Contaminated vegetable seeds have been identified as a potential source of foodborne bacterial pathogens. This study was undertaken to observe the behavior of Salmonella and enterohemorrhagic Escherichia coli (EHEC) on vegetable seeds, contaminated by direct contact with artificially inoculated soil, during germination. Sterile sandy soil inoculated with lyophilized cells of four individual strains of Salmonella or EHEC (three O157:H7 strains and one O104:H4 strain) was mixed with sanitized seeds (2 g) of alfalfa, fenugreek, lettuce, and tomato at 20°C for 1 h. The contaminated seeds were germinated on 1% water agar at 25°C for 9 days in the dark. Populations of Salmonella and EHEC on various tissues (seed coat, root, cotyledon, and stem, etc.) of sprouts and seedlings were determined every other day over the germination period. Overall, 70.4 and 72.4% of collected tissue samples (n = 544) tested positive for Salmonella and EHEC, respectively. In general, the mean populations of Salmonella and EHEC on sprout and seedling tissues increased with the prolongation of germination time. Seed coats had the highest bacterial counts (4.00 to 4.06 log CFU/0.01 g), followed by the root (3.36 to 3.38 log CFU/0.01 g), cotyledon (3.13 to 3.38 log CFU/0.01 g), and stem tissues (2.67 to 2.84 log CFU/0.01 g). On average, tissue sections of fenugreek sprouts and lettuce seedlings had significantly higher (P < 0.05) numbers of Salmonella and EHEC cells than that of alfalfa sprouts and tomato seedlings. Data suggest that the growth and dissemination of Salmonella and EHEC cells on alfalfa, fenugreek, lettuce, and tomato sprout and seedling tissues are influenced by the type of vegetable seeds and sprout and seedling tissues involved. The study provides useful information on the fate of two important foodborne bacterial pathogens on selected vegetable seeds, contaminated by direct contact with inoculated soil, during the germination process.
Vegetable seeds were contaminated via contact with pathogen-bearing sandy soil.
Pathogens on contaminated seeds were recovered from tissues of sprouts and seedlings.
Tomato and alfalfa tissues had lower pathogen counts than fenugreek and lettuce tissues.
Seed coats had higher pathogen counts than the root and cotyledon tissues.
Stem tissues had lower Salmonella/EHEC counts compared with all other tissues.
Regular and sufficient consumption of fresh produce is part of a healthy lifestyle (6). In the United States, per capita consumption of fresh vegetables in 2018 was 143.4 lb (36). Because a large portion of the fresh produce is eaten raw, foodborne outbreaks associated with the consumption of these products have occurred (26, 30). Fresh produce that has been linked to the outbreaks of human gastrointestinal infections included lettuce, tomato, and alfalfa and fenugreek sprouts (3, 8, 13, 39). Enterohemorrhagic Escherichia coli (EHEC) and Salmonella enterica are among the major bacterial causes of foodborne illnesses (42).
Fresh produce contamination by bacterial pathogens could occur at both pre- and postharvest stages in the production chain. Contaminated vegetable seeds could lead to the contamination of fresh produce, especially vegetable sprouts (20). Sprout seeds could become contaminated at any time during harvest, extraction, threshing, processing, or packaging. Cells of bacterial pathogens such as those of EHEC and Salmonella that have gained entry into or onto sprout seeds prior to germination are likely to grow and disseminate to the outer surfaces as well as the inner tissues of vegetable sprouts and seedlings (7, 22). Numerous studies have been conducted to understand the growth pattern of S. enterica and EHEC during seed germination, especially the germination of sprout seeds contaminated by immersion into bacterial suspensions (9, 17, 21). This inoculation approach is usually used to mimic the natural process of seed contamination by filthy water. In addition to this route, vegetable seeds could also be contaminated through pathogen infiltration and direct contact with soil or plant debris harboring pathogens. It is not yet clear whether human pathogens on vegetable seeds contaminated by different approaches would behave in a similar manner during seed germination because these approaches might introduce pathogen cells into different locations on or within a vegetable seed (32). We previously reported the fate of Salmonella and EHEC on vegetable seeds, contaminated by immersion into bacterial cell suspensions (11) or artificially internalized using vacuum infiltration (25) during germination. This study was undertaken to observe the behavior of selected Salmonella and EHEC on seeds of alfalfa, fenugreek, lettuce, and tomato, contaminated through direct contact with pathogen-bearing soil, and to observe the growth of pathogen cells on various tissues of sprouts and seedlings developed from the contaminated seeds during the germination process. The four types of vegetable seeds were selected based on the fact that alfalfa and fenugreek sprouts, tomato (fruit), and lettuce (leafy green) are commonly consumed fresh produce, which have all been previously linked to outbreaks of human gastrointestinal infections (3, 13, 19, 27, 28, 46).
MATERIALS AND METHODS
Vegetable seeds, bacterial strains, and growth conditions
Alfalfa (Medicago sativa, cv. unidentified), fenugreek (Trigonella foenum-graecum, cv. unidentified), lettuce (Lactuca sativa cv. Iceberg), and tomato (Solanum lycopersicum cv. Roma) seeds were obtained from a commercial source (Otis S. Twilley Seed Co. Inc., Hodges, SC), stored at 10°C, and used within a month.
Four S. enterica, three E. coli O157:H7, and one E. coli O104:H4 strain were used in the study (Table 1). Stock cultures of the bacterial strains were retrieved from storage at −70°C and grown on tryptic soy agar (TSA) at 37°C for 16 h. The resulting cultures were purified on bismuth sulfite agar (BSA), sorbitol MacConkey (SMAC) agar, and MacConkey (MAC) agar (BD, Sparks, MD), respectively. Spontaneous mutant cells resistant to 50 μg/mL nalidixic acid (MP Biomedicals, Santa Ana, CA) were selected and used throughout the study.
Preparation of freeze-dried Salmonella and EHEC cells
Overnight cultures (10 mL) of each Salmonella and EHEC strain grown in tryptic soy broth (BD) supplemented with nalidixic acid (50 μg/mL) were centrifuged at 5,000 × g for 5 min. The cell pellets were washed with 10 mL of sterile deionized water, and washed bacterial cells were resuspended in 20 mL of 10% sterile skim milk (Walmart, Bentonville, AR) to reach a final cell concentration of ca. 8 log CFU/mL. Each bacterial suspension (5 mL) in a 10-mL glass test tube (Fisher Scientific, Asheville, NC) was kept at −20°C in a static state for 24 h. The samples were then desiccated in a Free Zone Benchtop Freeze Dry System (Labconco, Kansas City, MO) at a condenser temperature of −40°C and a chamber pressure <0.05 mbar for 10 h. Freeze-dried bacterial cells were kept at −20°C until use.
Seed contamination through direct contact with artificially inoculated soil
Vegetable seeds described above were sanitized to inactivate the background microflora, according to the protocol described by Cui et al. (12), with some modification. Briefly, each type of vegetable seed (2 g) was sanitized with 10 mL of 20,000 ppm of sodium hypochlorite (pH 6.8; BD) at room temperature for 15 min with gentle mixing. Residual chlorine on sanitized seeds was neutralized with 10 mL of Dey-Engley broth (BD) for 10 min with gentle mixing and then rinsed twice, each with 10 mL of sterile deionized water for 1 min. The seeds were placed on sterile paper towels and dried in a biological safety cabinet (class II type A/B 3, Nuaire, Plymouth, MN) overnight at room temperature.
Twenty grams of sandy soil (Mosser Lee Co. Millston, WI) was autoclaved at 121°C for 15 min. Sterilized soil (water activity 0.34) was mixed with 106 CFU of freeze-dried, individual Salmonella or EHEC culture, in a Whirl-Pak bag (1 oz; Nasco, Fort Atkinson, WI), by agitation on an orbital platform shaker (model 3520, Lab-Line Instruments, Melrose Park, IL) at 200 rpm and room temperature for 12 h. The contaminated soils (20 g; 3.28 to 3.43 log CFU/g) were then mixed with sanitized vegetable seeds (2 g) by agitation under the same conditions described above for 1 h.
Germination of vegetable seeds and growth of sprouts and seedlings
For seed germination, 1% (w/v) water agar (BD) in sterile squared petri dishes with grid (10 by 10 cm; Electron Microscopy Sciences, Hatfield, PA) were used. Each type of vegetable seed contaminated with the inoculated sandy soil (n = 50) was placed with a proper spacing (10 seeds per plate) onto the water agar plate, using small, sterile curved forceps (Fisher Scientific). The petri dishes with the vegetable seeds were placed in transparent plastic boxes (Walmart), the bottom of which was covered with a layer of damp paper towels. The boxes were placed in a 25°C incubation room in the dark to allow the seeds to germinate for 9 days. Different tissues of sprouts and seedlings developed from each type of vegetable seed were taken every other day for microbiological analysis.
Sample preparation and microbiological analyses
On the first day of germination, pregerminated vegetable seeds were analyzed. On the third day of germination, developed sprouts and seedlings were carefully dissected using sterile forceps and scissors, and the seed coat–cotyledon (combination of seed coat and cotyledon due to difficulty in separating the two tissue sections), stem, and root tissues were collected. The roots were tissues with fibrils, and the boundary between root and stem tissues was determined by their positions in relation to the surface of water agar. The portion above the surface of water agar was taken as stem tissues and that beneath the surface of agar was the root tissues. On day 5 and onward, seed coat and cotyledon were analyzed separately.
An individual tissue section of sprouts and seedlings (cotyledon, root, or stem) developed from a composite sample of five vegetable seeds of an individual seed type (alfalfa, fenugreek, lettuce, or tomato) and inoculated with a single pathogen strain (each of the four Salmonella or E. coli strains) was ground, using a pestle, for 1 min in 5 mL of 0.1 M phosphate-buffered saline (pH 7.4) in a Whirl-Pak bag (1 oz; Nasco). Appropriate 10-fold serial dilutions of each sample were plated onto BSA, MAC, or SMAC with nalidixic acid to quantify the population of Salmonella, E. coli O104:H4, and E. coli O157:H7, respectively. Additionally, all samples were plated on TSA amended with nalidixic acid. When the numbers of cells dropped below the detection limit, enrichment was performed according to protocols outlined in the Bacteriological Analytical Manual (2, 15).
Two independent trials were conducted. Fisher's least significant difference test in the general linear model was conducted, using the Statistical Analysis Software (version 9.4, SAS Institute Inc., Carey, NC), to determine the difference in the cell population (log CFU per 0.01 g of seedling and sprout tissues) of various Salmonella or EHEC strains recovered from different sprouts and seedlings tissues. In addition, pathogen counts recovered at different sampling points during germination, from all sprout and seedling tissue sections developed from a single seed type, and from individual tissue section of all four types of sprouts and seedlings were also compared. All the tests were performed based on a confidence interval of 95%.
Overall statistical analysis
Among a total of 1,088 sprout and seedling tissue samples analyzed in the Salmonella (n = 544) or EHEC (n = 544) experiment (Table 2), 70.4% of the samples tested positive for Salmonella and 72.4% tested positive for EHEC (data not shown). Of Salmonella serotypes, Baildon had the highest mean cell population, followed by Stanley, Cubana, and Montevideo (Table 2A). E. coli O104:H4 BAA-2326 had the lowest recovery cell population from sprout and seedling tissues among the four EHEC strains used in the study. Among the three E. coli O157:H7 strains, F4546 had the highest mean cell population, followed by K4492 and then H1730.
Seed coat had the highest mean population of Salmonella (Table 2B). The mean Salmonella populations from the root and cotyledon samples were similar (P > 0.05), but they were significantly higher (P < 0.05) than those from the seed coat–cotyledon and stem samples. Pregerminated vegetable seeds had the lowest mean Salmonella population. A similar trend was observed with EHEC-contaminated samples, except that the mean cell populations on the seed coat–cotyledon samples were statistically similar to those from root and cotyledon samples.
On average, fenugreek sprouts and lettuce seedlings had significantly higher (P < 0.05) mean EHEC populations than alfalfa sprouts and tomato seedlings (Table 2C). The average cell population of the Salmonella strains was the highest on fenugreek sprouts, followed by lettuce seedlings, alfalfa sprouts, and tomato seedlings.
Pathogen population change over time during germination
The mean populations of EHEC on tissues of all sprouts and seedlings increased significantly (P < 0.05) with the prolongation of germination time (Table 2D). Its mean population increased from 0.06 log CFU/0.01 g on day 1 to 3.92 log CFU/0.01 g on day 9. A similar trend was observed with samples contaminated with Salmonella, except that there was no significant difference (P > 0.05) between the mean cell populations recovered on days 5 and 7. Mean Salmonella population was 0.15 log CFU/0.01 g on day 1 compared with 4.01 log CFU/0.01 g on day 9.
Detailed Salmonella and EHEC population change on tissues of sprouts and seedlings developed from individual seed types is shown in Figure 1. In general, cell populations of Salmonella and EHEC on tissues of fenugreek sprouts and lettuce seedlings increased as germination time increased (Fig. 1). The mean Salmonella population on alfalfa sprouts increased to ca. 3.7 log CFU/0.01 g on day 5 of germination and then decreased gradually (Fig. 1a). A significant increase in the mean population of Salmonella on tomato seedlings did not occur until day 7 of the germination process. The mean EHEC population recovered from tissues of alfalfa sprouts was stable after a significant growth on the first day of germination (Fig. 1b). A significant increase in EHEC population on tomato seedlings was observed from day 5 to day 7 and remained relatively stable thereafter.
Mean population of Salmonella and EHEC on individual tissue sections of each sprout and seedling
The average cell populations of all four Salmonella or EHEC strains from different tissue sections of each type of sprout and seedling over the 9-day germination period are shown in Tables 3 and 4, respectively. Pregerminated seeds of all vegetable types had the lowest mean populations of Salmonella (Table 3). Salmonella counts from fenugreek seed coats were significantly higher (P < 0.05) than those from other fenugreek tissues. The average Salmonella populations from fenugreek seed coat–cotyledon, cotyledon, and root samples were statistically similar (P > 0.05), but they were significantly higher than the cell population from the stem samples. A similar trend was found on lettuce tissues, except that similar populations of Salmonella were found on seed coat–cotyledon, cotyledon, root, and stem tissues. Similar Salmonella populations were also associated with all alfalfa tissue samples. Tomato seed coat–cotyledon tissues had significantly lower Salmonella population than other tomato tissues.
Similar to Salmonella-contaminated samples, all four types of pregerminated vegetable seeds had the lowest EHEC counts (Table 4). EHEC counts from seed coats of fenugreek sprouts and lettuce seedlings were significantly higher (P < 0.05) than those from other samples of the same seed types. The average cell populations from the seed coat–cotyledon and root samples were significantly higher than the populations from the cotyledon samples and the stem samples. Similar to what was found with Salmonella, there was no significant difference (P > 0.05) among EHEC populations recovered from the alfalfa tissue samples. The populations of EHEC on tomato seed coat, root, and cotyledon (1.33 to 2.25 log CFU/0.01 g) were not significantly different, but the recovery population from the seed coat (2.25 log CFU/0.01 g) was significantly higher than that from the stem (1.24 log CFU/0.01 g) and seed coat–cotyledon (0.30 log CFU/0.01 g) tissue samples.
Mean populations of individual Salmonella and EHEC strains on sprout and seedling tissues developed from individual seed type
The average cell populations of the four individual Salmonella or EHEC strains from all tissue samples of each type of vegetable seeds over the 9-day germination period are shown in Tables 5 and 6, respectively. Salmonella Baildon and Salmonella Stanley had the highest mean populations on tissues of fenugreek sprouts and lettuce seedlings, followed by alfalfa sprouts and then tomato seedlings (Table 5). A similar trend was found in the cell populations of Salmonella Montevideo, except that there was no significant difference between the cell populations recovered from alfalfa sprouts and tomato seedlings. Salmonella Cubana established the highest cell population on fenugreek sprouts, followed by lettuce seedlings; the cell populations associated with alfalfa sprouts and tomato seedlings were significantly lower (P < 0.05) than the other two types of sprout and seedling tissues.
The mean populations of the four individual Salmonella strains associated with tomato tissues were not significantly different (P > 0.05). Salmonella Montevideo had the lowest mean populations on lettuce tissues compared with the other three types of sprouts and seedlings. Salmonella Baildon and Salmonella Stanley had significantly higher mean cell populations than Salmonella Cubana on alfalfa and lettuce tissues. No significant differences were noticed among the mean populations of Salmonella serotypes Baildon, Stanley, and Cubana on fenugreek tissues.
All four individual EHEC strains established significantly higher cell populations on fenugreek and lettuce tissues than alfalfa and tomato tissues (Table 6). E. coli O104:H4 BAA-2326 had the lowest mean populations on all tested vegetable tissues compared with the three E. coli O157:H7 strains. Strain F4546 and K4492 had a significantly higher (P < 0.05) mean population than those of H1730 on fenugreek and lettuce tissue. Strain F4546 also had a significantly higher mean population than those of K4492 on alfalfa and tomato tissues, followed by those of H1730.
In the current study, S. enterica and EHEC cells were recovered from various tissues of sprouts and seedlings developed from vegetable seeds contaminated with soil-borne Salmonella or EHEC (Table 2). The precise mechanism of pathogen uptake by vegetable seeds from contaminated soil is unknown. However, it has been suggested that when vegetable seeds were exposed to soil-borne bacterial pathogens, seeds could be contaminated by pathogen cells either directly or by pathogen-borne soil particles (1). The latter circumstance is known as seed concomitant contamination, which could occur in several forms, including propagules and infected plant debris such as infected straw or pieces of chaff (1). The concomitant contamination by pathogen-bearing soil particles has been observed with some plant pathogens such as Verticillium albo-atrum in alfalfa seeds (37), Fusarium oxysporum f. sp. lycopersici in tomato seeds (1), and Fusarium solani f. sp. phaseoli in bean seeds (29). After establishing a close association with vegetable seeds, pathogen cells can grow and disseminate along the surface of vegetable seedlings and/or directly invade various tissues at any stages of the plant development (5, 23, 34, 45).
The present study used vegetable seeds contaminated by pathogen-bearing soil as one of the experimental materials. Recent studies conducted in our laboratory used two other approaches to inoculate vegetable seeds before germination, i.e., artificial inoculation of pathogens (same strains of Salmonella and EHEC) by immersing vegetable seeds into bacterial suspensions (11) or artificial internalization of pathogen cells into vegetables seeds using vacuum infiltration (25). Findings that were similar among the three studies were noticed, e.g., bacterial populations on sprout and seedling tissues increased as germination time increased and lower pathogen populations were associated with tomato seedling tissues compared to other sprout and seedling tissues. Dissimilar findings among the three studies were also noticed. First, the incidence of pathogen-positive sprout and seedling tissues developed from vegetable seeds contaminated through immersion into bacterial suspensions was much higher (ca. 92% (11)) than that by mixing seeds with contaminated soil (71% [present study]) and by pathogen infiltration into vegetable seeds (60% (25)). Second, cotyledon tissues along with seed coats had the highest mean level of recovered bacterial cells when pathogen cells were introduced to vegetable seeds by vacuum infiltration; whereas, in the other two circumstances, seed coats had the highest pathogen counts.
The present study found that the mean population of the four EHEC strains on tissues of fenugreek sprouts and lettuce seedlings increased significantly (P < 0.05) at each sampling point of the germination, and cell populations peaked on day 3 (Fig. 1b). However, in the study of Gomez-Aldapa et al. (17), the population of diarrheagenic E. coli on mung bean sprouts peaked on day 2 of the germination, one day sooner than what was observed in the present study. This difference could be attributed to the approach that the two studies used to contaminate vegetable seeds. In the present study, vegetable seeds were contaminated by direct contact with pathogen-bearing soil; Gomez-Aldapa et al. (17) submerged seeds into bacterial suspensions. It is known that imbibition of water by dry seeds is the initial step of the germination process. The submersion inoculation approach enables the influx of water into dry seeds, triggering the onset of the germination process. It was observed in the present study that lettuce seeds had no visible buds after 24 h of germination (results not shown). However, seeds were contaminated by submersion in a previous study of our laboratory; seedling buds as long as ca. 2 mm were noticed after the same length of germination (11). The submersion inoculation approach has been used in two other studies (7, 43), and the bacterial cell populations in these studies all peaked within 1 to 2 days of germination.
The present study found that seed coats, among evaluated sprout and seedling tissues, had the highest number of pathogen cells (Table 2). This suggests that most soilborne bacterial cells were associated with the surface of seeds, although some might have gained entrance into vegetable seeds through surface cracks. In addition to seed coats, higher pathogen populations were found on the root and cotyledon than stem tissues (Table 2). This observation was in agreement with a previous study by Warriner et al. (43), who found that the root tissues of mung bean sprouts had higher Salmonella Montevideo and E. coli populations than the stem tissues. Cooley et al. (10) also reported that E. coli O157:H7 and Salmonella Newport populations on Arabidopsis thaliana roots were higher than on the stem tissues. Seedling roots along with cotyledon are tissues related to seed exudation. Vegetable seed exudates contain a large amount of nutrients such as carbohydrates and amino acids, which support the growth of bacterial cells (33, 44). It has been reported that bacterial cells tend to move toward nutrient-rich regions in capillary tubes, in soil (35), and in plants (23, 47). On tissues with limited nutrients, such as those of stems, several cellular functions in bacteria are affected, and, thus, cells grow poorly (38, 43).
Results of the present study revealed that pathogen growth on tissues of tomato seedlings were significantly less profound (P < 0.05) than those on fenugreek sprouts and lettuce seedlings (Table 1). The precise reason for the observed population difference is currently unknown. However, the chemical composition of seed exudates might play a role (14). Tomato exudates contain organic acids and a range of polyphenols that may affect pathogen growth on tissues of tomato seedlings (41). Tu (40) found trace amount of reducing sugars and amino acids after 5 days of tomato seed germination, which could explain why a significant increase in the mean populations of Salmonella and EHEC cells on tomato tissues did not occur until 5 days into the germination process in the present study (Fig. 1). Furthermore, differences in the initial prevalence of pathogen-contaminated vegetable seeds and level of pathogen cells associated with vegetable seeds might be partially responsible for the observed phenomenon. A positive correlation between the initial and final levels of Salmonella on artificially contaminated alfalfa seeds by submersion inoculation was reported by Liao and Fett (24). Vegetable seeds used in this study varied in mass and size, as well as seed surface properties, which may have influenced the initial association of bacterial pathogen with vegetable seeds.
In the present study, the populations of the four Salmonella strains recovered from sprout and seedling tissues were significantly different (P < 0.05; Table 1). In a previous study, Klerks et al. (23) contaminated lettuce seedlings with strains of S. enterica of five different serotypes using vacuum infiltration and found that the population of individual Salmonella strains on lettuce leaves was significantly different. Howard and Hutcheson (21) reported that the growth of 13 S. enterica strains with eight different serotypes on germinating alfalfa seeds is serotype-independent. However, according to Olsen et al. (31), approximately 100 Salmonella serotypes are frequently associated with human infections. The current study included only a limited number of Salmonella strains with four different serotypes; thus, more research is needed to conclude whether Salmonella growth on germinating vegetable seeds is indeed serotype dependent.
The present study recovered Salmonella and EHEC cells from lettuce and tomato seedlings developed from contaminated seeds (Tables 2 through 6). However, the occurrence of pathogens cells on mature lettuce and tomato fruits will depend on their fate at later stage of plant development, which could be significantly influenced by environmental conditions (4). Franz et al. (16) found that lettuce leaves grown from 15 and 28% of the contaminated seedlings under greenhouse growth conditions (15°C with a relative humidity of 60%) tested positive for Salmonella and E. coli O157:H7, respectively, with an average cell population of 2 to 3 log CFU/g. Gu et al. (18) isolated Salmonella from tomato fruits grown from contaminated seedlings under greenhouse condition with an average temperature of 28°C. Furthermore, enteric bacterial pathogens could survive and even grow on plant surfaces after harvest, depending on temperature, water availability, level of tissue damage, and available nutrients (5, 10).
Results of the study suggest that Salmonella and EHEC from vegetable seeds contaminated through direct contact with artificially inoculated soil disseminated to various tissues of alfalfa, fenugreek, lettuce, and tomato sprouts and seedlings. The fate of pathogen cells on tissues of sprouts and seedlings was influenced by the type of vegetable seeds used, as well as the sprout and seedling tissues involved. The study fills the knowledge gaps in the current body of literature by providing information on the behavior of two major foodborne bacterial pathogens, introduced to vegetable seeds by direct contact with artificially inoculated soil, during the germination process.
This material is based upon work that is supported by the National Institute of Food and Agriculture, U.S. Department of Agriculture, under award no. 2014-67017-21705. The authors thank Dr. Yaa Asantewaa Kafui Klu for editorial assistance.