INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

In recent years, outbreaks of infections associated with raw and minimally processed fruits and vegetables have occurred with increased frequency (16). Factors thought to influence this increase include changes in agronomic practices and dietary habits and increased importation of fresh produce (2). Salmonella serotypes are estimated to cause approximately 1.5 million cases of foodborne infection each year in the United States, with 15,000 hospitalizations and 500 deaths (24). Foods of animal origin, such as poultry, eggs, meat, and dairy products, have been historically recognized as vehicles of Salmonella. However, salmonellosis has also been associated with consumption of tomatoes (13; R. V. Tauxe [Centers for Disease Control and Prevention], personal communication; R. C. Wood, C. Hedberg, and K. White, Abstr. Epidemic Intelligence Service 40th annual conference, 1991), seed sprouts (23, 26; C. A. Van Beneden, W. E. Keene, D. H. Werker, A. S. King, P. R. Cieslak, K. Hedberg, R. A. Strang, A. Bell, and D. Fleming, Abstr. 36th Intersci. Conf. Antimicrob. Agents Chemother., abstr. K47, 1996), watermelons (6, 11, 18), cantaloupes (12; A. A. Ries, S. Zaza, C. Langkop, R. V. Tauxe, and P. A. Blake, Abstr. 30th Intersci. Conf. Antimicrob. Agents Chemother., abstr. 915, 1990), and unpasteurized apple cider (10) and orange juice (14; K. Cook, D. Swerdlow, T. Dobbs, J. Wells, N. Puhr, G. Hlady, C. Genese, L. Finelli, B. Toth, D. Bodager, K. Pilot, and P. Griffin, Abstr. 36th Intersci. Conf. Antimicrob. Agents Chemother., abstr. K49, 1996).

In the farm-to-table production, processing, and distribution chain, there are various possible points of contamination of fruits and vegetables with disease-causing microorganisms. These include irrigation water, manure, wash water, handling by workers, and contact with contaminated surfaces (5, 28). It is essential that interventions be developed to prevent or minimize contamination of raw produce and to remove pathogens prior to consumption. To date, however, none of the chemical or physical treatments presently authorized by regulatory agencies for use to disinfect raw produce can be relied on to eliminate all types of pathogens from the surface or internal tissues (4). One of the keys to enable the selection of appropriate intervention steps to reduce populations of pathogenic microorganisms on fruits and vegetable is to identify sources of contamination and to characterize the ecology of pathogens as it is affected by agronomic and processing practices (4, 8, 9).

Tomato plants are grown for its edible fruit. The United States is one of the five leading tomato-producing countries. In 1985, per capita consumption of raw tomato fruit in the United States was 16.6 lb, increasing to 18.8 lb in 1995 (29). It is anticipated that the per capita consumption of raw tomatoes will continue to increase. Consumption of raw tomatoes has been epidemiologically linked to 176 cases of Salmonella enterica serotype Javiana infections in Illinois, Michigan, Minnesota, and Wisconsin in 1990 (Wood et al., Abstr. Epidemic Intelligence Service 40th annual conference). In 1993, tomatoes were identified as the vehicle for a multistate outbreak of S. enterica serotype Montevideo infection (13). More recently, S. enterica serotype Baildon was implicated in an outbreak with diced tomatoes in geographically separate areas of the United States (R. V. Tauxe [Centers for Disease Control and Prevention], personal communication).

Zhuang et al. (32) described conditions influencing survival and growth of S. enterica serotype Montevideo on the surfaces of intact tomatoes. Rapid growth occurred in chopped ripe tomatoes (pH 4.1 ± 0.1) at ambient temperature. S. enterica serotypes Enteritidis, Infantis, and Typhimurium were reported to grow in fresh-cut tomatoes (pH 3.99 to 4.37) at 22 and 30°C (3). Wei et al. (30) reported that S. enterica serotype Montevideo is able to multiply on wounded and cut tomatoes. The acidic pH (4.2 to 4.39 for ripe tomatoes and 4.33 to 4.52 for green tomatoes) did not completely inhibit growth. Weissinger et al. (31) reported that S. enterica serotype Baildon can grow in diced tomatoes (pH 4.40 ± 0.01); 0.79 log₁₀ CFU/g increased to 5.32 log₁₀ CFU/g and 7.00 log₁₀ CFU/g within 24 h at 21 and 30°C, respectively. However, after treatment with 200 µg of chlorine per ml, diced tomatoes initially containing 0.60 to 0.86 log₁₀ CFU of S. enterica serotype Baildon per g still harbored the pathogen (31). Chlorinated water is more effective in removing S. enterica serotype Montevideo from tomato skin or inactivating it than it is in doing so in internal core tissue (30).

Ercolani and Casolari (17) demonstrated possible internalization of bacteria into tomato fruits by spraying tomato blossoms with a suspension of a plant pathogen, Xanthomonas campestris pv. vesicatoria. Typical symptoms of disease were observed on tomato leaves one month after inoculation. The bacterium was isolated from the centers of fruits that did not show external symptoms of infection. A pathogenic isolate of Erwinia carotovora was injected into the centers of healthy cucumber fruits attached to the vine without causing disease (25). However, the bacterium was detected in the internal tissues of fruits harvested from the inoculated plants. Samish et al. (27) studied 10 fruits and vegetables grown on different farms and found that bacteria, mostly gram-negative motile rods, representatives of the Pseudomonadaceae and the Enterobacteriaceae, can occur within healthy, sound raw fruit tissues.

The fate of human pathogenic bacteria applied to tomato flowers or inoculated into tomato stems before or after fruits set has not been described. We hypothesized that pathogens such as Salmonella may migrate through tomato stems and internalize in tomato fruits. The possibility of internalization of Salmonella in tomato fruits developed from inoculated flowers was also considered. The objective of this study was to determine the fate of Salmonella inoculated into tomato stems and onto tomato flowers.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial cultures. Five Salmonella enterica serotypes were used: Montevideo (serogroup C₁) was isolated from a patient in a tomato-associated outbreak, Michigan (serogroup J) was isolated from cantaloupe, Poona (serogroup G) was isolated from a patient in a cantaloupe-associated outbreak, Hartford (serogroup C₁) was isolated from a patient in an orange juice-associated outbreak, and Enteritidis (serogroup D) was isolated from a patient in an egg-associated outbreak.

Inoculum preparation. Stock cultures maintained on brain heart infusion (BHI) agar (Difco Laboratories, Detroit, Mich.) at 4°C were transferred to BHI broth and incubated at 37°C. Cultures were transferred three times at 24-h intervals. Cells were harvested when the optical density at 600 nm (OD₆₀₀) reached ca. 1.0, which is equivalent to 10⁹ CFU/ml. Each culture was centrifuged at 10,000 × g for 5 min, washed with 0.01 M sodium phosphate buffer (pH 7.2) containing 0.85% sodium chloride (phosphate-buffered saline) twice, and resuspended in 5 ml of phosphate-buffered saline. Equal volumes of cell suspensions of each serotype were combined to form the inoculum for tomato plants.

Preparation of tomato plants for inoculation. `Better Boy' tomato plants were used. Young, healthy plants ca. 20 cm in height were purchased from a local market. Plants were grown in a greenhouse (6.7 m by 7.3 m) equipped with ridge vents, an evaporative cooler, and a gas heater. The temperature was maintained at 25°C. Each plant was transferred to potting soil (Scotts Miracle-Gro Products, Port Washington, N.Y.) in a 3.8-liter plastic pot on 6 June and placed in a polypropylene tray (Nalge Company, Rochester, N.Y.) to retain water that filtered through the soil after application of water to the soil surface. Water was applied daily, and Hoagland's nutrient solution (20) was applied weekly in volumes to maintain optimum soil moisture for plant growth and flowering and fruit development.

Inoculation procedures. Tomato plants were inoculated with Salmonella when they started to bloom, which was 57 to 76 days after transplanting. One hundred open flowers on eight plants were gently brushed by using a small paintbrush saturated with inoculum. Stems (1 to 2 cm in diameter) were inoculated with cell suspension at a location ca. 5 cm from the flower base. Inoculum (50 µl) containing ca. 7.5 log₁₀ CFU was deposited on the stem, which was then pricked with a 25-gauge syringe needle to facilitate contact with subsurface tissue. Stems were inoculated either before or after fruits set. Forty-nine stems were inoculated before fruiting, and 41 stems were inoculated when fruits were 1 to 2 cm in diameter. Uninoculated plants served as controls. Inoculated and control plants were grown in the same greenhouse.

Microbiological analysis. Tomatoes were harvested when subjectively judged to be "red ripe" and ready for consumption (Table 1). The weight of tomatoes ranged from 21 to 75 g. Each tomato was hand picked into a plastic zip-lock bag, which was sealed and transported to the laboratory for analysis within 1 h. Tomatoes were immersed in 70% ethanol for 2 min for surface disinfection and then dried in a laminar flow hood at 22°C ± 1°C for 30 min. Each tomato was placed in a stomacher bag containing 20 ml of 0.1% sterile peptone water at 37°C and hand rubbed for 2 min to dislodge surface populations of Salmonella that may have evaded contact with ethanol. The peptone wash water was surface plated (0.25 ml in quadruplicate and 0.1 ml in duplicate) on bismuth sulfite agar (BiSA; Difco). Plates were incubated at 37°C for 24 h before examination for presumptive colonies of Salmonella. The stem scar tissue and pulp of each tomato were analyzed for the presence of Salmonella. Tomatoes were removed from the peptone wash water, and the stem scar tissue was removed with a sterile scalpel. The stem scar tissue and remainder (pulp) of the tomato were separately placed in stomacher bags with 10 ml and 20 ml of sterile 0.1% peptone water, respectively, and pummeled at medium speed for 1 min with a Stomacher 400 instrument (Seward, London, United Kingdom). Four 0.25-ml portions and two 0.1-ml portions of homogenate were surface plated on BiSA. Plates were incubated at 37°C for 24 h before presumptive colonies were counted. Stem scar tissue and pulp homogenates were enriched by adding, respectively, 10 ml and 20 ml of universal preenrichment broth (Difco) and incubating the mixture at 37°C for 24 h. Cultures were streaked on BiSA and incubated at 37°C for 24 h before examination for presumptive Salmonella colonies.

Confirmation of presumptive Salmonella colonies by using PCR identification and DNA-based typing. PCR assays using HILA2 primer sets were conducted according to the procedure described by Guo et al. (19) to confirm presumptive isolates from inoculated tomatoes. PCR fingerprinting was also done to compare serotypes of isolates to those in the inoculum. The primer used for PCR fingerprinting was (5'-3') AAG TAA GTG ACT GGG GTG AGC G, based on a highly conserved, enterobacterial repetitive intergenic consensus sequence, which consists of 126 bp and appears to be restricted to transcribed regions of the genome, either in intergenic regions of polycistronic operons or in untranslated regions upstream or downstream of open reading frames (21). Crude DNA was prepared by boiling 20-h cultures of isolates in BHI broth for 10 min. One milliliter of culture was centrifuged at 12,000 × g for 2 min. Pellets were resuspended in 200 µl of sterile distilled water, boiled for 10 min, and centrifuged as described above. A 5-µl sample was used as a template for PCR. The 50-µl PCR mixture contained PCR buffer, deoxynucleoside triphosphates (0.4 mM each), primers (1 µM), Taq polymerase (1 U; Roche Diagnostics, Indianapolis, Ind.), and DNA template. PCRs were performed in a DNA thermal cycler 480 apparatus (Perkin Elmer, Norwalk, Conn.) using a cycle at 94°C for 5 min, followed by 40 cycles of 92°C for 45 s, 25°C for 1 min, and 68°C for 10 min, with a final extension at 72°C for 20 min. The PCR amplicons were analyzed by gel electrophoresis on 1% agarose (GIBCO BRL, Rockville, Md.) gel in TBE buffer (0.089 M Tris-borate, 0.002 M EDTA, pH 8.0). The gel was stained with ethidium bromide and visualized by using a Gel Doc System 2000 (Bio-Rad Laboratories, Hercules, Calif.).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Approximately 80% of the inoculated flowers abscised from plants within 10 days after inoculation, regardless of the site or time of inoculation relative to fruit set. About the same percentage of flowers abscised from uninoculated plants. This may be due to lack of pollination or other factors. Forty-three sound, red, ripe tomatoes were harvested and analyzed for the presence of Salmonella (Table 1). Thirteen of these tomatoes were from plants not inoculated with Salmonella, and 30 were from plants inoculated by either stem injection or flower brushing. Salmonella was not detected on tomatoes from uninoculated control plants or by direct plating samples of peptone wash water, stem scar tissue homogenates, or pulp homogenates of tomatoes from control or inoculated plants on BiSA agar. However, presumptive Salmonella was detected in enriched samples of peptone wash water, stem scar tissue, and pulp of tomatoes from inoculated plants. Salmonella was detected on or in tomatoes from plants receiving stem inoculation before or after flower set and on or in tomatoes that developed from inoculated flowers (Table 1).

Eleven of thirty tomatoes (37%) harvested from inoculated plants were positive for Salmonella. Of the tomatoes receiving stem inoculation before and after flower set, 43% and 40%, respectively, were positive for Salmonella. Twenty-five percent of the tomatoes produced from inoculated flowers were positive for the pathogen. Compared to pulp of the tomatoes (6 of 11 positive, 55%), a higher percentage of the surface (9 of 11 positive, 82%) and stem scar tissue (8 of 11 positive, 73%) harbored Salmonella.

Of the eleven Salmonella-positive tomatoes, three contained the pathogen only on the surface, as evidenced by detection of the pathogen in peptone wash water but not stem scar tissue or pulp (Table 2). Salmonella was detected only in stem scar tissue of two tomatoes, surface and stem scar tissue of one tomato, and all three sampling sites of five tomatoes, which indicates either systemic movement or cross-contamination at the time of removal of fruits from the plants. In only one tomato (SA2) was more than one serotype detected. Serotype Hartford was not detected in or on tomatoes. Of 11 Salmonella-positive tomatoes, serotype Poona was present in five, serotype Montevideo was present in four, and serotype Enteriditis and serotype Michigan were isolated in two.

All presumptive Salmonella isolates confirmed by PCR assay using the HILA2 primer set were subjected to PCR fingerprinting. The fingerprint results matched those of the serological tests. Fingerprint patterns of Salmonella isolates recovered from tomato samples matched the serotypes of Salmonella in the inoculum (Fig. 1).

View larger version (67K): [in this window] [in a new window]   FIG. 1.   DNA-based typing of presumptive Salmonella colonies isolated from tomato fruits. Lane 1, 100-bp DNA marker (GIBCO BRL); lanes 2 through 6, DNA profiles of serotype Enteriditis, serotype Hartford, serotype Michigan, serotype Poona, and serotype Montevideo, respectively; lanes 7 through 22, DNA profiles of stem scar tissue of tomato SA8, stem scar tissue of SA5, pulp of SA29, wash water from SB13, wash water from SB4, stem scar tissue of SB4, pulp of SB4, wash water from SA32, wash water from SA1, pulp of SA32, wash water from F30, wash water from F14, wash water from F14, stem scar tissue of F14, pulp of F14, and pulp of F14. See Table 2 for the key to tomato fruit numbers.

A range of 21 to 49 days elapsed between the date of inoculation and sampling (Table 1). Fingerprint patterns and dates of isolation indicate that serotype Montevideo was the most persistent, being isolated 49 days after inoculation, and the most dominant, being present on or in 4 of 11 (36%) tomatoes. Serotype Poona and serotype Michigan were detected on and in tomatoes 40 and 39 days postinoculation, respectively. Serotype Enteritidis was not detected in or on tomatoes harvested more than 27 days after plants were inoculated.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Tomato flowers feature style and anthers of similar heights, a characteristic that favors pollination. Tomato plants normally need bees or shaking of the plants for good pollination. In our study, most of the inoculated flowers aborted from the plant within 10 days. Chang and Pickett (C. J. Chang and P. A. Pickett, University of Georgia, personal communication) reported that pepper flowers brush-inoculated with Xanthomonas campestris pv. vesicatoria severed from plants 2 to 5 days after inoculation and were a port of entry to pepper seeds for X. campestris pv. vesicatoria. This bacterium did not move more than 9 cm from the injection site on the stem 56 days after inoculation, and only seeds collected from fruits produced from inoculated flowers contained the bacterium.

Observations on survival of Salmonella on or in tomato fruits (Table 1) indicate that the pathogen persisted on and in tomato plants from the time of inoculation at flowering through fruit ripening. Citric acid is the predominant acid in tomato fruit, and the pH of pulp of most cultivars is below 4.5. More than 90% of the weight of tomato fruits is water. As tomato fruits develop, the amount of sucrose decreases while starch and reducing sugars increase (22), which would favor nutrient availability for growth of Salmonella. Growth of serotypes Anatum, Senftenberg, and Tennessee at pH 4.05 under otherwise ideal conditions has been reported (15).

The large number of Salmonella cells (ca. 7.5 log₁₀ CFU) applied to the stems of tomatoes may not be realistic in terms of levels of contamination that may occur in the environment. Contamination caused by contact with fecal material containing large populations of Salmonella could, however, occur. A relationship between the number of Salmonella in the inoculum and the time required for death of all cells was not determined. The pathogen may not have been detectable on tomato fruits 49 days after inoculation of plants if a small number of CFU had been inoculated into stems or onto flowers. A relationship between level of inoculum and time required to achieve elimination of viable cells has yet to be determined.

The presence of Salmonella in peptone wash water, stem scar tissue, and pulp of tomatoes from inoculated plants indicates that flowers and injured stems could be possible routes of bacterial contamination of tomato fruits at different points during development and maturation. During plant growth, phytopathogens can penetrate the plant surface through natural openings such as stomata or leaf hydathodes or through wounds (1). Some bacteria enter blossoms through the nectarthodes or nectaries, which are similar to hydathodes. However, bacteria enter plants most often through wounds, and less frequently through natural openings. Plant pathogens may grow briefly on or in wounded tissue before advancing into healthy tissue (1). Injection of Salmonella into the tomato stem may introduce the pathogen into xylem, which has the principal role of transporting water and nutrients from the root to the extremities of the plant. Additionally, in the secondary xylem, the axial and ray parenchyma store nutrients and water (7) which sustain viability of plants and, possibly, promote survival of human pathogenic bacteria. The presence of epiphytal flora within tissue of fruits and vegetables through various pathways was reported by Samish et al. (27). By examining eight internal locations of tomatoes, they observed that bacteria are unevenly distributed in the fruit, and entry may be from the stem scar tissue through the core and into the endocarp. This study suggested that some epiphytal flora might reach internal tissue of tomatoes through natural apertures because of their small size and motility. It may be that bacteria enter fruit tissue more readily in the early stages of fruit development, at a time when various channels are not yet covered by corky or waxy materials (27). Broken trichomes on young fruits represent another site of entry of microorganisms.

Although Salmonella is a human pathogen, our study reveals its ability to survive on or in tomato fruits throughout the course of plant growth, flowering, and fruit development and maturation. Tomato stems and fruits are subject to mechanical injury in the field and during postharvest handling, which make them more susceptible to internalization of bacteria. Irrigation water, manure, soil, water used to prepare fungicides and insecticides, and human handling are potential sources of Salmonella. Interventions need to be applied to eliminate contamination of tomato fruits with Salmonella by preventing or minimizing its contact with tomato plants and fruits at all points from the farm to the consumer.