Streptococcus mutans membrane vesicles inhibit the biofilm formation of Streptococcus gordonii and Streptococcus sanguinis

Abstract

Streptococcus mutans, whose main virulence factor is glucosyltransferase (Gtf), has a substantial impact on the development of dental caries. S. mutans membrane vesicles (MVs), which are rich in Gtfs, have been shown to affect biofilm formation of other microorganisms. Streptococcus gordonii and Streptococcus sanguinis are initial colonizers of tooth surfaces, which provide attachment sites for subsequent microorganisms and are crucial in the development of oral biofilms. S. mutans and S. gordonii, as well as S. mutans and S. sanguinis, have a complex competitive and cooperative relationship, but it is unclear whether S. mutans MVs play a role in these interspecific interactions. Therefore, we co-cultured S. mutans MVs, having or lacking Gtfs, with S. gordonii and S. sanguinis. Our results showed that S. mutans MVs inhibited biofilm formation of S. gordonii and S. sanguinis but did not affect their planktonic growth; contrastingly, S. mutans ΔgtfBC mutant MVs had little effect on both their growth and biofilm formation. Additionally, there were fewer and more dispersed bacteria in the biofilms of the S. mutans MV-treated group than that in the control group. Furthermore, the expression levels of the biofilm-related virulence factors GtfG, GtfP, and SpxB in S. gordonii and S. sanguinis were significantly downregulated in response to S. mutans MVs. In conclusion, the results of our study showed that S. mutans MVs inhibited biofilm formation of S. gordonii and S. sanguinis, revealing an important role for MVs in interspecific interactions.

Introduction

Dental caries is a multifactorial disease, affecting 60–90% of children and the majority of adults (Du et al. 2018; Petersen et al. 2005). Biofilms are highly dynamic and structured microbial cell communities that adhere firmly to surfaces and are embedded in self-generated extracellular matrices (Flemming and Wingender 2010). Microbial colonization on tooth surfaces and the formation of cariogenic biofilms are key causes of dental caries (Hara and Zero 2010; Marsh and Zaura 2017; Pitts et al. 2017). Bacterial interactions in biofilms are essential for the development of multispecies microbial communities and for the transition from a healthy to a diseased state of the oral cavity (Hojo et al. 2009; Huang et al. 2011; Marsh and Zaura 2017).

Streptococcus mutans plays a crucial role in the development of dental caries (Pitts et al. 2017), given their abilities such as adhesion (Esberg et al. 2012), acid production (Lemos et al. 2005), acid resistance and biofilm formation (Hwang et al. 2014; Quivey et al. 2000). Streptococcus gordonii and Streptococcus sanguinis are both commensals in the oral cavity that are the initial colonizers of teeth surfaces, and are generally associated with lower levels of cariogenic S. mutans and improved dental health (Abranches et al. 2018; Becker et al. 2002; Caufield et al. 2000). There is a complex relationship of competition and cooperation between S. mutans and S. gordonii, as well as S. mutans and S. sanguinis. According to previous studies, S. mutans inhibits the growth of S. gordonii and S. sanguinis primarily by releasing bacteriocins, whereas S. gordonii and S. sanguinis hinder the growth of S. mutans mainly through H₂O₂ production (Becker et al. 2002; Ge et al. 2008; Kreth et al. 2005, 2008; Wang and Kuramitsu 2005).

Bacterial membrane vesicles (MVs), with a size range of 20–400 nm, carry a variety of cargo molecules, such as nucleic acids, proteins, enzymes, and toxins (Cao and Lin 2021; Toyofuku et al. 2019). S. mutans MVs have been successfully isolated and identified from a culture supernatant in 2014 (Liao et al. 2014), providing the foundation for research in this field. Glucosyltransferases (Gtfs) are the major proteins in S. mutans MVs; the primary role of these proteins is to utilize sucrose to form extracellular polysaccharides and aid biofilm formation of S. mutans (Bowen and Koo 2011; Cao et al. 2020; Rainey et al. 2019). Many oral bacteria, including those that do not synthesize Gtfs, can bind to Gtfs (Vacca-Smith and Bowen 1998). S. mutans and S. mutans ΔgtfBC mutants, as well as early colonizers on the tooth surface such as Streptococcus mitis, Streptococcus oralis, S. gordonii, and S. sanguinis, are highly inducible by the complex of MVs, Gtfs, and DNA to form Gtf-dependent biofilms (Senpuku et al. 2019). Our previous study also found that S. mutans MVs, which harbor Gtfs, can promote Candida albicans biofilm formation (Wu et al. 2020). These observations reveal an important role for S. mutans MVs in biofilm formation and interspecies interactions. However, it is unclear whether S. mutans MVs play a role in the interspecific interactions between S. mutans, S. gordonii, and S. sanguinis. Therefore, we questioned whether S. mutans MVs that contained or lacked Gtfs affected the growth and biofilm formation of S. gordonii and S. sanguinis.

In this study, we investigated the effects of S. mutans MVs on biofilm formation by S. gordonii and S. sanguinis by examining the biomass and surface structure of the biofilms. In addition, we analyzed the mechanism by which S. mutans MVs affect biofilm development of S. gordonii and S. sanguinis. Our findings will provide new insights into the interactions between S. mutans, S. gordonii, S. sanguinis, which may be a target for dental caries prevention.

Materials and methods

Bacterial strains and culture conditions

S. mutans UA159 (ATCC 700610), S. mutans UA159 ΔgtfBC mutant (Gong et al. 2018; Wu et al. 2020), S. gordonii (DL-1), and S. sanguinis (ATCC 10556) were grown in a brain heart infusion (BHI; Difco, Detroit, MI, USA) medium at 37 °C under anaerobic conditions (80% N₂, 10% H₂, and 10% CO₂).

Preparation of MVs

The preparations of S. mutans MVs and S. mutans ΔgtfBC mutant MVs were performed as per a previously described method with a few modifications (Liao et al. 2014). Briefly, the two S. mutans strains were incubated in 500 mL of BHI medium at 37 °C for 16 h. After centrifugation at 6000×g for 15 min at 4 °C, followed by 10,000×g for 15 min at 4 °C, most of the cells in the culture supernatants were removed. After being filtered through 0.22-μm filters (Millipore, MMAS, USA) to remove residual cells, the cell-free culture supernatants were concentrated using a 100 kDa Amicon ultrafiltration system (Millipore, MMAS, USA) at 4000×g for 30 min at 4 °C. The collected concentrate was subjected to an initial ultracentrifugation at 100,000×g for 70 min at 4 °C; the precipitate was resuspended in sterile phosphate-buffered saline (PBS) and subjected to ultracentrifugation under the same conditions. We then resuspended the precipitates, obtained from the second ultracentrifugation, in 2 mL sterile PBS for subsequent experiments. MV protein concentration was estimated using a BCA Protein Assay Kit (CWBIO, Beijing, China). Finally, the MVs were frozen at − 80 °C at 100 μg/mL until further experimentation.

Biofilm formation assay by crystal violet staining

Biofilms from S. gordonii and S. sanguinis were developed in 96-well plates, which were previously coated with artificial saliva at 4 °C for 16 h. S. gordonii and S. sanguinis were grown anaerobically in BHI broth overnight, and the overnight cultures were inoculated at a ratio of 1:50 into 0.25% BHIS (BHI medium supplemented with 0.25% sucrose) to form the bacterial solutions. Each well contained 100 μL of one of the bacterial solutions and 100 μL of PBS, S. mutans MVs, or S. mutans ΔgtfBC mutant MVs. The MVs were diluted to the desired concentrations with PBS. After 24 h, the biomass of the biofilms was calculated by crystal violet staining. In brief, supernatants and unbound bacteria were removed using three sterile PBS washes, after which 150 μL of absolute methanol was added to fix the biofilms for 15 min. The fixed biofilms were then stained with 0.1% (w/v) crystal violet for 15 min, and the stained areas were checked by gently washing with flowing water until no more dye was evident in the clean wells. Subsequently, the 96-well plates were allowed to dry naturally at about 23–27 °C. Crystal violet was solubilized in 95% ethanol at about 23–27 °C in dark for 30 min. Subsequently, the solubilized crystal violet was transferred to new 96-well plates to measure the absorbance of 95% ethanol solution at 595 nm using a spectrophotometer (Tecan, Reading, Switzerland). Preliminary screening confirmed that the effective concentrations of MVs for S. gordonii and S. sanguinis were 1 and 0.2 μg/mL, respectively.

Confocal laser scanning microscopy

Biofilm biomass was measured using confocal laser scanning microscopy (CLSM). The corresponding method was similar to the “Biofilm formation assay by crystal violet staining” method. We performed the experiment in three groups: S. gordonii or S. sanguinis with PBS; S. gordonii or S. sanguinis with S. mutans MVs; and S. gordonii or S. sanguinis with S. mutans ΔgtfBC mutant MVs. S. gordonii, and the S. sanguinis biofilms were developed on confocal dishes for 24 h, washed thrice with sterile PBS to remove the supernatants and floating cells, stained with 2.5 μM SYTO-9 (Invitrogen Corp., Carlsbad, CA, USA) in dark for 15 min, and observed using CLSM (Olympus, FV3000, Japan). The excitation wavelength of SYTO-9 was 488 nm. CLSM images were collected from three distinct regions of three biological samples. The biomass of S. gordonii and S. sanguinis were calculated using the COMSTAT image analysis system.

Scanning electron microscopy

The surface structure of the biofilm was observed using scanning electron microscopy (SEM) (Quanta 400F-FEI, Eindhoven, Netherlands). The biofilms were lightly washed with sterile PBS thrice to remove the supernatants and planktonic cells and then fixed with 2.5% (w/v) glutaraldehyde for ≥ 3 h. The fixed biofilms were then washed with sterile PBS four times for 20 min each time, followed by gradual dehydration for 15 min each with 30, 50, 70, 90 and 100% ethanol. The biofilm samples required an overnight and three dehydration cycles in 70 and 100% ethanol. The biofilms were then soaked in tert-butanol three times for 15 min each, dried overnight by lyophilization, and sputtered with gold. Finally, the biofilms were observed using SEM at 2000×, 5000×, and 10,000× magnification.

Biofilm eradication assay

Biofilm eradication assays were performed according to a modified protocol (Xi et al. 2019). The overnight cultures of S. gordonii or S. sanguinis were inoculated at a ratio of 1:100 in 200 µL of 0.25% BHIS in 96-well plates at 37 °C under anaerobic conditions. After 24 h, the medium was discarded from the wells and PBS was used to wash the preformed biofilm to remove planktonic bacteria. Then, 100 μL of PBS, S. mutans MVs, or S. mutans ΔgtfBC mutant MVs was added carefully to the wells. MV concentration was 1 μg/mL or 0.2 μg/mL. Next, 100 μL of 0.25% BHIS was added to each well. The plate was cultured for an additional 24 h. Subsequently, the biofilm biomass of each group was calculated using the crystal violet assay.

CFU counts

The effects of MVs on planktonic growth of S. gordonii or S. sanguinis at the concentration of 1 or 0.2 μg/mL were evaluated using CFU counts as described previously but with a few changes (Im et al. 2017). Biofilms were cultured according to “Biofilm formation assay by crystal violet staining”. After 24 h, to measure the colonies in the supernatant from the biofilm, the media of the wells were gently mixed as to avoid disturbing the biofilm and 100 μL of the grown culture was serially diluted from ten-fold to 10⁶-fold with PBS, and 10 μL was pipetted and placed on BHI agar plates at 37 °C anaerobically for 24 h. Finally, the number of colonies on each plate was counted.

Gene expression

Quantitative reverse transcriptase PCR (qRT-PCR) was used to identify the effect of MVs on the expression levels of genes related to biofilm development. S. gordonii and S. sanguinis biofilms were grown in 6-well plates with or without MVs, using 0.25% BHIS. After 24 h, biofilms were scraped off into 2 mL Eppendorf tubes and centrifugated at 12,000×g for 5 min at 4 °C. After being lysed for ≥2 h with 500 μL of 20 mg/mL lysozyme lysate, the bacterial cells were subjected to proteinase K treatment for at least 30 min. Total RNA was extracted from cell pellets using the miRNeasy Mini Kit (QIAGEN GmbH, Hilden, Germany) according to the manufacturer’s instructions, and RNA concentration and purity (A260/A280) were assessed using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Pittsburgh, PA, USA). RNA was reverse-transcribed using a PrimeScriptTM RT reagent kit (Takara Bio Inc., Otsu, Japan) according to the manufacturer’s instructions. The target RNA was amplified and quantified using the LightCycler 96 Real-Time System using 2× Super SYBR Green qPCR Master mix (ES Science, Shanghai, China). The primers used in this study are listed in Table 1. Finally, the 2^−ΔΔCT method was used to quantify the fold changes in gene expression.

Table 1 Primers used in this study

Statistical analysis

Biomass biofilms and CFU counts are expressed as mean ± standard deviation (SD) from at least three independent experiments. Statistical significance was evaluated using one-way ANOVA in GraphPad Prism 9 (GraphPad Software, San Diego, CA, USA). Differences between the control and MV-treated groups were compared using Dunnett's multiple comparison test. P < 0.05 was considered statistically significant.

Results

S. mutans MVs inhibit the biofilm formation of S. gordonii and S. sanguinis

Crystal violet staining was used to analyze the effects of S. mutans MVs on bacterial biofilm formation. S. gordonii and S. sanguinis were grown anaerobically in 0.25% BHIS containing a gradient concentration of MVs for 24 h, and biofilm biomass after incubation was quantified by crystal violet staining. The experimental results showed that when the S. mutans MVs concentration was approximately 1 μg/mL, the biofilm biomass of S. gordonii was significantly reduced compared to that of the control group. For S. sanguinis, the concentration of S. mutans MVs inhibiting biofilm formation was about 0.2 μg/mL (Additional file 1: Fig. S1). However, no significant differences were observed among S. mutans ΔgtfBC mutant MV-treated groups (Additional file 1: Fig. S2). We then added the same concentration of S. mutans MVs and S. mutans ΔgtfBC mutant MVs to 0.25% BHIS to develop biofilms of S. gordonii and S. sanguinis. The crystal violet staining assay results confirmed that S. mutans MVs significantly reduced biofilm formation by S. gordonii and S. sanguinis in contrast to the control group (Fig. 1a–d; P < 0.05). Contrastingly, the biofilms of S. gordonii and S. sanguinis treated with S. mutans ΔgtfBC mutant MVs were not significantly different from those of the control group (Fig. 1a–d; P > 0.05). CLSM also showed that S. mutans MVs reduced the aggregation and accumulation of S. gordonii and S. sanguinis (Fig. 1e). Compared with the control group, the biomass of the S. mutans MV-treated group was significantly reduced by approximately 0.6-fold (Fig. 1f, g; P < 0.05), whereas there was no significant change in the S. mutans ΔgtfBC mutant MV-treated group (Fig. 1f, g; P > 0.05), which was consistent with subsequent findings. Next, we used SEM to record morphological changes in S. gordonii and S. sanguinis biofilms. After treatment with S. mutans MVs, biofilms were significantly obstructed. The cells of the S. mutans MV-treated group were more dispersed than those of the control group under SEM, while the S. mutans ΔgtfBC mutant MV-treated group showed no significant changes (Fig. 2a, b). Finally, we explored whether S. mutans MVs could eradicate the established biofilms of S. gordonii and S. sanguinis. We applied the MVs to biofilms that had been formed for 24 h and continued to culture for another 24 h. However, the crystal violet assay showed that neither type of MV had a significant scavenging effect on the mature biofilm (Additional file 1: Fig. S3).

Fig. 1

Effects of MVs on the biofilm formation of S. gordonii and S. sanguinis. A, C Crystal violet staining applied to S. gordonii biofilm. B, D Crystal violet staining applied to S. sanguinis biofilm. E CLSM images of S. gordonii and S. sanguinis biofilms. Images were taken at 20× magnification. F The biomass of S. gordonii from CLSM images. G The biomass of S. sanguinis from CLSM images. The data are presented as mean ± SD from three independent experiments. (n = 3, **P < 0.01, ***P < 0.001)

Fig. 2

Morphological characteristics of S. gordonii and S. sanguinis biofilms after treatment with MVs. The bacteria and MVs were co-cultured for 24 h on glass coverslips. Images of S. gordonii and S. sanguinis biofilms were captured using SEM at 2000×, 5000×, and 10,000× magnifications

S. mutans MVs did not affect the planktonic growth of S. gordonii and S. sanguinis

Having observed the effects of S. mutans MVs on biofilm formation of S. gordonii and S. sanguinis, we questioned whether S. mutans MVs had an effect on the planktonic growth of S. gordonii and S. sanguinis. We measured the colonies in the supernatant from the biofilm experiments using the CFU count assay. However, no significant differences were found between the S. mutans MV-treated group, S. mutans ΔgtfBC mutant MV-treated group, and control group (Fig. 3a, b; P > 0.05). This indicated that S. mutans MVs does not affect the growth of S. gordonii or S. sanguinis under planktonic conditions.

Fig. 3

Effects of MVs on the planktonic growth of S. gordonii and S. sanguinis. A CFU counts of S. gordonii in the presence or absence of S. mutans MVs or S. mutans ΔgtfBC mutant MVs. B CFU counts of S. sanguinis in the presence or absence of S. mutans MVs or S. mutans ΔgtfBC mutant MVs. There were no significant differences among the groups. The data are presented as mean ± SD from three independent experiments (n = 3, P > 0.05)

S. mutans MVs inhibited the expression of virulence genes of S. gordonii and S. sanguinis

The next section of this study focused on the expression levels of virulence genes in S. gordonii and S. sanguinis. qRT-PCR was used to quantify the expression levels of GtfG, GtfP, and SpxB with S. mutans MVs treatment. When S. gordonii or S. sanguinis was incubated in 0.25%BHIS containing 1 μg/mL or 0.2 μg/mL S. mutans MVs, the transcription of Gtf genes related to biosynthesis of water-soluble and water-insoluble glucan decreased; this included the GtfG gene of S. gordonii and the GtfP gene of S. sanguinis (Vickerman et al. 1997; Xu et al. 2007), which were significantly downregulated 0.9-fold and 0.7-fold, respectively. Contrastingly, in S. mutans ΔgtfBC mutant MV-treated group, the expression level of GtfG did not change significantly, and GtfP was downregulated 0.8-fold (Fig. 4a, c; P < 0.05). S. gordonii and S. sanguinis depend on pyruvate oxidase, encoded by SpxB, to produce H₂O₂ which induces extracellular DNA (eDNA) release and cell aggregation (Itzek et al. 2011; Kreth et al. 2009), thereby promoting biofilm maturation. In our study, the SpxB gene was significantly downregulated in S. gordonii, 0.9-fold, and S. sanguinis, 0.6-fold, after S. mutans MVs treatment, whereas in the S. mutans ΔgtfBC mutant MV-treated group, the expression level of SpxB in S. gordonii showed little change and was downregulated by approximately 0.7-fold in S. sanguinis (Fig. 4b, d; P < 0.05).

Fig. 4

Effects of MVs on virulence genes of S. gordonii and S. sanguinis. A GtfG, B SpxB in S. gordonii, C GtfP, D SpxB in S. sanguinis. The data are presented as mean ± SD from three independent experiments (n = 3, ***P < 0.001)

Discussion

Given the assumption that vesicles cannot pass through the thick cell walls present in gram-positive bacteria, mycobacteria, and fungi, MVs research in these organisms was disregarded until recently. In 2009, MVs of the gram-positive bacterium, Staphylococcus aureus, were extracted from culture supernatants and was shown to enriched with virulence proteins (Lee et al. 2009).Subsequently, MVs secreted by gram-positive bacteria, such as Mycobacteria (Prados-Rosales et al. 2011) and Bacillus subtilis (Brown et al. 2014) were successively isolated, and the MVs of S. mutans were successfully isolated and identified in 2014(Liao et al. 2014). According to recent studies, Gtfs are the major proteins in the MVs of S. mutans (Cao et al. 2020; Senpuku et al. 2019). S. mutans generates at least three different Gtfs (Bowen and Koo 2011; Rainey et al. 2019), including GtfB, GtfC and GtfD, which are important virulence factors in S. mutans (Loesche 1986; Tsumori and Kuramitsu 1997). GtfB and GtfC are crucial in the metabolism of sucrose; GtfB mostly produces water-insoluble glucan, while GtfC produces both water-soluble and water-insoluble glucan (Bowen and Koo 2011). We isolated S. mutans MVs and S. mutans ΔgtfBC mutant MVs from culture supernatants, and the morphology of these MVs were observed to be “cup-shaped” and that S. mutans MVs had varied size ranges, which was consistent with previous studies on MVs isolated from S. mutans (Cao et al. 2020; Liao et al. 2014; Wu et al. 2020).

It is widely recognized that MVs are crucial for intercellular signal transduction and biofilm formation. S. mutans MVs affect the development of biofilms in a range of bacteria and fungi (Cao et al. 2020; Liao et al. 2014; Schooling and Beveridge 2006; Senpuku et al. 2019; Wang et al. 2015; Wu et al. 2022, 2020). Antagonism between beneficial commensals (such as S. gordonii and S. sanguinis) and cariogenic bacteria (such as S. mutans) is a major factor affecting the composition and ecology of supragingival biofilms (Huang et al. 2018). Therefore, we questioned whether S. mutans MVs that contained or lacked Gtfs had an impact on planktonic growth and biofilm formation of S. gordonii and S. sanguinis. Our findings demonstrate that S. gordonii and S. sanguinis biofilm development is inhibited by S. mutans MVs without having an impact on the planktonic growth. This is similar to our previous findings on S. mutans MVs, which have been shown to affect biofilm formation by C. albicans but not their growth (Wu et al. 2020). Similarly, S. aureus MVs also affect the biofilm formation of Acinetobacter baumannii, Enterococcus faecium and Klebsiella pneumonia, but not their growth (Im et al. 2017).

However, our findings are in contrast to those of a study conducted by Senpuku et al. showing that S. mutans MVs promote the biofilm formation of S. gordonii and S. sanguinis (Senpuku et al. 2019). This discrepancy may be due to differences in the strains, culture conditions, and detection methods employed. The Senpuku team used S. gordonii ATCC10558, whereas we used S. gordonii DL-1, which is commonly used in studies of interspecies interactions between S. mutans and S. gordonii (Huang et al. 2018; Ito et al. 2017; Kreth et al. 2008). In addition, in our biofilm formation experiments, S. gordonii and S. sanguinis were cultivated in BHI medium containing 0.25% sucrose, and after 24 h, the biomass of the biofilms was observed and measured by crystal violet staining and CLSM. Unlike our approach, Senpuku et al. developed biofilms in TSB medium and quantified the biofilm biomass by safranin staining after 16 h of incubation. Growth medium is one of the key factors affecting MVs production and contents (Klimentova and Stulik 2015). Changes in the composition of the growth medium altered the protein content and immunogenicity of Neisseria meningitidis vesicles (Tsolakos et al. 2010). For Francisella tularensis, the type of medium had some influence on the resulting bacterial phenotype (Hazlett et al. 2008). This suggests that the type of medium used may influence the effect MVs have on biofilm formation of S. gordonii and S. sanguinis. Another important difference was the extraction time of the MVs. We isolated MVs from the culture supernatants when S. mutans grew for 16 h, whereas the Senpuku team isolated MVs after 24 h. Liao et al. showed that S. mutans MVs from early exponential phase cultures contained 2.82-fold more eDNA than those prepared from overnight cultures (Liao et al. 2014). Tashiro et al. showed that vesicles secreted by Pseudomonas aeruginosa during the exponential and stationary phases exhibit distinct physicochemical properties; along with the growth transition, the characteristics of vesicles are changed, allowing a greater level of interaction with bacteria (Tashiro et al. 2010). Francisella novicida produces more vesicles in the early stationary phase than in the mid-logarithmic phase, and the protein profiles are different as well (McCaig et al. 2013). However, it is not yet known whether there are other differences in the MVs produced by S. mutans at the different growth stages, indicating a potential area for future research.

qRT-PCR results showed that the expression levels of GtfG in S. gordonii and GtfP in S. sanguinis were significantly downregulated in response to S. mutans MVs. GtfG and GtfP are important virulence factors in S. gordonii and S. sanguinis. GtfG, produced by S. gordonii, can synthesize both water-soluble and water-insoluble glucans and regulate the adhesion of S. gordonii (Vickerman et al. 1997). S. sanguinis has two Gtf genes, GtfA and GtfP (Xu et al. 2007), of which GtfP is the only one that produces glucans. When the GtfP is deleted, less water-soluble and water-insoluble glucan are produced, leading to a reduction in biofilm formation (Liu et al. 2017; Yoshida et al. 2014; Zhu et al. 2017). The expression levels of SpxB in S. gordonii and S. sanguinis were also significantly downregulated in response to S. mutans MVs. The H₂O₂ produced by SpxB can inhibit the growth of S. mutans and cause the release of eDNA, which is crucial for bacterial adhesion and aggregation during the initial stage of biofilm formation (Das et al. 2010, 2013; Kreth et al. 2008). qRT-PCR results also showed that under the treatment of S. mutans ΔgtfBC mutant MVs, the expression levels of GtfG and SpxB in S. gordonii did not change significantly compared with the control group, whereas the expression level of these genes in S. sanguinis decreased more significantly than those of S. mutans MV-treated group. Downregulation of GtfP and SpxB in S. sanguinis has also been observed in previous studies during simultaneous colonization with S. mutans under biofilm conditions (Lozano et al. 2019). This implies that S. mutans MVs may affect certain bacterial species differently. Gtfs may not be solely responsible for the reduced biofilm formation in S. gordonii and S. sanguinis. It is likely that other components of MVs are also involved; however, the exact mechanism is unknown. In general, the results indicate that one of the reasons for the decreased biofilm formation of S. gordonii and S. sanguinis when treated with S. mutans MVs may be the decrease in water-soluble and water-insoluble glucan. Furthermore, the reduction of eDNA release, caused by a decrease in H₂O₂ production, can cause a corresponding reduction in the matrix of the biofilm and make it less stable. However, since S. mutans MVs also contain significant amounts of Gtfs, which work with GtfG and GtfP to synthesize water-soluble and water-insoluble glucans using sucrose in the culture medium, we were unable to ascertain the amount of glucan reduction produced by GtfP and GtfG.

When we measured biofilm biomass using crystal violet staining, we found that biofilms in the S. mutans MV-treated group were more susceptible to shedding during PBS washing than the other groups, and this observation was especially prominent in S. gordonii biofilms. Therefore, we examined the expression levels of several adhesion-related genes of S. gordonii, including AbpA, AbpB and ScaA (Additional file 1: Fig. S4). AbpA and AbpB promote the binding of S. gordonii to the acquired pellicle, thus contributing to bacterial colonization and biofilm formation; this process becomes difficult for mutant strains lacking AbpA (Rogers et al. 2001; Tanzer et al. 2003). ScaA regulates the co-aggregation of S. gordonii cells (Zheng et al. 2012). Interestingly, we found that the expression levels of all these genes were significantly downregulated in response to S. mutans MVs. The downregulation of expression levels of these genes may help explain the reduced cell aggregation and biofilm formation observed in the S. mutans MV-treated group in our study. Taken together with the results of the biofilm eradication experiments, we hypothesized that S. mutans MVs may play an inhibitory role in the early stages of biofilm formation.

We successfully isolated S. mutans MVs and S. mutans ΔgtfBC mutant MVs under planktonic condition and applied them to both S. gordonii and S. sanguinis at the same concentration. We found that S. mutans MVs inhibited biofilm formation of S. gordonii and S. sanguinis but did not affect their planktonic growth. However, we experienced some unavoidable limitations across this study. Although the BCA assay is commonly used method to quantify MVs (Aytar Celik et al. 2022; Bitto et al. 2021), using this method to unify the concentration of two kinds of MVs solutions may have some influence on the experimental results because the protein content of the single MV in S. mutans and S. mutans ΔgtfBC mutant may be different. Furthermore, both in vitro and in vivo interactions of S. mutans MVs with mixed biofilms of S. gordonii and S. sanguinis is unclear. The effects of S. mutans MVs on species composition, spatial structure, and cariogenicity of mixed-species biofilms in vitro and in vivo are also unknown. In the future, it may be possible to isolate S. gordonii and S. sanguinis MVs and observe their influence on the growth and biofilm formation of S. mutans. This will undoubtedly provide deeper insights into the complicated interspecific interaction mechanism between S. mutans and S. gordonii, as well as S. mutans and S. sanguinis, from the perspective of MVs that will benefit our understanding of the functions of MVs.