The use of plant extracts and bacteriophages as an alternative therapy approach in combatting bacterial infections: the study of lytic phages and Stevia rebaudiana

Bacteriophages and their corresponding bacterial hosts were purchased from the German Collection of Microorganisms and Cell Cultures GmbH (Leibniz Institute DSMZ (Deutsche Sammlung von Mikroorganismen und Zellkulturen)). Pseudomonas syringae (DSM 21482) with phage Phi6 (DSM 21518), Escherichia coli (DSM 5695) with phage MS2 (DSM 13767) and E. coli (DSM 613) with phage T4 (DSM 4505) were selected as model microorganisms based on their different features, as presented below (Table 1).

Bacterium revival and phage propagation was conducted as described earlier (35). Briefly, the bacteria were revived from storage at -20°C in trypticase soy broth (BioMaxima, Lublin, Poland) with 10% (v/v) glycerol on Luria–Bertani (LB agar) (BioMaxima) by streaking glycerol stocks onto agar plates. The plates were then incubated for 24 h at 37°C for E. coli strains and for 48 h at 28°C for P. syringae. Phages were amplified by inoculating 50 mL of LB agar with colonies from bacterial stock plates and incubated (as described above) with agitation at 120 rpm in an orbital rotating shaker (ES-20 Shaker-Incubator; BioSan, Józefów, Poland) to reach optical density of 0.2 at 600 nm (OD_600nm). The optical density was measured using an Infinite 200 PRO NanoQuant microplate reader (Tecan, Männedorf, Switzerland). At this point the phages were added and samples were further incubated until lysis occurred. For MS2 and T4 lysate purification, samples were supplemented with 10% (v/v) chloroform, vortexed for 5 min and then centrifuged (model 5810 R; Eppendorf, Hamburg, Germany) at 5,000 rpm for 25 min at 4°C. The supernatant (phage lysate) was collected immediately and stored at 4°C for further use. For Phi6 lysate purification the sample was centrifuged at the first stage at 5,000 rpm for 15 min at 4°C and then sterilised through a 0.22 μm polyethersulfone (PES) filter. The lysate was collected and stored at 4°C for further use. The phages’ activities and titres were tested by a double-overlay agar plaque assay (19).

Preparation of S. rebaudiana extracts was carried out as described in our previous work (24) with some modifications. Fifty grams of dried stevia leaves (Flos, Mokrsko, Poland) were placed in glass bottles. The first bottle was intended for methanol extraction; therefore, 100 mL of 70% aqueous methanol (MeOH) was introduced. The second bottle intended for acetone extraction was supplemented with 100 mL of 70% aqueous acetone. Next, the samples were placed on a shaker (Ika, Staufen im Breisgau, Germany) and extracted at 150 rpm for 2 h at 70°C. The crude extracts were filtered through a Büchner funnel equipped with a cellulose filter. The extracts were then concentrated by evaporation at 50°C to obtain 10 mL of each aqueous solution. After the evaporation of methanol and acetone, the samples were diluted with 20 mL of water and sterilised through a 0.22 μm PES filter. The samples were then used for further experiments. At this stage the S. rebaudiana water solution of active compounds obtained by methanol extraction was marked “SRm”, and the S. rebaudiana water solution of active substances obtained by acetone extraction was marked “SRa”. Additionally, the dry mass of each extract was determined via moisture analyser (Radwag, Radom, Poland). For the next experiments, stock solutions of the extracts were serially diluted twofold in sterile, deionised water and kept at -20°C until further analysis.

In order to test the influence of S. rebaudiana SRa and SRm extracts on bacterial cells, a 96-well microplate dilution experiment was performed. For the microplate method, modified minimum inhibitory concentration (MIC) determination was used (42). The colonies from stock bacterial plates were used to inoculate Falcon tubes with 30 mL of LB agar and the tubes were shaken at 120 rpm and incubated at 37°C for E. coli strains and 28°C for P. syringae on the orbital rotating shaker until the cultures reached an OD_600nm of 0.2. Polystyrene flat-bottomed plates with 96 wells were filled with 50 μL of twofold serial dilutions of SRa and SRm (50%–0.003%) and 50 μL of bacterial suspensions. Sterile deionised water added in place of extracts was used as the positive control (as a bacteria growth control). The plates were then placed in incubators for 24 h at temperatures appropriate for the tested bacteria. Optical density values were measured using the Infinite 200 PRO NanoQuant microplate reader (Tecan, Männedorf, Switzerland). The experiment was conducted in triplicate.

For direct phage–extract interaction testing, a co-incubation experiment was performed (34). The test was carried out in order to reveal the influence of extracts on bacteriophages’ plaque-forming ability and titres. It was conducted in a static environment in order to detect basic interactions between the tested factors. Briefly, 12-well flat-bottomed polystyrene plates were initially filled with 1 mL of phage lysates (Phi6 and T4 at 10⁸ plaque-forming units (PFU)/mL and MS2 at 10⁹ PFU/mL). Then 1 mL of either SRa or SRm was added to reach final concentrations of 50.000–0.049% by twofold serial dilution. Extract-free deionised water with phage lysate was used as a positive control, i.e. as a phage control. The plates were then incubated at room temperature for 24 h in the dark. Later, samples were titrated_in tris–magnesium sulphate buffer (50 mM tris-HCl, 10 mM MgSO₄ at pH 7.5) by spotting 3 μL of each tenfold dilution onto an LB agar plate precoated with a top agar layer (7%) mixed with overnight bacterial culture, implementing the double-layer agar technique. The experiment was conducted in triplicate.

Phage-extract synography (an optically based real-time microtiter plate readout combined with a matrix-like heat map of treatment potencies to measure phage and antibiotic synergy (PAS)) was performed as described elsewhere (15) with minor modifications. For test culture preparation, 5 mL of the overnight culture was diluted in LB agar in order to achieve OD_600nm of 1 (approximately 1 × 10⁹ CFU/mL) and 100 μL of the suspension was transferred into each well of the 96-well flat-bottomed plates. The plates had previously been inoculated with 50 μL per well of phage solution at one of a range of concentrations (10²–10⁸ PFU/mL) and 50 μL per well of extract (25%–0.049%), creating a checkerboard assay. The plates were incubated for 18 h at 37°C for E. coli, and at 28°C for P. syringae. Bacterial turbidity was determined by measuring OD_600nm values using the Infinite 200 PRO NanoQuant microplate reader (Tecan, Männedorf, Switzerland). Afterwards, a resazurin assay was carried out for bacteria metabolic activity testing. The dye was added to the samples in wells to reach a final 1 mg/mL concentration, and the plates were further incubated in the dark for 3 h when the bacteria were E. coli and for 4.5 h when the bacterium was P. syringae. Next, fluorescence measurements were performed using a fluorescent plate reader (Synergy HTX, BioTek Instruments, Winooski, VT, USA) at 540 nm excitation and 590 nm emission. The experiment was conducted in triplicate.

Based on results from the static synography experiment without mixing conditions, combinations in which interesting phenomena were detected, particularly increased bacterium activity in the resazurin assay with simultaneous OD values showing a reduction in bacteria biomass were chosen. These combinations were selected in order to analyse the influence of combined treatments of phage extracts on bacterial host growth rate in real time in a dynamic environment, i.e. one with mixing conditions. Appropriate controls for result comparison (phage + extract maximal dose, extract maximal dose, and bacterial growth control) were also applied (Table 2).

In order to keep the experimental assumptions of the synography test, overnight bacterial host cultures were diluted in LB agar to achieve an OD_600nm of 1, and 5 mL of the suspensions were then transferred to 50 mL Falcon tubes. Afterwards, 2.5 mL of phage lysate was added to give a final titre of 10⁸ PFU/mL, along with 2.5 mL of SRa or SRm extract concentrations, in order to obtain the chosen final concentrations, e.g. 50% extract was added to obtain a final concentration of 12.5%. For extract-only treatments, 2.5 mL of sterile deionised water was added instead of phage. For the host bacterium growth control, 5 mL of sterile deionised water was added to 5 mL of the bacterial suspension. Samples were then shaken at 150 rpm and incubated for 16 h, E. coli at 37°C and P. syringae at 28°C, and real time OD_850nm values were measured using BioSan bioreactors (BS-010160-A04, BioSan, Riga, Latvia).

One-way (plant extract antimicrobial studies) and two-way (phage-extract co-incubation assay) analyses of variance were used to statistically analyse the results, along with Dunnett’s multiple comparisons test. Differences were considered significant at P ≤ 0.05. All statistical analyses were carried out using GraphPad Prism 8.01 (GraphPad Software, San Diego, CA, USA).

The tested plant extracts had dry masses of 32.8433% (328.433 g/L) in the case of SRa and 21.4921% (214.921 g/L) in the case of SRm. In order to test the effects of the extracts on bacterial cells, a microdilution assay was performed. The mixtures of plant extracts and bacterial cultures cultivated in static conditions produced varied effects that were strictly dose dependent (Fig. 1). The SRa extract resulted in a decreased bacterial biomass of E. coli DSM 613 and of E. coli DSM 5695, except when mixed at three concentrations: 6.25% and 3.125% raised the bacterial counts and 1.56% kept bacteria biomass at the control level of 0%.

Fig. 1.

The SRa extract addition caused a reduction in P. syringae DSM 21482 cell density when used at 0.19%, whereas the biomass was kept at the control level for most concentrations, specifically 50%, 12.5%, 6.25%, 3.125%, 1.56%, 0.78%, 0.097% and 0.024%, or was increased, this occurring at 25%, 0.39%, 0.048%, 0.012%, 0.006% and 0.003% (Fig. 1A). The SRm extract also showed an inhibitory effect on E. coli DSM 613 cells at the majority of the tested concentrations, except for 6.25%, 3.125%, 0.024%, 0.012%, 0.006% and 0.003%, at which SRm did not influence bacterial growth. A similar effect was observed for the SRm extract on E. coli DSM 5695, where the majority of the tested concentrations decreased the bacteria biomass, except for 12.5% and 6.25%, which propagated the bacteria’s growth, and 3.125%, which did not alter bacteria multiplication. The results were very similar to those obtained with the SRa extract. The addition of the SRm extract to P. syringae DSM 21482 cells at most concentrations induced no effect on the biomass of the bacilli, four concentrations being exceptions: 3.125% inhibiting and 0.012%, 0.006% and 0.003% stimulating bacteria growth (Fig. 1B).

Co-incubation showed varied results (Fig. 2). The addition of the SRa extract did not significantly influence MS2 phage counts, the 50%, 25% and 12.5% concentrations exerting a phage stimulating effect, however. When the SRa extract was added to the Phi6 phage lysate, the extract caused a phagicidal effect, where no phage plaques were present in any tested concentrations. The combination of this extract with the T4 phage resulted in mixed effects. Four SRa concentrations, which were 25%, 3.125%, 1.56% and 0.097%, did not alter phage counts significantly; two, namely 50% and 6.25%, decreased phage PFU/mL; and five stimulated the counts, these being 12.5%, 0.78%, 0.39%, 0.19% and 0.049% (Fig. 2A). Similarly to the SRa extract, the SRm extract did not significantly influence MS2 phage counts, with the exception of two concentrations, 50% and 25%, which raised the number of phage plaques. In contrast, the addition of the SRm extract to the Phi6 phage caused a significant phagicidal effect at all of the tested concentrations, which was also the case for the SRa extract. When the SRm extract was incubated with the T4 phage, only 0.049%, the lowest concentration, did not significantly alter the phage counts, whereas all higher concentrations caused a stimulating effect (Fig. 2B).

Fig. 2.

For most of the tested samples, the dark colour of the concentrated extract darkened the sample. This was taken into account when analysing the results – the extracts’ optical density backgrounds were removed from the data for clear results interpretation.

The synograms revealed complex relationships between the tested extracts and bacteriophages shaping their effect on their bacterial hosts (Fig. 3). Interestingly, in most of the cases the highest tested extract concentration (25%) caused increased growth of bacteria biomass; however, after 24 h of incubation in the stationary environment the cells were no longer active (Fig. 3A-A’–F-F’). This may indicate increased cell proliferation in the initial growth stage and earlier achievement of the stationary phase due to the presence of the extracts.

Fig. 3.

Specific interactions between the tested elements were also observed. The 0.78%, 0.39% and 0.19% concentrations of the SRa extract increased P. syringae cell activity, even though the OD values of the bacteria biomass were found to decline concomitantly with the decreasing concentration of the extract. Peak cell activity was detected in a mixture of 0.19% SRa and no phage. Interestingly, at an SRa concentration of 0.049% and maximum dose of the Phi6 phage at 10⁸ PFU/mL, the highest biomass reduction was detected, while cell activity grew by 44% compared to the control (Fig. 3A-A’). In general, the SRa extract influenced E. coli DSM 5695 by decreasing biomass proliferation, lowering the OD in parallel with the decreasing concentration of the extract, while at the same time cell activity rose almost to positive control levels. Here, an interesting phenomenon was also detected, when the highest biomass reduction was discovered with the SRa concentration at 0.19% and the MS2 level at 10⁸ PFU/mL, but the cell activity level remained at 70% of that of the control (Fig. 3B-B’). The SRa extract combined with T4 phage showed a similar influence on E. coli DSM 613 cells as on E. coli DSM 5695; however, more scattered phenomena regarding cell activity were detected. To summarise, bacterial activity in some cases was extract-concentration and phage-concentration independent, and this mainly applied to peaks in cell activity. However, in one case, with the SRa concentration at 0.049% and the T4 level at 10⁴ PFU/mL, the highest biomass reduction was seen, along with a notable decrease in cell activity level to only 20% of that of the control (Fig. 3C-C’).

When SRm extract was used on P. syringae cells, the results obtained were very similar to those of adding the SRa extract: the 1.56%, 0.78%, 0.39% and 0.19% concentrations increased bacteria cell activity, with OD values of the biomass dropping simultaneously and proportionally to the decreasing concentration of the extract. As with the SRa extract, also for the SRm extract peak cell activity was detected in the mixture of 0.19% and no phage. A further parallel with the findings for the SRa extract was that at an SRm concentration of 0.049% and maximum dose of phage, the highest biomass reduction was detected, while cell activity rose by 62% compared to the control (Fig. 3D-D’). A similar trend was also maintained for the results of the SRm extract and E. coli DSM 5695, in that they were comparable with those of the SRa extract under the same conditions. However, several cases of increased cell activity compared to the control were present. An interesting phenomenon and one similar to that of the SRa experiment was also observed when the SRm extract was investigated: the highest biomass reduction was discovered at an SRm concentration of 0.19% and MS2 phage concentration of 10⁸ PFU/mL, while the detected cell activity remained at 98% of that of the control (Fig. 3E-E’). Similar tendencies were also manifested by the SRm extract and E. coli DSM 613 to those presented by the SRa extract and this bacterium – scattered phenomena in the case of cell activity were also detected. However, the most interesting phenomenon observed concerned SRm at 0.019% and T4 at 10⁸ PFU/mL, which was the highest biomass reduction along with a decrease in the cell activity level to only 16% of that of the control (Fig. 3F-F’).

To test the course and lysis ability of the phages in real time, phage performance against bacteria was tested in the presence of the extracts at different concentrations, measured by changes in the OD of the samples that were cultivated in the bioreactors with agitation at 150 rpm (Fig. 4). The extract concentrations were chosen on the basis of results gained from the synography experiment and they were the concentrations in combinations in which the described phenomena were present with maximal treatment doses as controls. The experiment time was also shortened to 16 h because the synography experiment duration of >18 h revealed inactive cells. For most of the tested samples, the sample darkened because of the dark colour of the concentrated extract. This was taken into account when analysing the data – the extracts’ optical density backgrounds were removed from the curves for clear results interpretation.

Fig. 4.

The tested extracts influenced bacterial cell and phage activity in a varying manner; however, some dependencies were noted (Fig. 4). In general, high concentrations (25%) of the tested extracts resulted in decreased phage lytic efficiency in most cases, either through a stimulating effect on bacterial cells cancelling the lytic effect of the phages or through a deactivating action on the phage particles themselves. The combination of P. syringae and the SRa extract at a concentration of 25% resulted in a decrease in OD with or without phage Phi6, flatteining the growth curves – phage addition did not decrease the OD further. In addition, the curves were still in a rising stage, therefore bacterial cells were still multiplying. However, when the SRa concentration was 0.049%, this led to the Phi6 and control lysis curves being the same (Fig. 4A). The mixture of E. coli DSM 5695 and 25% SRa extract caused a clear, gradual decrease in ODs, showing the extract’s toxicity to bacteria. On the other hand, introducing the MS2 phage into this mixture resulted in a lytic effect lasting for up to 5 h of the incubation. Afterwards, bacterial growth rebounded, climbing rapidly. The reduction in SRa concentration to 0.19% led to the bacterial lysis curve being almost identical to the control lysis curve; however, lysis began a little later and was noted after 2.5 h vs after 2 h in the controls (Fig. 4A’). When the SRa extract at 25% was added to the E. coli DSM 613 cells, this visibly reduced the OD, flattening growth curves regardless of whether the T4 phage was used or not, which was similar in the case of P. syringae. However, in this case the curves were in the gradual reduction stage. When an SRa concentration of 0.049% was used to challenge the E. coli DSM 613 cells, it led to a slightly negative cell growth curve, and a contrasting one to the growth control curve (Fig. 4A”). The application of the SRm extract at 25% to the P. syringae suspension offered almost the same result as when the SRa extract was used. Regardless of the presence of the Phi6 phage, both growth curves were clearly flattened. However, after 12 h of incubation, the OD values started to fall gradually.

When the SRm concentration of 0.049% was applied, it also led to the Phi6 phage lysis curve being the same as the control lysis curve, as was also the case when SRa was the tested extract (Fig. 4B). A 25% concentration of SRm extract addition to the E. coli DSM 5695 caused a visible increased proliferation of the bacteria, stimulating the cells. When the MS2 phage was added into this mixture, it led to a very strong lytic effect from the 2.5 h point to the 7 h point of incubation and this was stronger and longer lasting than the lysis in the lysis control. However, strong bacterial proliferation was observed later. Lowering the SRm concentration to 0.19% delayed but elongated phage lysis compared to the control curve. Lysis was noted in the control from 2 h until 3.5 h, i.e. for 1.5 h; lysis induced by the tested sample was evident from 4.5 h until 6.5 h, i.e. for 2 h (Fig. 4B’). The addition of SRm at the concentration of 25% to the E. coli DSM 613 cells did not generate any specific effect on the bacteria, and the growth curve was almost identical to the bacterial control growth curve. When bacteriophage T4 was also added, what it mediated was similar to what the MS2 phage induced in E. coli DSM 5695: a marked strong lytic effect was observed. The effect was also longer and stronger than in the control – lytic activity was detected from the 2 h point of incubation, and no bacterial growth rebound was observed. Lowering the SRm concentration to 0.19% caused the curve to be highly comparable to the control growth curve until 9 h of incubation had passed. After that time, the curve began to fall, indicating the death of the bacterial cells (Fig. 4B”).