2.1 Bacterial strain and culture conditions

The strains used in the present study included the standard strain of S. mutans (UA159ATCC 700610) and clinical isolates. Clinical isolates were obtained from an epidemiological survey conducted in Guangdong Province, People's Republic of China, in 2015 (Yu et al., 2015). The survey was conducted among 5-year-old children. A total of 215 clinical strains were isolated from 215 children with different caries status (Zhu et al., 2018). From these isolates, 10 clinical strains were randomly selected. The S. mutans strains were grown in brain heart infusion (BHI) broth (Oxoid) overnight under anaerobic conditions (80% N₂, 20% CO₂) at 37°C. The optical density at 600 nm (OD₆₀₀) of overnight-cultured strains was measured using a microplate reader (Bio-tek, Epoch 2, America). UA159 suspensions (OD₆₀₀ = 0.7) were inoculated at 1:20 into fresh BHI in round-bottom 6-well plates to obtain planktonic cells; the same suspensions (OD₆₀₀ = 0.7) were inoculated at 1:20 into fresh BHI in flat-bottom 6-well plates and incubated for 4 h, 6 h, 12 h, and 24 h to monitor the dynamic biofilm formation process of S. mutans.

2.2 RNA extraction

Planktonic bacteria were collected by centrifugation at (13201 g) for 5 min. Biofilm bacteria were scraped from plates and centrifuged at (13201 g) for 5 min. Total RNA extraction was performed according to the method described by Zhu et al. (2017). Briefly, total RNA was stabilized with RNAprotect Bacteria Reagent (Qiagen) before extraction. Biofilms were harvested and washed three times with phosphate-buffered saline (PBS). The total RNA of biofilm cells was purified using a miRNeasy Mini Kit (Qiagen). A Thermo Scientific NanoDrop 2000 instrument (NanoDrop Technologies) and an Agilent 2100 system (Agilent Technologies) were used to assess RNA quality and quantity.

2.3 Quantitative real-time PCR (qRT-PCR)

The top twenty significantly differentially expressed sRNAs were selected as candidates from our sRNA library established in a previous study (Zhu et al., 2018). These candidate sRNAs were further analyzed between planktonic and biofilm conditions of S. mutans at 24 h by qRT-PCR. cDNA was synthesized using a Mir-X miRNA First-Strand Synthesis Kit (Takara and Clontech) according to the manufacturer's recommended protocol. qRT-PCR was performed using a LightCycler 96 Real-Time System and the SYBR Premix Ex Taq II Kit (Takara and Clontech). The primers used for sRNAs in this study are listed in Table 1. The reaction conditions were 95°C for 30 s followed by 40 cycles of 95°C for 15 s and 60–63°C for 30 s. The expressions of sRNAs were normalized to the 16S rRNA expression level.

The most highly expressed sRNA associated with biofilm formation from the analyzed 20 sRNAs was selected for further analysis. Five target mRNAs of the candidate sRNA were selected, and their possible functional roles were preliminarily explored. The primers used in the qRT-PCR analysis of the mRNAs are listed in Table 2. Synthesis of cDNA and qRT-PCR were performed as described above. The expression level of each gene was determined in triplicate. Expression levels were calculated using the ^2−ΔΔCt method (Livak & Schmittgen, 2001).

2.4 Crystal violet (CV) staining assay

The CV staining assay was used to evaluate the biofilm biomass of S. mutans (Weerasekera et al., 2016). Streptococcus mutans UA159 and the 10 clinical strains were incubated in flat-bottom 96-well plates under anaerobic conditions for 4 h, 6 h, 12 h, and 24 h. Then the contents of the 96-well plates were then removed, and the plates were washed three times with phosphate-buffered saline (PBS) to remove nonadherent cells. The washed biofilms were fixed with 95% methanol for 15 min and washed again. The biofilms were stained with 0.1% (wt/vol) CV solution for 15 min at room temperature. After thorough removal of the excess liquid, the remaining CV was dissolved in 200 μl of 95% ethanol for 15 min, and 100 μl of the sample was transferred to a new plate for OD₆₀₀ measurement.

2.5 Confocal laser scanning microscopy (CLSM)

For analysis of EPS production, 1 μM Alexa Fluor 647 (Invitrogen) and 2.5 μM SYTO 9 (Invitrogen) were used to label dextran and bacterial cells, respectively (Huang et al., 2017). COMSTAT was used to analyze the biomass of EPS (μm³/μm²). The three-dimensional architecture of the biofilms was reconstructed using Imaris 8.0.2 (Bitplane). Three independent experiments were performed for each condition, and images of five random fields were collected for each sample.

2.6 Bioinformatics analysis of candidate sRNAs

We predicted the structures of candidate sRNA using RNAfold (http://rna.tbi.univie.ac.at/cgi-bin/RNAWebSuite/RNAfold.cgi). According to sequence data for S. mutans UA159 (AE014133.2), functional annotation of sRNA was performed with the Kyoto Encyclopedia of Genes and Genomes analyses (KEGG) and Database for Annotation, Visualization and Integrated Discovery (DAVID) software (http://david.abcc.ncifcrf.gov/). The binding sites of sRNAs in putative target mRNAs were predicted by intaRNA (http://rna.informatik.uni-freiburg.de/IntaRNA/Input.jsp).

2.7 Statistical analyses

Each experiment was independently repeated three times. GraphPad Prism version 7.0a (GraphPad Software, San Diego, CA, USA) and IBM SPSS 24.0 (IBM, Armonk, NY, USA) were used to analyze the data. The means and standard deviations of all continuous variables were computed. The data were assessed for normal distribution and sphericity; an unpaired t test was used for two conditions, and repeated measures analysis of variance was used for multiple time points (p < 0.05). The Spearman rank correlation coefficient was applied with a p-value of <0.05 for correlation testing.

3.1 Screening for the most highly differentially expressed sRNAs associated with biofilm formation

To obtain the most relevant sRNAs associated with biofilm formation, we screened the top 20 differentially expressed sRNAs from our previous study in the standard strain of S. mutans (UA159, ATCC 700610). Among the 20 sRNAs, 18 were successfully detected with 14 were upregulated in cultures with biofilm status relative to those with planktonic status, and 4 were downregulated. Two sRNAs (sRNA0250 and sRNA0656) expressed so unstably after multiple repeated studies that the detection of these two sRNAs was not shown in Table 3. sRNA0426 was the most highly differentially expressed sRNA. Its expression was 5.87 times higher in the biofilm state than in the planktonic state (p < 0.001, Table 3).

3.2 Expression of sRNA0426 during biofilm formation in standard and clinical strains of S. mutans

To further verify the relationship between sRNA0426 and biofilm formation, we first evaluated the biofilm biomass by CV assays and measured the expression of sRNA0426 at 4 h, 6 h, 12 h, and 24 h in the standard strain of S. mutans. The biofilm biomass increased from 4 h to 24 h during the biofilm formation process in the standard strain (p < 0.05, Figure 1a). Also, expression of sRNA0426 changed dynamically during biofilm formation in the standard strain, gradually increasing from 4 h to 12 h and then decreasing slightly at 24 h, with a peak at 12 h (p < 0.001) (Figure 1b). We observed a similar trend in the clinical strains of S. mutans (Figure 2a,b). There was a positive correlation between sRNA0426 expression and biofilm biomass in the clinical strains at various times (4 h, 6 h, 12 h, and 24 h). From 4 h to 12 h, the correlation strengthened as the biofilm formation capability of the 10 clinical strains increased, although the correlation weakened from 12 h to 24 h (Figure 2c-f). The strongest correlation between sRNA0426 expression and biofilm formation capability was observed at 12 h (r = 0.8252, p = 0.0033) (Figure 2e).

3.3 Relationship between expression of sRNA0426 and EPS

EPS forms the core of the matrix scaffold and provides binding sites that promote the accumulation of microorganisms on the tooth surface and the establishment of pathogenic biofilms (Bowen, Burne, Wu, & Koo, 2018). Thus, to further explore the association between sRNA0426 and EPS, we specifically analyzed EPS by CLSM. According to the confocal micrographs of EPS, the bacteria became increasingly encased or surrounded by EPS with time, but no change was apparent from 12 h to 24 h (Figure 3a,b). The highest biovolume of EPS was observed at 12 h in the biofilm of both the standard and clinical strains of S. mutans (p < 0.05) (Figure 3c,d), and the trend of the change in the amount of EPS was consistent with the dynamic expression of sRNA0426 during biofilm formation. We then analyzed the relationship between expression of sRNA0426 and the amount of EPS in the biofilms of the 10 clinical strains at 4 h, 6 h, 12 h, and 24 h of culture. The results obtained with the clinical strains suggest that the expression level of sRNA0426 correlates positively with the amount of EPS present during biofilm formation. The strongest correlation was observed at 12 h (r = 0.8663, p = 0.0012) (Figure 4c). These results indicate that sRNA0426 may play a positive role in the production of EPS in S. mutans biofilms.

3.4 Functional annotation of sRNA0426 using bioinformatics analyses

Considering the importance of secondary structures in stabilizing sRNAs, the secondary structure of sRNA0426 was predicted using RNAfold. It is reported that sRNA0426 possesses a stem-loop structure with a ΔG value of −18.7 kcal/mol (Figure 5a). To the best of our knowledge, sRNA0426 is located on the antisense mRNA strand between SMU_1238c and SMU_1239 (Table A1). To explore the potential mechanism by which sRNA0426 regulates S. mutans biofilm formation, KEGG pathway annotation was used to investigate the sRNA0426 regulatory pathway, revealing eight pathways that are significantly regulated by sRNA0426 (p < 0.05) (Figure 5b). Specifically, most of the pathways are involved in biofilm formation, such as metabolic pathways, especially carbon metabolism. The results of the KEGG analysis of biological pathways for the other sRNAs are presented in Appendix B (https://doi.org/10.6084/m9.figshare.12310133). The KEGG analysis for the other differential sRNAs expression showed a potential similar pathway with sRNA0426 and that there might be several sRNAs involved in the biofilm regulatory network in S. mutans.

To determine whether similar putative sRNAs are present in other bacteria, we searched for sequences homologous of sRNA0426 using BLASTN. The results are shown in Figure 6. A sequence was only considered to be conserved when the coverage between the query and subject sequences was higher than 75% and the nucleotide identity was higher than 65% (E-value = 10⁻⁵, word = 11). The results suggest that sRNA0426 might be conserved in Streptococcus species, primarily in S. mutans strains and Streptococcus troglodytae (S. troglodytae). The genomes of 14 S. mutans strains were found to cover 100% of the sequence of sRNA0426. The S. mutans strain LAB761 and S. troglodytae separately cover 98.45% and 95.35% of the sequence respectively (Figure 6a). Furthermore, 14 streptococcus species including 105 strains have a 24%–27% query cover of sRNA0426. The BLASTN results of the representative 14 streptococcus species were shown in Figure 6b, and more details about the total 105 strains were presented in Appendix C (https://doi.org/10.6084/m9.figshare.12310133). We consider it specific seed sequences for the function of sRNA0426 in S. mutants.

3.5 Relationship of sRNA0426 and potential target mRNAs

To further explore the function of sRNA0426, we examined the association between sRNA0426 and five potential target mRNAs predicted by bioinformatics (GtfB, GtfC, GtfD, ComE, and CcpA) at 12 h, when the strongest correlation was observed between sRNA0426 and biofilm biomass together with EPS. According to the results, sRNA0426 expression showed a significantly positive relationship with GtfB, GtfC, ComE, and CcpA expression (p < 0.05) but no significant relationship with GtfD expression (Figure 7). Potential binding sites were also predicted by intaRNA (Figure A1). The presence of binding sites between potential target mRNAs and sRNA0426 provides evidence for a regulatory role of sRNA0426.