Interaction of Escherichia coli with Growing Salad ...

3 downloads 0 Views 2MB Size Report
consumption of minimally processed ready-to-eat salad vegetables is increasing ... ish (13, 15), alfalfa (18, 23, 24, 35), and bean sprouts (39). Furthermore, the ...
1790 Journal of Food Protection, Vol. 66, No. 10, 2003, Pages 1790–1797 Copyright q, International Association for Food Protection

Interaction of Escherichia coli with Growing Salad Spinach Plants KEITH WARRINER,1† FAOZIA IBRAHIM,1 MATTHEW DICKINSON,2 CHARLES WRIGHT,3 WILLIAM M. WAITES1* 1 Division

AND

of Food Sciences, 2Division of Plant Sciences, and 3Division of Agricultural Sciences, School of Biosciences, University of Nottingham, Sutton Bonington Campus, Loughborough, Leicestershire LE12 5RD, UK MS 03-39: Received 4 February 2003/Accepted 18 April 2003

ABSTRACT In this study, the interaction of a bioluminescence-labeled Escherichia coli strain with growing spinach plants was assessed. Through bioluminescence proŽ les, the direct visualization of E. coli growing around the roots of developing seedlings was accomplished. Subsequent in situ glucuronidase (GUS) staining of seedlings conŽ rmed that E. coli had become internalized within root tissue and, to a limited extent, within hypocotyls. When inoculated seeds were sown in soil microcosms and cultivated for 42 days, E. coli was recovered from the external surfaces of spinach roots and leaves as well as from surfacesterilized roots. When 20-day-old spinach seedlings (from uninoculated seeds) were transferred to soil inoculated with E. coli, the bacterium became established on the plant surface, but internalization into the inner root tissue was restricted. However, for seedlings transferred to a hydroponic system containing 102 or 103 CFU of E. coli per ml of the circulating nutrient solution, the bacterium was recovered from surface-sterilizedroots, indicating that it had been internalized. Differencesbetween E. coli interactions in the soil and those in the hydroponic system may be attributed to greater accessibility of the roots in the latter model. Alternatively, the presence of a competitive micro ora in soil may have restricted root colonization by E. coli. The implications of this study’s Ž ndings with regard to the microbiological safety of minimally processed vegetables are discussed.

The incidence of foodborne illness associated with the consumption of minimally processed ready-to-eat salad vegetables is increasing (2, 3, 19, 32, 36). To date, the majority of contamination of vegetables by human pathogenic bacteria has been considered to occur during postharvest handling (21). Nevertheless, it has been shown that salad vegetables can also be contaminated during cultivation via soil, water, animals, and harvest equipment (5, 29). Postharvest washing, typically with the use of sodium hypochlorite (at 100 to 200 ppm), is carried out to remove Ž eldacquired contamination, but it is now becoming established that this procedure achieves only a ,2-log reduction in bacterial counts (4, 5, 7). Although a diverse range of more potent disinfectant types have been applied, log reductions in bacterial counts achieved with these disinfectants on naturally contaminated produce have not been signiŽ cantly improved over those achieved with standard chlorine washes (27, 41). It is becoming evident that bacteria (including human pathogens) can be protected from the lethal effects of biocidal washes by virtue of their locations in protective areas of the plant. It has been proposed that bioŽ lms could be responsible for this protective effect (7, 9, 33). However, it has also been reported that Escherichia coli O157:H7 cells inoculated onto lettuce leaves survive biocidal washes when they are located in stomata or, to a greater extent, * Author for correspondence. Tel: 144 115 9516161; Fax: 144 115 9516162; E-mail: [email protected]. † Present address: Department of Food Science, University of Guelph, Guelph, Ontario, Canada N1G 2W1.

when they gain entry into the inner part of the leaf via cut or damaged areas (28, 34). The internalization of human pathogens via natural openings has also been reported for apples inoculated with E. coli O157:H7 (6) and tomatoes inoculated with Salmonella (42). Previous reports have suggested that human pathogens can also become internalized into the vascular systems of growing plants (14, 26), although this question remains open (22). Research has conŽ rmed that human pathogens can become internalized within sprouted seeds such as radish (13, 15), alfalfa (18, 23, 24, 35), and bean sprouts (39). Furthermore, the internalization of E. coli in lettuce (37), cabbage (38), and tomato plants (10–12) has been observed. However, to date, little work on other salad vegetable types has been carried out. The difŽ culties encountered in the decontamination of salad spinach (24) could possibly be attributable to internalized bacterial populations (i.e., endophytes). In this study, the interaction of E. coli with growing spinach plants was assessed. MATERIALS AND METHODS Bacterial strains and growth conditions. Bioluminescent E. coli P36 (nonverocytotoxic E. coli) was constructed from a slaughterhouse isolate by the Tn5 mini-transposon method. The mini-Tn5 plasmid (kindly donated by P. Hill, University of Nottingham), containing lux CDABE and a kanamycin resistance gene cassette, was maintained within E. coli lPir1. Competent E. coli S17-1 cells were transformed with a puriŽ ed plasmid preparation as described by Winson et al. (40). Conjugation between E. coli S17-1lpir and recipient cells (E. coli) was carried out with

J. Food Prot., Vol. 66, No. 10

E. COLI IN GROWING SPINACH

1791

FIGURE 1. Flow diagram of inoculation and spinach cultivation schemes. Spinach seeds were inoculated (A and B) and cultivated on the surface of solidiŽ ed hydroponic solution or within soil microcosms. Uninoculated seeds were cultivated in soil or in hydroponic solution inoculated with E. coli P36 (C and D).

a Whatman cellulose acetate membrane (0.45-mm pore size) overlaid onto a Luria-Bertani agar plate, which was subsequently incubated for 14 h at 378C (8). The selection of E. coli exconjugants was based on aliquots (50 to 200 ml) of the mating mixture that were plated onto chromogenic E. coli/coliform selective agar medium (Oxoid, Basingstoke, UK) containing kanamycin (30 mg/ml) and incubated overnight at 378C. The donor strain, E. coli S17-1, is deŽ cient in glucuronidase (GUS) activity and formed pink colonies. The recipient E. coli colonies were distinguished from the donor strain by the blue colonies formed via putative GUS activity. Bioluminescent colonies on agar plates were imaged with a Night-Owl image analyzer in conjunction with the manufacturer’s computer software (E. G. and G. Berthold, Munich, Germany). The bioluminescent phenotype was stable in the selected exconjugate, and growth characteristics were not signiŽ cantly different from those of the parental strain (39). Suspensions of E. coli P36 were prepared from overnight cultures grown aerobically at 378C in brain heart infusion broth (Difco Laboratories, Sparks, Md.). The cells were harvested by centrifugation (3,550 3 g for 10 min at 48C) and washed once in sterile maximum recovery diluent (MRD; Oxoid). The cell pellet was Ž nally resuspended in MRD to yield a suspension containing ca. 107 CFU/ml (A600 5 0.2). Inoculation of spinach seeds. Spinach (Spinacia oleracea L. cv. Sharan) seeds (Toza Seeds Ltd., Surrey, UK) were initially soaked in 4% (wt/vol) glycolic acid (Aldrich, Dorset, UK) for 30 min to stimulate germination. The seeds (in 20-g batches) were then washed three times in 2 liters of sterile distilled water before being submerged in E. coli suspensions (cell density, 107 CFU/ ml) for 20 min. The inoculated seeds were allowed to dry at room temperature for 8 h on sterile Ž lter paper. For the bacterial loading of inoculated seeds, three 1-g (ca. 150-seed) batches were with-

drawn and placed in 10 ml of MRD, and bacteria were released by vortexing for 1 min. Aliquots were plated onto brain heart infusion agar supplemented with kanomycin (30 mg/ml) and incubated at 378C for 24 h. To capture bioluminescence proŽ les of E. coli during germination, seeds were individually placed in glass tubes containing 10 ml of solidiŽ ed hydroponic nutrient solution (1:300 dilution hydroponic nutrient solution [Nutriculture Ormskirk, Lancashire, UK] containing 0.5% [wt/vol] agar). The tubes (n 5 10) were then incubated in the dark at 158C, and bioluminescent images were captured daily (Fig. 1A). In situ b -glucuronidase (GUS) stain. Whole 25-day-old spinach seedlings (cultivated on solidiŽ ed hydroponic media) derived from seeds inoculated with E. coli P36 were submerged for 15 min in sodium phosphate buffer (50 mM, pH 7.0) containing 0.3% (vol/vol) formaldehyde (16). The seedlings were washed three times in sterile distilled water and, Ž nally, transferred to a sodium phosphate buffer (50 mM, pH 7) containing 1 mM 5bromo-4-chloro-3-indoyl-b-D -glucuronide (X-GLUC; Sigma, Poole, Dorset, UK). The seedlings were incubated in the enzyme substrate solution for 8 h at 378C and were subsequently rinsed with 70% (vol/vol) ethanol. The blue-black stained sprouts (resulting from GUS activity) were then mounted on glass slides and viewed under a microscope. Inoculated seeds cultivated in soil microcosms. Inoculated seeds were sown into soil microcosms (6 cm2, 5 cm deep) containing John Innes no. 2 compost (Wessex Horticulture Products Ltd., Salisbury, Wiltshire, UK) (Fig. 1B). Soil samples (ca. 4 g each) were withdrawn from the microcosms, and bacterial counts were determined for three 1-g batches. The plants were cultivated for 35 days, and a further 4-g sample of soil was taken for mi-

1792

WARRINER ET AL.

J. Food Prot., Vol. 66, No. 10

FIGURE 2. Bioluminescence proŽ le for germinating spinach seeds inoculated with E. coli P36. Inoculated seeds were germinated on the surface of solidiŽ ed nutrient solution, and the bioluminescence images were obtained over a 25-day period. Images taken under illumination are shown on the right.

crobiological analysis. Plants were harvested intact and separated into three batches (three plants per batch). The roots and leaves from a batch of plants were separated with a sterile scalpel and placed in sterile stomacher bags. Surface bacteria and bacteria surviving surface sterilization were recovered as described below. Hydroponic cultivation of spinach plants. Uninoculated spinach seeds were germinated on damp Ž lter paper in the dark at 158C for 7 days (Fig. 1C). Germinated seeds were transferred to rockwool blocks (Growdan, Hedehusene, Denmark) predampened with a commercial hydroponic nutrient medium. Seedlings were allowed to develop in growth chambers maintained at 168C with illumination for 12 h per day. After the seedlings had developed sufŽ ciently (ca. 13 days), the plants were transferred to a nutrient Ž lm technique (NFT) hydroponic system within a con-

tained greenhouse. The system consisted of three polyvinyl chloride plastic channels 3 m long and 22 cm wide holding 17 plants in each trough. The nutrient solution, held in an 80-liter tank, was continuously circulated in a closed loop via a centrifugal pump. The bases of the plants were covered with a polyethylene sheet to prevent contact of spinach leaves with the underlying rockwool base substrate. Suspensions of E. coli in MRD were added to the nutrient solution in the holding tank to yield Ž nal cell densities of 103 or 102 CFU/ml. Control hydroponic systems containing no E. coli were set up in parallel. Periodically, 50 ml of hydroponic solution was withdrawn from the holding tank to determine bacterial counts. For this purpose, 0.1-ml aliquots of hydroponic solution were plated on tryptic soya agar (TSA) with and without kanamycin (30 g/ml) and incubated at either 308C for 48 h (for the

J. Food Prot., Vol. 66, No. 10

E. COLI IN GROWING SPINACH

1793

FIGURE 3. GUS stain of surface-sterilized spinach seedlings derived from seeds inoculated with E. coli P36 (A). The blue-black areas designate the presence of internalized E. coli. Control seedlings obtained from uninoculated seeds are shown for comparison (B). total viable count [TVC]) or 378C for 24 h (for E. coli). When E. coli counts had decreased to below the level of detection, 10 ml of hydroponic nutrient solution was passed through a sterile Ž lter, which was subsequently overlaid on a TSA-kanamycin plate and incubated for 24 h at 378C. Inoculated soil microcosms. Suspensions of E. coli were added to 3 kg of John Innes no. 2 compost (Wessex Horticulture Products) and distributed throughout the soil with manual agitation (Fig. 1D). The soil was then placed in soil microcosms (6 cm2, 5 cm deep), into which three 20-day-old spinach seedlings were sown. The soil microcosms were placed on a tray containing irrigation water whose level was maintained at 3-cm depth. Cultivation of the spinach was carried out over a 42-day period in contained greenhouses maintained at 20 to 268C under natural lighting conditions. Soil samples (ca. 4 g each) were withdrawn periodically from soil microcosms, and bacterial counts were determined for three 1-g batches. Microbiological analysis of spinach plants. The loosely attached bacteria on the surfaces of the plants were released by rinsing with sterile MRD. The samples were then surface sterilized for 10 min in a 4,000-ppm sodium hypochlorite solution. Residual chlorine was removed by rinsing samples Ž ve times in 500 ml of sterile distilled water. ConŽ rmation of surface sterilizationefŽ cacy was carried out by placing two treated leaves from a batch in contact with a TSA plate for 20 s. The plate was subsequently incubated at 308C for 48 h. Surface-sterilized plant samples were macerated by stomaching in MRD for 2 min. The TVCs for the different extracts were determined with the use of TSA (with incubation at 308C for 48 h). E. coli was enumerated on TSA supplemented with kanamycin (30 mg/ml) and incubated at 378C for 24 h. ConŽ rmation of E. coli was accomplished by visualizing bioluminescent colonies with a Night Owl image analyzer as described above.

RESULTS Interaction of E. coli with germinating spinach seeds. From the bioluminescence proŽ les obtained, it was clearly demonstrated that E. coli initially introduced onto the seed proliferated around the roots of the developing spinach plant (Fig. 2). When the spinach seedlings were removed after 25 days of cultivation, bioluminescent E. coli

remained attached, suggesting a close association between the bacterial cells and the plant roots. From plate counts of seedlings (two batches of 20 plants), E. coli was shown to be present both on the exterior of the plant (7.17 6 1.39 log CFU/g) and on inner plant tissues (4.03 6 0.95 log CFU/g) but not on any uninoculated controls. ConŽ rmation of the internalization of E. coli into the inner plant tissue was provided by the in situ GUS assay. For this assay, surface-sterilized seedlings were incubated in the presence of X-GLUC for 16 h at 308C. The majority of the precipitated substrate (resulting from the action of E. coli GUS) was observed in the roots of seedlings but also occasionally within the hypocotyls (Fig. 3). It was noted that the precipitate was not continuous from the roots to the hypocotyls. This Ž nding may imply that E. coli existed in aggregates within the plant or, alternatively, that its presence was due to the diffusion of hydrolyzed GUS substrate. No precipitate was found in control seedlings derived from uninoculated seeds, conŽ rming that internalized E. coli was responsible for the results observed (Fig. 3). When spinach plants derived from inoculated seeds were cultivated for 35 days in soil microcosms, E. coli could be recovered from the external (5.41 6 0.71 log CFU/g) and internal (1.93 6 0.08 log CFU/g) root structures but from only the exteriors of the leaves (5.20 6 1.32 log CFU/g). The E. coli counts for soil samples increased from below the level of detection (,1.25 log CFU/g) on day 1 to 2.99 6 0.31 log CFU/g by the end of the cultivation period. Interaction of E. coli with spinach cultivated in soil microcosms. Spinach seedlings (20 days postgermination) were transplanted into soil inoculated with E. coli. Although E. coli was introduced at a density of 102 CFU/g, E. coli counts progressively increased over the cultivation period, while the TVC remained relatively constant (Fig. 4). The E. coli counts for the spinach plant samples paralleled the soil counts in that numbers of cells generally increased during cultivation (Table 1). The majority of E.

1794

WARRINER ET AL.

J. Food Prot., Vol. 66, No. 10

FIGURE 4. Total viable counts ( ) and E. coli P36 counts (m) for soil during spinach cultivation. E. coli P36 was inoculated into soil microcosms, into which spinach seedlings (three plants per microcosm) were transplanted. Periodically during the cultivation period, soil samples (4 g each) were removed to determine TVCs and E. coli counts. Values are averages for three soil samples.

coli cells were present on the exterior part of the plant, although one surface-sterilized sample tested positive for E. coli on day 12. Toward the end of the cultivation period (day 32 onward), E. coli was recovered from all of the surface-sterilized spinach tested. On day 42, the remaining plants were harvested, and leaves were separated from the roots. E. coli was recovered from surface-sterilized root tissue (3.78 6 0.21 log CFU/g) but not from within leaves (,1.2 log CFU/g). Interaction with hydroponically cultivated spinach. A soil microcosm represents a complex system with a high abundance of nutrients and background micro ora that could potentially affect the interaction of E. coli with growing spinach plants. Therefore, a comparative study was carried out with hydroponically cultivated spinach. For this study, E. coli was introduced (at 102 or 103 CFU/ml) into the nutrient solution reservoir and circulated around rockwool blocks in which 20-day-old spinach seedlings were planted. The E. coli counts for the hydroponic nutrient solution declined progressively over the cultivation period,

FIGURE 5. Survival of E. coli P36 in nutrient solution in an NFT hydroponic system in the presence of rockwool blocks containing no spinach plants (l). Also illustrated is the relative survival of E. coli P36 in the absence of rockwool blocks (m). E. coli P36 was inoculated into hydroponic nutrient solution and circulated around the hydroponic system. Samples (50 ml each) were withdrawn periodically to determine E. coli levels. Values are averages for three samples.

and by day 8 these counts were below the level of detection (0.1 CFU/ml). This decrease was probably partly due to the poor nutrient environment of the hydroponic solution but was also due to the Ž ltration effect of rockwool blocks. This effect was conŽ rmed by determining the rate of decline in E. coli numbers in the presence and absence of rockwool blocks (containing no spinach seedlings). The decrease in E. coli counts was signiŽ cantly larger (P , 0.001) in the presence of rockwool blocks than in the absence of rockwool blocks (Fig. 5). E. coli counts for spinach plants (16 days after transplanting; average plant weight, 17 g) indicated that the bacterium was present on both the external and the internal root structures (as indicated by counts for surface-sterilized samples) at both inoculum levels used (Table 2). E. coli was also present on the external surfaces of (but not within) leaves cultivated in the presence of 103 CFU/ml (Table 2). DISCUSSION The interaction of E. coli with spinach was found to be dependent on the stage of introduction. When E. coli

TABLE 1. Bacterial counts for spinach plants cultivated in soil inoculated with E. coli P36a Total viable count (log CFU/g) Days after introduction

12 14 16 20 23 25 32 35

Wash

6.26 5.44 6.65 5.92 6.40 6.82 6.49 7.54

6 6 6 6 6 6 6 6

0.65 1.34 1.50 0.17 0.42 0.12 0.04 0.81

E. coli count (log CFU/g)

Extract

5.93 5.79 3.90 4.76 5.27 5.17 4.32 3.99

6 6 6 6 6 6 6 6

0.70 0.20 0.45 0.63 0.51 0.20 1.10 0.15

Wash

3.49 2.29 6.00 3.27 4.79 4.77 5.25 6.30

6 6 6 6 6 6 6 6

1.37 0.75 0.90 1.36 0.82 1.36 1.16 0.64

Extract

2.27b ND ND ND ND ND 2.16 6 0.04 2.91 6 0.81

Twenty-day-old spinach seedlings were transplanted into soil microcosms inoculated with E. coli P36 (102 CFU/g). Whole spinach plants (roots and leaves) were withdrawn from the soil microcosm periodically to assess the bacterial counts for surface samples (wash) and surface-sterilized samples (extract). Values presented are means 6 standard deviations for three plant samples. ND, not detected (,1.2 log CFU/g). b E. coli was recovered from one of three samples. a

J. Food Prot., Vol. 66, No. 10

E. COLI IN GROWING SPINACH

TABLE 2. Counts for spinach plants after 16 days of hydroponic cultivation in nutrient solution inoculated with E. coli P36a Sample type

Total viable count (log CFU/g)

Control (uninoculated) Roots Wash 6.35 6 2.53 Extract 4.59 6 1.46 Leaves Wash Extract

4.35 6 0.36 2.05 6 0.40

Initial inoculum of 102 CFU/g Roots Wash 6.64 6 0.63 Extract 5.33 6 0.99 Leaves Wash Extract

2.89 6 0.28 3.96 6 2.80

Initial inoculum of 103 CFU/g Roots Wash 7.16 6 0.63 Extract 5.35 6 1.41 Leaves Wash Extract a

3.03 6 0.21 2.79 6 0.25

E. coli count (log CFU/g)

ND ND ND ND

2.97 6 1.40 1.43 6 1.00 ND ND

2.24 6 1.06 1.64 6 0.71 1.91 6 0.50 ND

Twenty-day-old spinach seedlings were transplanted into a NFT hydroponic system. Nutrient solution with and without inoculated E. coli P36 was circulated around the blocks throughout. After 16 days of cultivation, the spinach plants were harvested and the roots were separated from the leaves with a sterile knife. Surface bacteria (wash) and bacteria derived from surface-sterilized samples (extract) were obtained as described in ‘‘Materials and Methods.’’ Values presented are means 6 standard deviations for three samples each containing two plants. ND, not detected (,1.2 log CFU/g).

was initially inoculated onto spinach seeds, its growth was sustained by the seed exudates released from germinating seeds, enabling it to become established both on the exterior and in the interior of the roots. This route of plant colonization is usually associated with rhizosphere endophyte bacteria, where the exudates released by the germinating seeds contain nutrients that initially attract bacteria and provide sources of energy and carbon for the colonization of roots (1, 13, 18, 20). Interestingly, Ji and Wilson (17) were able to predict the success of biocontrol bacteria in suppressing the plant pathogen Pseudomonas syringae by determining the nutritional similarity index (NOI). These authors tested 52 carbon sources known to be present in tomato plants against a range of bacteria. With the use of catabolic mutants of P. syringae with NOIs ranging from 0.07 to 0.90, a direct correlation between root colonization and NOI was found. E. coli (used as a control) had an NOI of 0.83 relative to P. syringae and was comparable to successful root colonizers. Although Ji and Wilson (17) did not determine the interaction of E. coli with tomato plants,

1795

the NOI model would suggest that the bacterium may compete successfully in the rhizosphere. The growth of E. coli and Salmonella on sprouted seeds (radish, alfalfa, and bean sprouts) is relatively well established (13, 15, 18, 23, 25, 35, 39). In addition, the growth of human pathogens on tomato and lettuce seedlings has also been reported (10– 12, 30, 37). From all these examples, it can be concluded that the internalization of E. coli into a diverse range of different seedling types is relatively common. Once established within the roots of spinach plants, E. coli might have been expected to migrate into the edible leaf portions of plants. Although the results of the in situ GUS assay indicated the presence of E. coli within hypocotyls, it is possible that X-GLUC hydrolyzed by cells in the root may have migrated into the vascular system prior to polymerization. Results obtained from the soil microcosm and hydroponic studies support the view that E. coli is restricted to colonization of the roots of plants, at least in mature plants. A similar pattern has been observed for cabbage crops accidentally irrigated with creek water contaminated with E. coli. In this case, E. coli was recovered from the roots of plants but not from the edible leaves (38). E. coli cell density in the plant environment has been reported to be a factor in determining the extent of plantcell interactions (30, 37). Solomon et al. (30) and Wachtel et al. (37) studied the extent to which E. coli O157:H7 interacted with lettuce when introduced at different cell densities. The conclusions of these authors were that at high cell densities (108 CFU/ml), E. coli O157:H7 became established on the roots and hypocotyls of plants. Conversely, when E. coli O157:H7 was introduced at lower cell densities (102 to 104 CFU/ml), the extent of colonization was comparatively low, although the pathogen was recovered from lettuce roots after a 10-day cultivation period (37). In the present study, the introduction of E. coli (at cell densities of 102 or 103 CFU/ml) into the hydroponic solution reservoir restricted colonization to the plant roots (Table 1). It is possible that the introduction of E. coli at higher cell densities may have resulted in a greater distribution of bacterial cells within the edible leaf portion of the plant. However, considering that such high cell densities would not occur in the natural environment, the current results may re ect a more accurate representation of the extent of the interaction of E. coli with growing spinach. It is interesting to note the differences in the interactions of E. coli with spinach plants cultivated in soil and with spinach plants cultivated in hydroponic systems. Although E. coli readily grew in the soil environment, the internalization into the roots of plants was delayed until the cultivation period was nearly complete. In contrast, although E. coli counts were lower for the hydroponic system, the bacterium readily became established on the interiors and exteriors of plant roots. Such differences between the soil and hydroponic models may re ect the role of competitive soil micro ora in preventing E. coli from becoming established on roots. It is known that the establishment of biocontrol bacteria in commercial crops is restricted by the presence of an adapted competitive micro ora (31). Therefore, considering that E. coli would be less adapted to the

1796

WARRINER ET AL.

soil environment, it would be relatively easily outcompeted. Alternatively, it is possible that in the hydroponic system, the roots of spinach were more accessible to E. coli for colonization. Although E. coli can become established in germinated seedlings, it is restricted to only the roots in mature spinach plants. Therefore, there is a low risk that the edible portion of the spinach leaves will harbor E. coli in the inner tissue. Nevertheless, it is possible that E. coli on the root maybe transferred to the inner leaf during harvesting. This possibility should be considered in the development of a hazard analysis critical control point scheme for the production of safe minimally processed salad vegetables. ACKNOWLEDGMENTS The authors acknowledge the United Kingdom Department of Environment, Food and Rural Affairs (Food Quality and Safety Programme, LINK FQS 33) for Ž nancial support and Dr. P. Hill (Division of Food Sciences, Nottingham University) for donating the bioluminescent strains used in this study.

REFERENCES 1. 2. 3.

4. 5. 6.

7.

8.

9. 10.

11.

12.

13.

14. 15.

Andrews, W. H., P. B. Mislivec, C. R. Wilson, et al. 1982. Microbial hazards associated with bean sprouting. J. AOAC 65:241–248. Beuchat, L. R. 1996. Listeria monocytogenes: incidence on vegetables. Food Control 7:223–228. Beuchat, L. R. 1998. Surface decontamination of fruits and vegetables eaten raw, a review. WHO/FSF/FQS 98.2. Food Safety Issues, Food Safety Unit, World Health Organization, Geneva. Beuchat, L. R., and J. H. Ryu. 1997. Produce handling and processing practices: special issue. Emerg. Infect. Dis. 3:459–465. Brackett, R. E. 1992. Shelf stability and safety of fresh produce as in uenced by sanitation and disinfection. J. Food Prot. 55:808–814. Buchanan, R. L., S. G. Edelson, R. L. Miller, and G. M. Sapers. 1999. Contamination of intact apples after immersion in an aqueous environment containing Escherichia coli O157:H7. J. Food Prot. 62: 444–450. Carmichael, I., I. S. Harper, M. J. Coventry, P. W. J. Taylor, J. Wan, and W. M. Hickey. 1999. Bacterial colonization and bioŽ lm development on minimally processed vegetables. J. Appl. Microbiol. 85: 45S–51S. De Lorenzo, V., and K. N. Timmes. 1994. Analysis and construction of stable phenotypes in Gram negative bacteria with Tn5- and Tn10 derived minitransposons, p. 386–405. In V. L. Clark and P. M. Bavoil (ed.), Methods in enzymology, vol. 235. Bacterial pathogenesis. Academic Press, London. Fett, W. F. 2000. Naturally occurring bioŽ lms on alfalfa and other types of sprouts. J. Food Prot. 63:625–632. Guo, X., J. Chen, R. E. Bracket, and L. R. Beuchat. 2001. Survival of salmonellae on and in tomato plants from the time of inoculation at  owering and early stages of fruit development through fruit ripening. Appl. Environ. Microbiol. 67:4760–4764. Guo, X., J. Chen, R. E. Bracket, and L. R. Beuchat. 2002. Survival of Salmonella on tomatoes stored at high relative humidity, in soil, and on tomatoes in contact with soil. J. Food Prot. 65:274–279. Guo, X., M. W. van Iersel, J. Chen, R. E. Bracket, and L. R. Beuchat. 2002. Evidence of association of salmonellae with tomato plants grown hydroponically in inoculated nutrient solution. Appl. Environ. Microbiol. 68:3639–3643. Hara-Kudo, Y., H. Konuma, M. Iwaki, et al. Potential hazard of radish sprouts as a vehicle of Escherichia coli O157:H7. J. Food Prot. 60:1125–1127. Haywood, A. C. 1974. Latent infections by bacteria. Annu. Rev. Phytopathol. 12:87–97. Itoh, Y., Y. Sugita-Konishi, F. Kasuga, et al. 1998. Enterohaemorrhagic Escherichia coli O157:H7 present in radish sprouts. Appl. Environ. Microbiol. 64:1532–1535.

J. Food Prot., Vol. 66, No. 10

16. Jefferson, R. A., T. A. Kavanagh, and M. W. Bevan. 1987. GUS fusions: b-glucuronidase as a sensitive and versatile gene fusion marker in higher plants. EMBO J. 6:3901–3907. 17. Ji, P., and M. Wilson. 2002. Assessment of the importance of similarity in carbon source utilisation proŽ les between the biological control agent and the pathogen in biological control of bacterial speck of tomato. Appl. Environ. Microbiol. 68:4383–4389. 18. Joce, R., D. G. Sullivan, C. Strong, B. Rowe, M. L. M. Hall, and E. J. Threlfall. 1990. A national outbreak of Salmonella Gold-Coast. Commun. Dis. Rep. 4:3–4. 19. Kaneko, K.-I., H. Hayashidani, Y. Ohtomo, et al. 1999. Bacterial contamination of ready-to-eat foods and fresh products in retail shops and food factories. J. Food Prot. 62:644–649. 20. Kim, J. S., M. Sakai, C. S. Yahng, and T. Matsuguchi. 2001. Comparison of the chemotaxis potential of bacteria isolated from spinach roots and non-rhizosphere soil. J. Microbiol. Biotechnol. 11:160– 163. 21. Lin, C.-M., S. Y. Fernado, and C.-I. Wei. 1996. Occurrence of Listeria monocytogenes, Salmonella spp., Escherichia coli and E. coli O157:H7 in vegetable salads. Food Control 7:135–140. 22. Lund, B. M. 1992. Ecosystems in vegetable foods. J. Appl. Bacteriol. 73:115S–126S. 23. Mahon, B. E., A. Ponka, W. N. Hall, et al. 1997. An international outbreak of Salmonella infections caused by alfalfa sprouts grown from contaminated seed. J. Infect. Dis. 175:876–882. 24. Pirovani, M. E., D. R. Guemes, J. H. Di Pentima, and M. A. Tessi. 2000. Survival of Salmonella harder after washing disinfection of minimally processed spinach. Lett. Appl. Microbiol. 31:143–148. 25. Ponka, A., Y. Anderson, A. Sitonen, B. de Jong, M. Johkola, and O. Haikapa. 1995. Salmonella in alfalfa sprouts. Lancet 345:462–463. 26. Samish, Z., and R. Etinger-Tulczynsha. 1962. Distribution of bacteria within the tissue of healthy tomatoes. Appl. Microbiol. 11:7–10. 27. Sapers, G. M., and G. F. Simmons. 1998. Hydrogen peroxide disinfection of minimally processed fruits and vegetables. Food Technol. 52:48–52. 28. Seo, K. H., and J. F. Frank. 1999. Attachment of Escherichia coli O157:H7 to lettuce leaf surface and bacterial viability in response to chlorine treatment. J. Food Prot. 62:3–9. 29. Shuval, H., Y. Lampart, and B. Fattal. 1997. Development of a risk assessment approach for evaluating wastewater re-use standards for agriculture. Water Sci. Technol. 35:15–20. 30. Solomon, E. B., S. Yaron, and K. R. Matthews. 2002. Transmission of Escherichia coli O157:H7 from contaminated manure and irrigation water to lettuce plant tissue and its subsequent internalisation. Appl. Environ. Microbiol. 68:397–400. 31. Strutz, A. V., B. R. Christie, and J. Nowak. 2000. Bacterial endophytes: potential role in developing sustainable systems of crop production. Crit. Rev. Plant Sci. 19:1–30. 32. Szabo, E. A., K. J. Scurrah, and J. M. Burrows. 2000. Survey for psychrotrophic bacterial pathogens in minimally processed lettuce. Lett. Appl. Microbiol. 30:456–460. 33. Takeuchi, K., and J. F. Frank. 2000. Penetration of Escherichia coli O157:H7 into lettuce tissues as affected by inoculum size and temperature and the effect of chlorine treatment on cell viability. J. Food Prot. 63:434–440. 34. Takeuchi, K., and J. F. Frank. 2001. Quantitative determination of the role of lettuce leaf structures in protecting Escherichia coli O157: H7 from chlorine disinfection. J. Food Prot. 64:147–151. 35. Taormina, P. J., and L. R. Beuchat. 1999. Behaviour of enterohemorrhagic Escherichia coli O157:H7 on alfalfa sprouts during the sprouting process as in uenced by treatments with various chemicals. J. Food Prot. 62:850–856. 36. Tauxe, R. V. 1997. Emerging foodborne diseases: an evolving public health challenge. Emerg. Infect. Dis. 3:425–434. 37. Wachtel, M. R., L. C. Whitehand, and R. E. Mandrell. 2002. Association of Escherichia coli O157:H7 with pre-harvest leaf lettuce upon exposure to contaminated irrigation water. J. Food Prot. 65: 18–25. 38. Wachtel, M. R., L. C. Whitehand, and R. E. Mandrell. 2002. Prevalence of Escherichia coli associated with a cabbage crop inadver-

J. Food Prot., Vol. 66, No. 10

tently irrigated with partially treated sewage wastewater. J. Food Prot. 65:471–475. 39. Warriner, K., S. Spaniolas, M. Dickinson, C. Wright, and W. M. Waites. Internalization of bioluminescent Escherichia coli and Salmonella Montevideo in growing bean sprouts. Submitted for publication. 40. Winson, M. K., S. Swift, P. J. Hill, et al. 1998. Engineering the luxCDABE genes from Photorhabdus luminescens to provide a bioluminescent reporter for constitutive and promoter probe plas-

E. COLI IN GROWING SPINACH

1797

mids and mimi-Tn5 constructs. FEMS Microbiol. Lett. 163:193– 202. 41. Zhuang, R. Y., and L. R. Beuchat. 1996. Effectiveness of trisodium phosphate for killing Salmonella Montevideo on tomatoes. Lett. Appl. Microbiol. 22:97–100. 42. Zhuang, R. Y., L. R. Beuchat, and F. J. Angulo. 1995. Fate of Salmonella Montevideo on and in raw tomatoes as affected by temperature and treatment with chlorine. Appl. Environ. Microbiol. 61: 2127–2131.