Inactivation of Escherichia coli O157:H7 and Salmonella on artificially or naturally contaminated mung beans (Vigna radiata L) using a stabilized oxychloro-based sanitizer

Authors


Keith Warriner, Department of Food Science, University of Guelph, Guelph, ON, N1G 2W1, Canada. E-mail: kwarrine@uoguelph.ca

Abstract

Aims:  To evaluate the efficacy of a stabilized oxychloro-based (SOC) sanitizer to decontaminate mung beans artificially or naturally contaminated with Escherichia coli O157:H7 or Salmonella.

Methods and Results:  Naturally contaminated beans were produced by introducing a five-strain cocktail of E. coli O157:H7 or Salmonella onto the flowers of growing mung bean plants. Escherichia coli O157:H7 was only sporadically recovered from sprout lots (three testing positive from 10 tested) derived from harvested beans. In contrast, Salmonella was recovered from 18 of 20 lots screened. Pathogens present on naturally contaminated seed could be successfully inactivated with SOC applied at 200 ppm for 24 h at 28°C. SOC treatment could also decontaminate artificially inoculated mung bean batches containing different levels of contaminated seed. SOC inactivated E. coli O157:H7, but not Salmonella introduced onto damaged (scarified) beans.

Conclusions:  SOC sanitizer could inactivate Salmonella or E. coli O157:H7 naturally or artificially introduced onto mung beans. However, the SOC treatment failed to inactivate Salmonella introduced onto damaged mung beans.

Significance and Impact of the Study:  SOC sanitizer represents an effective method for decontaminating undamaged mung beans.

Introduction

Over the last decade there has been numerous food-borne illness outbreaks linked to sprouted seeds such as alfalfa and bean sprouts (Fett et al. 2005). Salmonella is the most frequently encountered human pathogen on sprouts although cases involving Escherichia coli O157:H7 have also been reported (Mohle-Boetani et al. 2001; Fett et al. 2005). In the majority of outbreaks the seed used for sprout production has been shown to be the source of human pathogens (Anonymous 1999). Therefore, it was recommended that seeds should be treated with 20 000 ppm calcium hypochlorite prior to soaking and subsequently sprouting (Anonymous 1999). However, it is widely acknowledged that treatments based on hypochlorite or alternative sanitizing agents (e.g. peroxyacetic acid, hydrogen peroxide) can only reduce pathogen levels on seeds but cannot ensure complete elimination (Anonymous 1999; Wessinger and Beuchat 2000; Brooks et al. 2001). This is relevant considering that even relatively low levels of surviving pathogens (<0·1 CFU g−1) can grow to densities in excess of 6 log CFU g−1 within 48 h into the sprouting process (Holliday et al. 2001).

A successful decontamination method has been devised based on supplementing the soak water used in the early stages of sprout production with a stabilized oxychloro-based sanitizer (SOC) (Kumar et al. 2006). A previous study has demonstrated that SOC applied at 200 ppm for >19 h can consistently inactivate either E. coli O157:H7 or Salmonella introduced onto mung beans (Kumar et al. 2006). However, the efficacy of SOC to decontaminate naturally contaminated or damaged seed has yet to be determined. In both cases it is possible that low levels of pathogens can become located in protective sites within the seed inaccessible to sanitizing agents and significantly reduce the efficacy of decontamination treatments (Brooks et al. 2001; Charkowski et al. 2001; Fu et al. 2001; Howard and Hutcheson 2003; Winthrop et al. 2003; Montville and Schaffner 2005).

In the following study the efficacy of SOC sanitizer to inactivate pathogens (Salmonella and E. coli O157:H7) when introduced at low levels or on damaged mung beans was evaluated. In addition, batches of naturally contaminated mung beans were produced by introducing the pathogens onto the flowers of growing plants from which the seed pods emerged. The choice of flowers as the inoculation site was selected to mimic the effect of contaminated irrigation water which is considered to be an important source of human pathogens present on seeds (Anonymous 1999).

Materials and methods

Bacterial strains and preparation of inocula

Escherichia coli O157:H7 and Salmonella used in the study were composed of environmental, clinical, tomato or sprout isolates (Table 1). An aliquot (1 ml) of an overnight culture of the individual E. coli O157:H7 or Salmonella strains was transferred into 50 ml of tryptic soy broth (Difco, Sparks, MD, USA) and incubated for 24 h at 37°C. Bacterial cells were harvested by centrifugation (5500 g for 10 min, 4°C) and washed once in 0·8% saline. The final cell pellet was re-suspended in saline to a final cell density of 106 CFU ml−1. Equal volumes of the five different E. coli O157:H7 or Salmonella suspensions were then combined to produce a cocktail that was subsequently used to inoculate beans or flowers of mung bean plants.

Table 1. Escherichia coli O157:H7 and Salmonella strains/serovars used in the study
Escherichia coli O157:H7 strain*SourceSalmonella enterica serovarSource
  1. *Strains obtained from Canadian Research Institute for Food Safety, University of Guelph, Guelph, ON, Canada.

  2. †Strains donated by Dr C. Poppe, Health Canada, Guelph, ON, Canada.

E. coli O157:H7 C1033Water sedimentMeleagridis E1†Alfalfa sprouts
E. coli O157:H7 C1032SoilOranienburg C1†Alfalfa sprouts
E. coli O157:H7 C652ClinicalNewport C2†Alfalfa sprouts
E. coli O157:H7 C476ClinicalSenftenburg†Alfalfa sprouts
E. coli O157:H7 C477ClinicalMontevideo*Tomatoes

Inoculation of undamaged and damaged mung beans

Mung beans (Vigna radiata) were purchased from Mumms Seeds Ltd (Parkside, SK, Canada). Beans (250 g) were soaked in 250 ml of the five-strain cocktail of E. coli O157:H7 or Salmonella for 20 min. The seeds were then transferred to sterile filter paper within a biological safety cabinet and allowed to dry at ambient temperature for 48–60 h. The inoculated mung beans were then introduced into 100-g batches of uninoculated beans at different levels (1–50%, w/w) and manually mixed prior to sprouting. For example, 1 g of inoculated seeds was mixed with 99 g of non-inoculated seed to give an inoculation level of 1% (w/w).

For damaged seed, mung beans (in 5-g lots) were placed between sheets of abrasive paper and rubbed with moderate pressure for c. 10 s to produce visible scarification. The damaged mung beans (100 g) were then soaked in bacterial suspension as described above and sprouted directly without mixing with non-inoculated seed.

Cultivation and inoculation of mung bean plants

Mung beans (non-inoculated) were soaked in distilled water at 30°C overnight to stimulate germination prior to transplanting into plug trays containing commercial grade PGX soil (Professional plug and germination growing medium; Premier Horticulture Ltd, Dorval, QC, Canada). The trays were transferred to a sealed growth room maintained at 25°C with a relative humidity set at 90%. After 2 weeks the seedlings were transplanted into 2-l soil microcosms containing commercial grade BX soil (Professional general purpose growing medium, Premier Horticulture Ltd). The relative humidity within the growth room was maintained at 90% with 12/12 h day/night at 25/10°C respectively. Flowers started to appear on plants 7–8 weeks into the cultivation period. The calyx of individual flowers was inoculated with a single 0·1 ml aliquot of E. coli O157:H7 or Salmonella (106 CFU ml−1). The plants (12 inoculated with E. coli O157 and 24 with Salmonella) were cultivated for a further period (6–8 weeks) to enable seed pod development and ripening. Seed pods were manually harvested using sterile gloved hands and transferred to the laboratory in sterile bags. The beans were extracted from pods using a sterile scalpel blade and pooled prior to subdividing into 0·7–1·2 g lots (40 lots for Salmonella and 20 for E. coli O157:H7). Half the seed lots (for each pathogen type) were treated using SOC sanitizer with the remaining acting as non-treated controls.

SOC treatment and sprouting of mung beans

The different mung bean batches (i.e. naturally contaminated, damaged or those containing inoculated seed introduced at different levels) were soaked for 24 h at 28°C in 200 ppm SOC (commercial name Germin-8-or; Vernagene Ltd, Bolton, Lancashire, UK) using a ratio of 1 part beans (dry weight) to 5 parts sanitizer (Kumar et al. 2006). The seeds were then removed, rinsed with distilled water prior to germinating for a further 4 days at 28°C. The developing sprouts were watered daily by a 5-min soak in 500 ml of sterile distilled water. Control (non-decontaminated) mung bean batches were sprouted using the same method except the SOC solution was replaced with distilled water.

Microbiological analysis

For artificially inoculated mung beans, microbiological analysis was performed on both the inoculated seeds and sprouts. Duplicate samples (1 g) of inoculated seed were suspended in 9 ml of buffered peptone water (0·1%, Oxoid, Basingstoke, UK) and vortexed for 1 min. Salmonella was enumerated on Brilliant Green agar (BG; Oxoid) incubated at 42°C for 24 h. Escherichia coli O157:H7 was enumerated on Sorbitol Maconkey agar containing cefixime and tellurite (CT-SMAC; Difco) incubated at 37°C for 24 h.

For sprouts, duplicate 25 g samples were suspended in 225 ml buffered peptone water and stomached for 90 s at 230 rev min−1 (Model 400; A.J. Steward and Co., London, UK). The sprout homogenates were tested for the presence of Salmonella using the method described in the Canadian Compendium of Analytical Method, MFHPB 20 (Health Canada 2003). Sprouts were suspended in 225 ml buffered peptone water and incubated at 42°C for 24 h. An aliquot (0·1 ml) of the enriched culture was then inoculated into the centre of a semi-solid Rappaport–Vassiliadis plate (Oxoid) that was subsequently incubated at 37°C for 24 h. Cells from the outer perimeter of the growth halo (presumptive motile Salmonella) were streaked onto BG agar (Oxoid) and incubated at 37°C overnight. The Oxoid Salmonella Latex Test FT0203 was used for serological confirmation of typical colonies (i.e. red colonies surrounded by brilliant red zones).

Sprouts were enriched for E. coli O157:H7 using buffered peptone water containing 0·5% (w/v) sodium thioglycolate incubated at 37°C for 24 h (Sata et al. 2003). Aliquots (10 μl) of the enriched culture were then streaked onto CT-SMAC that was subsequently incubated at 37°C for 24 h. Typical colonies (colourless) were confirmed as E. coli O157:H7 using the Oxoid E. coli O157 Latex Test DR0620M.

Sprouts from mung beans derived from inoculated plants were screened for the presence of pathogens using the same methods highlighted above. However, because of sprouting of the small bean lots the entire batch of sprouts (8–13 g) was analysed.

DNA typing of Salmonella

Enterobacterial Repetitive Intergenic Consensus (ERIC)-PCR (Versalovic et al. 1991) was used to identify which of the five strains of Salmonella serovar present on sprouts upon completion of the sprouting process. Up to five colonies from the positive selective agar plates were picked onto Luria–Bertani (LB; Difco) agar that was subsequently incubated at 37°C for 24 h. Single colonies were then suspended in 0·2 ml of TE buffer heated at 100°C to lyse cells. Cell debris were removed by centrifugation (13 000 g) and the supernatant, containing DNA, decanted into a new Eppendorf tube. ERIC-PCR was performed using the method described by Kumar et al. (2006). DNA patterns were analysed and dice similarity coefficients calculated using Molecular Analyst Software, Bio-Rad Fingerprinting II version 3·0 (Bio-Rad Laboratories, Hercules, CA, USA).

Bioassay for antimicrobial metabolites produced by Salmonella

A bioassay was performed (as described by Fett 2006) to determine the antagonistic activity exhibited by individual Salmonella serovars used in the study. Briefly, aliquots (20 μl) of an overnight culture of the test Salmonella was spotted onto LB agar and incubated at 30°C for 24 h. The agar plate was then exposed to chloroform vapour for 1 h at room temperature within a fume hood (Protector, VWR, Mississauga, ON, Canada). Each plate was then overlaid with 6 ml molten water agar (6 g agar l−1) containing 60 μl of an overnight suspension of the appropriate Salmonella serovar. The agar was allowed to solidify prior to incubating plates at 30°C for 24 h. The plates were then removed and visually inspected for zones of inhibition.

Data analysis

Qualitative data were statistically analysed using 2 × 2 contingency tables (s-plus, Insightful Corp., New York, NY, USA). In all cases the significance level was set at P ≤ 0·05.

Results

Inactivation of E. coli O157:H7 and Salmonella inoculated onto undamaged or damaged mung beans

When the inoculated beans were introduced into batches of non-inoculated beans and sprouted the subsequent bean sprouts tested positive for Salmonella or E. coli O157:H7 (Table 2). However, no pathogens were recovered from sprouts derived from SOC-treated beans (Table 2). Escherichia coli O157:H7 introduced onto damaged beans could also be successfully inactivated by treating with 200 ppm SOC. However, SOC treatment failed to decontaminate damaged seeds inoculated with Salmonella, with all the subsequent sprout samples testing positive for the enteric pathogen (Table 2).

Table 2.   The effect of introducing contaminated mung beans at different levels on the decontaminating efficacy of stabilized oxychloro-based sanitizer treatment
Level of contaminated mung beans introduced†TreatmentEscherichia coli O157:H7 (positive/number tested)¶Salmonella (positive/number tested)¶
  1. *Significantly (P < 0·05) greater number of sprout batches derived from non-treated mung beans tested positive for pathogens compared with those receiving SOC treatment.

  2. †Mung beans were inoculated with a five-strain cocktail of either Salmonella or E. coli O157:H7. The inoculated mung beans (containing 103–104 CFU g−1) were introduced to batches of non-inoculated beans (1–50%, w/w) prior to treatment.

  3. ‡Mung bean batches (n = 3) soaked in distilled water for 24 h prior to sprouting.

  4. §Mung bean batches (n = 3) soaked in 200 ppm SOC for 24 h prior to sprouting.

  5. ¶At the end of the 4-day sprouting process, bean sprouts (25 g) were enriched for either E. coli O157:H7 or Salmonella.

  6. ††Bean sprouts derived from scarified mung beans inoculated with Salmonella or E. coli O157:H7 and sprouted over a 4-day period.

1%Not treated‡3/3*3/3*
SOC treated§0/30/3
10%Not treated3/33/3
SOC treated0/30/3
50%Not treated3/33/3
SOC treated0/30/3
100%Not treated3/33/3
SOC treated0/30/3
Damaged seed†† 100%Not treated8/88/8
SOC treated0/88/8

Decontamination of naturally contaminated seed

Mung beans derived from plants inoculated with E. coli O157:H7 were sporadically contaminated with the pathogen. When the beans was sprouted without applying SOC sanitizer, only three of the 10 lots were found to be contaminated with E. coli O157:H7 (Table 3). However, none of the sprouts derived from mung beans subjected to SOC treatment tested positive for the pathogen (Table 3).

Table 3. Salmonella or Escherichia coli O157:H7 recovered from bean sprouts produced from mung beans derived from plants in which the human pathogens were introduced onto flowers
 Mung bean treatment*Sprouts positive by enrichment (positive/number tested)§
  1. *Mung beans derived from inoculated plants were pooled and segregated into equal lots.

  2. †Mung beans were soaked in distilled water for 24 h prior to sprouting.

  3. ‡Mung bean batches (n = 3) soaked in 200 ppm SOC for 24 h prior to sprouting.

  4. §At the end of the 4-day sprouting process, bean sprouts (8–13 g) were enriched for either E. coli O157:H7 or Salmonella.

Escherichia coli O157:H7Non-treated† 3/10
SOC treated‡ 0/10
SalmonellaNon-treated18/20
SOC treated 0/20

The majority of mung bean lots (18 of 20) derived from Salmonella-inoculated plants tested positive for the bacterium when sprouted over a 4-day period (Table 3). However, no Salmonella was recovered from sprouts produced from beans treated with SOC (Table 3).

Through DNA typing, Salmonella enterica serovar Meleagridis was found to be the only serovar of the five introduced that was recovered from sprouts derived from naturally contaminated, non-sanitized, mung beans. A bioassay was performed to determine if serovar Meleagridis may have produced an antimicrobial compound to inhibit the growth of the other Salmonella within the cocktail. However, no zones of inhibition were noted for any of the combinations of Salmonella applied (results not shown).

Discussion

The study has demonstrated that the efficacy of SOC to inactivate Salmonella or E. coli O157:H7 was independent of pathogen levels introduced into batches of undamaged beans. It has been reported that human pathogens present at low levels in the presence of a high level of endogenous microflora express stress proteins that provide enhanced tolerance to inimical processes (Dodd and Aldsworth 2002; Komitopoulou et al. 2004). Obviously, in the present case even if stress responses were induced within Salmonella and E. coli O157:H7 this did not affect the efficacy of SOC treatment.

Stabilized oxychloro-based treatment was less effective in inactivating Salmonella introduced onto damaged mung beans. This is in agreement with other works that have reported on the difficulty in decontaminating scarified or damaged seed (Charkowski et al. 2001). It is likely that by being located within crevices deep within the seed the Salmonella were protected from the antimicrobial effects of SOC. The greater susceptibility of E. coli O157:H7 to SOC relates to the lower tolerance of the pathogen to the sanitizer. In previous studies it was noted that SOC sanitizer applied at 100 ppm could ensure inactivation of E. coli O157:H7 on mung beans compared with 200 ppm required to inactivate Salmonella (Kumar et al. 2006). Therefore, in the current study it was likely that SOC could penetrate into the bean interior but at an insufficient concentration to inactivate Salmonella. Regardless of this fact, the results support the view that scarification or using damaged seed in sprout production should be avoided (Anonymous 1999).

From studies using naturally contaminated seed it was clearly evident that Salmonella could become established on beans to a greater extent compared with E. coli O157:H7. The result would suggest that E. coli O157:H7 has a lower level of persistence on plants or unable to become associated with developing beans. This may explain why food-borne illness outbreaks linked to sprouts are more commonly associated with Salmonella as opposed to E. coli O157:H7 (Barak et al. 2002). However, it should be noted that in a previous study performed by Cooley et al. (2003) it was reported that E. coli O157:H7 could contaminate Arabidopsis seed to a greater extent compared with S. enterica serovar Newport when introduced onto growing plants. Therefore, it is likely that the ability of human pathogens to become established on seeds is plant and strain dependent. In this respect it was interesting to note that from the five Salmonella serovars introduced onto the flowers of plants only serovar Meleagridis was recovered from the subsequent sprouts. In a previous study, the same serovar was recovered from sprouts derived from mung beans inoculated with the same Salmonella serovar combination (Kumar et al. 2006). This may suggest that Meleagridis has physiological attributes that enhance interaction with sprouted seeds or elicits antagonistic effects against other Salmonella. The latter seems unlikely as none of the Salmonella used in the current study exhibited antagonistic affects in the agar plate bioassay. Therefore, it is possible that Meleagridis could exhibit greater tolerance to environmental stress, higher attachment strength and/or greater growth rates on sprouting seeds. Similar attributes have been implicated in the establishment of Salmonella on sprouting alfalfa (Howard and Hutcheson 2003; Barak et al. 2005). Whether such factors are associated with Meleagridis is unclear at present but is worthy of further study.

In the current study it was found that SOC treatment could be used to decontaminate naturally contaminated seed. Despite the sporadic occurrence of E. coli O157:H7 on mung beans it is questionable whether the pathogen was present on the seed prior to decontamination. However, the fact that no Salmonella was recovered from control sprouts strongly indicates that the SOC treatment was a success. In previous studies with naturally contaminated seeds it was reported that 1800–2000 ppm calcium hypochlorite was insufficient to ensure the elimination of Salmonella (Fett 2002). Given that the SOC sanitizer could achieve complete elimination of Salmonella when applied at 200 ppm underlines the effectiveness of the treatment compared with the recommended hypochlorite-based method. Nevertheless, it is acknowledged that seeds can be contaminated via various routes other than irrigation water. In this respect further studies are warranted on the efficacy of SOC treatment to decontaminate seeds derived from sprout-related outbreaks where the levels and spatial distribution of human pathogens may differ.

Acknowledgements

The authors wish to thank the Ontario Ministry of Agriculture, Food and Rural Affairs (OMAFRA) Food Safety Program for their financial support (Project FS 030308). We also wish to thank Dr C. Poppe of Health Canada for donating the Salmonella strains used in the current study.

Ancillary