Lentisk berry oil (LBO) was obtained from P. lentiscus fruits from Mediflora^® (Pula, Cagliari, Italy) by using ripe berries harvested in winter in the southern Sardinian region (specific weight = 890 Kg/m³). After a cycle of light dehydration by air for 20 days, the drupes were cold-pressed following an already described method (Mezni et al., 2014b). The oil was then filtered and stored in stainless steel containers at 20°C until bottling. Prior to use, the sample was centrifuged at 12,000 rpm for 15 min and the supernatant was used for the experiments.

A set of bacteria/yeasts described in human tissues as pathogens, commensals, or probiotics was used: S. aureus ATCC 6538 (American Type Culture Collection), Staphylococcus hominis human clinical isolate NC5, Pseudomonas aeruginosa ATCC 27853, Bacillus clausii isolated from a commercial product, Enterogermina^® (Senesi et al., 2001). Streptococcus agalactiae human clinical isolate NC2, Streptococcus intermedius DSM 20573 (German Collection of Microorganism and cell culture), Streptococcus mitis human clinical isolate NC1, S. mutans CIP103220 (Collection Institut Pasteur), Streptococcus pyogenes, human clinical isolate NC4, Streptococcus salivarius, strain K12 isolated from a commercial product (Bactoblis) corresponding to ATCC strain^® BAA-1 024 and S. salivarius strain M18 isolated from a commercial product (Carioblis^®). Prior to use, these strains were stored at -80°C in a tube containing Muller Hinton or Schaedler Broth (Microbiol, Uta, Cagliari) with 20% glycerol. Furthermore, three different Candida spp. were assayed: Candida albicans clinical oral isolate BF01, Candida glabrata clinical isolate BF02 and Candida krusei clinical isolate BF03. The yeast colonies were identified with an API ID32C AUX system (Biomerieux, St. Louis, MO, United States) and maintained at -80°C in Sabouraud Broth/Glycerol 20% prior to use. For each strain, a growth curve in the corresponding liquid medium plus 10% glycerol was performed. The culture at the mid- lag phase was stored at -80°C prior to use. These bacterial suspensions were used as standardized inoculum to minimize the variation in experimental conditions and ensure reproducible experimental data.

As a first step, an agar diffusion test (Kirby-Bauer) was used to evaluate bacterial resistance/susceptibility (Barry et al., 1979). The protocol was modified due to the high oil density. For each bacterium strain, 15 mL of agarized medium (Microbiol, Uta, Cagliari) at 55°C was added to a 90 mm Petri dish and, prior to agar solidification, four sterile iron rivets, Ø 10 mm diameter and 2 mm thick (Firm, Milan, Italy), were put into the agar mixture and then removed from the medium when it was cold. In these conditions, the wells contained 50 μL of oil. Each strain was inoculated onto the plate surface using a sterile swab with bacterial 5^∗10⁷ (CFU) standardized inoculum. Three wells were used for the oil testing and two for a negative control. The Petri dishes were incubated in air at 37°C for 24 h for the aerobic strains and in 5% CO₂ at 37°C for the Streptococcal species. After incubation, the diameter of the possible inhibition alone was measured and the experiment was performed in triplicate. In this experiment, Schaedler agar was used for Streptococci, Muller Hinton agar for aerobic bacteria and Sabouraud agar for yeasts (Microbiol, Uta, Italy). Broth dilution and antibiofilm tests were only used for bacterial genera that showed sensitivity to LBO.

The minimum inhibitory concentration and minimum bactericide concentration, MIC and MBC respectively, were performed by the micro-dilution method following CLSI procedures (Mushtaq et al., 2004). The method was executed in sterile Nunc^TM Microwell^TM 96-well microplates (Thermo Fisher Scientific) and each well was coated with twofold serial dilutions of lentisk oil with Schaedler Broth (from 50 to 0.04%) until a final volume of 0.20 mL for each well. For each strain, 1^∗10⁷ CFU/mL liquid suspension was prepared from the standardized inoculum in Schaedler broth; 0.02 mL of this suspension were put into each well to obtain a final concentration of 10⁶ CFU/mL. The cultures were incubated at 37°C and 5% CO₂ for 24 h. To determine the MBC, 150 μl of dilution representing the MIC and two of the more concentrated LBO wells were plated in Schaedler Agar at 37° with 5% CO₂; after 24 h, the colony-forming units (CFUs) were enumerated. The MBC is the lowest concentration that demonstrates a reduction (such as 99.9%) in CFU/mL when compared to the MIC dilution.

Using the same experimental conditions, we tested the antibacterial activity of serial dilutions of oleic acid, from 500 to 0.5 μg/mL. The MIC and MBC were carried out with oleic acid, 18:1 (n-9) (Sigma-Aldrich, St. Louis, MO, United States).

Lentisk berry oil’s ability to inhibit biofilm formation, MBIC (Subramenium et al., 2015), was evaluated following the modified crystal violet staining protocol described by Montana University Center for Biofilm Engineering¹. Each Streptococcus strain was cultured in triplicate on 96-well microplates with different concentrations of LBO in Schaedler Broth; after 7 days at 37°C with 5% CO₂, the medium was discarded and the wells were gently washed three times with a 0.9% NaCl solution; then 0.1 mL of a 0.1% of crystal violet solution was added to each well; after 10 min the dye was discarded, followed by three washes with 0.9% NaCl. After an air-drying procedure at 25°C for 15 min, 0.3 mL of 30% acetic acid were added to each well. The plates were read with a microplate reader at 550 nm (SLT-Spectra II, SLT Instruments, Germany). The MBIC represented the lowest concentration showing an absorbance comparable with the negative control (sample without bacteria), in the same concentration series the data showing a Standard deviation (SD) within 10% of the mean value were considered significant.

We developed a growth system to increase the bacterium-LBO contact area by means of an emulsion bioreactor (Figure 1). The system contained 300 mL of Schaedler Broth/LBO in a biphasic oil/medium phase. Three different LBO concentrations were studied: 15, 30, and 50% LBO respectively. This system produced a number of physical conditions able to promote the growth of the Streptococcus strains analyzed in this work; the temperature was maintained at 37°C and the air contained 5% CO₂. The reactor was kept in continuous agitation by a lab horizontal mixer (Continental Instruments, Italy) with a rotation speed (β) of 15 rpm (Denotti et al., 2009). Single significative LBO resistant strains, such as S. pyogenes, S. salivarius K12, S. agalactiae and S. mutans, were inoculated with an inoculum of 1^∗10⁶ CFU/mL.

In these physical conditions, individual oil droplets can be considered as micro incubators containing a considerable number of bacteria on their surface. The adhesion force for the bacterial cell-droplets (ΔE) depends on droplet radius (R), interfacial tension (γ), and the contact angle (𝜃), with: [ΔE = -πR2γ (1-|cos𝜃|)₂] (Dewey et al., 2014). ΔE also depends on the number of droplets and this is proportional to the bascule rotation rate [ΔE ∝ β], which in this case was 60 cycles/min. Every 30 min, 2 mL of medium (Shaedler Broth) and 0.1 mL of oil fraction were sampled and used for growth curve measurement and for S. mutans inhibitory activity detection, respectively. The fatty acid profile in the LBO was measured at T0 and after 8 h using 10 mL of LBO specimens. The growth curves were performed in triplicate by reading the absorbance = 550 nm of the medium by a spectrophotometer (Prixma 5100, Italy, optical path length = 10 mm). For each value series, the max absorbance standard deviation (SD_max) accepted was ≠20% of the mean value.

The oral pathogen S. mutans was used as a target bacterium to evaluate whether S. salivarius K12 or S. pyogenes produce antibacterial metabolites from LBO, when cultured in a biphasic oil/medium emulsion. Four emulsion cultures of S. pyogenes and S. salivarius K12 containing different LBO concentrations (0, 15, 30, or 50%) provided the oil sample aliquots. These were then used at 0 and after 8 h (T0, T8) of incubation to establish the inhibitory activity against S. mutans. The antibacterial assay was carried out following the described Kirby-Bauer test by using 50 μL of the bioreactor oil in every well. The growth inhibition zone was assessed after 24 h with 5% CO₂, and the control (Schaedler broth plus bacterium) was analyzed under the same conditions.

At (T0) and after 8 h of incubation an aliquot (10 mL) of the bioreactor medium, inoculated with S. pyogenes, was used to analyze the fatty acid composition. Fatty acid standards and solvents of the highest purity available were used and purchased from Sigma–Aldrich (Milan, Italy). All the other chemicals used in this study were of analytical grade.

Aliquots (3 mg) of P. lentiscus oils dissolved in EtOH (2 mg/mL solutions) were subjected to mild saponification in the dark at room temperature for 14 h (Rosa et al., 2012). The saponifiable fraction with fatty acids was collected, the solvent was evaporated, and a portion of the dried residue was dissolved in CH₃CN/0.14% CH₃COOH (v/v). Analyses of fatty acids were carried out with an Agilent Technologies 1100 liquid chromatograph equipped with a diode array (DAD) detector and an Infinity 1260 evaporative light scattering detector (ELSD) (Agilent Technologies, Palo Alto, CA, United States). Unsaturated (detected at 200 nm) and saturated (detected with ELSD) fatty acids were separated with an XDB-C18 Eclipse column (150 mm, 4.6 mm, 3.5 μm Ø particle size) equipped with a Zorbax XDB-C18 Eclipse (12.5 mm 4.6 mm, 5 μm Ø particle size) guard column (Agilent Technologies), with a mobile phase of CH₃CN/H₂O/CH₃COOH (75/25/0.12, v/v/v), at a flow rate of 2.3 mL/min (Rosa et al., 2012). Fatty acid identification was made using standard compounds and the conventional UV spectra, generated with the Agilent OpenLAB Chromatography data system. Calibration curves of all the compounds (correlation coefficients > 0.995) were constructed using standards and were found to be linear for the DAD detector and exponential for ELSD.

To evaluate the lipase activity of non-sensitive native LBO strains (S. agalactiae, S. pyogenes, S. salivarius, and S. mutans) a lipolytic enzyme test was performed, using the oil/phenol red procedure described by Lee et al. (2015), modified to be used with Streptococci and LBO.

The test was executed in a Petri plate (Ø = 30 mm) containing: phenol red (0.05% w/v), LBO (0.1% v/v) in Schaedler Agar. Each strain was cultured with Schaedler Broth/LBO (50%) in a shake emulsion culture at 37°C for 24 h. After centrifugation at 6000 rpm for 10 min, the bacteria were recovered and re-suspended in a 0.9% saline solution. 0.1 mL of 0.5 McFarland standard suspension corresponding to 10⁸ CFU/mL was inoculated onto the center of the plate containing a fissure obtained through the rivet method. After 24 h of incubation at 37°C and 5%, the CO₂ yellow alone diameter was measured (ø mm) and reported as a semi-quantitative lipase assay. Lipase activity was calculated with the subsequent formula:

[Lipase act = Ø mm _Lbo – Ø mm _control]

where: Ø mm _Lbo = the diameter of the yellow alone around the fissure in the medium with LBO and Ø mm _control was the diameter in the same medium without LBO.

We analyzed and compared in silico 31 different Myosin cross reactive antigen sequences (MCRA) for Streptococcus spp., available in the UniProt data bank and in the NCBI Protein database. The alignments of multiple sequences were analyzed using the Clustal Omega program². The phylogenetic tree was generated using GeneBee-NET software and Phylogeny.fr³ according to a neighbor joining algorithm (Brodskii et al., 1995). Table 1 shows the MCRA accession number relative to the Streptococci used in this work.

Evaluation of the statistical significance of differences was performed using Pearson’s chi-square test for antimicrobial activity experiments and one-way analysis of variance (one-way ANOVA), followed by the Bonferroni multiple comparison test for chemical data, using GraphPad InStat software (San Diego, CA, United States).

An initial evaluation using Kirby-Bauer analysis showed that yeasts, Candida spp. and Gram-negative bacteria did not demonstrate appreciable sensitivity to this extract. The same also held true for S. aureus, S. hominis and B. clausii, but an unusual antimicrobial profile was observed within the Streptococcus genus. In fact, native LBO resulted active against two different strains: S. mitis and S. intermedius (inhibition diameter 12≠1 and 15≠1 mm respectively), while it appeared to be inactive against S. salivarius and other Streptococci. This first experiment suggested a possible differential activity within Streptococcus spp. and, from this point on, we investigated their interaction with LBO. The antimicrobial-anti biofilm analysis evaluated in microplates through the use of a liquid medium partially reflected the results suggested by the Kirby-Bauer procedure with substantial differences amongst the native LBO non-sensitive strains. In fact, by comparing different LBO concentrations with the studied Streptococci, we were able to observe three types of “sensitivity motifs” (Table 2). These values were thus in accordance with the MIC values obtained with oleic acid and were correlated with the lipase–hydrolase activity, as described later. The S. salivarius K12 and M18 probiotic strains confirmed their non-sensitivity to lentisk, group I (MIC and MBIC > 50%). Other Streptococci followed two trends:

(i) High MIC value [50%] and a biofilm structure observable from [25–3%] in LBO dilution, group II.

(ii) Very sensitive strains characterized by low LBO MICs [6–3%] and low LBO MBIC [<4%], group III.

By comparing the diffusion alone obtained with the studied Streptococci, great activity was observed after 12–24 h with S. pyogenes (Ø = 7 ± 1 mm diffusion), while the other Streptococci showed a low activity (Ø = 5 ± 1 mm). This was why we chose to study the HPLC profile of FFAs with S. pyogenes (Table 2).

The presence of LBO in the bioreactor medium imposed a different growth trend as shown in Figure 2. Group I, S. salivarius K12 presented a classic curve comparable to growth without oil, whereas the partially sensitive LBO group II, S. pyogenes, S. agalactiae, S. mutans, showed an interesting trend: the observed “bell curve” could indicate possible inhibition after an initial growth, which suggests that secondary metabolites may act as antimicrobials. These curves were designed by using the medium value of three experiments and the max Standard Deviation observed (SD_max) was = 20%, from the absorbance medium value. In these conditions, the first point of decrease in growth on the curve was observed at 7/8 h for S. pyogenes and S. mutans, while it occurred at 9–10 h for S. agalactiae, indicating a possible representative concentration of these hypothetical antibacterial metabolites (Figure 2).

Quali-quantitative information on the individual fatty acids that compose the lipid classes of P. lentiscus oil used in the growth experiment with S. pyogenes was obtained by HPLC-DAD/ELSD analyses after lipid saponification. The values of the main saturated and unsaturated fatty acids (UFAs) (expressed as mg/g of oil) of P. lentiscus oil are reported in Figure 3A. The oil was characterized by a high level of oleic acid C18:1 n-9 (465.5 ± 2.0 mg/g of oil extract), linoleic acid C18:2 n-6 (254.2 ± 5.5 mg/g of oil extract), and palmitic acid C16:0 (170.6 ± 19.5 mg/g of oil extract), with a low amount of α-linolenic (C18:3 n-3) and palmitoleic (C16:1 n-7) acids. The oil also showed a detectable amount of FFAs, and values of the free form in the range of 5–12% were measured using HPLC for the UFAs. After incubation with S. pyogenes, LBO added at different amounts (15, 30, and 50%) to the culture medium, was separated and subjected to lipid saponification for the determination of the total fatty acid profile and directly analyzed for the quantification of FFA. No significant differences were observed in the total fatty acid profile of oil samples after microorganism incubation as compared to the control oil (LBO). However, a significant increase (133% of CO, P < 0.01) in FFA content (as the sum of free C16:0, C18:1 n-9, C18:2 n-6, and C18:3 n-3) was observed in 50% of oil samples versus the native oil control (Figure 3B).

Streptococcus mutans inhibition was observed with oil sampled after S. pyogenes culture and shows an approximately third level polynomial curve function (R² = 1) (Figure 4). The relationship between the percentage of lentisk oil and the diameter (mm) of the inhibition zone could be represented as follows: [Y = -7^∗10^-4X³+5^∗10^-2X²-0.6^∗X+3^∗10^-12], where: Y = medium of diameter inhibition (Ø mm) and x represents the percentage of lentisk oil in the bioreactor. Neither the respective liquid medium (Shaedler broth) fraction, nor any of the LBOs coming from S. salivarius showed an appreciable antibacterial activity. The represented curve was performed by using the values of three experiments.

We observed the complete inhibition of S. intermedius and S. mitis (group III) at a low concentration of oleic acid (0.5 μg/mL), while the S. salivarius K12 (group I) showed a MIC > 500 μg/mL. For partially sensitive strains, group II, i.e., S. pyogenes, the MIC values were comprised between 250 and 0.5 μg/mL. These results indicate a striking difference in sensitivity to oleic acid amongst the analyzed Streptococci at concentration ranges of oleic acid from 500 to 0.5 μg/mL, as described by other authors (Speert et al., 1979; Speert and Wannamaker, 1980) (Table 2).

All MCRA protein sequences relative to different S. salivarius strains, extracted from the NCBI Data Bank, were comprised in the same identity group. At the same time, this group showed a considerable aminoacidic identity distance in comparison with other protein sequences published for other Streptococci (Figure 5).