Antimicrobial activity of a novel Spanish propolis against planktonic and sessile oral Streptococcus spp

Increased bacterial resistance to traditional antimicrobial agents has prompted the use of natural products with antimicrobial properties such as propolis, extensively employed since ancient times. However, the chemical composition of propolis extracts is extremely complex and has been shown to vary depending on the region and season of collection, due to variations in the flora from which the pharmacological substances are obtained, being therefore essential for their antimicrobial activity to be checked before use. For this purpose, we evaluate the in vitro antimicrobial and anti-biofilm activity of a new and promising Spanish ethanolic extract of propolis (SEEP) on Streptococcus mutans and Streptococcus sanguinis, responsible, as dominant ‘pioneer’ species, for dental plaque. Results reveal that S. sanguinis is more sensitive to SEEP, slowing and retarding its growth considerably with lower concentrations than those needed to produce the same effect in S. mutans. SEEP presents concentration- and time-dependent killing activity and, furthermore, some of the subinhibitory concentrations employed increased biofilm formation even when bacterial growth decreased. Mono and dual-species biofilms were also inhibited by SEEP. Findings obtained clearly show the relevance of using biofilm and subinhibitory concentration models to determine optimal treatment concentrations.


Results and discussion
Antibacterial testing. The minimal inhibitory concentration (MIC) and minimal bactericidal concentration (MBC) values of SEEP estimated for S. mutans were 240 µg/mL (0.4%) and 480 µg/mL (0.8%), respectively. For S. sanguinis, the measured MIC and MBC values were lower, i.e., 60 µg/mL (0.1%) and 120 µg/mL (0.2%), respectively. The biological activity of SEEP detected was not influenced by the presence of ethanol in the propolis solutions, as a minimum solvent concentration of around 12.5% was required to inhibit the growth of both strains. MIC and MBC values of Ampicillin sodium (positive control) for S. mutans were 0.16 µg/mL and 0.31 µg/mL, respectively, and 0.08 µg/mL and 0.31 µg/mL for S. sanguinis, values close to those already published 21,22 .
It is difficult to compare MIC and MBC results from different studies due to the chemical variability of the propolis samples employed and/or the different methods used for their evaluation. Nevertheless, an increase in the bactericidal concentration to eradicate the growth of S. mutans with regard to MIC was also observed with propolis from other regions such as Argentina, MIC: 50 µg/mL, MBC: 460 µg/mL 23  Growth kinetics. The growth curves obtained were described by three key parameters: starting growth time, starting growth rate and maximum growth rate (Fig. 1a), all of which were found to be affected by the presence of SEEP.
The control culture starting growth time points of S. mutans and S. sanguinis lasted approximately 8 h and 5 h, respectively (Fig. 1b). In both cases, a delay in time to start growth was observed when using concentrations between 1/2 and 1/8 subinhibitory concentrations (sub-MICs). At 1/2 MIC, both strains needed twice the time taken by the controls to begin to grow. The starting and exponential growth rates were normalized with respect to the control values in order to better compare the two strains. As can be seen in Fig. 1c, the starting growth rate of both strains appears equally affected by sub-MICs, decreasing exponentially with the increase of the concentration of propolis relative to their MICs. In terms the maximum growth rate (Fig. 1d), however, sub-MICs affected both strains differently. In this latter case, the rate of S. mutans was found to decline from the beginning, whereas the S. sanguinis rate maintained similar values to that of the control until a SEEP concentration of 1/16 MIC was reached. From this concentration on, the maximum growth rate of S. sanguinis decreased faster than it did for S. mutans, and finally reached slightly lower values at the highest SEEP concentration studied. Sub-MIC concentrations therefore delay the onset and initial growth rate of both strains equally, but they affect both strains differently once the exponential growth phase is reached. Importantly, this later observation could potentially be useful for the design of optimal dosage regimes and suggests that determination of the typically reported MIC and specific maximum growth rate values alone is probably incomplete.
Sub-MICs can delay the onset of the lag phase and slow down the exponential phase once it has begun. The slowdown effect on S. mutans growth was also observed with natural honey in a study in which potential antibacterial properties were examined 27 . This feature may be considered as an advantageous property for dental plaque control in the early stages of biofilm formation. Indeed, different studies have already proposed propolis as a treatment of dental caries and mouth infections 28 , e.g., as a component in toothpaste oral cavity 29,30 . In any case, before it can be recommended for routine application in dentistry clinics or in the manufacture of oral products for home use, guidelines for quality control of this natural product should be developed 8 .
Lethality curves assay. Time-to-kill assay is a kinetic method of determining whether bacterial killing is concentration and/or time dependent, i.e., it assesses the ability to kill in relation to time and with different fixed concentrations of antimicrobial agents 31 . Figure 2 shows the evolution of control cultures growing exponentially (100% of relative luminescent units, RLU) after being placed in contact with propolis for different periods of time with different concentrations of SEEP. These data show that S. mutans control cultures grew during the first four hours analysed. Those cultures in contact with MICs (Fig. 2), and sub-MICs (data not shown), began to decrease after two hours of contact, www.nature.com/scientificreports/ although their metabolic activity up to that point was statistically higher than that of the controls (P < 0.01). At 2 × MIC bacterial growth was only observed in the first hour of contact and, after four hours, growth was significantly reduced by more than half (P < 0.01). The most effective propolis concentration was 4 × MIC (0.96 mg/ mL), which inhibited growth from the moment of contact (P < 0.01) and produced a significant bactericidal effect after 4 h. A similar study on optimal concentrations of Korean propolis against isolates of mutans streptococci revealed a bacteriostatic effect on S. mutans ATCC 25175 at 4 × MIC (0.14 mg/mL) after 24 h 32 , but given the lack of standardization of extraction methods and in vitro tests, interpretation of results such as this is complicated 8 .
In the case of S. sanguinis, the most effective propolis concentration was also 4 × MIC (0.24 mg/mL), which significantly inhibits growth from the first hour of contact (P < 0.01), i.e., the bactericidal effect was observed earlier. In this case, there was no significant increase in growth in the first hours of contact with MICs and sub-MICs. At 24 h of growth, all propolis concentrations tested showed no growth. www.nature.com/scientificreports/ Time-kill, therefore, was concentration and exposure time dependent. Experiments on bacterial killing kinetics of SEEP using planktonic organisms have shown that the extracts might act differentially in the tested strains. SEEP acts faster and in lower concentrations against an exponentially growing culture of S. sanguinis as compared to S. mutans. As reported by 33 , this experimental observation suggests that only concentrations greater than or equal to 2 × MIC of SEEP inhibit bacterial metabolic reactions that precede cell doubling.

Influence of subinhibitory concentrations on biofilm bacterial formation. The values of growth
and biofilm formation obtained after 24 h in the presence of the different sub-MICs of propolis studied are shown in Fig. 3. Both strains show a similar pattern of growth and biofilm formation concerning the control inocula. In particular, it was found that growth of the treated cells was significantly lower than that of the control cultures. Biofilm formation, on the other hand, did not show significant differences with respect to the control values.
However, when the Slime Index (SI) is considered according to Eq. (1), which represents the relationship between bacterial growth and biofilm formation, it is found that the values obtained show a significant difference with regard to the last two sub-MICs compared to the control (100%) in the case of S. mutans: 1/8 MIC = 151%**, 1/16 MIC = 137%*, while in the case of S. sanguinis only the first two sub-MICs showed significant differences: 1/2 MIC = 204%**, 1/4 MIC = 167%** (* 0.01 ≤ P ≤ 0.05 and ** P < 0.01). This means that the increase in biofilm formation is higher than expected based on the bacterial growth produced. This increase is caused by lower propolis sub-MICs in the case of S. mutans and by higher ones in the case of S. sanguinis. Figure 4 shows representative SEM images of biofilms after 24 h of untreated (control) and SEEP-treated cells. It is observed that S. sanguinis produces large clusters of bacteria that could cause an increase in the slime index at higher MICs (1/2 × and 1/4 × MIC), as previously mentioned. S. mutans biofilm does not show such large bacterial clusters, it is more homogeneously distributed and, at the same time, there is a higher amount of bacteria in the 1/8 × and 1/16 × MIC. Previous studies on coaggregating and non-coaggregating oral bacterial have highlighted the involvement of positive cooperativity in the coaggregation process due to streptococci, preferably  www.nature.com/scientificreports/ binded to a coaggregate rather than to available free different bacteria 34 . Future studies in this respect should be carried out to consider the differences found in the structure of the biofilms produced by the strains here studied. The sub-MICs of some antiseptic compounds, such as chlorhexidine, sodium fluoride 35 or triclosan 36 are known to increase the expression of genes related to S. mutans biofilm formation. This observation has also been demonstrated for other related species, such as for Streptococcus pyogenes treated with fluoroquinolones 37 or Streptococcus sobrinus treated with several antibiotics 38 . However, it has been shown that natural compounds such as emodin 39 , tea polyphenols 35 , leaves of Dodonaea viscosa var. angustifolia 40 or farnesol 41 can decrease S. mutans biofilm at sub-MICs, but no previous studies have analysed the relationship between growth and biofilm formation (expressed as Slime Index, SI) for these two Streptococcus strains. The present study demonstrates that some sub-inhibitory concentrations, also of propolis, can increase biofilm formation even when bacterial growth has decreased.
Activity on mature biofilms. The influence of SEEP on monobacteria and dual-species mature biofilms is shown on Fig. 5 and Table 1, respectively. The propolis extract at 100 × MIC (S. mutans: 24 mg/mL and S. sanguinis: 6 mg/mL) was the most significantly (P < 0.01) effective concentration, substantially decreasing the metabolic bacterial activity of the biofilms in 1 h, the shortest time tested (Fig. 5a). S. mutans mature biofilm was also significantly (P < 0.01) eradicated with 10 × MIC (2.4 mg/mL) after 4 h of contact; however, 10 × MIC (0.6 mg/ mL) against S. sanguinis needed at least 8 h of contact with SEEP to significantly (P < 0.01) eliminate its mature biofilm. As in the time-to-kill assay with the exponentially growing bacteria, in monobacteria mature biofilm a significant increase in activity was also observed after the first hours of contact with the lowest concentrations studied (S. mutans: P < 0.01; S. sanguinis 0.01 ≤ P ≤ 0.05). These low concentrations may initially stimulate bacterial activity, although after some time they are able to decrease viability in the biofilms.
It seems, therefore, that 4 h of contact with propolis is sufficient to totally inhibit activity of S. mutans (see Fig. 5a). Specifically, 0.96 mg/mL (4 × MIC) is needed to inhibit activity of planktonic bacteria on exponential growth (see results on "Lethality curves assay"), and 2.4 mg/mL (10 × MIC) to inhibit activity of bacteria already included in a mature biofilm. Duarte et al. 42 , however, evaluating activity of a Brazilian propolis extract against a  www.nature.com/scientificreports/ S. mutans UA159 5-day-old biofilm found that 0.8 mg/ml acting during 4 h did not have any major killing activity against this older biofilm, although acid production of the biofilms was notably reduced.
In the case of S. sanguinis, the lowest concentration tested that inhibited activity of planktonic bacteria was 0.12 mg/mL (2 × MIC) after 4 h of contact (see Fig. 5b), while the mature biofilm needed 0.6 mg/mL for 8 h to be eliminated (see, Fig. 5b). Such concentrations may be considered harmless to humans when applied in the oral cavity 43 . Furthermore, the Spanish propolis extract studied for these experiments does not contain toxic metal compounds 10 , which could cause propolis toxicity 44 . However, special care should be taken with the use of lower concentrations as results have shown an increase in metabolic activity of both sessile and planktonic bacteria during the first hours of contact.
The viability of dual-species biofilm treated with SEEP differs from that of single-species biofilms in that the former begins to be significantly affected already with the MIC of S. mutans (Table 1). Nevertheless, the lowest concentration required to totally inhibit dual-species mature biofilms in the shortest possible contact time was the same as that for the biofilm of the most resistant strain, S. mutans, i.e., 2.4 mg/mL (10 × MIC) of SEEP during 4 h of contact. Furthermore, in this case, no significant increase in viability was observed with the lower concentrations studied. Thus, the results indicate that SEEP also inhibited dual-species biofilm activity. Due to the antagonistic relationship of S. mutans and S. sanguinis within a biofilm 16,17 , further bacterial composition analyses in dual-species biofilms are necessary to evaluate the proportion of each strain when both coexist in the same biofilm that is treated with and without an antimicrobial compound 45 .
The combination of mechanisms for the observed ability of SEEP to reduce biofilm formation and to removal mature biofilms is unknown. It has been proposed that the diterpenes and triterpenes of Mediterranean propolis exhibit very strong antimicrobial activity against S. mutans 46 , such as flavonoids of Anatolian propolis 47 , benzophenones of Brazilian geopropolis 48,49 or Apigenin and tt-farnesol also identified in Brazilian and Chilean propolis [50][51][52] . SEEP is a Mediterranean-type propolis that does not have the typical pattern of "poplar type" propolis. Its composition is characterized by a high amount of polyphenol (205 ± 34 mg GAE/g), flavonoids (127 ± 19 mg QE/g), and phenolic compounds found in extra virgin olive oil such as vanillic acid, 1-Acetoxypinoresinol, p-HPEA-EA and 3,4-DHPEA-EDA, never before detected in propolis samples 10 . Previously, the interaction of SEEP and/or its molecules with bacterial cells has been studied from a physicochemical approach. In such studies, it has been detected that propolis induces substantial changes in volumetric charge density, electrophoretic smoothness and degree of hydrophobicity of the external surface layer of the cells (both gram-positive and gram-negative). According to the results of this research, the proposed mechanism of action of SEEP against bacteria appears to be initially physical, producing structural damage to the membrane/wall 53 . Afrasiabi et al. 54 also observed scattering of the S. mutans biofilm structure by propolis nanoparticles, which can may cause the loss of membrane potential, required for bacterial viability.
Future studies of SEEP compounds separately will help to better understand the mechanism of action against oral biofilm formation, and thus help to avoid the defensive effect produced by the natural extract at subinhibitory concentrations. Since nutrients can come from dead bacterial cells, inefficient SEEP concentrations could lead to adhesion and proliferation of the surviving bacteria 55 . As other authors note 8 , using such purified compounds from propolis would be easier to standardize and would provide a selective pressure which is generally associated with the emergence of bacterial resistance.
The studies carried out to date concerning SEEP and its potential action against the in vitro oral bacteria studied here guarantee the effectiveness of this natural product and contribute to the continuation of pharmacokinetic and pharmacodynamic investigations 56 which will test the efficacy in vivo of this natural substance.

Conclusions
The following conclusions can be drawn from the present study: • The Spanish propolis is effective against oral Streptococcus strains, S. mutans and S. sanguinis, responsible, as dominant 'pioneer' species, for dental plaque. • S. sanguinis is more susceptible to SEEP, as it slows and retards its growth considerably with lower concentrations than those necessary to cause the same effect in S. mutans. • It is the first time that the relationship between growth and biofilm formation, expressed as Slime Index, has been analysed for these two Streptococcus strains. In particular, this study demonstrates that some of the SEEP sub-inhibitory concentrations employed increase biofilm formation to the detriment of bacterial growth. In addition, SEM images reveal significant clustering among the S. sanguinis cells, while S. mutans yields relatively homogeneous biofilms. • Mono and dual-species biofilm are inhibited by this ancestral compound and findings obtained clearly show the relevance of using biofilm models and sub-MICs to determine antibacterial activity and their optimal treatment concentrations of the tested extract. • Further studies are required to isolate the most active molecules of SEEP against these and other oral bacteria and to elucidate their mechanisms of action, with a perspective for their use either alone or as an adjuvant in the treatment of oral infections. www.nature.com/scientificreports/ Ottoweg, Darmstadt. Germany), hereinafter TSBs, to obtain final concentrations of 12.5% to 0.025%, which correspond with 7.688 mg to 0.015 mg of dry weight of propolis per milliliter 10 .

Materials and methods
Bacterial strains and growth conditions. The strains used in the present study, i.e., S. mutans ATCC 25175 and S. sanguinis ATCC 10556, come from The American Type Culture Collection (ATCC). The strains were inoculated in blood agar plates (OXOID LTD., Basingstoke, Hampshire, UK) and incubated at 37 °C in an incubator with 5% of CO 2 (New Brunswick™ Galaxy® 170S, Eppendorf AG, Hamburg, Deutschland, Germany) to obtain cultures. Subsequently, they were cultivated in Brain Heart Infusion Agar plates (BHA) (OXOID LTD., Basingstoke, Hampshire, UK) under the same conditions mentioned above to refresh colonies. Both species were routinely grown overnight at 37 °C and 5% CO 2 in TSBs. After this time, each bacterial suspension was adjusted to 82% of transmittance, equivalent to 0.086 of optical density (O.D.), at 492 nm wavelength. A spectrophotometer (Helios epsilon Model, Thermiospectronic, Waltham, MA, USA) was used for the measurement and subsequently, the suspension was diluted 1/100 or 1/10 in TSBs to obtain approximately 10 6 or 10 7 CFU/mL, respectively and thus to be used as inocula in the remaining experiments. Growth kinetics. Bacterial growth curves of S. mutans and S. sangunis were obtained by turbidity measurements. Specifically, cells were cultivated, with and without propolis, as described above in 96-well microtitre plates, and the optical density of each of the samples was recorded at regular intervals from zero time to 24 h. Once determined, growth curves were described by the following parameters: starting growth time, starting growth rate and maximum growth rate (Fig. 1a). The first parameter describes the time at which the absorbance value first differs significantly from zero. After this point in time, cells were found to experience a gradual continuous growth before entering the exponential phase. The slope of these data points, i.e., those preceding the exponential phase, was estimated and defined as the starting growth rate. Finally, maximum growth rate is defined as the highest speed of growth detected along the growth curve and corresponds to the slope of the curve at the exponential phase. The data presented represent the mean ± SD of at least three-independent curves obtained from different cultures.
Lethality curves assay. Time-to-kill assays were performed in 96-well microtiter plates containing 100 μL of approximately 10 6 CFU/mL cultures that were grown for 12 h to reach the beginning of stationary grown phase. At this point, before adding the different SEEP concentrations, the first measure of metabolic activity of bacteria (zero time) was made. The procedure to measure bacterial killing was by ATP bioluminescence assay which allows determination of the number of viable microbial cells in cultures based on the quantification of adenosine triphosphate, which is a chemical form of energy of all living cells, measured by a luminometer and expressed as relative luminescent unit 58 .
To carry out the following measurements, 50 μL of each well was replaced, taking care not to cause any contamination, with 50 μL of fresh medium with and without SEEP. After 1, 2, 4 and 24 h of contact, viability measurements were performed by adding 100 μL of the reagent BacTiter-Glo™ (Promega Corporation, Madison, WI, USA) to the wells and allowed to interact for 5 min in the dark. Finally, the contents were transferred to 96-well white polystyrene flat-bottomed microtiter plates (Greiner Bio-One, Frickenhausen, Germany) and light emission (luciferin-luciferase reaction) was measured in a fluorescence microplate reader (FLx800; Bio-Tek Instruments, Inc. Winooski, VT, USA). In this assay, the surfaces without bacteria were included as negative controls and propolis-free medium served as positive controls. The experiments were carried out at least three times with independent cultures in order to confirm reproducibility.
Anti-biofilm activity of SEEP. Biofilm formation was performed in 96-well microtiter plates. The initial inoculum was approximately 10 6 CFU/mL and was added to the wells containing 100 µl of TSBs with and without propolis. Positive and negative controls were also included as described above. Bacterial growth was measured after 24 h of incubation as O.D. at 492 nm using a spectrophotometer ELx800. Each well was aspirated, and samples were carefully washed twice with phosphate buffered saline (PBS) using a suction pump (Model FTA-2i, Biosan SIA. Riga, Latvia). The biofilms adhered to the bottom of the wells were heat-fixed in a Pasteur Heraeus electronic oven (C.R. Maré, S.A., Barcelona Activity of SEEP on mature biofilms. Mature biofilms were obtained from inocula (200 μL; 10 7 CFU/ mL) of S. mutans, S. sanguinis and mixed cultures incubated at 37 °C for 24 h in a 96-well white polystyrene flatbottomed microtiter plates. The 1-day-old biofilm was washed twice with sterile PBS using a suction pump to remove nonadherent cells. These biofilms, established for 24 h, were subsequently treated with different propolis concentrations and the control with TSBs. After incubation during different contact times, viability of bacteria in biofilm was measured by bioluminescence reaction 60 . First, the supernatant was removed and washed twice with PBS to remove non-adherent bacteria, and then 200 μL of BacTiter-Glo™ reagent prepared according to the manufacturer's instructions was added to each well. Results were expressed as percentages relative to control at zero time without SEEP treatment.
Scanning electron microscopy (SEM). Bacterial suspensions were incubated at 37 °C for 24 h (5% CO 2 ) with TSBs and TSBs containing the different subinhibitory concentrations of SEEP studied on glass coverslips circles within 12-well microtitre plates (BioLite 12 Well Multidish, Thermo Fisher Scientific, Rochester, NY. USA). After 1, 4, 8 and 24 h of incubation, the covers were carefully washed twice with sterile TSBs to eliminate the non-adherent bacteria.
The growth biofilms were fixed at room temperature with 3% vol/vol glutaraldehyde (PanreacQuímica SAU, Barcelona, Spain) for approximately 15 h and dehydrated in a series of ethanol solutions (30,50,70, 90 and 100% vol/vol) for 1 h each. The samples were then dried in a vacuum chamber, coated with a thin layer of gold (≤ 5 nm) using an EMITECH K575K (Quorum Technologies Ltd., West Sussex, UK) sputter coater, and finally, the image was captured with a Scanning Electron Microscope (HITACHI S-4800, Hitachi High-Technologie, Tokyo, Japan).

Statistical analysis.
Data are presented as mean and standard deviation is determined across the duplicates with at least three independent experiments. Statistical analyses were performed using SPSS 22.0 (SPSS Inc., Chicago, IL, USA). In all experiments, to compare multiple means of treated samples versus control a one-way analysis of variance test (ANOVA) followed by Tukey's post hoc test or by Dunnett's T3 post hoc test when Levene's test revealed unequal variance. When the data were not normally distributed, the Kruskal-Wallis and multiple-comparison tests were used to compare values.

Data availability
Correspondence and request form datasets generated and analysed during the current study should be addressed to MLNP and MCFC.