E. coli O157:H7 strains 932 (clinical isolate from patients who had hemorrhagic colitis), ATCC 35150 (human feces isolate; acid resistant) and F4637 (sprout isolate) were used in this study. All three virulent isolates expressed genes for Shiga toxins 1 and 2 (Stx-1 and Stx-2). A cocktail of these three strains (in equal proportions) was prepared and used in all experiments. Cultures were prepared in tryptic soy broth with 0.6% yeast extract (TSB; Difco, Becton Dickinson, Sparks, MD). The cells were incubated overnight (18 to 20 h) at 37˚C to reach an optimal concentration of ~9 logs CFU/ml. After antimicrobial treatments, all viable organisms were enumerated by diluting in buffered peptone water (BPW; Difco, Becton Dickinson), plating on Cefixime-Tellurite-Sorbitol MacConkey agar (CT-SMAC; Difco, Becton Dickinson) and incubating at 37˚C for 18 - 24 h.

The test compounds consisted of six essential oils and active components and six powder extracts. The essential oils of cinnamon, lemongrass, and oregano were obtained from Lhasa Karnak Herb Co. (Berkeley, CA). The essential oil components carvacrol (>98%), trans-cinnamaldehyde (99.5%), and citral (>96%) were obtained from Sigma-Aldrich Co. (St. Louis, MO). The powder extracts used were green tea polyphenols (≥95%, LKT Laboratories, Inc., Minnesota, MN), apple skin extract (Apple Poly^TM, standardized to 80% polyphenols/5% phloridzin, Apple Poly LLC, Morrill, NE), black tea extract, caffeinated and decaffeinated (hot water infused and freeze-dried English Breakfast tea from Coffee Bean Direct, Titusville, NJ), grapeseed extract (MegaNatural^® Gold, >90% polyphenols, Polyphenolics Inc., Madera, CA), and grape pomace extract (MegaNatural^® GSKE, >80% polyphenols and >0.075% anthocyanins, Polyphenolics Inc., Madera, CA). Initially all antimicrobials were tested for their antimicrobial activity in vitro in sterile phosphate-buffered saline pH 7.4 (PBS). The most effective antimicrobials from the essential oil and the plant extract groups were further tested by applying them directly onto foods implicated in E. coli O157:H7 outbreaks. Hence, oregano oil and green tea polyphenols were chosen to be tested on romaine lettuce and spinach.

We wanted to select virulent strains of E. coli O157:H7 for the antimicrobial studies. Hence, nine strains were tested for the presence of Shiga-toxin Stx-1 and Stx-2 genes. Multiplex PCR was conducted to detect them using the protocol and primers designed by Blanco, et al. [13]. The Stx1 and Stx2 oligonucleotide sequence (5'-3') were: VT1-A forward CGCTGAATGTCATTCGCTCTGC (Eurofins, MVG Operon, Huntsville, AL) and VT1-B reverse C’GTGGTATAGCTAC TGTCACC (Eurofins); VT2-A forward CTTCGGTATCCTATTCCCGG (Eurofins) and VT2-B reverse CTGCTGTGACAGTGACAAAACGC (Eurofins). The nine E. coli O157:H7 isolates were 960212, 960218, F4546, F4637, and F5853 isolated from sprouts, 1388 from apple juice, 932 and ATCC 35150 from humans, and ATCC 43895 from beef. The DNA was extracted by suspending a loop of bacteria harvested from TSA in 250 µl of sterile water and incubating it at 100˚C for 5 min, using a thermocycler (Bio-Rad Hercules, CA) to release the DNA. This was used in Multiplex PCR. Multiplex PCR was determined with 150 ng of the primers; 0.2 mM of each deoxynucleotide (dATP, dGTP, dCTP, and dTTP) (Promega BioSciences, LLC. San Luis Obispo, CA); 10 mM Tris-HCl; 1.5 mM MgCl₂; 50 mM KCl (Promega BioSciences); and 1 U of DNA Taq polymerase (Promega BioSciences), and 7 μl of extracted DNA in a final volume of 30 µl. The PCR procedure of Blanco, et al. (2013) [13] was used. To visualize the amplicons, electrophoresis was performed by preparing 2% agarose gel (Difco, Becton Dickinson) in TBE buffer (89 mM Tris, 89 mM boric acid, 2 mM EDTA) (Promega BioSciences), and running at 130 V for 30 min. The gel was stained with ethidium bromide (Promega BioSciences). The size of DNA was analyzed by UV fluorescence using a 100 bp DNA ladder Molecular Marker (Promega BioSciences).

The essential oils and their active components (oregano oil, cinnamon oil, lemongrass oil, carvacrol, cinnamaldehyde, and citral) and the plant powder extracts (green tea polyphenols, apple skin, black tea, decaffeinated black tea, grape pomace, and grapeseed) were dissolved in sterile PBS (Difco, Becton Dickinson). Serial dilutions were prepared at 0.3%, 0.2%, and 0.1% for the essential oil/active components and 4%, 3%, 2%, and 1% for the dehydrated plant extracts from stock solutions in micro-centrifuge tubes (90 µl). The test organism (10 µl; approximately 3 - 4 log CFU/ml) was added to each tube, mixed well, and sampled immediately (0 min). Surviving bacterial populations were also enumerated after incubation for 1, 3, 5 and 24 h. All samples were serially diluted in BPW, plated on CT-SMAC, and incubated for 18 - 24 h at 37˚C.

Romaine lettuce and spinach leaves from local grocery stores were sectioned into 10-g pieces. The pieces were exposed to UV light under a biohood for 30 min. Lettuce and spinach samples were then placed into sterile plastic bags and inoculated by dipping in E. coli O157:H7 culture of ~6 log CFU/ml for 2 min. After inoculation, samples were dried for 1 h under a biohood. The inoculated leaf samples were then immersed for 10 min in 0.5% oregano oil or 3% green tea solutions prepared as described above. After treatment, the samples were stored at 4˚C for 7 days. Survivors were enumerated on days 0, 1, 3 and 7. The samples were stomached (Seward, Ltd, London, UK) in 90 ml BPW at normal speed for 1 min. Samples were diluted in BPW, plated on CT-SMAC, and incubated at 37˚C for 18 to 24 h. Appropriate controls (uninoculated and untreated inoculated lettuce and spinach samples) were done for each test. All experiments were repeated three times.

Each experiment was repeated three times and duplicate agar plates were used in each repeat giving 2 data points for each value, meaning duplicate numbers were obtained for each time point per repeat. An average of the duplicate values was taken for each repeat. The average values from the three repeats were then used to calculate the final mean and standard deviation.

As already mentioned, Shiga-toxin-producing E. coli is the leading cause of hemorrhagic colitis (HC) and hemolytic uremic syndrome (HUS), mainly in infants and the elderly. Virulent E. coli O157:H7 is characterized by the presence of Stx-1, Stx-2, intimin (eae-ξ), and ehxA genes [13].

Figure 1 shows the amplicons isolated from nine E. coli O157:H7 strains. All isolates contained both Shiga toxin genes 1 (516 bp) and 2 (302 bp), except ATCC 43895 (beef) that only showed the genes for Stx-1. One purpose of this study was to select the virulent isolates that had both Stx-1 and Stx-2 genes for

investigating the antimicrobial effect of plant compounds. Based on this analysis, the following E. coli O157:H7 strains were selected for evaluation: 932, ATTC 35150, and F4637, isolated from humans, humans, and sprouts, respectively. We do not know whether in addition to inactivating/inhibiting the growth of the bacteria, the plant-based antimicrobials also concurrently inhibit the toxicological properties of the toxins by modification of receptor-binding sites [14].

Figure 2 and Figure 3 show the results of the antimicrobial effect of each plant compound against E. coli O157:H7. Figure 2(a) and Figure 2(b) illustrate the

strong bactericidal activity of oregano oil and carvacrol, as evidenced by no detectable survivors at time 0 and thereafter, at all tested concentrations. Figure 2(c) and Figure 2(d) show that the bacteria exposed to cinnamon oil and cinnamaldehyde at different concentrations also did not survive (no detectable levels) starting at 1 h and thereafter. Figure 2(e) and Figure 2(f) depict the activity of lemongrass oil and citral, respectively; exposure to both these compounds resulted in no detectable survivors at time 0 at a concentration of ≥0.2%. A 2-log reduction was observed at a concentration of 0.1% at time 0 and no detectable survivors starting at 1 h and thereafter, respectively. The results show that the essential oils and their active compounds have the potential to inactivate E. coli O157:H7 rapidly and could be applied in foods for decontaminating them.

Figure 3(a) demonstrates the strong antimicrobial activity of green tea polyphenols as indicated by no detectable survivors at 1 h and thereafter for concentrations ≥ 2%, and for a 1% concentration by about a 1.5 log reduction and no detectable survivors at 1 and 3 h, respectively. Figure 3(b) depicts the bactericidal effect of apple skin extract that showed about a 2.5-log reduction after 1 h at 4%, and no detectable survivors starting at 3 h for 2%, 3%, and 4% concentrations, respectively. Black tea (Figure 3(c)) did not show any effect at 0 and 1 h at all concentrations, a >2-log reduction at 3 h at ≤2%, no detectable survivors at 3 h at 3%, and no detectable survivors at 24 h at all concentrations. Similar to black tea, decaffeinated black tea (Figure 3(d)) did not induce any change in the bacterial population at 0 and 1 h at all concentrations; but a ~1 - 2-log reduction was observed at all concentrations at 3 h, no survivors at 5 h at 3%, and finally no detectable survivors at 24 h at all tested concentrations. Figure 3(e) shows that exposure to grapeseed extract resulted in no detectable survivors at 3 h at all tested concentrations. Figure 3(f) shows that exposure of E. coli O157:H7 to grape pomace extract resulted in no detectable survivors at 3 and 5 h at ≤2% concentration, >1.5-log reduction at 5 h and no detectable survivors at 24 h, at 3% and 4% concentrations. One possible explanation for the better survival of E. coli O157:H7 at 3 and 4% concentrations of grape pomace extract at 3 and 5 h could be that the organism was able to utilize the nutrients available in the extract, which could also provide a protective effect on the pathogen. Verification of such an effect by conducting additional investigations is beyond the scope of this study; however, this observation provides additional area for future research.

Among the essential oils, oregano oil and carvacrol exhibited the best antimicrobial activity against E. coli O157:H7. Among the plant extracts, green tea polyphenols exhibited the highest antimicrobial activity. Black tea had better antimicrobial activity than decaffeinated black tea and grapeseed extract demonstrated better antimicrobial activity than grape pomace extract on E. coli O157: H7 in vitro.

Figure 4 and Figure 5 show the antimicrobial activity of oregano oil at a concentration of 0.5% and green tea extract at 3% against E. coli O157:H7 on romaine lettuce and on spinach. Oregano oil (0.5%) on day 0 induced > 2 log reductions on romaine lettuce and >3 log reductions on spinach. For both leafy greens, no survivors were detected on day 1.

The results obtained from green tea were similar to those of oregano oil. At day 0, green tea (3%) induced a ~1.5 log reduction on lettuce and ~1 log reduction on spinach; no survivors were detected for both types of leafy greens on day 1. These results demonstrate that plant compounds can be used to decontaminate and protect produce during refrigerated storage.

There have been several related reports on the antimicrobial properties of plant essential oils and their bioactive compounds, as well as those of apple polyphenols, grape and tea compounds, that complement the current study.

Friedman, et al. [15] determined the quantitative antimicrobial activities of more than 100 essential oils and their bioactive compounds against four foodborne pathogens, expressed as bactericidal activity (BA50) values, defined as the concentration of the antimicrobial that inhibited 50% of the bacteria under the test conditions. This widely cited paper serves as a resource for selecting antimicrobials with high activities in the cell assays (in vitro) for evaluation against efficacy on contaminated salads and meats. An investigation by Moore-Neibel, et al. [16] examined the antimicrobial effectiveness of lemongrass essential oil on organic leafy greens, romaine and iceberg lettuces, and mature and baby spinach inoculated with Salmonella Newport and demonstrated the concentration- and time-dependent inactivation of the pathogen. In another study on salad leaves, Todd, et al. [17] found that washing organic baby and mature spinach and iceberg and romaine lettuces with cinnamon oil inactivated antibiotic-resistant Salmonella Newport bacteria, indicating the potential of cinnamon oil as a practical treatment option. This was a positive progress because inhibiting the growth of antibiotic resistant pathogens is a major challenge for food microbiology [18].

Appearance of the food product can be an important factor in consumer choice. Yossa, et al. [19] demonstrated the efficacy of cinnamaldehyde-containing washes against E. coli O157:H7 and Salmonella enterica on iceberg lettuce without any effect on the color and texture of the leaves, maintaining their visual appearance for the consumer.

Poimenidou, et al. [20] reported that vinegar, lactic acid or oregano aqueous extract alone or in combination can serve as alternative washing solutions to chlorine against E. coli O157:H7 on fresh cut spinach and lettuce with acceptable appearance.

In an alternative strategy, Zhu, et al. [21] discovered that apple, carrot and hibiscus edible films containing the antimicrobials carvacrol and cinnamaldehyde inactivated Salmonella Newport on organic Romaine and Iceberg lettuce, and mature and baby spinach in sealed salad bags. Films with 3% carvacrol showed the highest reduction. A related study found these antimicrobial edible films also inactivated E. coli O157:H7 on organic leafy greens in sealed plastic bags [22]. The inactivation of the bacteria is most likely induced by vapors released from the antimicrobials incorporated into the films.

A similar approach has also been used for incorporating apple polyphenols into edible films to examine their effect on microbial activity and it was shown that the films inhibited the growth of L. monocytogenes but not S. enterica serotype Hadar [23] [24]. Human studies have shown other beneficial health activities for apple polyphenols, in which they ameliorated hyperglycemia [25] and protected against UV radiation-induced skin pigmentation [26].

Although not relevant to the present study, it is of interest that Wang, et al. [27] found that applying a mixture of acetic and lactic acids on lettuce reduced the presence of E. coli O157:H7 and L. monocytogenes, as well as aerobic mesophilic and psychrotrophic bacteria, coliforms, molds, and yeasts during storage.

Grape compounds have also previously been shown to have beneficial effects against microorganisms. For example, Peixoto, et al. [28] and Friedman [29] reported that the bioactive phenolic compounds in grape pomace, a byproduct of wine production composed of seeds, skin, and stems, exhibited antibacterial and other health benefits.

Black and green tea extracts, such as those used here, are often cited for their health benefits. Friedman, et al. [30] reported that both black tea theaflavins and green tea catechins inhibited the growth of Bacillus cereus at nanomolar levels and that some were more active than medicinal antibiotics; freshly prepared teas were more active than day-old teas. Tea compounds also showed strong antiviral [31] and anti-parasitic trichomonad properties [32].

The results of the present study on salads and these related cited studies indicate that some of the test compounds can act as broad-spectrum antibiotics against nonresistant and resistant pathogens in cell assays (in vitro) and on contaminated leafy greens.