The strain E. faecium FC12 used in this study and belonging to the Department of Agricultural and Food Sciences (University of Bologna) collection was isolated from a traditional Italian cheese. It was previously described as a high tyramine producer (39, 40). The strain was maintained in BHI medium (Oxoid, Basingstoke, UK) and added with 20% (w/v) glycerol at −80°C until use. Before each experiment, the strain was pre-cultivated two times in BHI medium for 24 h at 30°C.

The PEs and EOs used in this experiment, obtained from R. fruticosus leaves and J. oxycedrus needles, have been previously described and characterized in the study of Barbieri et al. (38).

The in vitro antimicrobial activity of these plant derivatives against the target strain was assessed to determine the MIC by broth microdilution method using microtiter plates (Corning Incorporated, NY, USA) according to the method of Arioli et al. (41). In particular, 198 μl of BHI broth inoculated with E. faecium FC12 (cell load approximately 6 log CFU/ml) were placed into 200 μl microtiter wells. Then, 2 μl of each plant derivative, previously dissolved in ethanol, was added to each well to obtain final concentrations ranging from 0 to 3 mg/ml. The data were collected after 48 h of incubation at 30°C. The MIC was defined as the lowest concentration of plant derivatives which was able to prevent visible microbial growth in the well.

The target microorganism was inoculated in BHI broth with a final concentration (approximately 6 log CFU/ml), and sub-lethal concentrations of plant derivates (50% of MIC) were added to evaluate their effect on the growth kinetics. In particular, 1 mg/ml of extract was used in each condition, except for R. fruticosus EO, whose concentration was 0.75 mg/ml. Culturability was assessed by plate counting at a specific time interval during the 96 h incubation at 20°C. Appropriate decimal dilutions were plated into BHI agar medium and incubated at 30°C for 48 h. The analyses were performed in triplicate, and the data obtained from plate counting were modeled with the Gompertz equation as modified by Zwietering et al. (42):

where y is the cellular load at time t, k is a constant representing the initial cellular load (6 log CFU/ml), the parameter A represents the difference between the maximum cellular load reached and the initial cell concentration, μ_max is the maximum log CFU/ml rise rate in exponential phase (h⁻¹), and λ is the lag phase duration.

Since, in some samples, an initial decrease in cell counts was observed, a double-peaked Gompertz equation was used to fit also the first part of the curves (43). In this case, the model was as follows:

where the parameters were defined with the subscript number 1 or 2 to differentiate the first phase of decrease (A₁, μ_max1, and λ₁) and the second growth phase (A₂, μ_max2, and λ₂). Data modeling was performed using STATISTICA software (Statsoft Italia, Vigonza, Italy).

The formation of tyramine and 2-phenylethylamine in the different tested conditions was monitored by collecting samples at a defined time (0, 24, 48, 72, and 96 h). The cell-free supernatants were analyzed using an HPLC Agilent Instrument 1260 Infinity with the automatic injector (G1329B ALS 1260, loop of 20 μl), equipped with a UV detector (G1314F VWD 1260) set at 254 nm. A C18 Waters Spherisorb ODS-2 (150 × 4.6 mm, 3 μm) column was used for the chromatographic separation. Before the injection, the samples were subjected to a dansyl chloride derivatization (Sigma Aldrich, Gallarate, Italy) according to the method of Martuscelli et al. (44). The BAs were expressed as mg/L with reference to the BAs standard calibration curve. All the samples were analyzed in triplicate.

To test cell viability, the samples were collected after 24, 48, and 72 h of incubation and analyzed in triplicate with a flow cytometer Accuri C6 (BD Biosciences, Milan, Italy), following the protocol reported by Arioli et al. (41). Before analyses, samples were diluted (if needed) in filtered PBS, and the cells were stained with SYBR-Green I (1X) and propidium iodide (7.5 μM) at 37°C for 15 min. This dual staining allowed to differentiate three sub-populations corresponding to different physiological states: live, injured, and dead cells. The data obtained were analyzed using the BD ACCURITM C6 software version 1.0 (BD Biosciences, Milan, Italy).

The PEs and the EOs of R. fruticosus and J. oxycedrus, previously characterized by Barbieri et al. (38), were first tested to define the MIC of the plant derivatives against E. faecium FC12. MIC was 1.5 mg/ml for R. fruticosus EO and 2 mg/ml for J. oxycedrus EO and both the PEs.

To clearly understand the effects of the EOs and PEs on E. faecium FC12, cells were inoculated in BHI and incubated at 20°C. The culture media were added with sub-lethal concentrations of plant derivatives (50% of MIC), aiming to activate a metabolic response due to the stress conditions. The growth dynamics were monitored by plate counting and compared with the control grown in the absence of plant derivatives. Figure 1 reports the experimental points and the corresponding fitted models, which show two different kinetics trends. The control and the samples containing PEs presented a typical growth curve characterized by three distinct steps: lag, exponential, and stationary phases. In contrast, the samples containing EOs showed an initial part in which cells decreased their culturability. After that, cell concentration increased, reaching an exponential phase and a subsequent stationary phase. In the first case, the experimental data were fitted with the classical Gompertz equation (42), while in the second case, a double-peaked Gompertz equation was chosen (38, 43). Table 1 reports the parameters estimated for the models used. In the table, the maximum cell concentration attained according to the models, and some diagnostics of fitting are reported as well.

E. faecium FC12 cells in the control rapidly started their exponential phase (the lag phase estimated length was 0.11 h) and reached the maximum cell concentration (approx. 9 log CFU/ml) in < 24 h, with a maximum growth rate (μ_max) of 0.309 h⁻¹. The Gompertz parameters estimated for the control and the sample containing R. fruticosus PE were similar, and the consequent models were almost completely superimposable, indicating the absence of any antimicrobial effect of this extract. However, the PE of J. oxycedrus determined a marked increase of the lag phase (approximately 24 h), a reduced μ_max value (0.069 h⁻¹), and a maximum growth of 8.11 log CFU/ml, resulting in almost 1 log unit lower than the control.

In the previous study, the same PEs were more effective in reducing the growth kinetics of L. monocytogenes when used at 50% of the MIC. However, in this case, a relevant inhibition of E. faecium was obtained only with J. oxycedrus PE, which contains not only high amounts of vanillic acid (10.51 mg/l), apigenin (7.66 mg/l), and rutin (6.95 mg/l) but also gallic acid and p-hydroxybenzoic acid (38). The antimicrobial potential of the molecules found in these PEs has been reviewed by Oulahal and Degraeve (45). Rutin, apigenin, and gallic acid showed good antimicrobial activity against E. faecalis (46, 47). Taviano et al. (48) observed a remarkable antimicrobial activity against Enterococcus hirae of methanolic extracts of J. oxycedrus berries, containing rutin, apigenin, cupressoflavone, amentoflavone, and methyl-biflavone, among the others. The R. fruticosus PE was rich in rutin, chlorogenic acid, astrignin, and caffeic acid but showed no antimicrobial activity against E. faecium under the conditions adopted in this study. Interesting antimicrobial effects on E. faecalis were described in an extract from R. fruticosus (rich in ellagic acid), which was absent in the PE used in the present study (49). In addition, Krzepiłko et al. (25) reported the antimicrobial activity of several extracts of R. fruticosus buds against the same species. E. fecalis showed high susceptibility to quercetin in combination with gallic acid and chlorogenic acid, while insufficient effects were obtained with rutin (50).

The mechanism of action responsible for the antimicrobial activity of these molecules is not completely clear. The literature suggests that some compounds (quercetin, gallic acid, and chlorogenic acid) may induce an imbalance of the redox potential, while others (apigenin and rutin) could interfere with enzymes of DNA replisome or perturbate membrane permeability and functions (50, 51). Recently, Xu et al. (52) observed a depolarization and permeabilization of the membrane and an alteration of the intracellular enzymatic activities in S. aureus when treated with a phenolic extract from Taraxacum officinale that contained rutin, caffeic acid, and chlorogenic acid.

Both EOs determined an initial decrease in E. faecium culturability. In the samples containing the J. oxycedrus EO, a reduction of more than 1 log unit of the initial population was observed after 10 h, followed by a cell load increase (λ₂ estimated at 17.12 h). However, in this second phase, the μ_max (0.160 h⁻¹) and the final cell concentration in the stationary phase (8.34 log CFU/ml) were lower when compared to the control.

The main components of the J. oxycedrus EO used in this trial were terpenes and terpenoids. Limonene (13.6%), α-pinene (10.8%), and manoyl oxide (8.4%) were the main constituents, followed by 3-carene, α-curcumene, and 4(15),5-muuroladiene (38). Recently, an EO from Juniperus phoenicea, whose main constituents were α-pinene, δ-3-carene, and β-caryophyllene, showed good antimicrobial activity against a strain of E. faecalis, while Najar et al. (53) reported a variable inhibitory effect of J. oxycedrus EOs, depending on the geographical area. Similarly, E. faecalis was the most susceptible bacterial species to an EO from Juniperus horizontalis leaves, containing p-cymene and linalool as the major constituents (54). However, no studies regarding the effects of specific EO constituents on enterococci are available.

J. oxycedrus EO and R. fruticosus leaves EO reduced the initial culturability of E. faecium, although the successive growth started later (λ₂ estimated at 33.77 h) and slowly (μ_max 0.095 h⁻¹), with a final predicted cell concentration that reached 8.23 log CFU/ml. R. fruticosus EO's major constituents were geraniol (13.7%), phytol (4.9%), β-citronellol (4.6%), linalool (4.1%), and β-ionone (3.7%). In addition, the EO contained methyl-salicylate (1.3%) and citral (2.4%) (38). Even if many of these compounds have been studied in relation to their antimicrobial activity (55), no data are reported concerning their effect on enterococci. The antimicrobial effects of terpenes, terpenoids, and phenylpropanoids are strictly dependent on their chemical structure and the presence of specific functional groups. For this reason, the mechanisms of action can be extremely different from one compound to another. However, the first requirement for EO active components relies on the possibility to solubilize in cell membranes which, in many cases, are the major target of their bacteriostatic or bactericidal effects (22, 55–57).

E. faecium FC12 has already been studied for its tyrosine decarboxylase activity (39, 40). For this reason, the formation of tyramine was monitored during growth with or without the PEs and EOs (Figure 2). As already observed for the growth curves, the control and sample containing R. fruticosus PE showed almost the same behavior. In fact, in both cases, a relevant amount of tyramine (approximately 120 mg/l) was present after 24 h, when the cells had already reached their stationary phase. The concentration slightly increased after 48 h (approximately 150 mg/l) and then remained constant.

The J. oxycedrus PE delayed the tyramine formation, with only 6 mg/l detected after 24-h incubation. This concentration increased together with the cell growth, reaching an amount comparable with the control and sample containing R. fruticosus PE after 96 h. When J. oxycedrus EO was added to the growth medium, tyramine concentration was 33.5 mg/l after 24 h incubation (with an estimated cell concentration of 5.8 log CFU/ml). Then, it increased up to 118.2 mg/l after 48 h (when the maximum cell load of 8.1 log CFU/ml was reached), where it stabilized for the subsequent 48 h. A similar maximum concentration was obtained in the samples containing R. fruticosus EO, which, however, showed a slower accumulation kinetic, following the growth curve reported in Figure 1.

It is known that, when tyrosine is depleted, the tyrosine decarboxylase of enterococci can use another aromatic amino acid as substrate (phenylalanine), producing 2-phenylethylamine (10, 18, 40). For this reason, this amine was also determined (Figure 3). As expected, it was produced only after tyramine concentration reached its maximum. In the control, 85.3 mg/l of 2-phenylethylamine was found at the end of incubation (96 h). Despite the growth data and tyramine production, in samples containing R. fruticosus PE, 2-phenylethylamine accumulation had a trend similar to the control during the first 48 h of incubation, while lower values were measured after 96 h (63.8 mg/l). In contrast, the presence of J. oxycedrus EO delayed the production of this amine after 48 h, when tyramine reached its maximum concentration. Even J. oxycedrus PE and R. fruticosus EO showed the same trend, with the latter EO determining the lowest accumulation of 2-phenylethylamine after 96 h, confirming its antimicrobial effect.

As already observed, detectable tyramine accumulation by E. faecium started in the last part of the exponential growth phase, while 2-phenylethylamine was produced by cells in the stationary phase following the depletion of tyrosine (18). The availability of precursors depends on the media composition. Previous experiments showed that E. faecium FC12 grown in the same medium (BHI) under optimal conditions (37°C without antimicrobials) determined a final ratio of 3:1 between 2-phenylethylamine and tyramine (40). In this case, after 96 h of incubation, its concentration still had an increasing trend in all the conditions, including the control.

To evaluate the impact of EOs and PEs on E. faecalis FC12 viability, an FCM protocol (38, 58) was used, and the data obtained after 0, 24, 48, and 72 h of incubation at 20°C were compared with those derived from plate counting. In particular, the total cell number, expressed as a log of cells stained with SYBR-Green I independently of their physiological state, was compared with cell culturability, expressed as log CFU/ml as predicted by the models (Figure 4A). In addition, for each condition and sampling time, the percentages of viable, injured, and dead cells detected after the dual staining with SYBR-Green I and propidium iodide were also reported (Figure 4B).

In the control, culturability (plate counting) and total cells (FCM) did not show significant differences. The viable cells were 94% at time 0 and ranged between 96 and 99% during incubation. Similar values were obtained in the sample containing R. fruticosus PE, although a slight decrease in viable cells was observed after 24 and 48 h. With the addition of J. oxycedrus PE, the number of culturable and total cells at 24 h did not increase compared to the initial concentration. Even though both values were not significantly different (6.13 log CFU/ml vs. 6.30 total cells/ml), viable cells represented 71% of the total population (corresponding to a concentration of 6.15 log cells/ml), while injured cells were 26%. After 48 h, the culturable cells increased to 7.52 log CFU/ml, whereas the total cell number was 8.73 log total cells/ml. This difference can be related to the presence of viable but non-culturable cells resulting from J. oxycedrus PE treatment that induced cell damage. This is supported by the fact that the percentage of cells recognized as live (i.e., with intact cell membrane) still decreased (63%, corresponding to 8.53 log live cells/ml), and dead cells increased (approximately 12% of the total population). At the end of incubation (72 h), total cells and culturable cells were again not significantly different, and a higher number of viable cells (95%) was reached. Therefore, J. oxycedrus PE induced temporary cell damage resulting in a slower growth kinetic, namely an increase in lag phase duration and a decrease in growth rate. The same effect with the same PE was already reported in a previous study carried out on L. monocytogenes (38).

Data concerning the J. oxycedrus EO showed comparable numbers of culturable and total cells during all the incubation. The percentage of live cells ranged from 87.8 to 98.0% log CFU/ml, and the values of viability detected by FCM (log live cells/ml) were substantially the same, indicating an exponential growth between 48 and 72 h.

A different behavior was observed in the presence of R. fruticosus EO. The number of total cells detected by FCM did not significantly change between time 0 h and time 24 h (6.17 vs. 6.35 total cells/ml). However, the cells recognized as live corresponded to 3.7% of the total population, while 93.2% resulted as injured. Thus, live cells were 4.92 log cells/ml, a value comparable with the cells detected by plate counting (4.83 log CFU/ml). After 48 h, almost all the cells recovered (92.3%), with 7.02 log live cells/ml compared to the predicted cell concentration that was lower (6.17 log CFU/ml). In fact, at this time, a great part of live cells resulted unculturable with the traditional plate counting. After 72 h, the data from FCM and plate counting were similar, being 7.66 log live cells/ml and 7.75 log CFU/ml, respectively. The effect of R. fruticosus EO on E. faecium during the first hours of incubation was in line with the results described by Barbieri et al. (38) on L. monocytogenes, indicating the ability of this EO to induce relevant cell damage. However, while in their study, damages were lethal up to 48 h (more than 99% of dead cells), those observed in E. faecium cells were transient since the bacterium was able to recover within the same time frame. Although the differences observed could depend on the specific constituents of the extracts and their interaction with the bacteria, the exact mechanism of cell recovery activated by E. faecium upon exposure to these plant derivatives is still unknown and requires further studies.