and ASV abundances were normalized to the sample with the fewest sequences. Sequencing data are available in the National Centre for Biotechnology Information (NCBI) SRA database under the access number PRJNA 1,255,202.

Statistical and bioinformatics analysis

Microbial community analysis was performed using QIIME2 (v2020.6). Venn diagrams were generated to display shared OTUs among groups. Alpha diversity (Shannon and Chao1 indices) was calculated to assess microbial richness and evenness. Statistical comparisons between groups were conducted using one-way ANOVA in R (v3.5.3). The threshold for statistical significance was set at p < 0.05. Beta diversity was evaluated via weighted and unweighted UniFrac distances and visualized with two-dimensional PCoA using the ade4 and ggplot2 R packages (v2.15.3). Group differences in community structure were assessed using PERMANOVA (Adonis) and ANOSIM in QIIME2.

Differences in bacterial relative abundances at the genus and species levels were analyzed using the Kruskal–Wallis test and one-way ANOVA. Linear discriminant analysis effect size (LEfSe) was employed to identify taxa associated with specific groups (LDA score > 4). Functional annotation was conducted with PICRUSt2 (v2.1.2-b). Species correlation matrices were generated using Spearman’s rank correlation, and microbial co-occurrence networks were visualized with Graphviz. Filters applied included: (1) correlation coefficient > 0.6, (2) exclusion of self-links, and (3) removal of nodes with relative abundance < 0.005%.

Results

Participant characteristics

A total of 18 participants were enrolled, including 8 in the caries-free (CF) group and 10 in the caries-active (CA) group. The mean age was 32.6 ± 4.2 years in the CF group and 33.4 ± 3.8 years in the CA group, with no significant difference between groups (p = 0.61). The sex distribution (5 males/3 females in CF vs. 6 males/4 females in CA) and the mean number of teeth present (28.1 ± 1.2 in CF vs. 27.6 ± 1.5 in CA) were also comparable (both p > 0.05).

In contrast, the mean DMFT score was 0.0 ± 0.0 in the CF group and 6.2 ± 1.1 in the CA group, confirming the presence of active caries in the latter (p < 0.001). The mean Plaque Index was significantly lower in the CF group (0.6 ± 0.2) compared with the CA group (1.9 ± 0.4, p < 0.001), indicating poorer oral hygiene among caries-active individuals. A summary of participant characteristics is presented in Table 1.

Table 1.

Participant characteristics of caries-free (CF) and caries-active (CA) groups

Characteristic	CF (n = 8)	CA (n = 10)	p-value
Age (years, mean ± SD)	32.6 ± 4.2	33.4 ± 3.8	0.61
Sex (M/F)	5/3	6/4	0.89
Teeth present (mean ± SD)	28.1 ± 1.2	27.6 ± 1.5	0.47
DMFT (mean ± SD)	0.0 ± 0.0	6.2 ± 1.1	< 0.001
Plaque Index (mean ± SD)	0.6 ± 0.2	1.9 ± 0.4	< 0.001

Overview of the salivary Microbiome

A total of 4,407,122 raw reads were generated from 54 saliva samples. After quality filtering and chimera removal, 3,791,654 effective reads were retained for downstream analysis. In the caries-free (CF) group, 933, 944, and 1,119 amplicon sequence variants (ASVs) were identified in the baseline (CF.B), 30-minute (CF.30 M), and 2-hour (CF.2 H) samples, respectively, with 423 ASVs shared among all three time points, accounting for 14.12% of the total ASVs. In the caries-active (CA) group, 1,105, 1,262, and 1,828 ASVs were identified in CA.B, CA.30 M, and CA.2 H, respectively, with 568 shared ASVs (13.54%) (Fig. 1A). In both CF and CA groups, the number of observed ASVs increased following sucrose rinsing, with a greater increase observed as the duration progressed. Notably, the rate of ASV increase was more pronounced in the CA group (Fig. 1B).

Fig. 1.

Analysis of ASVs before and after sucrose rinsing. A Venn diagram showing microbial richness based on ASV numbers. Each circle represents a group, and the overlapping areas indicate shared ASVs. B Alterations in ASV counts in response to sucrose challenge in caries-free and caries-active salivary microbiomes. (ASVs, amplicon sequence variants; CF.B, caries-free saliva before sucrose rinsing; CF.30 M, after 30 min; CF.2 H, after 2 h; CA.B, caries-active saliva before rinsing; CA.30 M, after 30 min; CA.2 H, after 2 h)

Salivary Microbiome profiles before and after sucrose rinsing

Alpha diversity was assessed using Shannon’s index and Chao1 index. No significant differences in Shannon’s index were observed between pre- and post-rinse samples in either CF or CA groups (p > 0.05) (Fig. 2A). Similarly, the Chao1 index, which estimates species richness, including low-abundance taxa, showed no significant changes (p > 0.05) (Fig. 2B). This suggests that while the number of ASVs increased, these were likely rare or low-abundance taxa that contributed little to overall community evenness.

Fig. 2.

Diversity of the salivary microbiome before and after sucrose rinsing. a, b Shannon index (A) and Chao1 index (B) showing alpha diversity. Outliers are shown as black points. Principal coordinate analysis (PCoA) plots based on weighted (C) and unweighted (D) UniFrac distances. Each point represents a sample; percentages denote the variation explained by each principal component. Samples from the same group share the same color. (PCoA, principal coordinate analysis)

Beta diversity was evaluated using UniFrac distance-based principal coordinate analysis (PCoA). Weighted UniFrac distances did not show clear separation among the three CF or CA groups, indicating minimal variation in microbial community structure following sucrose rinsing (Fig. 2C). Similarly, unweighted UniFrac analysis, which emphasizes the presence or absence of rare taxa, also showed no distinct clustering among time points (Fig. 2D).

Top 10 most abundant salivary genera

A total of 111 bacterial genera and 413 species were identified, covering 96.3% and 61.5% of all sequences, respectively. Both CF and CA groups shared a similar microbial composition. The 10 most abundant genera were Streptococcus, Prevotella, Neisseria, Alloprevotella, Lautropia, Haemophilus, Veillonella, Leptotrichia, Pseudomonas, and Ralstonia. Most genera exhibited a decreasing trend after sucrose rinsing, with the exception of Streptococcus and Leptotrichia, which increased in abundance (Fig. 3). However, none of the genera showed statistically significant differences in relative abundance before and after sucrose exposure.

Fig. 3.

Relative abundance of the top 10 bacterial genera before and after sucrose rinsing. No statistically significant differences in relative abundance were observed between pre- and post-rinsing samples

Distinctive salivary bacteria before and after sucrose rinsing

No significant differences in bacterial composition were observed between baseline and 30-minute samples in either group. However, Actinomyces graevenitzii showed a significantly higher relative abundance in CF.2 H compared to CF.B (p < 0.05) (Fig. 4A). In the CA group, both Actinomyces graevenitzii and Schaalia odontolytica were significantly more abundant in CA.2 H than in CA.B (p < 0.05) (Fig. 4B). While no significant biomarkers were identified among the CF subgroups, LEfSe analysis in the CA group revealed that the order Actinomycetales and family Actinomycetaceae significantly differed across time points. Furthermore, Actinomycetaceae abundance increased progressively with longer sucrose exposure (Fig. 4C, D).

Fig. 4.

Species with significant changes in relative abundance following sucrose rinsing. A Statistically different species between CF.2 H and CF.B. B Statistically different species between CA.2 H and CA.B. Bars represent mean abundance; confidence intervals are shown at right (95% CI). C Discriminative biomarkers identified between CA.B and CA.2 H by LDA Effect Size (LEfSe). D Discriminative biomarkers among CA.B, CA.30 M, and CA.2 H by LEfSe, with LDA scores > 4. Bar length represents magnitude of impact. (LDA, linear discriminant analysis; LEfSe, LDA Effect Size)

Ecological network analysis of the microbiome

Ecological network analysis was performed to visualize changes in microbial interactions in response to sucrose rinsing. After 30 min, both CF and CA networks showed increased node size and reduced connectivity. At 2 h, the number of connecting lines increased while node size decreased. The network in the CA group demonstrated greater complexity and higher intermodular connectivity compared to the CF group (Fig. 5A).

Fig. 5.

Ecological network analysis of microbial interactions in saliva samples before and after sucrose rinsing. A Network inferences of microbial relationships in saliva with and without sucrose rinsing. Each node represents an ASV, and each edge represents a significant pairwise association. Red lines indicate synergistic relationships, while green lines represent antagonistic relationships. B Graph density (GD) in sucrose-rinsed and unrinsed groups, calculated as the ratio of observed edges to the total possible edges. C Average degree (AD) in sucrose-rinsed and unrinsed groups, calculated as the total number of node connections divided by the number of nodes. D Average path length (APL) in sucrose-rinsed and unrinsed groups, calculated as the sum of the shortest distances between all node pairs divided by the number of node pairs

Graph density (GD), defined as the ratio of actual to possible connections, was higher in the CA group than in the CF group at baseline and 2 h post-rinse. Interestingly, GD was lower in the CA group at 30 min post-rinse (Fig. 5B). Average degree (AD), representing the mean number of connections per node, was higher in the CA group at baseline but decreased after 30 min, falling below that of the CF group. At 2 h, the CA group again showed a higher AD, with a larger increase than before rinsing (Fig. 5C).

Average path length (APL), measuring the mean shortest distance between nodes, was consistently lower in the CA group at all time points, indicating more closely connected microbial associations within the salivary community of caries-active individuals (Fig. 5D).

Discussion

Dental caries arises from dysbiosis of the oral microbial community, a process strongly influenced by environmental factors such as excessive sucrose consumption. While previous cross-sectional studies have compared the salivary microbiome profiles of caries-free (CF) and caries-active (CA) individuals (Hurley et al. 2019; Jiang et al. 2016), relatively few have investigated how these microbiomes respond dynamically to a sucrose challenge under in situ conditions. The present study therefore examined the short-term salivary microbiome responses to sucrose rinsing in CF and CA participants.

The use of amplicon sequence variants (ASVs) generated by the DADA2 pipeline provided high-resolution insights compared with traditional OTU-based clustering (Callahan et al. 2016). Following sucrose exposure, the number of observed ASVs increased more markedly in the CA group, suggesting a heightened reactivity of their salivary microbiota. However, no significant changes were detected in the most abundant genera in either group. This apparent discrepancy is explained by the fact that newly observed ASVs likely represent rare or low-abundance taxa that contribute little to overall community evenness. This interpretation is consistent with our alpha diversity findings: both Shannon and Chao1 indices remained unchanged post-rinsing, indicating that short-term sucrose exposure alters microbial abundance rather than richness or evenness (Callahan et al. 2016; Xu et al. 2018; Zijnge et al. 2010).

Beta diversity analyses similarly showed no clear clustering by time point, suggesting that a single sucrose rinse does not drastically restructure salivary microbiota at the community level. Nevertheless, caries-active individuals exhibited more pronounced ecological responses, reflected by increased ASV counts and distinct enrichment of specific taxa. Actinomyces graevenitzii increased significantly in both groups after two hours, whereas Schaalia odontolytica was uniquely enriched in CA participants. Formerly classified within the Actinomyces genus, S. odontolytica has been associated with caries lesions, pericoronitis, and implant failure (Nouioui et al. 2018; Bartlett et al. 2022; Sarkonen et al. 2005; Sixou et al. 2003; Young et al. 2001). Its selective enrichment in caries-active saliva following sucrose rinsing highlights its potential utility as a biomarker for caries susceptibility.

Ecological network analysis provided further insight into microbial interactions. In the CA group, sucrose rinsing increased graph density (GD) and average degree (AD), while average path length (APL) remained consistently shorter than in CF individuals, suggesting more closely connected microbial associations. These changes reflect intensified microbial interactions and reduced antagonism, which may facilitate a shift toward a cariogenic network structure (Marsh 1994). Such findings suggest that even in the absence of gross diversity changes, caries-active microbiota demonstrate more dynamic ecological reorganization in response to dietary sugar.

The detection of Pseudomonas and Ralstonia among the most prevalent genera requires careful interpretation. These taxa are not generally recognized as core members of the oral microbiome and are often regarded as environmental or opportunistic organisms. Their presence in sequencing-based datasets may reflect transient colonization, low-abundance background taxa, or potential contamination introduced during sample collection or DNA extraction (Salter et al. 2014). Similar observations have been reported in other salivary microbiome studies, where Pseudomonas and Ralstonia were occasionally detected at low levels but were not consistently associated with oral disease (Leão et al. 2023). In our dataset, Pseudomonas and Ralstonia were among the most abundant genera detected in certain samples. However, these genera are generally regarded as environmental or opportunistic taxa that may appear sporadically due to technical factors or transient oral exposure (Salter et al. 2014; Schulze-Schweifing et al. 2014). Importantly, they are not consistently associated with oral disease, and their presence in this study likely reflects background variability rather than a biologically meaningful role in caries pathogenesis.

Some limitations should be acknowledged. The relatively small sample size (n = 18) limits statistical power and generalizability, and the study should be interpreted as exploratory. Larger cohorts are needed to validate the discriminatory value of S. odontolytica and other potential microbial markers. Moreover, the two-hour observation period captures only the initial response; longer-term studies extending to 24 h or several days are necessary to track microbial succession and stabilization. The use of a 20% sucrose rinse, selected to approximate concentrations found in sweetened beverages and confectionery, provides a physiologically relevant challenge, but responses to lower concentrations or different sugar types warrant investigation. Unmeasured host factors, such as dietary habits and salivary flow rates, may also have influenced microbial responses and should be addressed in future studies. In addition, no radiographs were taken to confirm caries status, which may have limited the detection of early non-cavitated lesions.

In conclusion, this study demonstrates that the salivary microbiome of caries-active individuals exhibit more pronounced responses to sucrose rinsing compared with caries-free individuals. These responses included greater increases in ASV counts, enhanced microbial network complexity, and selective enrichment of Schaalia odontolytica, which may serve as a potential biomarker for caries risk. While Actinomyces species are often considered commensal, their role in dysbiosis and caries pathogenesis requires further exploration. Together, these findings provide new ecological insights into the early microbial events underlying caries development and highlight the importance of extending such studies to larger populations and longer-term dietary challenges.

Abbreviations

rRNA

Ribosomal RNA

ASV

Amplicon sequence variants

PCoA

Principal coordinates analysis

LDA

Linear discriminant analysis

LEfSe

LDA effect size

Author contributions

YW and LW designed the study. YW, XL, SH, and LL collected and analyzed the data. LW supervised the study. YW, XL and LW wrote and reviewed the manuscript. All authors contributed to the article and approved the submitted version.

Funding

This work was supported by Tianjin Health Commission Science and Technology Talent Cultivation Project (KJ20089). The Program of Tianjin Municipal Natural Science Foundation (23JCYBJC01480).

Data availability

The sequencing data are available in the National Centre for Biotechnology Information (NCBI) (https://www.ncbi.nlm.nih.gov/) SRA database under the access number PRJNA 1255202.

Declarations

Ethics approval and consent to participate

This study was approved by the Ethics Committee of Tianjin Stomatological Hospital (Tianjin, China; approval number: PH2020-B-013) and was conducted in accordance with the Declaration of Helsinki (General Assembly of the World Medical Association 2014).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Not applicable.