Journal of Bioscience and Bioengineering VOL. xx No. xx, 1e5, 2015 www.elsevier.com/locate/jbiosc
Inactivation of Shiga toxin-producing Escherichia coli O104:H4 using cold atmospheric pressure plasma Matthias Baier,1 Traute Janßen,2 Lothar H. Wieler,2 Jörg Ehlbeck,3 Dietrich Knorr,4 and Oliver Schlüter1, * Leibniz Institute for Agricultural Engineering, Max-Eyth-Allee 100, 14469 Potsdam-Bornim, Germany,1 Centre for Infection Medicine, Institute of Microbiology and Epizootics, Freie Universität Berlin, Robert-von-Ostertag-Str. 7-13, 14163 Berlin, Germany,2 Leibniz Institute for Plasma Science and Technology, Felix-Hausdorff-Straße 2, 17489 Greifswald, Germany,3 and Department of Food Biotechnology and Food Process Engineering, Berlin University of Technology, Königin-Luise-Straße 22, 14195 Berlin, Germany4 Received 19 August 2014; accepted 6 January 2015 Available online xxx
From cultivation to the end of the post-harvest chain, heat-sensitive fresh produce is exposed to a variety of sources of pathogenic microorganisms. If contaminated, effective gentle means of sanitation are necessary to reduce bacterial pathogen load below their infective dose. The occurrence of rare or new serotypes raises the question of their tenacity to inactivation processes. In this study the antibacterial efficiency of cold plasma by an atmospheric pressure plasma-jet was examined against the Shiga toxin-producing outbreak strain Escherichia coli O104:H4. Argon was transformed into non-thermal plasma at a power input of 8 W and a gas flow of 5 L minL1. Basic tests were performed on polysaccharide gel discs, including the more common E. coli O157:H7 and non-pathogenic E. coli DSM 1116. At 5 mm treatment distance and 105 cfu cmL2 initial bacterial count, plasma reduced E. coli O104:H4 after 60 s by 4.6 ± 0.6 log, E. coli O157:H7 after 45 s by 4.5 ± 0.6 log, and E. coli DSM 1116 after 30 s by 4.4 ± 1.1 log. On the surface of corn salad leaves, gentle plasma application at 17 mm reduced 104 cfu cmL2 of E. coli O104:H4 by 3.3 ± 1.1 log after 2 min, whereas E. coli O157:H7 was inactivated by 3.2 ± 1.1 log after 60 s. In conclusion, plasma treatment has the potential to reduce pathogens such as E. coli O104:H4 on the surface of fresh produce. However, a serotype-specific adaptation of the process parameters is required. Ó 2015, The Society for Biotechnology, Japan. All rights reserved. [Key words: Plasma-jet; Escherichia coli O157:H7; Antibacterial treatment; Food safety; Fresh produce; Corn salad]
Outbreaks of food-borne illness due to pathogenic microorganisms were traditionally associated with the consumption of protein-rich foods of animal origin. Over the last decades, awareness has risen for the growing role of fresh produce as vehicle for hazardous microbes. Although percentages of detected pathogens remain higher for foods of animal origin such as ground meat, also fruits and vegetables harbor a latent risk to consumers’ health (1,2). Despite their lower frequency, produce-related outbreaks caused, e.g., by Salmonella and Escherichia coli O157:H7 are consistently large (3) and recent estimations of outbreaks and numbers of cases assume an essential share of this group of foods as a responsible source for food-borne illness (4). The largest outbreak of E. coli O157:H7 involving more than 12,000 cases and 12 fatalities was linked to raw radish sprouts in Japan in 1996 (5). While most frequently outbreaks are reported to be caused by E. coli O157:H7, in 2011 a large outbreak of Shiga toxin-producing E. coli O104:H4 occurred in Europe, a strain with strong enteroaggregative features. In contrast to O157:H7, this strain does not seem to originate from the typical cattle reservoir (6). The outbreak was epidemiologically traced back to fenugreek sprouts as the most likely vehicle. This strain belongs to the HUSEC041 clone of sequence type 678 (ST678) and exhibited a combination of typical features of both Shiga toxin-producing E. coli (STEC) and
* Corresponding author. Tel.: þ49 331 5699 613; fax: þ49 331 5699 849. E-mail address: [email protected] (O. Schlüter).
enteroaggregative E. coli (EAEC). In addition to HUSEC041, the outbreak strain O104:H4 also produced extended-spectrum blactamase (ESBL) (7). Latest research was able to proof this strain to display a particular biofilm phenotype, in that it produces high amounts of curli fimbriae, which are known to cause strong inflammation (8). These properties led to an extraordinary high number of 855 cases of hemolytic-uremic syndrome (HUS), 2987 cases of acute gastroenteritis, and 53 fatalities (9). Despite the principally higher risk potential of meat products, they have the advantage of an efficient antimicrobial reduction by appropriate heating during their preparation. Heat sensitive fresh produces, destined to be eaten raw, lack such an effective step and are limited to only minor reductions by washing procedures. Whereas use of disinfecting chemicals harbors the risk of residues, non-thermal techniques such as irradiation demand highest levels of safety at work or face low consumer acceptance. A promising physical approach is the application of cold or non-thermal nonequilibrium plasma (NTP). NTP can be generated by subjecting a process gas to a strong electric field at atmospheric pressure or pressures both above and below ambient. This leads to the partial ionization of the process gas molecules and the concomitant formation of other reactive chemical species, such as ions and radicals, heat (gas temperature less than 500 K), and UV light, which all together potentially react with the microbes adhesive to the food surfaces (10). Studies of the antimicrobial efficacy of different plasma sources on STEC have already been conducted. Critzer et al. (11)
1389-1723/$ e see front matter Ó 2015, The Society for Biotechnology, Japan. All rights reserved. http://dx.doi.org/10.1016/j.jbiosc.2015.01.003
Please cite this article in press as: Baier, M., et al., Inactivation of Shiga toxin-producing Escherichia coli O104:H4 using cold atmospheric pressure plasma, J. Biosci. Bioeng., (2015), http://dx.doi.org/10.1016/j.jbiosc.2015.01.003
2
BAIER ET AL.
treated E. coli O157:H7 with the exhaust of a one atmosphere uniform glow discharge plasma (OAUGDP) on cv. Red Delicious apples and obtained inactivation of over 2 logs after 2 min exposure. Dried air, ionized by a gliding arc plasma inactivated E. coli O157:H7 on the surface of cv. Golden Delicious apples by 3.4 log cfu ml1 after 1 min (12). Direct treatment of apple juice using a needle-to-plate arrangement of electrodes was reported by Montenegro et al. (13). This treatment in liquid resulted in more than 5 logs reduction of E. coli O157:H7 after 40 s. Focused on E. coli O157:H7 only, the potential of NTP on other non-O157 STEC strains is still unknown, as is the actual share of these strains in all cases of food-borne illness. Difficulties in clear analytical distinctions from non-pathogenic commensal E. coli limit the gain of accurate information on non-O157 STECs and specific outbreak investigations are likely to remain the main source for further knowledge (14). In this context E. coli O104:H4, which has not been detected by monitoring from 2009 to 2011 (1,15e17), exhibits an example of a rare non-O157 STEC of extremely high virulence. It was shown that plasma can be used to treat a leafy green such as corn salad under gentle conditions, reaching 3 log cycles reductions of non-pathogenic E. coli DSM 1116 after 20 s (18). Particularly with regard to the special aggregative ability of E. coli O104:H4 (19), the question was raised if cold plasma is capable to achieve effective reductions of this serotype as authors previously reported on E. coli O157:H7. Therefore, the aim of this study was to determine the antibacterial efficacy of an atmospheric pressure plasma-jet on the enteroaggregative STEC outbreak strain E. coli O104:H4. For purposes of comparison, a serovar of the much more common E. coli O157:H7 was included in this study. Treatment distance and initial count were varied in a set of basic tests on a polysaccharide-gel and promising process parameters were applied on artificially inoculated corn salad leaves.
FIG. 1. Experimental setup of the atmospheric pressure plasma-jet.
J. BIOSCI. BIOENG., MATERIALS AND METHODS Atmospheric pressure plasma-jet The atmospheric pressure plasma-jet was generated at the tip of a pin-type electrode mounted concentrically in a quartz capillary (Fig. 1). Argon with 0.1% oxygen were delivered by a gas supply unit and passed into the capillary (inner diameter 1.6 mm) at a flow rate of 5 slm. A DC power supply (system power: 8 W at 220 V, 50/60 Hz) delivering a high frequency (HF) voltage (1.1 MHz, 2e6 kVpp) was coupled to the pin-type electrode. The process gas was transformed into plasma at the tip of the centered electrode and driven out of the capillary into the surrounding air. The plasma had a length of up to 10 mm and thus, enabled intense direct plasma application at 10 mm treatment distance or below. At 12 mm or more, a gap to the sample surface emerged, which inhibited direct interaction of glowing plasma filaments and samples, resulting in an attenuated semi-direct treatment mode. A detailed description on the device and its plasma parameters is given in Weltmann et al. (20). The plasma-jet (kinpen, neoplas tools GmbH, Greifswald, Germany) gained the CE marking for meeting EU consumer safety, health or environmental requirements. Bacterial strains The E. coli strains used in this study were the HUS outbreak strain RKI II-2027 (ST678, O104:H4, kindly provided by Angelika Fruth, Robert KochInstitut Wernigerode), isolated from a patient in 2011, the STEC reference strain EDL933 (ST11, O157:H7), isolated in 1983, and the non-pathogenic E. coli reference strain DSM 1116 (also referred to as ATCC 9637, CCM 2024, NCIB 8666, or NRRL B766). All strains were stored at 80 C in glycerol in brain heart infusion (BHI). Gel samples Luria Broth (LB) agar was used for culturing the E. coli strains at 37 C for 24 h. A colony was taken and transferred to 5 ml nutrient broth (NB). This pre-culture was bred shaking at 37 C for 24 h. Optical density (OD) was measured at 600 nm and 100 ml NB in a 300 ml Erlenmeyer flask were inoculated with preculture corresponding to an OD of 0.07. The bacteria were grown at 37 C for 18 h while stirring at 125 rpm. The bacteria then were harvested by centrifugation
FIG. 2. Plasma treatment (8 W, 5 slm argon, 0.1% oxygen) of E. coli strains inoculated on gel discs at 17 mm distance to the plasma nozzle outlet at (A) z6 106 cfu/cm2 and (B) z3 106 cfu/cm2. Different letters indicate significant (p < 0.05) differences between means.
Please cite this article in press as: Baier, M., et al., Inactivation of Shiga toxin-producing Escherichia coli O104:H4 using cold atmospheric pressure plasma, J. Biosci. Bioeng., (2015), http://dx.doi.org/10.1016/j.jbiosc.2015.01.003
VOL. xx, 2015 (4 C, 6 min, 4000 g) and re-suspended in phosphate buffered saline (PBS) to concentrations of 107 up to 109 CFU/ml. Small gel discs (A ¼ 1 cm2) of a 1% polysaccharide gel (Gelrite, Carl Roth, Karlsruhe, Germany) were cut out using a sterile cork borer. Gel discs were inoculated with 10 ml of the bacterial suspension and let dry under the safety hood until the suspension fluid was absorbed by the gel disc (z20 min). The samples were put on sterile glass slides and treated under the plasma-jet. After plasma treatment, gel discs were transferred to 2 ml Eppendorf vials, 1 ml PBS was added and bacteria were resuspended by shaking on an Eppendorf Thermomixer at 750 rpm for 5 min. Samples were serially diluted and plated out on LB plates. All treatment times were performed in triplicate and dilution series were plated twofold. Plates were incubated at 37 C and colonies were counted after 24 h.
PLASMA TREATMENT OF SHIGA TOXIN-PRODUCING E. COLI
3
by Duncan’s multiple range test (p < 0.05). In figures, the mean variability of data was indicated by the standard deviation and significant difference of groups of values was marked using different letters.
RESULTS
Statistical analysis All data were statistically analyzed (ANOVA) with WinSTAT (R. Fitch Software, Staufen, Germany). Significant differences were determined
Basic tests of the antibacterial capacity of the plasma-jet were performed on small gel discs (A z 1 cm2), as illustrated in Fig. 1. Bacteria were subjected to plasma in a gentle semi-direct mode at 17 mm distance to the plasma-jet outlet (Fig. 2), and in an intense direct mode at 5 mm distance (Fig. 3). Experiments were started at initial bacterial loads in the range of 106e107 cfu cm2. As observed in a set of pretests using pure argon plasma, obtained inactivation kinetics at 17 mm showed a continuous higher amount of survivors of the Shiga toxin-producing outbreak strain E. coli O104:H4 compared to the far more common E. coli O157:H7 (Fig. 2). Especially after the first 30 s treatment time, E. coli O157:H7 proved to be more susceptible to plasma leading to 1.7 0.4 log reduction, whereas E. coli O104:H4 only slowly declined by 1 log cycle after 1 min. The higher initial load of almost 107 cfu cm2 (Fig. 2A) featured a pronounced tailing without notable further inactivation after 2 min and seemed to mark the upper limit of efficacy of the plasma device at 17 mm. Inoculation of half the amount of bacteria (Fig. 2B) improved the inactivation of E. coli O104:H4 and allowed for a further log cycle to 2.0 0.1 logs after 2 min and 3.3 0.8 logs of E. coli O157:H7. In comparison, the non-pathogenic surrogate
FIG. 3. Plasma treatment (8 W, 5 slm argon, 0.1% oxygen) of E. coli strains inoculated on gel discs at 5 mm distance to the plasma nozzle outlet at (A) high initial bacterial load (z108 cfu cm2) and (B) low initial load (z105 cfu cm2). Different letters indicate significant (p < 0.05) differences between means.
FIG. 4. Plasma treatment (8 W, 5 slm argon, 0.1% oxygen) of E. coli strains inoculated on corn salad leaves (A) at 13 mm distance to the plasma nozzle outlet and 105 cfu cm2 and (B) at 17 mm distance and 104 cfu cm2. Different letters indicate significant (p < 0.05) differences between means.
Biological samples Corn salad was purchased at a local supermarket. Leaves were cut from the plants and soil and dirt was removed by rinsing with tap water. Retained water was dabbed off using tissue paper and single leaves were placed on sterile glass slides. An area of 1 cm2 was marked for inoculation. Bacterial suspension (5 ml) was inoculated on the upper side of 1% polysaccharide gel cylinders which were cut out of a 50 ml gel plate with a thin cork borer. After approximately 10 min, the suspension fluid was absorbed by the gel cylinder. Immediately before plasma treatment, a gel cylinder was gently grabbed with sterile forceps and the inoculated side was rubbed over the marked area of a corn salad leaf, simulating a contact contamination during post-harvest handling. Directly after treatment, leaves were transferred to 50 ml centrifugation tubes. PBS was added (10 ml) and bacteria were resuspended by shaking on a lab mixer for 1 min. Samples were serially diluted and plated out on CHROMagar (Mast Diagnostics, Reinfeld, Germany) supplemented with ampicillin (100 mg/ml) to prevent growth of indigenous bacteria of the lettuce surface. Plates were incubated at 37 C and colonies were counted after 24 h. All treatment times were performed in triplicate and dilution series were plated twofold.
Please cite this article in press as: Baier, M., et al., Inactivation of Shiga toxin-producing Escherichia coli O104:H4 using cold atmospheric pressure plasma, J. Biosci. Bioeng., (2015), http://dx.doi.org/10.1016/j.jbiosc.2015.01.003
4
BAIER ET AL.
J. BIOSCI. BIOENG.,
E. coli DSM 1116 could be reduced by 3.1 0.5 log and showed a susceptibility to plasma very similar to E. coli O157:H7. Further tests at lower initial counts down to 105 cfu cm2 did not lead to more inactivation at 17 mm (data not shown). In order to test the potential for higher and faster inactivation, treatment was intensified by reducing the distance to 5 mm, thus, enabling direct interaction of the plasma and the bacteria inoculated on the gel discs. Additionally, treatment time was increased to examine, if the weak inactivation at 107 cfu cm2 (Fig. 2A) can simply be overcome by longer exposure to the plasma. At high initial level (Fig. 3A), extended treatment for 4 min led to minor reductions of all three strains by 0.8e1.2 log cycles. Inactivation of E. coli O104:H4 was not significant and differences in susceptibility between the strains as observed in Fig. 2 disappeared. This lower efficiency after 4 min and the alignment of curves support the observation made in Fig. 2 that the antibacterial efficiency of the plasma-jet decreases when increasing the initial bacterial count. In contrast, at 105 cfu cm2 complete inactivation of all three strains was achieved after 2 min (Fig. 3B). Single samples without any cfu left were already obtained after 15 s, resulting in the high standard deviations between 15 and 45 s. Most efficient reduction of 3 log cycles was achieved within the first 15 s treatment time. Further inactivation below 10 cfu cm2 depended on the strain and was reached for E. coli O157:H7 (by 4.5 0.6 log), E. coli DSM 1116 (by 4.4 1.1 log), and E. coli O104:H4 (by 4.6 0.6 log) after 30 s, 45 s, and 60 s, respectively. Consequently, treatment at 5 mm resulted in pronouncedly improved inactivation of all strains at the moderate initial bacterial count of 105 cfu cm2 and enabled to overcome the shoulder of residual 103 cfu cm2 (Fig. 2B) down to complete inactivation. On the basis of these results the next step was to find out if relevant inactivation is also achievable under near-natural conditions. Therefore, the supporting medium of gel discs was exchanged with corn salad leaves. In consideration of the much less amount of available water on the leaf surface, possible effects of the dry argon gas flow without plasma ignition were examined to find out whether cell viability of E. coli O104:H4 is lost due to this additional drying stress. The sole argon gas flow, however, did not reduce the inoculated E. coli O104:H4 on the corn salad leaves (Fig. 4A), whereas after plasma ignition, the strain was inactivated by 1.9 0.1 log cycles after 30 s. Extended treatment for up to 2 min showed a pronounced tailing and did not lead to further significant reduction. According to the results gained on the gel discs (Fig. 2), a lower initial bacterial load below 105 cfu cm2 (Fig. 4B) was tested as a potential factor for an improved inactivation on the grooved leaf surface. At 104 cfu cm2 initial count and 17 mm distance, inactivation of E. coli O104:H4 slowed down at the level of residual 103 cfu cm2 (Fig. 4B). In contrast to 105 cfu cm2, extended treatment at 104 cfu cm2 resulted in further decline and ended at the detection limit after 3.3 1.1 log cycles inactivation. E. coli O157:H7 did not show a similar shoulder, but was faster inactivated by 3.2 1.1 log cycles to the detection limit after 60 s. A steep initial decline such as at 5 mm (Fig. 3B) was not observed. Moreover, the shape of the curves resembled those at 106 cfu cm2 on the gel discs (Fig. 2B) made at the same treatment distance of 17 mm.
DISCUSSION Plasma proved to be capable to reduce pathogenic as well as non-pathogenic E. coli examined in this study. Achieved inactivation and efficiency strongly depended on the experimental conditions. Intense treatment at 5 mm in direct contact with plasma filaments, enabling maximum interaction with the full range of antibacterial plasma components, led to the most efficient and complete inactivation of 105 cfu cm2 of all strains. At the high load
of 108 cfu cm2, however, the used plasma device seemed to be limited in its capacity to minor reductions after long treatment duration of 4 min. Similar results and a decreasing inactivation efficiency of plasma at higher initial cell concentrations were reported by Fernández et al. (21) on Salmonella enterica. In the gentle treatment mode at 17 mm, showing a gap of some millimeters between the tip of the glowing plasma and the sample surface, differences in susceptibility between the E. coli strains became apparent and inactivation showed a biphasic characteristic of a faster first part followed by slower inactivation or even tailing. The STEC outbreak strain O104:H4 with both STEC and enteroaggregative properties turned out to be more stable against plasma on gel discs (Fig. 2) and on lettuce leaves (Fig. 4B). This stability seemed to be more pronounced the lower the initial load and the softer the plasma treatment was. A possible cause might be found in outer differences of its aggregative adherence fimbriae (AAF) acting as a potential barrier for plasma species. The morphological properties of this strain might have supported the cells to entangle themselves, leading to shelter effects for deeper layers of cells by the topmost layer. This may be due to the particular strong expression of curli-fimbriae (9). Reductions down to 103 cfu cm2 after 30 s and to z10 cfu cm2 after 2 min on corn salad are promising results on the way to an actively enhanced food safety of heat-sensitive fresh produce. The low infective dose of STEC (14) and the ability of E. coli O104:H4 to enter a stress-induced state of viable but non-culturable (22), however, remain serious issues, which need further research on residual or hidden survivors and improvements to reach full inactivation. Inactivation of the STEC strain O157:H7 on corn salad after 1 min was very similar to that on apples reported by Niemira and Sites (12). It also strongly resembled the response to plasma of the nonpathogenic E. coli DSM 1116 on corn salad (18). In contrast to E. coli O104:H4, the detection limit of E. coli O157:H7 was reached after 1 min. Using the same plasma-jet in a previous study, treatment durations at 17 mm for up to 1 min proved to be suitable for gentle application of plasma on corn salad leaves without detrimental effects (18). This inactivation success, therefore, constitutes an essential step of harmonizing antibacterial efficiency with the limits of quality retention. The plasma did not affect the bacteria following a log-linear mechanism. Moreover, the main part of the inactivation was obtained within the initial 30 s. Critzer et al. (11) also observed biphasic kinetics of plasma inactivation of E. coli O157:H7 on apples. Inactivation success of 3 logs using the argon plasma-jet, however, resembled more the findings of Niemira and Sites (12) using the gas flow of a gliding arc plasma. This higher efficiency may be simply attributed to the shorter distance of plasma generation zone and sample surface as observed in the present study at 5 mm in comparison to 17 mm (see Figs. 2B and 3B). The gel discs as the supporting medium for bacteria are a useful auxiliary means for exploration of efficient process parameters and limits of efficacy for an individual plasma device to be tested. As a simplified model system, high numbers of samples can be analyzed in the absence of drying stress, competing indigenous micro-flora or other inhibiting factors. The role of available water on the sample surface for inactivation efficiency of plasma, however, remained unclear in this study. The moist surface of gel discs was expected to yield improved inactivation, similar to the findings on agar medium compared to filter paper (23). Equivalent amounts of log cycles reduction, though, seemed to be delayed by 1 min on the gel discs (Fig. 2B) compared to the waxy surface of corn salad (Fig. 4B). Whereas the sole dry gas flow without plasma ignition did not reduce the bacteria (Fig. 4A), beneficial effects due to additional drying stress on the corn salad leaves cannot be excluded as the cause for faster inactivation. In contrast, surficial warming of samples during treatment was not considered to have an impact on bacteria.
Please cite this article in press as: Baier, M., et al., Inactivation of Shiga toxin-producing Escherichia coli O104:H4 using cold atmospheric pressure plasma, J. Biosci. Bioeng., (2015), http://dx.doi.org/10.1016/j.jbiosc.2015.01.003
VOL. xx, 2015
PLASMA TREATMENT OF SHIGA TOXIN-PRODUCING E. COLI
Using the same plasma-jet and treatment conditions as in Fig. 4B, temperature on the corn salad leaves did not exceed 35 C (18). The similar susceptibility to plasma of E. coli O157:H7 and E. coli DSM 1116 constitutes useful information which can help to reduce microbiological efforts and accelerate plasma research. As most new plasma sources are constructed in physics laboratories as prototypes, they often lack appropriate housings in terms of hygienic design. For this early stage the non-pathogenic surrogate strain proved to be suitable for choosing and testing plasma sources under safe work conditions. Examinations of the pathogen, therefore, can be postponed to advanced stages and limited to adjustments for optimized process settings. Notwithstanding the promising results on E. coli O157:H7, improvements for a faster inactivation of E. coli O104:H4 and especially overcoming the shoulder at 103 cfu cm2 on the produce surface are highly desirable. The antibacterial efficiency of a plasma-jet system can be enhanced by increasing the plasma power as results on E. coli DSM 1116 show (24). A concomitant rise in process temperature and potentially detrimental impacts on product quality in turn limit the maximum of applicable process power (25). A shorter treatment distance down the range of the plasma discharge also increases the antimicrobial efficiency and led to the fastest reductions in this study (Fig. 3B). Measurements show a steep increase of detected UV emission at distances lower than 12 mm, but also a rise in temperature to over 60 C at the outlet of the quartz capillary (20). How near a sample may be placed to the plasma generation zone particularly depends on the specific characteristics of the sample surface and needs to be defined by quality measurements. Using the example of thin corn salad leaves, treatment at 17 mm proved to be a gentle and suitable setting whereas shorter distances were accompanied with increasing quality loss (18). The latter may still be an option for more robust surfaces such as the pericarps of melons or in the field of plasma disinfection of abiotic surfaces like work spaces and conveyors of food production sites. Eventually, the composition of the process gas can be varied, and effects of relative humidity tested, to screen for the optimum of inactivation efficiency. In conclusion, plasma treatment is capable of reducing the STEC outbreak strain E. coli O104:H4. Differences of its response to plasma on gel compared to corn salad leaves emphasize the need to account for matrix effects of any biological surface to be disinfected. Treatment times, necessary to reduce E. coli O104:H4 below its infective dose, need to be adapted to the specific commodity and its expected level of initial bacterial count. Gentle treatment intensities, suitable for corn salad leaves, are more effective on E. coli O157:H7 than on E. coli O104:H4. Inactivation efficiency needs to be improved to overcome delayed microbial reductions and tailing effects. Finally, numbers of samples should be increased as high as possible to ensure the reliable killing of low residual contaminations of the highly infective pathogens below standard sampling. ACKNOWLEDGMENTS The authors thank Susanne Klocke of ATB Potsdam-Bornim for her excellent technical assistance. This work was supported by the German Federal Ministry of Food, Agriculture and Consumer Protection (BMELV) and the Federal Office for Agriculture and Food (BLE) within the innovation program (FKZ 28-1-63.003e07). References 1. Hartung, M. and Käsbohrer, A. (Eds.): Erreger von Zoonosen in Deutschland im Jahr 2009. Bundesinstitut für Risikobewertung, Berlin (2011). 2. Sivapalasingam, A., Friedman, C. R., Cohen, L., and Tauxe, R.: Fresh produce: a growing cause of outbreaks of foodborne illness in the United States, 1973 through 1997, J. Food Prot., 67, 2342e2353 (2004).
5
3. Warriner, K., Huber, A., Namvar, A., Fan, W., and Dunfield, K.: Recent advances in the microbial safety of fresh fruits and vegetables, Adv. Food Nutr. Res., 57, 155e208 (2009). 4. Scharff, R. L.: Health-related costs from foodborne illness in the United States. The Produce Safety Project at Georgetown University. The Pew Charitable Trusts, Philadelphia. http://www.pewtrusts.org/w/media/legacy/uploadedfiles/phg/ content_level_pages/reports/PSPScharff20v9pdf.pdf (2010). 5. Michino, H., Araki, K., Minami, S., Tayaka, S., Sakai, N., Ono, A., and Yanagawa, H.: Massive outbreak of Escherichia coli O157:H7 infection in school children in Sakai City, Japan, associated with consumption of white radish sprouts, Am. J. Epidemiol., 150, 787e796 (1999). 6. Wieler, L. H., Semmler, T., Eichhorn, I., Antao, E. M., Kinnemann, B., Geue, L., Karch, H., Guenther, S., and Bethe, A.: No evidence of the Shiga toxinproducing E. coli O104:H4 outbreak strain or enteroaggregative E. coli (EAEC) found in cattle feaces in northern Germany, the hotspots of the 2011 HUS outbreak area, Gut Pathog., 3, 17 (2011). 7. Bielaszewska, M., Mellmann, A., Zhang, W., Köck, R., Fruth, A., Bauwens, A., Peters, G., and Karch, H.: Characterisation of the Escherichia coli strain associated with an outbreak of haemolytic uraemic syndrome in Germany, 2011: a microbiological study, Lancet, 7, 671e676 (2011). 8. Richter, A. M., Povolotsky, T. L., Wieler, L. W., and Hengge, R.: Cyclic-di-GMP signalling and biofilm-related properties of the Shiga toxin-producing 2011 German outbreak Escherichia coli O104:H4, EMBO Mol. Med., 6, 1622e1637 (2014). 9. Robert Koch Institute: Report: final presentation and evaluation of epidemiological findings in the EHEC O104:H4 outbreak, Germany 2011. Robert Koch Institute, Berlin (2011). 10. Keener, K. M.: Atmospheric non-equilibrium plasma, Encycl. Agric. Food Biol. Eng., 1, 1e5 (2008). 11. Critzer, F. J., Kelly-Winterberg, K., South, S. L., and Golden, D. A.: Atmospheric plasma inactivation of foodborne pathogens on fresh produce surfaces, J. Food Prot., 70, 2290e2296 (2007). 12. Niemira, B. A. and Sites, J.: Cold plasma inactivates Salmonella Stanley and Escherichia coli O157:H7 inoculated on golden delicious apples, J. Food Prot., 71, 1357e1365 (2008). 13. Montenegro, J., Ruan, R., Ma, H., and Chen, P.: Inactivation of E. coli O157:H7 using a pulsed nonthermal plasma system, J. Food Sci., 67, 646e648 (2002). 14. Grant, M. A., Hedberg, C., Johnson, R., Harris, J., Logue, C. M., Meng, J., Sofos, J. N., and Dickson, J. S.: The significance of non-O157 Shiga toxinproducing Escherichia coli in food, Food Prot. Trends, 1, 33e45 (2011). 15. Hartung, M. and Käsbohrer, A. (Eds.): Erreger von Zoonosen in Deutschland im Jahr 2010. Bundesinstitut für Risikobewertung, Berlin (2012). 16. Hartung, M. and Käsbohrer, A. (Eds.): Erreger von Zoonosen in Deutschland im Jahr 2011. Bundesinstitut für Risikobewertung, Berlin (2013). 17. European Food Safety Authority (EFSA): ECDC The European Union summary report on trends and sources of zoonoses, zoonotic agents and food-borne outbreaks in 2011, EFSA J., 11, 3129 (2013). 18. Baier, M., Görgen, M., Ehlbeck, J., Knorr, D., Herppich, W. B., and Schlüter, O.: Non-thermal atmospheric pressure plasma: screening for gentle process conditions and antibacterial efficiency on perishable fresh produce, Innov. Food Sci. Emerg., 22, 147e157 (2014). 19. Scheutz, F., Møller-Nielsen, E., Frimodt-Møller, J., Boisen, N., Morabito, S., Tozzoli, R., Nataro, J. P., and Caprioli, A.: Characteristics of the enteroaggregative Shiga toxin/verotoxinproducing Escherichia coli O104:H4 strain causing the outbreak of haemolytic uraemic syndrome in Germany, May to June 2011, Euro Surveill., 16, pii=19889 (2011). 20. Weltmann, K. D., Kindel, E., Brandenburg, R., Meyer, C., Bussiahn, R., Wilke, C., and von Woedtke, T.: Atmospheric pressure plasma jet for medical therapy: plasma parameters and risk estimation, Contrib. Plasm. Phys., 49, 631e640 (2009). 21. Fernández, A., Shearer, N., Wilson, D. R., and Thompson, A.: Effect of microbial loading on the efficiency of cold atmospheric gas plasma inactivation of Samonella enterica serovar Typhimurium, Int. J. Food Microbiol., 152, 175e180 (2012). 22. Aurass, P., Prager, R., and Flieger, A.: EHEC/EAEC O104:H4 strain linked with the 2011 German outbreak of haemolytic uremic syndrome enters into the viable but non-culturable state in response to various stresses and resuscitates upon stress relief, Environ. Microbiol., 13, 3139e3148 (2011). 23. Huang, C.: An experimental investigation of low temperature plasma sterilization, treatment, and polymerization processes, PhD thesis. University of Missouri-Columbia (2006). 24. Fröhling, A., Baier, M., Ehlbeck, J., Knorr, D., and Schlüter, O.: Atmospheric pressure plasma treatment of Listeria innocua and Escherichia coli at polysaccharide surfaces: inactivation kinetics and flow cytometric characterization, Innov. Food Sci. Emerg., 13, 142e150 (2012). 25. Baier, M., Foerster, J., Schnabel, U., Ehlbeck, J., Knorr, D., Herppich, W. B., and Schlüter, O.: Direct non-thermal plasma treatment for the sanitation of fresh corn salad leaves. Evaluation of physical and physiological effects and antimicrobial efficacy, Postharvest Biol. Technol., 84, 81e87 (2013).
Please cite this article in press as: Baier, M., et al., Inactivation of Shiga toxin-producing Escherichia coli O104:H4 using cold atmospheric pressure plasma, J. Biosci. Bioeng., (2015), http://dx.doi.org/10.1016/j.jbiosc.2015.01.003
Inactivation of Shiga toxin-producing Escherichia coli O104:H4 using cold atmospheric pressure plasma Matthias Baier,1 Traute Janßen,2 Lothar H. Wieler,2 Jörg Ehlbeck,3 Dietrich Knorr,4 and Oliver Schlüter1, * Leibniz Institute for Agricultural Engineering, Max-Eyth-Allee 100, 14469 Potsdam-Bornim, Germany,1 Centre for Infection Medicine, Institute of Microbiology and Epizootics, Freie Universität Berlin, Robert-von-Ostertag-Str. 7-13, 14163 Berlin, Germany,2 Leibniz Institute for Plasma Science and Technology, Felix-Hausdorff-Straße 2, 17489 Greifswald, Germany,3 and Department of Food Biotechnology and Food Process Engineering, Berlin University of Technology, Königin-Luise-Straße 22, 14195 Berlin, Germany4 Received 19 August 2014; accepted 6 January 2015 Available online xxx
From cultivation to the end of the post-harvest chain, heat-sensitive fresh produce is exposed to a variety of sources of pathogenic microorganisms. If contaminated, effective gentle means of sanitation are necessary to reduce bacterial pathogen load below their infective dose. The occurrence of rare or new serotypes raises the question of their tenacity to inactivation processes. In this study the antibacterial efficiency of cold plasma by an atmospheric pressure plasma-jet was examined against the Shiga toxin-producing outbreak strain Escherichia coli O104:H4. Argon was transformed into non-thermal plasma at a power input of 8 W and a gas flow of 5 L minL1. Basic tests were performed on polysaccharide gel discs, including the more common E. coli O157:H7 and non-pathogenic E. coli DSM 1116. At 5 mm treatment distance and 105 cfu cmL2 initial bacterial count, plasma reduced E. coli O104:H4 after 60 s by 4.6 ± 0.6 log, E. coli O157:H7 after 45 s by 4.5 ± 0.6 log, and E. coli DSM 1116 after 30 s by 4.4 ± 1.1 log. On the surface of corn salad leaves, gentle plasma application at 17 mm reduced 104 cfu cmL2 of E. coli O104:H4 by 3.3 ± 1.1 log after 2 min, whereas E. coli O157:H7 was inactivated by 3.2 ± 1.1 log after 60 s. In conclusion, plasma treatment has the potential to reduce pathogens such as E. coli O104:H4 on the surface of fresh produce. However, a serotype-specific adaptation of the process parameters is required. Ó 2015, The Society for Biotechnology, Japan. All rights reserved. [Key words: Plasma-jet; Escherichia coli O157:H7; Antibacterial treatment; Food safety; Fresh produce; Corn salad]
Outbreaks of food-borne illness due to pathogenic microorganisms were traditionally associated with the consumption of protein-rich foods of animal origin. Over the last decades, awareness has risen for the growing role of fresh produce as vehicle for hazardous microbes. Although percentages of detected pathogens remain higher for foods of animal origin such as ground meat, also fruits and vegetables harbor a latent risk to consumers’ health (1,2). Despite their lower frequency, produce-related outbreaks caused, e.g., by Salmonella and Escherichia coli O157:H7 are consistently large (3) and recent estimations of outbreaks and numbers of cases assume an essential share of this group of foods as a responsible source for food-borne illness (4). The largest outbreak of E. coli O157:H7 involving more than 12,000 cases and 12 fatalities was linked to raw radish sprouts in Japan in 1996 (5). While most frequently outbreaks are reported to be caused by E. coli O157:H7, in 2011 a large outbreak of Shiga toxin-producing E. coli O104:H4 occurred in Europe, a strain with strong enteroaggregative features. In contrast to O157:H7, this strain does not seem to originate from the typical cattle reservoir (6). The outbreak was epidemiologically traced back to fenugreek sprouts as the most likely vehicle. This strain belongs to the HUSEC041 clone of sequence type 678 (ST678) and exhibited a combination of typical features of both Shiga toxin-producing E. coli (STEC) and
* Corresponding author. Tel.: þ49 331 5699 613; fax: þ49 331 5699 849. E-mail address: [email protected] (O. Schlüter).
enteroaggregative E. coli (EAEC). In addition to HUSEC041, the outbreak strain O104:H4 also produced extended-spectrum blactamase (ESBL) (7). Latest research was able to proof this strain to display a particular biofilm phenotype, in that it produces high amounts of curli fimbriae, which are known to cause strong inflammation (8). These properties led to an extraordinary high number of 855 cases of hemolytic-uremic syndrome (HUS), 2987 cases of acute gastroenteritis, and 53 fatalities (9). Despite the principally higher risk potential of meat products, they have the advantage of an efficient antimicrobial reduction by appropriate heating during their preparation. Heat sensitive fresh produces, destined to be eaten raw, lack such an effective step and are limited to only minor reductions by washing procedures. Whereas use of disinfecting chemicals harbors the risk of residues, non-thermal techniques such as irradiation demand highest levels of safety at work or face low consumer acceptance. A promising physical approach is the application of cold or non-thermal nonequilibrium plasma (NTP). NTP can be generated by subjecting a process gas to a strong electric field at atmospheric pressure or pressures both above and below ambient. This leads to the partial ionization of the process gas molecules and the concomitant formation of other reactive chemical species, such as ions and radicals, heat (gas temperature less than 500 K), and UV light, which all together potentially react with the microbes adhesive to the food surfaces (10). Studies of the antimicrobial efficacy of different plasma sources on STEC have already been conducted. Critzer et al. (11)
1389-1723/$ e see front matter Ó 2015, The Society for Biotechnology, Japan. All rights reserved. http://dx.doi.org/10.1016/j.jbiosc.2015.01.003
Please cite this article in press as: Baier, M., et al., Inactivation of Shiga toxin-producing Escherichia coli O104:H4 using cold atmospheric pressure plasma, J. Biosci. Bioeng., (2015), http://dx.doi.org/10.1016/j.jbiosc.2015.01.003
2
BAIER ET AL.
treated E. coli O157:H7 with the exhaust of a one atmosphere uniform glow discharge plasma (OAUGDP) on cv. Red Delicious apples and obtained inactivation of over 2 logs after 2 min exposure. Dried air, ionized by a gliding arc plasma inactivated E. coli O157:H7 on the surface of cv. Golden Delicious apples by 3.4 log cfu ml1 after 1 min (12). Direct treatment of apple juice using a needle-to-plate arrangement of electrodes was reported by Montenegro et al. (13). This treatment in liquid resulted in more than 5 logs reduction of E. coli O157:H7 after 40 s. Focused on E. coli O157:H7 only, the potential of NTP on other non-O157 STEC strains is still unknown, as is the actual share of these strains in all cases of food-borne illness. Difficulties in clear analytical distinctions from non-pathogenic commensal E. coli limit the gain of accurate information on non-O157 STECs and specific outbreak investigations are likely to remain the main source for further knowledge (14). In this context E. coli O104:H4, which has not been detected by monitoring from 2009 to 2011 (1,15e17), exhibits an example of a rare non-O157 STEC of extremely high virulence. It was shown that plasma can be used to treat a leafy green such as corn salad under gentle conditions, reaching 3 log cycles reductions of non-pathogenic E. coli DSM 1116 after 20 s (18). Particularly with regard to the special aggregative ability of E. coli O104:H4 (19), the question was raised if cold plasma is capable to achieve effective reductions of this serotype as authors previously reported on E. coli O157:H7. Therefore, the aim of this study was to determine the antibacterial efficacy of an atmospheric pressure plasma-jet on the enteroaggregative STEC outbreak strain E. coli O104:H4. For purposes of comparison, a serovar of the much more common E. coli O157:H7 was included in this study. Treatment distance and initial count were varied in a set of basic tests on a polysaccharide-gel and promising process parameters were applied on artificially inoculated corn salad leaves.
FIG. 1. Experimental setup of the atmospheric pressure plasma-jet.
J. BIOSCI. BIOENG., MATERIALS AND METHODS Atmospheric pressure plasma-jet The atmospheric pressure plasma-jet was generated at the tip of a pin-type electrode mounted concentrically in a quartz capillary (Fig. 1). Argon with 0.1% oxygen were delivered by a gas supply unit and passed into the capillary (inner diameter 1.6 mm) at a flow rate of 5 slm. A DC power supply (system power: 8 W at 220 V, 50/60 Hz) delivering a high frequency (HF) voltage (1.1 MHz, 2e6 kVpp) was coupled to the pin-type electrode. The process gas was transformed into plasma at the tip of the centered electrode and driven out of the capillary into the surrounding air. The plasma had a length of up to 10 mm and thus, enabled intense direct plasma application at 10 mm treatment distance or below. At 12 mm or more, a gap to the sample surface emerged, which inhibited direct interaction of glowing plasma filaments and samples, resulting in an attenuated semi-direct treatment mode. A detailed description on the device and its plasma parameters is given in Weltmann et al. (20). The plasma-jet (kinpen, neoplas tools GmbH, Greifswald, Germany) gained the CE marking for meeting EU consumer safety, health or environmental requirements. Bacterial strains The E. coli strains used in this study were the HUS outbreak strain RKI II-2027 (ST678, O104:H4, kindly provided by Angelika Fruth, Robert KochInstitut Wernigerode), isolated from a patient in 2011, the STEC reference strain EDL933 (ST11, O157:H7), isolated in 1983, and the non-pathogenic E. coli reference strain DSM 1116 (also referred to as ATCC 9637, CCM 2024, NCIB 8666, or NRRL B766). All strains were stored at 80 C in glycerol in brain heart infusion (BHI). Gel samples Luria Broth (LB) agar was used for culturing the E. coli strains at 37 C for 24 h. A colony was taken and transferred to 5 ml nutrient broth (NB). This pre-culture was bred shaking at 37 C for 24 h. Optical density (OD) was measured at 600 nm and 100 ml NB in a 300 ml Erlenmeyer flask were inoculated with preculture corresponding to an OD of 0.07. The bacteria were grown at 37 C for 18 h while stirring at 125 rpm. The bacteria then were harvested by centrifugation
FIG. 2. Plasma treatment (8 W, 5 slm argon, 0.1% oxygen) of E. coli strains inoculated on gel discs at 17 mm distance to the plasma nozzle outlet at (A) z6 106 cfu/cm2 and (B) z3 106 cfu/cm2. Different letters indicate significant (p < 0.05) differences between means.
Please cite this article in press as: Baier, M., et al., Inactivation of Shiga toxin-producing Escherichia coli O104:H4 using cold atmospheric pressure plasma, J. Biosci. Bioeng., (2015), http://dx.doi.org/10.1016/j.jbiosc.2015.01.003
VOL. xx, 2015 (4 C, 6 min, 4000 g) and re-suspended in phosphate buffered saline (PBS) to concentrations of 107 up to 109 CFU/ml. Small gel discs (A ¼ 1 cm2) of a 1% polysaccharide gel (Gelrite, Carl Roth, Karlsruhe, Germany) were cut out using a sterile cork borer. Gel discs were inoculated with 10 ml of the bacterial suspension and let dry under the safety hood until the suspension fluid was absorbed by the gel disc (z20 min). The samples were put on sterile glass slides and treated under the plasma-jet. After plasma treatment, gel discs were transferred to 2 ml Eppendorf vials, 1 ml PBS was added and bacteria were resuspended by shaking on an Eppendorf Thermomixer at 750 rpm for 5 min. Samples were serially diluted and plated out on LB plates. All treatment times were performed in triplicate and dilution series were plated twofold. Plates were incubated at 37 C and colonies were counted after 24 h.
PLASMA TREATMENT OF SHIGA TOXIN-PRODUCING E. COLI
3
by Duncan’s multiple range test (p < 0.05). In figures, the mean variability of data was indicated by the standard deviation and significant difference of groups of values was marked using different letters.
RESULTS
Statistical analysis All data were statistically analyzed (ANOVA) with WinSTAT (R. Fitch Software, Staufen, Germany). Significant differences were determined
Basic tests of the antibacterial capacity of the plasma-jet were performed on small gel discs (A z 1 cm2), as illustrated in Fig. 1. Bacteria were subjected to plasma in a gentle semi-direct mode at 17 mm distance to the plasma-jet outlet (Fig. 2), and in an intense direct mode at 5 mm distance (Fig. 3). Experiments were started at initial bacterial loads in the range of 106e107 cfu cm2. As observed in a set of pretests using pure argon plasma, obtained inactivation kinetics at 17 mm showed a continuous higher amount of survivors of the Shiga toxin-producing outbreak strain E. coli O104:H4 compared to the far more common E. coli O157:H7 (Fig. 2). Especially after the first 30 s treatment time, E. coli O157:H7 proved to be more susceptible to plasma leading to 1.7 0.4 log reduction, whereas E. coli O104:H4 only slowly declined by 1 log cycle after 1 min. The higher initial load of almost 107 cfu cm2 (Fig. 2A) featured a pronounced tailing without notable further inactivation after 2 min and seemed to mark the upper limit of efficacy of the plasma device at 17 mm. Inoculation of half the amount of bacteria (Fig. 2B) improved the inactivation of E. coli O104:H4 and allowed for a further log cycle to 2.0 0.1 logs after 2 min and 3.3 0.8 logs of E. coli O157:H7. In comparison, the non-pathogenic surrogate
FIG. 3. Plasma treatment (8 W, 5 slm argon, 0.1% oxygen) of E. coli strains inoculated on gel discs at 5 mm distance to the plasma nozzle outlet at (A) high initial bacterial load (z108 cfu cm2) and (B) low initial load (z105 cfu cm2). Different letters indicate significant (p < 0.05) differences between means.
FIG. 4. Plasma treatment (8 W, 5 slm argon, 0.1% oxygen) of E. coli strains inoculated on corn salad leaves (A) at 13 mm distance to the plasma nozzle outlet and 105 cfu cm2 and (B) at 17 mm distance and 104 cfu cm2. Different letters indicate significant (p < 0.05) differences between means.
Biological samples Corn salad was purchased at a local supermarket. Leaves were cut from the plants and soil and dirt was removed by rinsing with tap water. Retained water was dabbed off using tissue paper and single leaves were placed on sterile glass slides. An area of 1 cm2 was marked for inoculation. Bacterial suspension (5 ml) was inoculated on the upper side of 1% polysaccharide gel cylinders which were cut out of a 50 ml gel plate with a thin cork borer. After approximately 10 min, the suspension fluid was absorbed by the gel cylinder. Immediately before plasma treatment, a gel cylinder was gently grabbed with sterile forceps and the inoculated side was rubbed over the marked area of a corn salad leaf, simulating a contact contamination during post-harvest handling. Directly after treatment, leaves were transferred to 50 ml centrifugation tubes. PBS was added (10 ml) and bacteria were resuspended by shaking on a lab mixer for 1 min. Samples were serially diluted and plated out on CHROMagar (Mast Diagnostics, Reinfeld, Germany) supplemented with ampicillin (100 mg/ml) to prevent growth of indigenous bacteria of the lettuce surface. Plates were incubated at 37 C and colonies were counted after 24 h. All treatment times were performed in triplicate and dilution series were plated twofold.
Please cite this article in press as: Baier, M., et al., Inactivation of Shiga toxin-producing Escherichia coli O104:H4 using cold atmospheric pressure plasma, J. Biosci. Bioeng., (2015), http://dx.doi.org/10.1016/j.jbiosc.2015.01.003
4
BAIER ET AL.
J. BIOSCI. BIOENG.,
E. coli DSM 1116 could be reduced by 3.1 0.5 log and showed a susceptibility to plasma very similar to E. coli O157:H7. Further tests at lower initial counts down to 105 cfu cm2 did not lead to more inactivation at 17 mm (data not shown). In order to test the potential for higher and faster inactivation, treatment was intensified by reducing the distance to 5 mm, thus, enabling direct interaction of the plasma and the bacteria inoculated on the gel discs. Additionally, treatment time was increased to examine, if the weak inactivation at 107 cfu cm2 (Fig. 2A) can simply be overcome by longer exposure to the plasma. At high initial level (Fig. 3A), extended treatment for 4 min led to minor reductions of all three strains by 0.8e1.2 log cycles. Inactivation of E. coli O104:H4 was not significant and differences in susceptibility between the strains as observed in Fig. 2 disappeared. This lower efficiency after 4 min and the alignment of curves support the observation made in Fig. 2 that the antibacterial efficiency of the plasma-jet decreases when increasing the initial bacterial count. In contrast, at 105 cfu cm2 complete inactivation of all three strains was achieved after 2 min (Fig. 3B). Single samples without any cfu left were already obtained after 15 s, resulting in the high standard deviations between 15 and 45 s. Most efficient reduction of 3 log cycles was achieved within the first 15 s treatment time. Further inactivation below 10 cfu cm2 depended on the strain and was reached for E. coli O157:H7 (by 4.5 0.6 log), E. coli DSM 1116 (by 4.4 1.1 log), and E. coli O104:H4 (by 4.6 0.6 log) after 30 s, 45 s, and 60 s, respectively. Consequently, treatment at 5 mm resulted in pronouncedly improved inactivation of all strains at the moderate initial bacterial count of 105 cfu cm2 and enabled to overcome the shoulder of residual 103 cfu cm2 (Fig. 2B) down to complete inactivation. On the basis of these results the next step was to find out if relevant inactivation is also achievable under near-natural conditions. Therefore, the supporting medium of gel discs was exchanged with corn salad leaves. In consideration of the much less amount of available water on the leaf surface, possible effects of the dry argon gas flow without plasma ignition were examined to find out whether cell viability of E. coli O104:H4 is lost due to this additional drying stress. The sole argon gas flow, however, did not reduce the inoculated E. coli O104:H4 on the corn salad leaves (Fig. 4A), whereas after plasma ignition, the strain was inactivated by 1.9 0.1 log cycles after 30 s. Extended treatment for up to 2 min showed a pronounced tailing and did not lead to further significant reduction. According to the results gained on the gel discs (Fig. 2), a lower initial bacterial load below 105 cfu cm2 (Fig. 4B) was tested as a potential factor for an improved inactivation on the grooved leaf surface. At 104 cfu cm2 initial count and 17 mm distance, inactivation of E. coli O104:H4 slowed down at the level of residual 103 cfu cm2 (Fig. 4B). In contrast to 105 cfu cm2, extended treatment at 104 cfu cm2 resulted in further decline and ended at the detection limit after 3.3 1.1 log cycles inactivation. E. coli O157:H7 did not show a similar shoulder, but was faster inactivated by 3.2 1.1 log cycles to the detection limit after 60 s. A steep initial decline such as at 5 mm (Fig. 3B) was not observed. Moreover, the shape of the curves resembled those at 106 cfu cm2 on the gel discs (Fig. 2B) made at the same treatment distance of 17 mm.
DISCUSSION Plasma proved to be capable to reduce pathogenic as well as non-pathogenic E. coli examined in this study. Achieved inactivation and efficiency strongly depended on the experimental conditions. Intense treatment at 5 mm in direct contact with plasma filaments, enabling maximum interaction with the full range of antibacterial plasma components, led to the most efficient and complete inactivation of 105 cfu cm2 of all strains. At the high load
of 108 cfu cm2, however, the used plasma device seemed to be limited in its capacity to minor reductions after long treatment duration of 4 min. Similar results and a decreasing inactivation efficiency of plasma at higher initial cell concentrations were reported by Fernández et al. (21) on Salmonella enterica. In the gentle treatment mode at 17 mm, showing a gap of some millimeters between the tip of the glowing plasma and the sample surface, differences in susceptibility between the E. coli strains became apparent and inactivation showed a biphasic characteristic of a faster first part followed by slower inactivation or even tailing. The STEC outbreak strain O104:H4 with both STEC and enteroaggregative properties turned out to be more stable against plasma on gel discs (Fig. 2) and on lettuce leaves (Fig. 4B). This stability seemed to be more pronounced the lower the initial load and the softer the plasma treatment was. A possible cause might be found in outer differences of its aggregative adherence fimbriae (AAF) acting as a potential barrier for plasma species. The morphological properties of this strain might have supported the cells to entangle themselves, leading to shelter effects for deeper layers of cells by the topmost layer. This may be due to the particular strong expression of curli-fimbriae (9). Reductions down to 103 cfu cm2 after 30 s and to z10 cfu cm2 after 2 min on corn salad are promising results on the way to an actively enhanced food safety of heat-sensitive fresh produce. The low infective dose of STEC (14) and the ability of E. coli O104:H4 to enter a stress-induced state of viable but non-culturable (22), however, remain serious issues, which need further research on residual or hidden survivors and improvements to reach full inactivation. Inactivation of the STEC strain O157:H7 on corn salad after 1 min was very similar to that on apples reported by Niemira and Sites (12). It also strongly resembled the response to plasma of the nonpathogenic E. coli DSM 1116 on corn salad (18). In contrast to E. coli O104:H4, the detection limit of E. coli O157:H7 was reached after 1 min. Using the same plasma-jet in a previous study, treatment durations at 17 mm for up to 1 min proved to be suitable for gentle application of plasma on corn salad leaves without detrimental effects (18). This inactivation success, therefore, constitutes an essential step of harmonizing antibacterial efficiency with the limits of quality retention. The plasma did not affect the bacteria following a log-linear mechanism. Moreover, the main part of the inactivation was obtained within the initial 30 s. Critzer et al. (11) also observed biphasic kinetics of plasma inactivation of E. coli O157:H7 on apples. Inactivation success of 3 logs using the argon plasma-jet, however, resembled more the findings of Niemira and Sites (12) using the gas flow of a gliding arc plasma. This higher efficiency may be simply attributed to the shorter distance of plasma generation zone and sample surface as observed in the present study at 5 mm in comparison to 17 mm (see Figs. 2B and 3B). The gel discs as the supporting medium for bacteria are a useful auxiliary means for exploration of efficient process parameters and limits of efficacy for an individual plasma device to be tested. As a simplified model system, high numbers of samples can be analyzed in the absence of drying stress, competing indigenous micro-flora or other inhibiting factors. The role of available water on the sample surface for inactivation efficiency of plasma, however, remained unclear in this study. The moist surface of gel discs was expected to yield improved inactivation, similar to the findings on agar medium compared to filter paper (23). Equivalent amounts of log cycles reduction, though, seemed to be delayed by 1 min on the gel discs (Fig. 2B) compared to the waxy surface of corn salad (Fig. 4B). Whereas the sole dry gas flow without plasma ignition did not reduce the bacteria (Fig. 4A), beneficial effects due to additional drying stress on the corn salad leaves cannot be excluded as the cause for faster inactivation. In contrast, surficial warming of samples during treatment was not considered to have an impact on bacteria.
Please cite this article in press as: Baier, M., et al., Inactivation of Shiga toxin-producing Escherichia coli O104:H4 using cold atmospheric pressure plasma, J. Biosci. Bioeng., (2015), http://dx.doi.org/10.1016/j.jbiosc.2015.01.003
VOL. xx, 2015
PLASMA TREATMENT OF SHIGA TOXIN-PRODUCING E. COLI
Using the same plasma-jet and treatment conditions as in Fig. 4B, temperature on the corn salad leaves did not exceed 35 C (18). The similar susceptibility to plasma of E. coli O157:H7 and E. coli DSM 1116 constitutes useful information which can help to reduce microbiological efforts and accelerate plasma research. As most new plasma sources are constructed in physics laboratories as prototypes, they often lack appropriate housings in terms of hygienic design. For this early stage the non-pathogenic surrogate strain proved to be suitable for choosing and testing plasma sources under safe work conditions. Examinations of the pathogen, therefore, can be postponed to advanced stages and limited to adjustments for optimized process settings. Notwithstanding the promising results on E. coli O157:H7, improvements for a faster inactivation of E. coli O104:H4 and especially overcoming the shoulder at 103 cfu cm2 on the produce surface are highly desirable. The antibacterial efficiency of a plasma-jet system can be enhanced by increasing the plasma power as results on E. coli DSM 1116 show (24). A concomitant rise in process temperature and potentially detrimental impacts on product quality in turn limit the maximum of applicable process power (25). A shorter treatment distance down the range of the plasma discharge also increases the antimicrobial efficiency and led to the fastest reductions in this study (Fig. 3B). Measurements show a steep increase of detected UV emission at distances lower than 12 mm, but also a rise in temperature to over 60 C at the outlet of the quartz capillary (20). How near a sample may be placed to the plasma generation zone particularly depends on the specific characteristics of the sample surface and needs to be defined by quality measurements. Using the example of thin corn salad leaves, treatment at 17 mm proved to be a gentle and suitable setting whereas shorter distances were accompanied with increasing quality loss (18). The latter may still be an option for more robust surfaces such as the pericarps of melons or in the field of plasma disinfection of abiotic surfaces like work spaces and conveyors of food production sites. Eventually, the composition of the process gas can be varied, and effects of relative humidity tested, to screen for the optimum of inactivation efficiency. In conclusion, plasma treatment is capable of reducing the STEC outbreak strain E. coli O104:H4. Differences of its response to plasma on gel compared to corn salad leaves emphasize the need to account for matrix effects of any biological surface to be disinfected. Treatment times, necessary to reduce E. coli O104:H4 below its infective dose, need to be adapted to the specific commodity and its expected level of initial bacterial count. Gentle treatment intensities, suitable for corn salad leaves, are more effective on E. coli O157:H7 than on E. coli O104:H4. Inactivation efficiency needs to be improved to overcome delayed microbial reductions and tailing effects. Finally, numbers of samples should be increased as high as possible to ensure the reliable killing of low residual contaminations of the highly infective pathogens below standard sampling. ACKNOWLEDGMENTS The authors thank Susanne Klocke of ATB Potsdam-Bornim for her excellent technical assistance. This work was supported by the German Federal Ministry of Food, Agriculture and Consumer Protection (BMELV) and the Federal Office for Agriculture and Food (BLE) within the innovation program (FKZ 28-1-63.003e07). References 1. Hartung, M. and Käsbohrer, A. (Eds.): Erreger von Zoonosen in Deutschland im Jahr 2009. Bundesinstitut für Risikobewertung, Berlin (2011). 2. Sivapalasingam, A., Friedman, C. R., Cohen, L., and Tauxe, R.: Fresh produce: a growing cause of outbreaks of foodborne illness in the United States, 1973 through 1997, J. Food Prot., 67, 2342e2353 (2004).
5
3. Warriner, K., Huber, A., Namvar, A., Fan, W., and Dunfield, K.: Recent advances in the microbial safety of fresh fruits and vegetables, Adv. Food Nutr. Res., 57, 155e208 (2009). 4. Scharff, R. L.: Health-related costs from foodborne illness in the United States. The Produce Safety Project at Georgetown University. The Pew Charitable Trusts, Philadelphia. http://www.pewtrusts.org/w/media/legacy/uploadedfiles/phg/ content_level_pages/reports/PSPScharff20v9pdf.pdf (2010). 5. Michino, H., Araki, K., Minami, S., Tayaka, S., Sakai, N., Ono, A., and Yanagawa, H.: Massive outbreak of Escherichia coli O157:H7 infection in school children in Sakai City, Japan, associated with consumption of white radish sprouts, Am. J. Epidemiol., 150, 787e796 (1999). 6. Wieler, L. H., Semmler, T., Eichhorn, I., Antao, E. M., Kinnemann, B., Geue, L., Karch, H., Guenther, S., and Bethe, A.: No evidence of the Shiga toxinproducing E. coli O104:H4 outbreak strain or enteroaggregative E. coli (EAEC) found in cattle feaces in northern Germany, the hotspots of the 2011 HUS outbreak area, Gut Pathog., 3, 17 (2011). 7. Bielaszewska, M., Mellmann, A., Zhang, W., Köck, R., Fruth, A., Bauwens, A., Peters, G., and Karch, H.: Characterisation of the Escherichia coli strain associated with an outbreak of haemolytic uraemic syndrome in Germany, 2011: a microbiological study, Lancet, 7, 671e676 (2011). 8. Richter, A. M., Povolotsky, T. L., Wieler, L. W., and Hengge, R.: Cyclic-di-GMP signalling and biofilm-related properties of the Shiga toxin-producing 2011 German outbreak Escherichia coli O104:H4, EMBO Mol. Med., 6, 1622e1637 (2014). 9. Robert Koch Institute: Report: final presentation and evaluation of epidemiological findings in the EHEC O104:H4 outbreak, Germany 2011. Robert Koch Institute, Berlin (2011). 10. Keener, K. M.: Atmospheric non-equilibrium plasma, Encycl. Agric. Food Biol. Eng., 1, 1e5 (2008). 11. Critzer, F. J., Kelly-Winterberg, K., South, S. L., and Golden, D. A.: Atmospheric plasma inactivation of foodborne pathogens on fresh produce surfaces, J. Food Prot., 70, 2290e2296 (2007). 12. Niemira, B. A. and Sites, J.: Cold plasma inactivates Salmonella Stanley and Escherichia coli O157:H7 inoculated on golden delicious apples, J. Food Prot., 71, 1357e1365 (2008). 13. Montenegro, J., Ruan, R., Ma, H., and Chen, P.: Inactivation of E. coli O157:H7 using a pulsed nonthermal plasma system, J. Food Sci., 67, 646e648 (2002). 14. Grant, M. A., Hedberg, C., Johnson, R., Harris, J., Logue, C. M., Meng, J., Sofos, J. N., and Dickson, J. S.: The significance of non-O157 Shiga toxinproducing Escherichia coli in food, Food Prot. Trends, 1, 33e45 (2011). 15. Hartung, M. and Käsbohrer, A. (Eds.): Erreger von Zoonosen in Deutschland im Jahr 2010. Bundesinstitut für Risikobewertung, Berlin (2012). 16. Hartung, M. and Käsbohrer, A. (Eds.): Erreger von Zoonosen in Deutschland im Jahr 2011. Bundesinstitut für Risikobewertung, Berlin (2013). 17. European Food Safety Authority (EFSA): ECDC The European Union summary report on trends and sources of zoonoses, zoonotic agents and food-borne outbreaks in 2011, EFSA J., 11, 3129 (2013). 18. Baier, M., Görgen, M., Ehlbeck, J., Knorr, D., Herppich, W. B., and Schlüter, O.: Non-thermal atmospheric pressure plasma: screening for gentle process conditions and antibacterial efficiency on perishable fresh produce, Innov. Food Sci. Emerg., 22, 147e157 (2014). 19. Scheutz, F., Møller-Nielsen, E., Frimodt-Møller, J., Boisen, N., Morabito, S., Tozzoli, R., Nataro, J. P., and Caprioli, A.: Characteristics of the enteroaggregative Shiga toxin/verotoxinproducing Escherichia coli O104:H4 strain causing the outbreak of haemolytic uraemic syndrome in Germany, May to June 2011, Euro Surveill., 16, pii=19889 (2011). 20. Weltmann, K. D., Kindel, E., Brandenburg, R., Meyer, C., Bussiahn, R., Wilke, C., and von Woedtke, T.: Atmospheric pressure plasma jet for medical therapy: plasma parameters and risk estimation, Contrib. Plasm. Phys., 49, 631e640 (2009). 21. Fernández, A., Shearer, N., Wilson, D. R., and Thompson, A.: Effect of microbial loading on the efficiency of cold atmospheric gas plasma inactivation of Samonella enterica serovar Typhimurium, Int. J. Food Microbiol., 152, 175e180 (2012). 22. Aurass, P., Prager, R., and Flieger, A.: EHEC/EAEC O104:H4 strain linked with the 2011 German outbreak of haemolytic uremic syndrome enters into the viable but non-culturable state in response to various stresses and resuscitates upon stress relief, Environ. Microbiol., 13, 3139e3148 (2011). 23. Huang, C.: An experimental investigation of low temperature plasma sterilization, treatment, and polymerization processes, PhD thesis. University of Missouri-Columbia (2006). 24. Fröhling, A., Baier, M., Ehlbeck, J., Knorr, D., and Schlüter, O.: Atmospheric pressure plasma treatment of Listeria innocua and Escherichia coli at polysaccharide surfaces: inactivation kinetics and flow cytometric characterization, Innov. Food Sci. Emerg., 13, 142e150 (2012). 25. Baier, M., Foerster, J., Schnabel, U., Ehlbeck, J., Knorr, D., Herppich, W. B., and Schlüter, O.: Direct non-thermal plasma treatment for the sanitation of fresh corn salad leaves. Evaluation of physical and physiological effects and antimicrobial efficacy, Postharvest Biol. Technol., 84, 81e87 (2013).
Please cite this article in press as: Baier, M., et al., Inactivation of Shiga toxin-producing Escherichia coli O104:H4 using cold atmospheric pressure plasma, J. Biosci. Bioeng., (2015), http://dx.doi.org/10.1016/j.jbiosc.2015.01.003
