Synergistic antimicrobial activities of aqueous extract derived from olive byproduct and their modes of action

Background Plant‑derived antimicrobials (PDAs) are considered a viable alternative to synthetic antimicrobial agents. Diverse antimicrobial mechanisms of PDAs significantly reduce the risk of developing antimicrobial resistance. Utilization of PDAs also offers economic and environmental advantages, as they can be derived from agricultural byproducts, such as olive pomace. Results In this study, a green, water‑based, ultrasound‑assisted extraction (UAE) was deployed to obtain aqueous olive pomace extract (OPE) from dry olive pomace. Total phenolic content, extraction yield, chemical compositions, and antimicrobial activities of OPE were evaluated. In addition, the potential synergistic interaction between the phe‑ nolic components in OPE and the antimicrobial mechanisms underlying the synergistic interaction were character‑ ized. The results show that ca. 25 mg GAE/g of extraction yields were achieved by the UAE of dry olive pomace. Based on the high‑performance liquid chromatography (HPLC) analysis, diverse phenolic compounds such as gallic acid (GA), hydroxytyrosol (HT), and 4‑hydroxyphenylacetic acid (4‑HPA) were identified in OPE. OPE exhibited strong antimicrobial activities, and 0.2 mg GAE/mL of OPE achieved > 5 log reductions of Escherichia coli O157:H7 and Listeria innocua cells within 30 min of treatment. A 3D isobologram analysis demonstrated that OPE exhibited strong syn‑ ergistic antimicrobial activities, compared to those of individual phenolic components (GA, HT, or 4‑HPA), showing interaction index ( γ ) of 0.092 and 0.014 against E. coli O157:H7 and L. innocua , respectively ( γ < 1: synergistic activity). Antimicrobial mechanism analyses revealed that phenolic components in OPE exerted strong synergistic activities through diverse modes of action, and increased levels of oxidative stress, membrane damage, and decreased levels of metabolic activities were observed in the OPE‑treated bacterial cells. Conclusions These findings demonstrate an approach for valorizing agricultural byproducts to develop plant byproduct‑based antimicrobials with strong synergistic activities. Multiple modes of action of this byproduct extract may enable the control of diverse microbes in food and agriculture systems.


Introduction
Natural antimicrobials such as plant-derived antimicrobials (PDAs) have received increasing attention as a viable alternative to conventional antimicrobials [1][2][3].Compared to synthetic antimicrobial agents, PDAs are known to possess a significantly reduced risk of adverse health and environmental impacts [3,4].Diverse antimicrobial components in PDAs can exert antimicrobial activities through different modes of action, markedly lowering the chance of bacterial pathogens to adapt and develop resistance [4,5].In addition, the use of PDAs can be economically desirable since diverse antimicrobial phytochemicals can be obtained from agricultural byproducts [6,7].For example, olive byproducts such as olive pomace, olive mill wastewater, (OMWW), and olive leaves are considered a rich source of antimicrobial phenolics [8][9][10][11].Despite the merits of PDAs, the extraction of PDAs from plant by-products using conventional solvent-based methods, such as maceration, percolation, and Soxhlet extraction, often involves extensive use of organic solvents that can impact the environment and pose potential health risks to workers involved in the extraction process [12,13].Therefore, the development of an effective extraction method to obtain diverse PDAs from spent plant sources using a simple and environmentally friendly process is needed.
Ultrasound-assisted extraction (UAE) is considered one of the simple and cost-effective extraction methods [14,15].The cavitation bubbles and acoustic waves generated during UAE may cause disruption and fragmentation of the plant cell components and facilitate the diffusion of the phytochemicals from the plant cells to the extractant [14,16,17].In addition, the hydrophobic surfaces of the cavitation bubbles can lead to shorter extraction time and higher extraction yields of diverse classes of phytochemicals and, in turn, reduce the use of organic solvents for extraction [15,18].For example, it has been reported that water-based UAE can improve the extraction yields of polyphenols and carotenoids from spent plant sources by up to 6-25% [18].Another previous study reported that UAE of grapefruit byproducts resulted in ca.50% and 66% higher total phenolic content (TPC) and total antioxidant activity (TAA) compared to the conventional solid-liquid extraction [19].The authors also reported that similar TPC and TAA could be achieved by ethanolfree, water-based UAE compared to ethanol-based UAE.Recently, significant efforts have been made to extract a variety of phenolic compounds from different olive byproducts using UAE.Studies have shown that waterbased UAE of olive byproducts can extract higher levels of water-soluble phenolic compounds, such as hydroxytyrosol and diverse phenolic acids, compared to organic solvent-based UAE [20,21].These suggest that waterbased UAE can facilitate the extraction of diverse compounds from plant by-products.
Antimicrobials derived from plant byproducts can be used either as crude extracts or pure compounds after extensive separation of the crude extracts [22].However, due to the compositional diversity, crude extracts can exert antimicrobial activities through diverse modes of action [22,23].It has been suggested that diverse antimicrobial components in the crude extract can act on different target sites of a microorganism including cell wall, cell membrane, metabolic enzymes, protein synthesis, and genetic systems, and often induce synergistic inactivation of the pathogens [2,4,5,24].Furthermore, compared to pure compounds, the preparation of crude extracts is considered more energy-efficient since it does not require time-and energy-intensive separation and purification steps [12,25].Previously, we evaluated the antimicrobial potential of different C18 fractions of OPE obtained using water-based UAE [26].Based on our observation, stronger antimicrobial activities were exhibited by the water-eluted C18 fractions compared to those of methanol-eluted C18 fractions.Although the results seemed promising, a considerable knowledge gap still exists in understanding the antimicrobial activities of the crude aqueous extract due to its complex chemistry.Currently, there is limited understanding of the modes of antimicrobial activity of diverse active compounds in extracts.Furthermore, there is a lack of evaluation of the potential antimicrobial synergisms between the active components of the extract and an understanding of the antimicrobial mechanisms underlying the synergistic activities between the active components [22].
The objective of the present study was to explore the potential antimicrobial synergisms and the mechanisms underlying the synergistic activities between the phenolic components of the aqueous extract obtained by the UAE of olive pomace.To achieve this goal, we evaluated the UAE in extracting phenolic compounds from dry olive pomace obtained in California and characterized the chemical composition of the aqueous OPE.The antimicrobial activity of the OPE was evaluated using Escherichia coli O157:H7 and Listeria innocua as model Gram-negative and Gram-positive bacteria, respectively.The synergistic antimicrobial potential of OPE was evaluated and compared to that of the individual phenolic components of OPEs.Lastly, the antimicrobial mechanisms underlying the synergistic activities between the phenolic components of OPE were explored.The findings presented in this study will provide useful insights into valorizing the olive by-products as a sustainable source of an antimicrobial agent using the green extraction method.The results of this study will also highlight the antimicrobial synergistic activities of the extract compared to pure components of the extract.

Preparation of olive pomace extract (OPE) using ultrasound-assisted extraction (UAE)
Fresh Arbequina olive pomace was collected from California Olive Ranch (Artois, CA, USA) in 2019.The germplasm accession for Arbequina olives is available from the USDA National Germplasm Repository (Davis, CA, USA) (https:// www.ars.usda.gov/ pacifi c-west-area/ davis-ca/ natl-clonal-germp lasm-rep-tree-fruit-nut-cropsgrapes/).Dried olive pomace was obtained after steamblanching, pitting, drum-drying, and milling the collected olive pomace samples following the procedures described by Zhao et al. [21].Aqueous extracts from the dried olive pomace were obtained using an ultrasound-assisted extraction (UAE).Briefly, 1, 2, 4, or 8 g of dried olive pomace was immersed in 20 mL of Milli-Q water, and the mixture was bath-sonicated for 30 min (Branson 2510, 100 W, 42 kHz).After bath sonication, the mixture of olive pomace and the extract was centrifuged at 5000×g for 10 min, and the supernatant was recovered.The collected crude extract of olive pomace was filtered through a polyethersulfone (PES) filter (pore size: 0.45 μm) to remove the residual pomace particles and stored at 4 °C until further use.

Total phenolic content of OPE
The total phenolic content of the OPE was determined using the Folin-Ciocalteu assay as described by Singleton and Rossi [27] and Saucier and Waterhouse [28] with slight modifications.Briefly, OPE samples were tenfold diluted in Milli-Q water, and 20 µL of the diluted OPE sample was mixed with 1.58 mL of Milli-Q water.A blank sample was prepared by adding 20 µL of Milli-Q water instead of OPE.Then, 100 µL of Folin-Ciocalteu reagent and 300 µL of 20% sodium carbonate solution were added to the mixture, and the mixture was incubated at 40 °C in the dark for 45 min.After incubation, the absorbance was measured at λ max = 765 nm using a UV-Vis spectrophotometer (GENESYS 10S Series, Thermo Scientific).Gallic acid solution (0.625-10 mg/L) was used as a standard solution for a calibration curve, and the total phenolic content was expressed in mg gallic acid equivalents per mL extract (mg GAE/mL).The extraction yield of UAE in obtaining phenolic compounds from dry olive pomace was calculated as follows:

High-performance liquid chromatography (HPLC) analysis of OPE
The concentration of the selected phenolic compounds in the extract was measured using an Agilent 1290 high-performance-liquid-chromatography (HPLC) system (Santa Clara, USA) with an Agilent 1290 diodearray detector (DAD).An analytical C18 column (Eclipse Plus, 4.6 mm × 250 mm, 5 μm, Agilent Technologies) was used for separation.The injection volume of the extract was 20 µL and the flow rate of elution buffers was 1.0 mL/min.Elution buffers include mobile phase A (3% acetic acid aqueous solution) and mobile phase B (50% methanol and 50% acetonitrile) were used in different ratios as follows: 5% of B at 0-25 min; 30% of B at 25-35 min; 35% of B at 35-40 min; 40% of B at 40-50 min; 70% of B at 50-55 min; 100% of B at 55-60 min.The DAD was set to absorbance wavelengths at 280 nm for the detection of hydroxytyrosol, tyrosol, 4-HPA, vanillic acid, vanillin, o-coumaric acid, oleuropein, pinoresinol, cinnamic acid.Similarly, DAD was set at 320 nm for caffeic acid, p-coumaric acid, ferulic acid, apigenin-7-glucoside, apigenin, and verbascoside and at 365 nm for rutin, luteolin-7-glucoside, and luteolin, respectively.The retention times for each of the selected phenolic components in the analyzed pool were compared to the standards at the concentration of 50 mg/L.

Bacterial strains
A Shiga toxin-negative strain of Escherichia coli O157:H7 (ATCC 700728, Manassas, VA, USA) and Listeria innocua (TVS451, Manassas, VA, USA) were selected as model bacterial strains for Gram-negative and Grampositive foodborne pathogens, respectively.E. coli O157:H7 and L. innocua strains were kindly provided by Dr. Linda Harris (University of California, Davis) and Dr. Trevor Suslow (University of California, Davis), respectively.Both selected strains were modified with a rifampicin-resistant plasmid.To activate the bacterial strains, the cryopreserved stocks were transferred to TSB and incubated at 37 °C with constant shaking at 200 rpm overnight.The overnight cultures (10 µL) were transferred to TSB (10 mL) and cultured three times at 37 °C with 24-h intervals.Each bacterial culture was then streaked on TSA, and the TSA plates were incubated at 37 °C for 48 h.The agar plate with E. coli O157:H7 or L. innocua colonies was then stored at 4 °C prior to use.Before each experiment, a single bacterial colony on the TSA plate was transferred into TSB and incubated at 37 °C with constant shaking at 200 rpm overnight to achieve the stationary phase cultures (ca.9.3 log CFU/ mL).The overnight E. coli O157:H7 or L. innocua culture was harvested, washed twice with PBS, and resuspended in PBS to prepare a bacterial inoculum.

Antimicrobial activities of OPE
The antimicrobial activities of OPE were evaluated at different concentrations.Briefly, 10 µL of E. coli O157:H7 or L. innocua suspension (ca.8.3 log CFU/mL) was inoculated in 1 mL of DW containing 0.1, 0.2, and 0.5 mg GAE/mL of OPE and incubated at room temperature for 0, 30, and 60 min, respectively.After incubation, the bacterial populations in different concentrations of OPE were enumerated using a plate count assay.The bacterial suspensions were serially diluted in PBS, surface-plated on TSA, and incubated at 37 °C for 48 h before counting the colonies.The theoretical detection limit of the plate count assay was 1.0 log CFU/mL.The inhibitory concentration of OPE determined based on this assay was used for the synergistic antimicrobial analysis in the following section.

Synergistic antimicrobial activities of phenolic components in OPE
The antimicrobial activities of OPE and its phenolic components were evaluated against E. coli O157:H7 and L. innocua, respectively.In this assay, the extended treatment time, for up to 60 h, was selected to assess the antimicrobial activities of pure phenolic components of OPE.Hydroxytyrosol (HT) and 4-hydroxyphenylacetic acid (4-HPA) were selected as representative, water-soluble, phenolic components of OPE based on the concentration in the OPE (Table 1).Gallic acid (GA) was also included as a representative water-soluble phenolic component based on its synergistic antimicrobial potential with HT and/or 4-HPA observed in the previous study [26].The concentration of OPE and each phenolic solution was adjusted to 0.1 mg GAE/ mL, which was considered an inhibitory concentration of OPE based on the previous results (Fig. 2).A 10 µL of E. coli O157:H7 or L. innocua suspension (ca.8.3 log CFU/mL) was inoculated in 1 mL of OPE, GA, HT, or 4-HPA solution to achieve ca.6.3 log CFU/mL as the initial inoculum level and incubated at room temperature for up to 60 h.The bacterial populations in OPE or the solution of individual phenolic compounds were monitored using a plate count assay as described above.
The pH of 0.1 mg GAE/mL of OPE, GA, HT, and 4-HPA was 5.28 ± 0.02, 3.89 ± 0.15, 5.00 ± 0.02, and 3.80 ± 0.04, respectively (data not shown).A 3D isobologram analysis was performed using OPE and its phenolic components (GA, HT, and 4-HPA) to evaluate the synergistic antimicrobial activities of the phenolic compounds in OPE.Isobologram is a graphical analysis widely used in the field of pharmacology and toxicology that enables a quantitative measurement of the interactions between two or more treatments, such as a combination of drugs, on a biological system [29,30].In this study, the conventional isobologram approach was adapted based on the treatment time required to achieve a 1-log reduction of target pathogens as described by Huu Nguyen et al. [31] with slight modifications.Briefly, a 1-log isobole plane was constructed by plotting the treatment time required for each phenolic component of OPE (GA, HT, or 4-HPA) to achieve a 1-log reduction of the tested bacteria on the x, y, and z-axis, respectively.For example, 13.5, 58.0, and 95.0 h were plotted on the x, y, and z-axis, respectively, to construct a 1-log isobole plane for GA, HT, and 4-HPA against E. coli O157:H7.The treatment time required for OPE to achieve 1-log reduction of the tested bacteria was plotted as a single point (red dot) on the same 3D plot with the 1-log isobole plane.For example, a single point (0.9 h, 0.9 h, 0.9 h) for OPE treatment was plotted on the 3D plot with the 1-log isobole plane.It is considered a synergistic, additive, and antagonistic effect if the plotted point lies below, on, and above the 1-log isobole plane, respectively.The interaction index (γ) of the phenolic components in OPE was also calculated as follows [32]: where a, b, and c are the treatment time required for GA, HT, and 4-HPA to achieve 1-log reduction of the tested bacteria, and x is the treatment time required for OPE to achieve 1-log reduction.It is considered a synergistic, additive, and antagonistic effect if γ < 1, γ = 1, and γ > 1, respectively.For example, when 13.5, 58.0, 95.0, and 0.9 h were required for GA, HT, 4-HPA, and OPE to achieve 1-log reduction of E. coli O157:H7, respectively, the interaction index is calculated as 0.092.

Antimicrobial mechanisms of OPE Oxidative stress
The potential of OPE or its phenolic components (GA, HT, and 4-HPA) to generate oxidative stress on bacterial cells was assessed by measuring the intracellular thiol contents of the bacterial cells after the treatment.Intracellular thiol, such as glutathione, is a comprehensive biomarker that is indicative of a redox balance in a cell [35].A decrease in free thiol contents within the cell can indicate that oxidative stress has been generated in the cells.The intracellular thiol contents of E. coli O157:H7 cells after the treatment were extracted using a bead-beating method.In this assay, a relatively higher number of E. coli O157:H7 cells (ca.9.3 log CFU/mL) were treated to extract a sufficient level of thiol content for detection using a fluorescence assay [36].Briefly, E. coli O157:H7 cells were treated with 0.1 mg GAE/mL of OPE, GA, HT, or 4-HPA for 30 min and recovered after centrifuging at 13,000×g for 2 min.E. coli O157:H7 cells treated with DW were included as a control.The collected bacterial pellets were resuspended in an ice-cold PBS, transferred into a Lysing Matrix B containing 0.1 mm diameter silica beads (MP Biomedials, Irvine, CA, USA), and homogenized three times using a FastPrep ® -24 instrument (MP Biomedials, Irvine, CA, USA) at a speed of 6.5 m/s for 60 s with 5 min of ice incubation between the cycles.The supernatant was recovered after centrifuging at 13,000×g and 4 °C for 15 min, and the total thiol content in the bacterial lysate was quantified using the Measure-iT ™ Thiol Assay Kit (Invitrogen, Waltham, MA, USA) following the manufacturer's instruction.

Membrane damage
The effect of OPE and the phenolic components (GA, HT, and 4-HPA) on the membrane integrity of E. coli O157:H7 cells was evaluated using a selective medium plating technique (SMPT) [33,34].SMPT involves the use of TSA supplemented with 3% (w/v) NaCl (TSAN) as a selective medium to discriminate the membrane-intact and membrane-compromised bacterial cells.Membranecompromised but culturable cells are not expected to grow and form colonies on TSAN but grow and form colonies on TSA without NaCl.E. coli O157:H7 culture (ca.6.3 log CFU/mL) was incubated in OPE, GA, HT, or 4-HPA (0.1 mg GAE/mL, respectively) for 30 min to induce sublethal damage on the bacterial cell membrane.
After the treatment, the bacterial cells in each treatment solution were serially diluted in PBS, surface-plated on TSA with and without 3% NaCl, and incubated at 37 °C for 48 h before counting the colonies.The number of membrane-damaged bacterial cells was determined by subtracting the number of E. coli O157:H7 colonies formed on TSAN from that of E. coli O157:H7 colonies formed on TSA without NaCl.The theoretical detection limit of the plate count assay was 1.0 log CFU/mL.

Metabolic interference
The effect of OPE and other phenolic components (GA, HT, and 4-HPA) on metabolic activities of E. coli O157:H7 cells was evaluated using a resazurin-based metabolic assay as described by de Oliveira et al. [35] with minor modifications.Briefly, OPE and the phenolic solutions (0.5 mg GAE/mL) were prepared in TSB containing 50 µM resazurin and inoculated with bacterial cells (final population: ca.5.3 log CFU/mL).Since this assay was conducted in the presence of growth media (TSB), a higher concentration of OPE and its phenolic components was applied as described above.The inoculated samples were then dispensed into a 96-well microtiter plate and incubated at 37 °C for 12 h.During the incubation, the fluorescence intensity of each sample was monitored using a fluorescence plate reader

Statistical analysis
Statistical analysis was performed using the Graph-Pad Prism software V.9.5.1 (Graphpad Software, Inc., La Jolla, CA, USA).All experiments were performed at least in triplicates.The significant differences between treatments were determined through one-way analysis of variance (ANOVA) followed by Tukey's pairwise comparisons with a 95% confidence interval (p < 0.05) unless otherwise stated in the corresponding section.

Total phenolic contents of OPE
The extraction yield of phenolic compounds from dry olive pomace using UAE was measured based on the total phenolic contents of OPE at different sample-tosolvent ratios (1-8 g/20 mL) (Fig. 1).The results show that the UAE process achieved ca.1.22 ± 0.06, 2.47 ± 0.06, 4.20 ± 0.05, and 5.66 ± 0.05 mg GAE/mL of total phenolic content in OPE (Fig. 1a), which corresponds to an extraction yield of ca.24.43 ± 1.16, 24.69 ± 0.58, 20.98 ± 0.27, and 14.14 ± 0.12 mg GAE/g with a sample-to-solvent ratio of 1, 2, 4, and 8 g dry olive pomace per 20 mL DW, respectively (Fig. 1b).The extraction efficiency of waterbased UAE in recovering phenolic compounds from olive pomace has been evaluated by Goldsmith et al. [15].The authors reported that ca.35% higher extraction yields were achieved by using UAE (75 min; 250 W; 2 g/100 mL), compared to the control extraction method without ultrasound.The extraction yield in this study was 19.71 ± 1.41 mg GAE/g, and this increased yield was attributed to the enhanced antioxidant properties of OPE.It is also noteworthy that much lower power (100 W), less amount of water (20 mL), and shorter extraction time (30 min) were used in our study to achieve a similar level of extraction yields reported by Goldsmith et al. [15].These results suggest that water-based UAE can recover a significant amount of phenolics from plant by-products.The individual phenolic compounds in OPE were identified using HPLC-DAD.Figure 1c shows an example of chromatogram of OPE.In total, 24 different phenolic compounds were identified using the phenolic standards, and the concentration of each phenolic compound was determined by calculating the peak area (Table 1).Water-soluble phenolic compounds such as gallic acid (GA) (0.224 ± 0.007 mg/g sample), hydroxytyrosol (HT) (1.992 ± 0.043 mg/g sample), tyrosol (0.463 ± 0.009 mg/g sample), and 4-hydroxyphenylacetic acid (4-HPA) (1.711 ± 0.056 mg/g sample) were identified, and less polar, water-soluble molecules such as verbascoside (0.835 ± 0.003 mg/g sample), rutin (0.771 ± 0.017 mg/g sample), and oleuropein (0.816 ± 0.011 mg/g sample) were also found in the OPE.Similar results were demonstrated in a recent study conducted by Zhao et al. [21].The authors reported that nineteen phenolic compounds and five phenolic sugar derivatives were identified from the aqueous extract of olive pomace.These results suggest that UAE enables rapid extraction of various phenolic compounds including water-soluble and sparingly water-soluble compounds present in the olive pomace.

Antimicrobial activities of OPE
Figure 2 shows the populations of E. coli O157:H7 and L. innocua cells incubated in different concentrations of OPE (0, 0.1, 0.2, and 0.5 mg GAE/mL) for 0, 30, and 60 min.The number of E. coli O157:H7 cells (ca.6.38 log CFU/mL) treated with 0.2 mg GAE/mL showed > 5 log reductions within 30 min and decreased below the detection limit (1.0 log CFU/mL) within 60 min (Fig. 2a).Similarly, populations of L. innocua cells (ca.6.47 log CFU/ mL) incubated at 0.2 mg GAE/mL of OPE showed > 5 log reduction within 30 min and decreased below the detection limit within 60 min (Fig. 2b).A 0.1 mg GAE/mL of OPE did not induce rapid inactivation of tested bacterial strains in 30 min, however, exhibited a significant inhibitory activity after 60 min of treatment.These results suggest that aqueous extract obtained from olive pomace possesses strong antimicrobial activities, and 0.1 mg GAE/mL was selected as an inhibitory concentration of OPE for the following synergistic antimicrobial analysis.
Antimicrobial properties of the crude extracts obtained from diverse types of olive by-products, such as olive mill wastewater (OMWW) [37][38][39] and olive leaves [40][41][42] have been widely reported.However, only a limited number of studies have investigated the potential antimicrobial activities of the aqueous extracts obtained from olive pomace.A recent study conducted by Nunes et al. [43] evaluated the antimicrobial properties of the crude paste obtained after high-pressure extraction (200-300 bar) followed by lyophilization of the olive pomace.The obtained crude olive pomace paste exhibited measurable antibacterial activities, and the minimal inhibitory concentrations (MICs) were determined as ca.2.18-4.79 and 1.09-3.83mg GAE/mL against E. coli and Staphylococcus aureus, respectively.The authors reported that, among olive pomace pastes obtained from four different olive cultivars, the olive paste with the highest HT content showed the lowest MIC values while the one with the lowest HT content exerted the highest MIC values against both tested bacterial strains.More recently, Zhao et al. (2023) [26]

evaluated the antimicrobial activities of four different fractions of aqueous OPE obtained with a C18 column using different eluents (eluted syrup [ES], acidified water [AW], 35% methanol [ME35], and 70% methanol [ME70] fractions).
At the same GAE level, more polar C18 fractions, such as ES and AW fractions, showed stronger antimicrobial activities compared to the less polar fractions, such as ME35 and ME70 fractions.This suggests that the chemical properties of the solvent might play a crucial role in determining the chemical composition as well as the antimicrobial activities of the extracts.For example, the composition analysis of C18 fractions revealed that high levels of HT, GA, and tyrosol-glucoside were identified from ES fractions, while high levels of tyrosol-glucoside, tyrosol, oleuropein, and verbascoside were found in 70ME fractions.In this study, Fig. 2 Antimicrobial activities of OPE against (a) Escherichia coli O157:H7 and (b) Listeria innocua.E. coli O157:H7 and L. innocua cells were incubated at different concentrations of OPE (0, 0.1, 0.2, and 0.5 mg GAE/mL) for up to 60 min, and the bacterial cells incubated in DW were included as a control.The results represent the mean values and their standard deviations (n = 3).The detection limit of direct plating (dashed line) was 1.0 log CFU/mL, and the asterisks indicate that the number of viable bacteria was below the detection limit GA, HT, and 4-HPA were selected as representative, water-soluble phenolic components of OPE for further antimicrobial analysis.

Synergistic antimicrobial activities of the phenolic components in OPE
Synergistic antimicrobial activities of the phenolic compounds in OPE were evaluated by assessing the inactivation kinetics of E. coli O157:H7 and L. innocua in OPE and its phenolic components (GA, HT, or 4-HPA), respectively (Fig. 3).Notably, OPE showed much stronger antimicrobial activities than its phenolic components at the equivalent GAE levels against both tested bacterial strains.The populations of E. coli O157:H7 (ca.6.5 log CFU/mL) inoculated in 0.1 mg GAE/mL of OPE significantly (p < 0.05) decreased to ca. 5.7 log CFU/mL within 1 h and decreased below the detection limit (1.0 log CFU/ mL) within 4 h (Fig. 3a).In addition, populations of L. innocua cells (ca.6.5 log CFU/mL) in 0.1 mg GAE/mL of OPE showed > 5 log reduction within 1 h and decreased below the detection limit within 1.5 h (Fig. 3c).In contrast, 3.5, 5.6, and 5.8 log CFU/mL of E. coli O157:H7 cells and 1.4, 3.0, and 3.6 log CFU/mL of L. innocua cells remained culturable in 0.1 mg GAE/mL of GA, HT, and 4-HPA solutions even after 60 h of incubation, respectively (Fig. 3a, c).
The synergistic antimicrobial activities of the phenolic components in OPE were further evaluated by performing an isobologram analysis.Figure 3b, d illustrate the 3D isobologram of OPE and its phenolic components (GA, HT, and 4-HPA) against E. coli O157:H7 (Fig. 3b) and L. innocua (Fig. 3d), respectively.A 1-log isobole plane (shaded area within dashed boundaries) was constructed on the isobologram by plotting the time required for 0.1 mg GAE mg/mL of GA, HT, or 4-HPA to achieve 1-log reduction of the tested bacterial strains.The time required for OPE (0.1 mg GAE/mL) to achieve a 1-log reduction of the tested strains was also plotted as a single Fig. 3 Synergistic antimicrobial activities of phenolic components in OPE: (a) E. coli O157:H7 and (c) L. innocua cells were incubated in 0.1 mg GAE/mL of OPE, GA, HT, or 4-HPA for up to 60 h.1-log isobole plane (shaded area within dashed boundaries) was constructed based on the time required for GA, HT, or 4-HPA to achieve 1-log reduction of (b) E. coli O157:H7 and (d) L. innocua cells, and the time required for OPE to achieve 1-log reduction of the tested bacteria was plotted as a single point (red dot).The results represent the mean values and their standard deviations (n = 3).The detection limit of direct plating (dashed line) was 1.0 log CFU/mL point (red dot) on the same isobologram.The results in Fig. 3b demonstrate that OPE required 15.6, 64.4, or 105.6-fold shorter treatment time to achieve a 1-log reduction of E. coli O157:H7 than the time required for GA, HT, or 4-HPA to achieve the same log reduction of E. coli O157:H7.The phenolic compounds in OPE even showed stronger antimicrobial synergisms against L. innocua.Figure 3d illustrates that OPE required 160, 226.7, or 266.7-fold shorter treatment time to achieve a 1-log reduction of L. innocua compared to the time required for GA, HT, or 4-HPA to achieve the same log reduction of L. innocua.The strong antimicrobial synergism was also confirmed by calculating the interaction index (γ) of the phenolic compounds in OPE.The interaction index of the phenolic compounds in OPE was determined as ca.0.092 ± 0.008 and 0.014 ± 0.003 against E. coli O157:H7 and L. innocua, respectively (γ < 1: synergistic activity).
Taken together, these results demonstrate that the component in OPE exhibited strong synergistic activities compared to individual phenolic compounds at the same concentration level, and this synergistic antimicrobial activity significantly enhanced the rate of inactivation of the tested bacteria.It is also noteworthy that OPE exhibited stronger antimicrobial activities despite having a higher pH level (5.28 ± 0.02) than those of GA (3.89 ± 0.15), HT (5.00 ± 0.02), or 4-HPA (3.80 ± 0.04) solution at 0.1 mg GAE/mL.This is somewhat contradictory to the well-known fact that organic acids, including phenolic acids, tend to exert stronger antimicrobial activities at pH levels lower than their pK a values (e.g.pK a of GA ≈ 4.21) [44][45][46].This indicates that the components in OPE exerted strong synergistic activities possibly through diverse modes of action when they were applied simultaneously as a crude extract.Therefore, potential mechanisms underlying the synergistic antimicrobial activities of OPE components were further investigated in the following section.
A limited number of studies have been conducted to evaluate the potential antimicrobial synergisms between the components of phenolic extracts obtained from olive byproducts.A study conducted by Tafesh et al. [47] demonstrated the synergistic antimicrobial activities between HT (isolated from the ethanolic extract of OMWW) and some other pure compounds including GA and ascorbic acid.Based on their results, the combination of GA/HT and ascorbic acid/ HT showed strong synergistic antimicrobial activities against E. coli, Streptococcus pyogenes, Klebsiella pneumoniae, and S. aureus, respectively.The authors reported that GA and ascorbic acid showed strong synergistic antimicrobial activities only when combined with HT, suggesting that HT might be the major bioactive compound that contributes to the antimicrobial synergisms.However, to the best of our knowledge, this is the first study reporting the strong antimicrobial synergisms exerted between the phenolic components of the crude OPE and its quantitative evaluation using isobologram analysis.

Antimicrobial mechanisms Oxidative stress
The potential of OPE and its phenolic components (GA, HT, or 4-HPA) to generate oxidative stress in bacterial cells was evaluated by measuring the intracellular thiol content of E. coli O157:H7 cells after the treatment.Intracellular thiol content is one of the widely used biomarkers that are indicative of the oxidative stress induced in bacterial cells [48]. Figure 4a shows changes in the intracellular thiol contents of E. coli O157:H7 cells after 30 min of treatments with 0.1 mg GAE/mL of OPE or its phenolic components.E. coli O157:H7 cells treated with OPE showed significantly (p < 0.05) lower thiol contents compared to those treated with DW (control) and other phenolic components, and only ca.0.60 µM of thiol contents were found in the cells after the OPE treatment.Interestingly, E. coli O157:H7 cells treated with phenolic components of OPE also showed significantly (p < 0.05) lower intracellular thiol contents compared to control, and ca.0.92, 1.17, and 1.04 µM of thiol contents remained after the treatment with GA, HT, and 4-HPA, respectively.No detectable amount of thiol content was found in OPE, GA, HT, or 4-HPA (< 0.05 µM).These results indicate that, although each phenolic component exerted oxidative stress on the bacterial cells, a significantly higher degree of oxidative stress has been induced by OPE.The pro-oxidant activities induced by simple organic acids have been reported in the previous study [49].de Oliveira et al. [35] reported that 1 mM of GA or 5 mM of LA induced a significant reduction in intracellular thiol contents of E. coli O157:H7 cells, however, no further reduction of thiol contents was observed by the combined treatment of GA and LA.This comparison suggests that the combined effect of bioactives in OPE can induce a synergistic reduction in intracellular thiol levels compared to the individual phenolic compounds.

Membrane damage
The potential effects of OPE and its phenolic components on the membrane integrity of bacterial cells were evaluated against E. coli O157:H7 using SMPT.Figure 4b shows the number of membrane-compromised E. coli O157:H7 cells after 30 min of treatment with 0.1 mg GAE/mL of OPE or its phenolic components (GA, HT, or 4-HPA).OPE induced a significantly (p < 0.05) higher degree of membrane damage to the cells compared to those treated with DW (control), and ca.93.5% of E. coli O157:H7 cells were assumed to be membrane-compromised after the OPE treatment.In contrast, those treated with the single phenolic component did not show any significant (p > 0.05) difference in the number of membrane-compromised cells compared to the control.These results indicate that the components of OPE may have synergistically induced the membrane damage of E. coli O157:H7 cells.The potential of phenolic compounds to induce membrane damage in bacterial cells has been documented by others.For example, Campos et al. [49] evaluated the effect of hydroxycinnamic acids (p-coumaric acids, caffeic acids, and ferulic acids) and hydroxybenzoic acids (p-hydroxybenzoic, protocatechuic, gallic, vanillic, and syringic acids) on the cell membrane permeability of lactic acid bacteria.The authors reported that all the tested phenolic acids induced ion leakages (potassium and phosphate) and proton influx in a concentrationdependent manner.However, in the previous study conducted by Oliveira et al. [34], the authors reported that no significant (p > 0.05) membrane damage was induced on E. coli O157:H7 cells with 10 mM GA (ca.1.7 mg/mL) and 1 mM ferulic acid (ca.0.19 mg/mL).It is worth noting that, in the present study, significant membrane damage was induced when 0.1 mg GAE/mL of phenolic acids were applied simultaneously as in OPE, further highlighting the antimicrobial synergism in the OPE.

Inhibition of metabolic activities
The inhibitory effects of OPE and its phenolic components (GA, HT, or 4-HPA) on the metabolic activities of E. coli O157:H7 were evaluated using a resazurinbased metabolic assay.The metabolic activities of E. coli O157:H7 cells were monitored during incubation in TSB supplemented with 0.5 mg GAE/mL of OPE or its phenolic components for up to 12 h.Figure 4c shows the fluorescence intensity of resorufin molecules, a reduced form of resazurin, formed by the metabolic activities of E. coli O157:H7 cells.The fluorescence intensity of all samples reached its peak after 3-5 h of incubation, however, the height of these peaks was significantly affected by the treatment.Based on our observation, the metabolic activities of E. coli O157:H7 cells were strongly suppressed in TSB supplemented with OPE and showed ca.3.8-fold lower peak height compared to those incubated in TSB (control).Compared to the single phenolic components, E. coli O157:H7 cells treated in OPE showed ca.1.7, 3.2, and 3.2-fold lower peak heights than those treated in GA, HT, and 4-HPA, respectively.This result indicates that the phenolic components in OPE can synergistically induce irreversible damage to the metabolic activities of E. coli O157:H7 cells.Previously, the effect of 1 mM GA (ca.0.17 mg/mL), 5 mM lactic acid (ca.0.45 mg/mL), and their combined treatment (1 mM GA + 5 mM lactic acid) on the metabolic activities of E. coli O157:H7 have been evaluated by de Oliveira et al. [35].The authors reported that about 0.4-0.9h of delay was observed for E. coli O157:H7 cells to achieve maximum fluorescence peak when they were treated with GA, lactic acid, or GA + lactic acid compared to the control cells.The delay in time indicates that recoverable damages were induced by GA and lactic acid on the metabolic activities of the bacterial cells.However, no significant synergistic inhibition was exhibited by the combined treatment of GA (1 mM) and lactic acid (5 mM).These results suggest that the components of OPE exerted strong synergistic activities and induced irreversible damage to the metabolic activities of E. coli O157:H7 cells.
Taken together, the results of this study suggest that OPE can exert strong antibacterial activities by synergistically interrupting the redox balance, membrane integrity, and metabolic activities of the bacterial cells, ultimately, leading to cell death.These findings provide useful information in utilizing PDAs, in the form of aqueous, crude extract, from agricultural byproducts using UAE, a green extraction method.Furthermore, our study provides valuable insights in developing a quantitative method to evaluate the potential synergisms between the antimicrobial components of PDAs as well as in elucidating the antimicrobial mechanisms underlying the synergistic activities.
In future studies, the potential differences in the chemical composition and antimicrobial activities of the extracts obtained from different olive cultivars could be further investigated.In addition, untargeted spectroscopy and metabolomics approaches can be used to further characterize the chemical compositions of OPE.These compositional details may improve our understanding of the mechanisms for the synergistic antimicrobial activity.In this study, OPE synergistically inhibited the growth of the bacterial pathogens through diverse modes of action.However, the potential antimicrobial activities of OPE against a broader class of microorganisms, including spoilage bacteria and plant pathogenic fungi, may be evaluated for more extensive applications of OPE.

Conclusions
In this study, a plant byproduct-based antimicrobial was developed using olive pomace, and its synergistic antimicrobial activities were evaluated.Aqueous olive pomace extract (OPE) was obtained using water-based, ultrasound-assisted extraction (UAE).The extraction yield of phenolic compounds with UAE was ca. 25 mg GAE/g.Based on the HPLC-DAD analysis, 24 different phenolic compounds were identified in OPE.The OPE (0.2 mg GAE/mL) showed strong antimicrobial activities both against Gram-negative and -positive bacteria and achieved more than 5 log reductions of E. coli O157:H7 and L. innocua cells within 30 min.Gallic acid (GA), hydroxytyrosol (HT), and 4-hydroxyphenylacetic acid (4-HPA) were selected as three representative, watersoluble phenolic components of OPE for the evaluation of the synergistic antimicrobial activities.Components in OPE showed strong synergistic activities (interaction index [γ] < 1) against E. coli O157:H7 (γ: 0.092) and L. innocua (γ: 0.014), compared to the equivalent GAE level of individual phenolic components (GA, HT, or 4-HPA).Such synergistic activity was achieved through diverse modes of action.Bacterial cells treated with OPE exhibited a higher degree of oxidative stress, membrane damage, and inhibition of metabolic activities compared to those treated with a single phenolic component.The results of this study suggest a strong potential of utilizing plant byproduct-derived antimicrobials, such as OPE, as a promising substitute for conventional antimicrobials.This could lead to the application of OPE in the food industry as a natural preservative or in the agriculture sector for developing new biopesticides, thereby enhancing sustainability in the food and agriculture system.

=
Weight of the extracted phenolic compounds (mg GAE)Initial weight of dry olive pomace (g) .

Fig. 4
Fig.4 Effects of OPE and its phenolic components (GA, HT, and 4-HPA) on (a) redox balance, (b) membrane integrity, and (c) metabolic activities of E. coli O157:H7 cells.Potential membrane damage and oxidative stress induced on E. coli O157:H7 cells were evaluated after 30 min of incubation in 0.1 mg GAE/mL of OPE, GA, HT, or 4-HPA.E. coli O157:H7 cells incubated in DW were included as a control.Metabolic activities of E. coli O157:H7 cells were monitored during 12 h of incubation in TSB supplemented with 0.5 mg GAE/mL of OPE, GA, HT, or 4-HPA.E. coli O157:H7 cells incubated in TSB were included as a control.The results represent the mean values and their standard deviations (n ≥ 3).Different lowercase letters indicate statistically significant differences determined by one-way ANOVA followed by Tukey's pairwise comparison with a 95% confidence interval (p < 0.05)

Table 1
Phenolic compositions of OPE identified using HPLC-DAD

Table 1
(continued) (SpectraFluor Plus, Tecan Group Ltd., Sunrise, Austria) with excitation/emission wavelengths at 530/580 nm.Bacterial cells in TSB with 50 µM resazurin (without OPE, GA, HT, or 4-HPA) were included as a positive control, and OPE, GA, HT, or 4-HPA in TSB with 50 µM resazurin (without bacterial cells) was used as a blank for each measurement.