Introduction

Outbreaks of foodborne illnesses, which are related to the consumption of fresh produce, have rapidly increased during recent years and were frequently caused by Listeria monocytogenes and enterohemorrhagic E. coli (EHEC) [6, 31, 38]. Rangel et al. [25] identified fresh produce as a transmission vehicle for E. coli O157:H7 in 21 % of the foodborne outbreaks in the United States from 1982 to 2002. From 2000 to 2004, E. coli O157:H7 was identified as the second most common pathogen in fresh produce-related illness outbreaks [2]. Beuchat et al. [6] reported that E. coli O157 outbreaks have been linked to the consumption of apple cider, lettuce, radish, alfalfa sprouts, and other mixed salads. The largest outbreak of illness in Germany occurred in May–July 2011 and was caused by E. coli O104:H4-contaminated fenugreek sprouts with a total of 3126 cases of illness [13]. In 2013, 6043 confirmed cases of EHEC infections were reported in the European Union, which was 5.9 % higher than in 2012. The most commonly reported EHEC serogroup in 2013 was, as in previous years, O157 with 48.9 % of cases [12].

Listeria monocytogenes is widely distributed on raw fruits and vegetables and on plant material [7]. It is able to survive fresh produce processing and can grow on fresh produce stored at refrigerated temperatures [8, 18]. While L. monocytogenes causes moderate gastroenteritis in healthy adults, listeriosis can be severe in susceptible individuals including young, elderly, and immunocompromised people as well as pregnant women [33]. In 2013, the European Union reported 1763 confirmed human cases of listeriosis, which represented an 8.6 % increase compared to 2012, with a total of 13 L. monocytogenes outbreaks resulting from the consumption of ready-to-eat foods [12].

These outbreaks show that processing of fresh produce, as well as the washing steps before produce consumption, could contribute to decrease the potential of infection and to maintain human health. Washing procedures aim at removing dirt, soil, pesticide residues and other debris, reducing microorganisms responsible for quality loss, limiting browning, lowering product temperature and removing plant cell exudates that may support microbial growth [16, 30, 43]. Many studies have been published regarding the washing and sanitizing treatments for reduction of the microbial load on fresh-cut produce. A wide range of chemical sanitizers have been tested with various degrees of effectiveness. Fett et al. [15] achieved the reduction of E. coli O157:H7, Salmonella enterica sv. Anatum F4317, S. enterica sv. Infantis F4319, S. enterica sv. Newport H1275, and S. enterica sv. Stanley HO558 on laboratory-inoculated mung bean seed by chlorine treatment. In another study, the effectiveness of lactic acid and hydrogen peroxide against E. coli O157:H7, Salmonella enterica sv. Enteritidis and L. monocytogenes was demonstrated [19]. Furthermore, chlorine (100 mg/L), citric acid [0.5 % (v/v)], and ascorbic acid [0.5 % (v/v)] were applied against E. coli and L. monocytogenes on iceberg lettuce and resulted in a reduction of approximately 1.0 log10 colony forming units (CFU)/g [1].

Lauroyl arginate ethyl (N α-lauryl-l-arginine ethyl ester, LAE®) is a surfactant produced synthetically through the reaction of l-arginine, hydrochloric acid, ethanol, thionyl chloride, sodium hydroxide, lauryl chloride, and deionized water [22, 28]. Due to its cationic surfactant nature, it damages the bacterial cell membrane by reacting with oppositely charged proteins or enzymatic systems, causing their denaturation. This results in growth inhibition of the microorganisms or finally in bacterial cell death [22, 27]. In the food industry, it is mainly used as an antimicrobial agent against molds, Gram-positive and Gram-negative bacteria. For example, Porto-Fett et al. [23] treated frankfurters with LAE and investigated the effect on L. monocytogenes in the sausage packages. When frankfurters were treated with LAE, numbers of L. monocytogenes decreased by approximately 2.0 log10 CFU/package within 2 h [23]. In another study, chocolate and pasteurized milk were treated with different concentrations of LAE and the reduction of the bacterial growth was observed [39]. LAE has been approved by the United States Food and Drug Administration [36] and the European Food Safety Agency [14] as “Generally Recognized as Safe” (GRAS) and represents a potential alternative food preservative to inhibit microbial growth in selected applications.

Currently, there are no reports on the effectiveness of LAE used as wash water additive in lettuce processing. Consequently, we investigated the use of LAE for the reduction of L. monocytogenes and E. coli O157:H7 on the surfaces of ready-to eat lettuce. As a model, we used oak leaf lettuce, because it is a frequently consumed lettuce cultivar in Germany, and in addition it is commercially available throughout the whole year.

In this context, we evaluated the attachment of E. coli O157:H7 on oak leaf lettuce surfaces after deletion of the hcpA and fliC genes, which are known to be required for attachment of EHEC on baby spinach leaves [29]. The hcpA gene encodes the pilin subunit of the hemorrhagic coli pilus (HcpA) [42]. The fliC gene encodes the major subunit of flagellin, which polymerizes to form the filaments of bacterial flagella. In addition to their role in motility, flagella in general mediate other functions associated with pathogenicity such as adherence and invasion of epithelial cells [17, 40]. Besides colonization of the intestinal tract, EHEC can also employ their flagella to survive in the environment, making leafy vegetables a unique vehicle in the human disease process [40]. Currently, there is no information about their impact on the attachment of EHEC on the leaf surfaces of oak leaf lettuce.

Materials and methods

Preparation of the lauroyl arginate ethyl solution

Lauroyl arginate ethyl (N α-lauryl-l-arginine ethyl ester, LAE®, CAS registration number 60372-77-2), was distributed by VEDEQSA SA (Barcelona, Spain) as Mirenat NSM®, a white powder consisting of 14.5 % (w/w) LAE and 85.5 % (w/w) maltodextrin. 68.9 mg of Mirenat NSM® was diluted in 100 mL of autoclaved ultra-pure water (Milli-Q system, Millipore, Bedford, MA) to obtain a final LAE concentration of 100 mg/L.

Bacterial strains, plasmids and growth conditions

The bacterial strains and plasmids used in this study are listed in Table 1. The E. coli strains were routinely cultivated in Luria–Bertani (LB) broth composed of 10 g/L Bacto™ tryptone, 5 g/L yeast extract, and 10 g/L NaCl, pH 7.5, under aeration at 37 °C with 180 rpm unless otherwise mentioned. For the solid medium, 1.5 % (w/v) Bacto™ agar was added to the broth prior to autoclaving. When needed, ampicillin-sodium salt (Carl Roth GmbH & Co KG) and kanamycin-sulfate (Carl Roth GmbH & Co KG) were added after autoclaving in final concentrations of 100 and 50 μg/mL, respectively. L. monocytogenes DSM 20600T was routinely grown in Brain Heart Infusion (BHI) broth in a rotary shaker at 180 rpm or on BHI agar (both Merck KGaA) at 37 °C for 48 h under aerobic conditions. Polymyxin–acriflavine–lithium chloride–ceftazidime–aesculin–mannitol (PALCAM) agar plates (Merck KGaA) were used for selective growth of L. monocytogenes.

Table 1 Bacterial strains and plasmids used in this study
Full size table

Construction of E. coli EDL933∆fliC and EDL933∆hcpA deletion mutants

The deletion of the fliC and hcpA genes in EHEC strain EDL933 was carried out by site directed mutagenesis according to Datsenko and Wanner [10]. The target genes were exchanged by an aph kanamycin resistance cassette using plasmids pKD4 and pKD46. Subsequently, the aph kanamycin resistance cassette was removed using the temperature-sensitive plasmid pCP20 (Table 1).

Electrocompetent EDL933 cells were prepared as described in the pulse controller instruction manual (Bio-Rad Laboratories GmbH). For transformation, 40 µL of the electrocompetent cells was thawed and mixed with 30-300 ng of respective plasmid DNA. The mixture was transferred in an ice-cold Gene Pulser/MicroPulser Cuvette (Bio-Rad Laboratories GmbH) and pulsed in the Gene Pulser Xcell™ Electroporation Systems device (Bio-Rad Laboratories GmbH) at 25 µF, 200 Ω, 2.5 kW for 5 ms. Following this, 1 mL of pre-warmed SOC-medium was added immediately. SOC-medium contained 20 g/L Bacto™ tryptone, 5 g/L yeast extract, 0.5 g/L NaCl and 2.5 mM KCl, pH 7. After autoclaving, the medium was adjusted to final concentrations of 5 mM MgCl2, 5 mM MgSO4, and 20 mM glucose (filter sterilized). The bacterial suspension was incubated for 1 h at 37 °C, and then spread plated on LB agar containing appropriate antibiotics. The deletions were confirmed by PCR with primer pairs surrounding the position were the deletion had taken place (Table 2) and by DNA-sequencing using a CEQ 8000 genetic analysis system (Beckmann Coulter GmbH) according to standard protocols. Oligonucleotides and PCR programs used for deletion mutagenesis and confirmation experiments are given in Table 2.

Table 2 Oligonucleotides and PCR conditions used in this study
Full size table

Preparation of fluorescent E. coli EDL933 wild type and mutants

Electrocompetent E. coli EDL933 cells and the respective fliC and hcpA mutants were prepared, and then transformed with plasmid pWRG435 (Table 1) as described above. The plasmid pWRG435 was designed by Bender et al. [4] and is composed of the plasmid pFPV25.1 backbone [37], encoding an ampicillin resistance gene as selective marker [37], as well as a red fluorescence protein variant (TagRFP-T) gene as an insert [4]. Plasmid pWRG435 was used for constitutive expression of the red fluorescence protein (RFP) and for clear differentiation of E. coli EDL933 from the autochthonous microbiota by red fluorescence. Transformants were grown in LB broth with ampicillin (100 µg/mL) and could be detected under UV light at 302 nm as red fluorescent colonies on LB agar containing ampicillin (100 µg/mL). One transformant of each experiment was selected, and designated as EDL933/pWRG435, EDL933ΔfliC/pWRG435, and EDL933ΔhcpA/pWRG435 (Table 1).

The stability of the plasmids was examined in LB medium with and without ampicillin. Overnight cultures of EDL933/pWRG435 and the respective mutants previously grown in the presence of ampicillin were inoculated into LB medium without ampicillin. After 24 h at 37 °C and 180 rpm, and the cultures were serially diluted in 0.9 % NaCl solution (w/v), and spread plated onto LB agar with and without ampicillin, respectively. The plates were incubated for 24 h at 37 °C and the total numbers of colonies as well as numbers of red fluorescent colonies were determined.

Evaluation of the antimicrobial effect of LAE as wash water additive

Heads of red oak leaf lettuce (Lactuca sativa L. var. crispa) were purchased from a local supermarket. Outer leaves were removed, and lettuce leaves were cut into pieces with a weight of approximately 10 g using a sterile scalpel.

The bacterial strains used for inoculation experiments are listed in Table 1. The inoculum preparation and the experimental design were carried out as follows: 10 mL overnight cultures of each bacterial strain were prepared as described above. After centrifugation at 4000×g for 10 min at 4 °C, the cell pellets were washed with 10 mL of 0.9 % NaCl solution (w/v) and centrifuged again. Each cell pellet was resuspended in 5 mL 0.9 % NaCl solution (w/v), and the optical density at 600 nm (OD600) was measured. In a preliminary experiment, the cell concentration of each bacterial strain was correlated with the OD600. The bacterial suspension was then serially diluted in 0.9 % NaCl (w/v) to achieve the desired cell concentration of approx. 9.0 log10 CFU/mL. Each leaf sample was inoculated with 10 spots of 10 µL of the respective bacterial suspension at different sites on the upper leaf surface, resulting in a final concentration of approx. 7.0 log10 CFU/g lettuce, and incubated for 15 min under a clean bench (NuAire, IBS Tecnomara) for drying. After that, the lettuce leaves were further processed I) without washing, II) with washing in 100 mL sterile deionized water for 120 s at 6 °C, and III) with washing in 100 mL sterile deionized water containing LAE in a final concentration of 100 mg/L for 120 s at 6 °C. LAE was subsequently removed by rinsing the leaves with 100 mL sterile deionized water for approximately 10 s (Fig. 1).

Fig. 1
figure 1

Flowchart of the experimental design of this study

Full size image

The leaves were then transferred into sterile stomacher bags with lateral filter and the nine-fold weight of a 0.9 % (w/v) NaCl solution was added. Each sample was treated for 120 s using a Stomacher 400 Circulator (Seward Ltd.). Decimal dilutions of each homogenized sample solution were prepared in the same diluent. E. coli EDL933 wild type and mutants were estimated by spread plating in duplicate on LB agar containing the appropriate antibiotics. Plates were incubated aerobically at 37 °C for 24 h. Furthermore, the wash water, with and without LAE, as well as the rinsing water were analyzed in duplicate on the same media directly after processing. Colonies of E. coli EDL933/pWRG435 and the mutants could be clearly differentiated from the autochthonous microbiota by red fluorescence after incubating for another 24 h at 4 °C.

Listeria monocytogenes viable counts were estimated by spread plating on PALCAM agar. Plates were incubated aerobically at 37 °C for 72 h before colonies were counted. L. monocytogenes colonies were approximately 2 mm in diameter, gray-green with a black sunken center and a black halo on a cherry-red background. As a control, the microbial loads of the lettuce leaves were determined without bacterial inoculation. In addition, each bacterial suspension used for lettuce inoculation was serially diluted in 0.9 % NaCl solution (w/v) and spread plated on the appropriate agar (see above) to ensure that the adjusted cell concentration of approx. 9.0 log10 CFU/mL was achieved.

Colony forming units were counted, and their weighted arithmetic means were calculated. All experiments were performed independently at least at three different times per investigated bacterial strain. The findings from representative experiments are given in the results.

Statistical evaluation

For the inoculation assays, the averaged values of three technical replicates were used for calculation of the standard error of the three biological replicates derived from three different days. Data evaluation was performed using SAS software version 9.1 (SAS Institute, Cary, NC, USA). For statistical analysis demonstrating significant differences (α = 0.05), Tukey’s multiple range test was used.

Results

Antimicrobial efficacy of lauroyl arginate ethyl against L. monocytogenes

Initially, the microbial load of lettuce samples was investigated. Initial total aerobic counts on unwashed, non-inoculated lettuce leaves were 6.0 ± 0.3 log10 CFU/g. The inoculum of L. monocytogenes DSM 20600T contained 9.1 ± 8.3 log10 CFU/mL, resulting in a contamination of approximately 7.1 log10 CFU/g lettuce. Counts of L. monocytogenes on unwashed, non-inoculated red oak leaf lettuce were either below the detection limit of 2.0 log10 CFU/g (2 samples) or amounted to 2.4 log10 CFU/g oak leaf lettuce (one sample). After the inoculation, L. monocytogenes counts on red oak leaf lettuce were 6.6 ± 6.0 log10 CFU/g on PALCAM agar.

After inoculation and washing with cold water, counts of Listeria were 5.7 ± 5.1 log10 CFU/g on PALCAM agar. This means a reduction of 0.9 ± 0.1 log10 CFU/g. Washing with LAE reduced viable counts to 3.1 ± 0.3 log10 CFU/g oak leaf lettuce. The L. monocytogenes counts in the process water after washing without LAE were 5.4 ± 4.6 log10 CFU/mL. After washing with LAE, L. monocytogenes was not detected in the process water (detection limit of 1.0 log10 CFU/mL). In summary, it could be demonstrated that LAE is an effective antimicrobial agent to reduce viable counts of L. monocytogenes on red oak leaf lettuce. Therefore, LAE represents an effective alternative to traditional washing techniques for fresh-cut lettuce.

Antimicrobial efficacy of lauroyl arginate ethyl against E. coli strain EDL933

Members of the family Enterobacteriaceae are part of the lettuce phyllosphere [26]. To differentiate the inoculated EHEC strains from the autochthonous oak leaf lettuce microbiota under UV light, the strain EDL933 and its isogenic fliC and hcpA mutants were constructed to produce a red fluorescent protein (see above). This additional selection marker was necessary because additional tests had shown that there are still some ampicillin resistant bacteria in the lettuce microbiota, which could falsify the results, if ampicillin was used as sole selection marker (data not shown). In a preliminary investigation, the stability of the plasmid in the transformant was evaluated. Plasmid pWRG435 was present in all cells (100 %) under ampicillin selection. Without ampicillin, approximately 90 % of the EHEC cells retained the plasmid and the ability to form red fluorescent colonies on LB agar. Therefore, all experiments with EDL933/pWRG435 and the mutants were performed under ampicillin selection.

E. coli O157:H7 strain EDL933/pWRG435 was used for inoculation of oak leaf lettuce. The inoculum contained 9.2 ± 8.2 log10 CFU/mL on LB agar with ampicillin (100 µg/mL), resulting in a contamination of approximately 7.2 log10 CFU/g lettuce. The unwashed, non-inoculated lettuce leaves were free of red fluorescent E. coli (data not shown). Viable counts of EDL933/pWRG435 after inoculation without washing were 7.0 ± 5.8 log10 CFU/g. After inoculation and washing with cold water, counts of EHEC amounted to 6.1 ± 4.7 log10 CFU/g, meaning a reduction of 0.9 ± 0.0 log10 CFU/g. Washing with LAE reduced EHEC counts by 2.6 ± 0.2 log10 CFU/g. In the process water without LAE, viable counts of EHEC were 5.8 ± 3.9 log10 CFU/mL. After the LAE treatment EDL933/pWRG435 could not be detected in the process water (detection limit of 1.0 log10 CFU/mL).

Impact of gene deletion in E. coli O157:H7 strain EDL933 on the adherence to the lettuce leaf surface

In order to investigate the influence of particular adherence factors of EDL933/pWRG435 on the ability to adhere on the lettuce surface, as well as the effectivity of the washing process, isogenic mutants of the EHEC strain EDL933 with deletion of the fliC and the hcpA gene were constructed, transformed with pWRG435, and tested in the washing procedures.

The inoculum of EDL933ΔfliC/pWRG435 contained 9.2 ± 8.1 log10 CFU/mL on LB agar with ampicillin (100 µg/mL), resulting in a contamination of approximately 7.2 log10 CFU/g lettuce. Viable counts of EDL933ΔfliC/pWRG435 after inoculation without washing were 7.1 ± 5.8 log10 CFU/g. Based on the initial inoculum, the fliC deletion mutant adhered less than the parental strain, which is demonstrated in a ~30 % higher recovery as compared to the wild-type strain (7.0 ± 5.8 log10 CFU/g).

The inoculum of E. coli O157:H7 strain EDL933ΔhcpA/pWRG435 contained 9.1 ± 8.1 log10 CFU/mL on LB agar with ampicillin (100 µg/mL), resulting in a contamination of approximately 7.1 log10 CFU/g lettuce. Viable counts of EDL933ΔfliC/pWRG435 after inoculation without washing were 6.8 ± 5.9 log10 CFU/g.

Interestingly, the hcpA deletion mutant was detected approximately 18 % less than the wild-type strain.

The effect of washing with and without LAE to the isogenic mutants was quite similar to the results described above. Washing in sterile cold water led to a reduction between 0.7 ± 0.1 log10 for E. coli EDL933ΔhcpA and 1.0 ± 0.1 log10 for the other strains tested. The LAE treatment resulted in a reduction of approximately 2.7 log10 for the wild-type strain and the hcpA isogenic mutant. The viable counts of EDL933ΔfliC/pWRG435 were reduced by 3.5 ± 0.3 log10 CFU/g after LAE treatment.

Discussion

In the present study, with a conventional washing step the viable counts of L. monocytogenes and E. coli O157:H7 on oak leaf lettuce leaves were reduced by approximately 1.0 log10 CFU/g. The treatment of lettuce wash water with 100 mg/L of LAE resulted in an initial reduction of 3.5 ± 0.3 log10 CFU/g of L. monocytogenes compared to untreated controls. The reduction of E. coli O157:H7 by LAE amounted to 2.7 log10 CFU/g. Comparable results were observed in a number of other studies, which evaluated the bactericidal efficacy of LAE on the same pathogens in other food systems. Soni et al. [32] reported a L. monocytogenes reduction of approximately 2 log10 CFU/g in Queso Fresco cheese after treatment with 200 ppm LAE. In “Lyoner style” sausages, LAE used in concentrations of 2 mg/mL reduced L. innocua by 2.1 ± 0.9 log10 CFU/g when added as aqueous solution [35]. Taormina and Dorsa [34] investigated the short-term efficacy of LAE against L. monocytogenes on vacuum-packaged frankfurters and noted a ~ 2 log CFU/package reduction within 48 h. In 2 % reduced fat milk the growth of E. coli O157:H7 was reduced by approximately 2 log CFU/mL when treated with 750 ppm LAE [20].

Besides its strong antimicrobial characteristics, LAE offers some further advantages, which were interesting for the development of new applications in food preservation [3]. LAE showed fast antimicrobial activity, which indicates that only a short exposure time is needed to reduce the viability of bacteria [3]. This is especially interesting in lettuce processing, where average washing, and thus exposure times, are approximately 2 min. Another interesting observation of this study was the effect of LAE on the wash water viable counts. These could be reduced by at least 4.1 log10 CFU/mL. Due to the reduction of L. monocytogenes and E. coli O157:H7 below the detection limit in the wash water, LAE represents an effective alternative to traditional washing methods for fresh-cut lettuce and could be beneficial in a bioeconomic view.

The present study was conducted to address the potential involvement of the H7 flagella from E. coli O157:H7 strain EDL933 and its hemorrhagic coli pilus to attach to lettuce leaf surfaces. Various studies have shown that enteric bacteria, such as E. coli, have many molecular and physiological mechanisms that allow them to colonize and survive in the plant environment [24]. E. coli O157 attached to lettuce leaves by several mechanisms involving the type III secretion system (T3SS), flagella, pilli, curli and other outer membrane proteins [5]. Our experiments on the recovery of attached E. coli to lettuce leaf surfaces showed that the isogenic mutants of E. coli EDL933 displayed a different behavior in attachment than the wild-type strain. Recovery of strain EDL933 amounted to (7.0 ± 5.8 log10 CFU/g), while strain EDL933ΔfliC was recovered at (7.1 ± 5.8 log10 CFU/g), which indicates an inferior attachment on the leaf surface in comparison to the parental strain. This is in accordance with Saldaña et al. [29] who reported that an EDL933ΔfliC isogenic mutant adhered 72 % less to spinach leaves than the wild-type strain. Xicohtencatl-Cortes [40] investigated the interaction of E. coli O157:H7 with baby spinach and lettuce leaves. They found that EDL933ΔfliC mutants were significantly hampered in their adherence to leaf surfaces. They quantified colony forming units of E. coli O157:H7 by plating on selective MacConkey-sorbitol agar. The ΔfliC mutants were reduced by approximately 2 log10 CFU/mL in their ability to adhere at lettuce leaves.

The hemorrhagic coli pilus (HCP) is essentially a type 4 pilus, which is associated with a number of cellular functions, like adhesion to host cells, bacterial aggregation, biofilm and microcolony formation, cell signaling and DNA uptake [9, 41]. The inactivation of the major subunit, the hcpA gene, in EHEC O157:H7 reduces adherence to bovine epithelial and human intestinal cells [42]. In our study, lack of hcpA affected the recovery of EDL933ΔhcpA/pWRG435 in comparison to the wild-type strain in an opposite way. Deletion of hcpA resulted in 18 % (6.8 ± 5.9 log10 CFU/g) less recovery of the mutant than the wild-type strain, meaning a tighter adherence to the surfaces of oak leaf lettuce. This was not expected and can actually not be explained. Regarding the recovery in the wash water, there were no significant differences between the wild-type strain and EDL933ΔhcpA/pWRG435. Saldaña et al. [29] reported that hcpA deletion mutants were significantly hampered (49 % reduction) in adherence to spinach leaves in comparison to the parental strain and that HCP of E. coli strain EDL933 was required for spinach leaf adherence. They quantified colony forming units of E. coli O157:H7 by plating on selective MacConkey-sorbitol agar. The use of different media could be a reason for the different recovery compared to our study.

In conclusion, the attachment of E. coli O157:H7 and L. monocytogenes to cut produce surfaces underlines the need for effective wash treatments to reduce potential contamination of pathogens. The present in vitro study revealed that LAE possesses strong and fast antimicrobial activity against both foodborne pathogens on lettuce. While LAE is known to possess an antibacterial effect in other food commodities, it was not clear if that is also the case in produce. The findings of this study indicate that LAE can enhance the food safety and extend product shelf life by inhibiting foodborne pathogens. Therefore, the knowledge gained in this study will be helpful in developing novel intervention strategies for fresh produce wash treatments.