2016 Volume 22 Issue 3 Pages 349-357
Outbreaks of food-borne illness caused by Listeria monocytogenes in or on fresh produce are sometimes reported. Tomatoes have been considered as one of the most implicated vehicles for produce-associated outbreaks. In the present paper, using tomato plants and three isolates of L. monocytogenes showing different serotypes (1/2a, 1/2b, 4b), viability and injury of L. monocytogenes in soil and in or on tomato plants during cultivation was investigated. Soil was artificially contaminated with L. monocytogenes at levels of 2, 4, 6 or 8 log CFU/g, followed by cultivation of tomato plants in the contaminated soils. The population of L. monocytogenes in the soil decreased to less than the detection limit (< 2 log CFU/g) 6 – 8 weeks after the start of cultivation. L. monocytogenes was not detected in any harvested fruit after a 16-week cultivation, although it was detected qualitatively from soil samples. Artificial injury of the root did not induce contamination of tomato fruit by L. monocytogenes via vessels. Thus, the possibility of internalization or contamination of tomato fruit by L. monocytogenes from contaminated soil is considered quite low during cultivation. However, L. monocytogenes on the fruit surface survived up to 2 – 3 weeks. Regardless of contamination level, hygienically inappropriate handling by workers might lead to contamination of tomato fruit by bacteria such as L. monocytogenes and thus to food poisoning.
Outbreaks of foodborne illnesses associated with fresh produce and salads including minimally processed fruits and vegetables have been reported in developed countries (Beuchat et al., 1996; De Roever, 1998; Sivapalasingam et al., 2004; Kozak et al., 2013). In 2011, outbreaks caused by the unusual enteroaggregative verocytotoxin-producing Escherichia coli O104:H4 were reported from incriminated bean and seed sprouts (including fenugreek, mung beans, lentils, adzuki beans and alfalfa)i). Furthermore, a Salmonella Agona outbreak linked to papayas from Mexicoii) and Listeria monocytogenes infection associated with cantaloupe (McCollum et al., 2013) were also reported in the United States in 2011. Outbreaks caused by Salmonella Poona infection in cucumber were reported in the United States in October, 2015iii). Substantially higher consumption of raw or minimally processed fruits and vegetables in developed countries such as the United States cause outbreaks of foodborne illness (Sivapalasingam et al., 2004). In Japan, both the consumption of ready-to-eat vegetables, packaged salad and pre-cut salad and foodborne illnesses associated with this fresh produce have been increasing in recent years (MAFF, 2007). According to the Ministry of Health, Labour and Welfare's 2010 report, approximately 4% of cases of foodborne illnesses were caused by vegetables in Japan.
In the farm-to-table steps, including cultivation, processing, and distribution, there are several possible points where fresh produce might be contaminated with foodborne pathogens. In the cultivation step, irrigation water, improperly composted manure, wild and domestic animals, wash water, and handling by workers are considered some of the possible contamination points (Beuchat et al., 1997; Tauxe et al., 1997). Because pathogens may survive in improperly composted manure and soil, they may contaminate fresh produce. Furthermore, inappropriate handling by workers with contaminated water might lead to bacterial contamination of edible parts of fresh produce. Water used for irrigating, mixing with pesticides, washing produce, or making processed foods is presumed to be one of the contamination sources (Jacobsen and Bech, 2012; Berger et al., 2010). If pathogens are directly attached to the edible parts of fresh produce, they have the possibility to survive and might trigger a food-poisoning incident.
L. monocytogenes is widely distributed in the environment; thus there is a possibility that L. monocytogenes contaminates food materials and the materials then become a source of foodborne illness (Heisick et al., 1989; Arumugaswamy et al., 1994; Prazak et al., 2002; Vahidy et al., 1992). Tomatoes have been considered to be one of the most common vehicles for produce-associated pathogens. In fact, tomato was a suspected vehicle for L. monocytogenes in food poisoning incidents in hospitals in the Boston area (Ho et al., 1986). Although the route of contamination was not identified in the incidents, cultivation steps are a suspected contamination source. During cultivation, the possibility of contamination of fresh produce by pathogenic bacteria from soil is unclear. To our knowledge, no researchers have investigated contamination of tomato plants by L. monocytogenes. Therefore, it is important to investigate this possibility.
It generally takes several months to cultivate vegetables. If pathogenic bacteria have contaminated soil or the surface of the edible parts of vegetables, the bacteria might be exposed to low nutritional, dry, or other unfavorable physiological conditions. Thus, the number of viable and healthy bacterial cells may decrease over time. Injured bacterial cells have the possibility of recovery after improvement in environmental conditions. Afterward, the bacteria might proliferate again, leading to food poisoning. Thus, it is also necessary to assess the degree of injury of bacterial cells that have contaminated the edible parts or their surrounding environments such as soil and water. There have been some reports regarding injured cells of L. monocytogenes treated with heat (Busch and Donnelly, 1992; Uyttendaele et al., 2008) or high pressure (Jantzen et al., 2006) in the food industry. To our knowledge, injury to L. monocytogenes inoculated onto fresh produce has not been investigated, so it is important to quantify both the injured cells and the normal cells that might be viable.
Furthermore, in our previous study, contamination routes of a pathogenic bacterium, Salmonella Enteritidis, to fruit and leafy vegetables were investigated (Mishima et al., 2012; Honjoh et al., 2014). The purposes of the present paper were to clarify the possibility of contamination of tomato by L. monocytogenes via soil and water and to determine the viability of and injury to the bacteria in soil or on the surface of the edible parts of vegetables during cultivation.
Bacterial strains and culture conditions Different types of Listeria monocytogenes showing 1/2a, 1/2b, or 4b serotypes in our laboratory stock (Liu et al., 2012) were used. Three strains of L. monocytogenes were separately inoculated in 100 mL of Luria broth (BD Diagnostics, Sparks, MD, USA) and incubated with shaking at 37°C overnight. The cultured bacterial cells were harvested by centrifugation at 8,000 × g for 5 min and resuspended in phosphate buffered saline (PBS) at concentrations of approximately 9 log CFU/mL by adjusting to an OD600 of 0.3. Cell suspensions of the three strains were mixed equally. Then, the mixed suspension was diluted with PBS to cell concentrations of 3, 5, or 7 log CFU/mL and used for further experiments.
Preparation of contaminated soil with L. monocytogenes Commercially available soil (peat-moss/perlite/vermiculite) was purchased from Oishi Bussan Co. Ltd., Tokyo, Japan. About 5 kg of soil was autoclaved at 121°C for 20 min before inoculation. Approximately 100 g of soil was mixed with 50 mL of L. monocytogenes mixture at a cell concentration of 3, 5, 7, or 9 log CFU/mL. The concentrations of L. monocytogenes in soil were approximately 2, 4, 6, or 8 log CFU/g after mixing. These contaminated soils were then put into plastic pots (diameter: 7.5 cm).
Disinfection of seeds Seeds of Solanum lycopersicum cv. Micro-Tom were provided by the Gene Research Center, University of Tsukuba (Tsukuba, Ibaraki, Japan). For disinfection, the seeds were dipped in 70% ethanol for 10 – 15 s, and then in 0.2% sodium hypochlorite for 10 min. After disinfection, the seeds were rinsed three times with sterile water.
Growth conditions of tomato plants The disinfected seeds were sown in the contaminated soil in pots and grown at 25°C with a photoperiod consisting of 16-h light and 8-h dark in a growth chamber (Model CLE-303: TOMY SEIKO Co. Ltd, Tokyo, Japan). As regular maintenance, 20 to 30 mL sterile tap water was added directly to soil without contacting leaves or stems every two or three days and 20 – 30 mL of nutrient solution (0.2% Hyponex; Hyponex, Osaka, Japan) was applied to the plants once per week.
Sampling of soil and tomato plants During cultivation, approximately 5 g soil samples were analyzed for L. monocytogenes every two weeks. After 16 weeks, tomato plants were harvested to determine the viable bacteria and L. monocytogenes counts on fruit, stems/leaves, and roots. Stems and leaves were discriminated from roots by cutting the stems with scissors 2 cm above the soil.
Irrigation with contaminated water to keep high levels of L. monocytogenes in soil To examine the prevalence of L. monocytogenes on growing tomato plants in 150 g of soil, prepared at 8 log CFU/g as described above, 20 mL of approximately 9 log CFU/mL of cell suspension of L. monocytogenes in sterile water was prepared and added to the soil by irrigation every two weeks. The population size of L. monocytogenes in the soil was determined from collected soil (5 g) just before irrigation every two weeks except at the start of the investigation. The initial level of L. monocytogenes was determined from soil collected after contamination. Tomato plants were cultivated in the soil for 16 weeks and fruit was harvested for detection of L. monocytogenes.
Determination of viable counts of L. monocytogenes in soil and plants Samples of soil, tomato fruit and stems/leaves were each transferred to stomacher bags, diluted ten-fold dilution with PBS, and homogenized for 30 s using a stomacher. Homogenized samples were serially diluted with PBS, and then the diluted samples (100 µL) were spread onto tryptic soy agar (TSA, BD Diagnostics) containing 0.6% dried yeast extract (YE; Nissui Pharmaceutical Co., Ltd., Tokyo, Japan), CHROMagar listeria (CHROM, Kanto Chemical Co., Inc., Tokyo Japan), CHROM containing 4% NaCl (CHROM+NaCl), or CHROM containing 0.2% phenyl ethyl alcohol (CHROM+PEA) in duplicate and incubated at 37°C for 48 h. All colonies on TSA containing 0.6% YE (TSAYE) were counted as total viable bacteria. On the other hand, colonies of L. monocytogenes that were light blue with halos on CHROM were recognized as positive; thus, positive colonies on CHROM were counted as the sum of the number of injured and normal (uninjured) cells. Positive colonies on CHROM+NaCl or CHROM+PEA were counted as uninjured normal cells of L. monocytogenes. Thus, the values obtained by subtracting the number of positive colonies grown on CHROM+NaCl or CHROM+PEA from the number on CHROM were judged as equivalent to the number of injured cells of L. monocytogenes.
Qualitative detection of L. monocytogenes in soil or plants As a qualitative test, samples underwent primary or secondary enrichment before plating onto CHROM. Half Fraser broth (HFB; Oxoid, Thermo Fisher Scientific Inc. Tokyo, Japan) was used for primary enrichment. Samples (e.g., 1 g) were suspended in 9 volumes of HFB and homogenized for 30 s using a stomacher, a handy-type homogenizer, or a vortex mixer. Then, the homogenized samples were incubated at 30°C for 48 h. One loopful of the culture in HFB was spread onto CHROM for confirmation of growth of L. monocytogenes. For secondary enrichment, 1 mL of the culture in HFB was inoculated into 9 mL of Fraser broth (FB; Oxoid, Thermo Fisher Scientific Inc. Tokyo, Japan) and the inoculated samples were incubated at 37°C for 48 h. Then, one loop of the culture in FB was spread onto CHROM for confirmation of growth of L. monocytogenes. Positive colonies were assayed by colony direct PCR with hlyA primer set 7 (forward primer: 5′-AAATCATCGACGGCAACCT, reverse primer: 5′-GGACGATGTGAAATGAGC) described in the method of Liu et al. (2012) for confirmation of L. monocytogenes.
Influence of root injury on contamination of tomato fruit To investigate whether injury of roots promotes invasion of L. monocytogenes via vessels into fruit, the main root of tomato plants were cut off 1 cm below the soil surface at flowering time (8 weeks after start of cultivation). Then, the plants were replanted in soil contaminated with L. monocytogenes at approximately 7 log CFU/g. After 8 weeks of cultivation, fruit was harvested. To investigate invasion of L. monocytogenes, fruit without calyces was disinfected with 0.02% sodium hypochlorite, rinsed twice with sterile water twice, then 70% ethanol, and then washed twice with sterile water, each for 1 min. Surface-disinfected fruits as well as non-disinfected parts (stem/leaves and roots) were used for qualitative testing for L. monocytogenes.
Survival of L. monocytogenes on the surface of tomato fruit during cultivation To investigate the viability of and injury to L. monocytogenes on the surface of tomato fruit, 5 µL L. monocytogenes suspensions (approximately 4, 6, 8, or 10 log CFU/mL) were spotted onto two nearby areas of the surface of ripened tomato fruit after 14 weeks of cultivation. Cultivation was continued and, at 0, 1, 2, and 3 weeks after inoculation, only the inoculated parts of the fruits were excised with a knife for determination of viability of and injury to L. monocytogenes.
Viable counts of L. monocytogenes in soil To investigate injury to and survival of L. monocytogenes in soil during cultivation of tomato plants, cell suspensions of L. monocytogenes were mixed into soil at an initial contamination level of 2, 4, 6, or 8 log CFU/g, and then the contaminated soil was used for tomato cultivation. Changes in viable counts of L. monocytogenes in soil were determined by analyzing aliquots of soil samples every two weeks during cultivation of tomato over 16 weeks and by counting positive colonies derived from the samples on CHROM agar plates. Total viable bacterial counts were determined by counting colonies on TSAYE plates. As shown in Fig. 1, the population of L. monocytogenes at the initial inoculum of 2 and 4 log CFU/g in soil slightly increased within two weeks and then decreased gradually to the detection limit (2 log CFU/g) or lower after 6 weeks. L. monocytogenes counts in 6 and 8 log CFU/g-inoculated soils decreased less than the detection limit (2 log CFU/g) after 6 and 8 weeks, respectively. On the other hand, total viable bacterial counts in all soil samples increased to 7 – 8 log CFU/g and the counts remained similar for 16 weeks.
Changes in viable counts of L. monocytogenes (LM) in initially contaminated soils during tomato cultivation. ○, Total viable bacterial counts on TSAYE; ●, LM counts on CHROM; △, LM counts on CHROM+PEA; ▲, LM counts on CHROM+NaCl. Data are means ± SD (n=3).
Injured cells of L. monocytogenes in soil Injury of L. monocytogenes was also investigated by two types of agar plates in addition to CHROM; CHROM+PEA and CHROM+NaCl agar plates were used for discrimination of normal (uninjured) cells from injured cells, and colonies that grew on both types of plates were regarded as normal cells. As shown in Fig. 1, at any inoculation level, the difference in colony number between CHROM and CHROM+PEA or CHROM+NaCl was approximately 1 log at two weeks after the start of cultivation, suggesting that most of the detected L. monocytogenes cells were in an injured state. After 16 weeks of cultivation of tomato plants, qualitative tests of L. monocytogenes in soil samples were carried out (Table 1). Although L. monocytogenes was not detected from any primary enrichment cultures of any soil, it was detected in some of the 6 and 8 log CFU/g inoculated soil samples after secondary enrichment culture. These results showed that L. monocytogenes survives at a low level in soil for a long period (16 weeks).
Possibility of L. monocytogenes contamination of tomato plants from soil To study the possibility of contamination of tomato plants with L. monocytogenes, parts of tomato plants that had been cultivated in soil contaminated with the bacterium at different initial concentrations were analyzed. Initial concentrations of L. monocytogenes in the soil were 2, 4, 6, or 8 log CFU/g. After a 16-week cultivation, fruit, stems/leaves, and roots were harvested for L. monocytogenes assay. Counts of L. monocytogenes were less than the detection limit (<2 log CFU/g) from most parts (fruits and roots) of plants except for one stem/leaf sample, regardless of initial contamination level (Table 2). The contamination level of the stem/leaf sample was 3.97 log CFU/g. Furthermore, L. monocytogenes was not detected from any fruits, even using a qualitative test (Table 3), suggesting that the likelihood of tomato fruit becoming contaminated from L. monocytogenes originating from contaminated soil early in cultivation is quite low.
Possibility of L. monocytogenes contamination of tomato cultivated in highly contaminated soil Contamination of L. monocytogenes was investigated in tomato plants that were cultivated in highly contaminated soil. For this purpose, the soil was irrigated with 20 mL of highly contaminated water (9 log CFU/mL) every two weeks and the contamination level of the soil was checked just before and one day after irrigation with water contaminated with L. monocytogenes (Fig. 2). Although L. monocytogenes counts were around 6 log CFU/g one day after irrigation, the counts rapidly decreased to 2 – 4 log CFU/g after two weeks. Furthermore, most detected L. monocytogenes cells were considered injured, because the difference in colony numbers between CHROM and CHROM+PEA or CHROM+NaCl was approximately 1 or 2 log at several measuring points. Thus, even if soil was irrigated with water containing fresh bacterial cells every two weeks, most of the L. monocytogenes cells in soil returned to an inactive or injured state after two weeks.
Changes in viable counts of L. monocytogenes in continually contaminated soil during tomato cultivation. ○, Total viable bacterial counts on TSAYE; ●, LM counts on CHROM; △, LM counts on CHROM+PEA; ▲, LM counts on CHROM+NaCl. Data are means ± SD (n=3).
After a 16-week cultivation, L. monocytogenes counts from all parts (fruit, stems/leaves, and roots) of tomato plants were lower than the detection limit (<2 log CFU/g) (Table 4). Furthermore, as Table 5 shows, no L. monocytogenes was detected from harvested fruit of plants that had been cultivated in highly contaminated soil using a qualitative test. These results suggest that the possibility of translocation of L. monocytogenes from soil to tomato fruit during cultivation is very low.
Possibility of internalization of L. monocytogenes into fruit via injured roots To investigate the possibility of internalization of L. monocytogenes into tomato fruit via injured roots, the main roots of 8-week-old tomato plants were cut off. The plants with injured roots were transplanted to contaminated soil with a concentration of L. monocytogenes of approximately 7 log CFU/g. After a further 8-week cultivation, fruit was harvested and analyzed for L. monocytogenes by a qualitative test. As shown in Table 6, the fruit was not contaminated even though the roots had been artificially injured. Furthermore, L. monocytogenes was not detected from stems or leaves or from the injured roots.
Survival of L. monocytogenes on the surface of tomato fruit To investigate viability and injury of L. monocytogenes on the surface of tomato plants during cultivation, L. monocytogenes counts were investigated for three weeks after inoculation (Fig. 3). L. monocytogenes counts on CHROM decreased during the three weeks. In the case of low inoculation levels (2 or 4 log CFU/inoculated part), the counts were less than the lower detection limit (<1 log CFU/inoculated part) after three weeks. In the case of high inoculation levels (6 or 8 log CFU/inoculated part), survival of L. monocytogenes was confirmed at countable levels of approximately 2 log CFU/inoculated part. By plating sample onto CHROM+PEA or CHROM+NaCl plates, injury to L. monocytogenes was investigated at the same time. No signicant differences between colony numbers on CHROM+ PEA and CHROM+NaCl were found. On the other hand, differences between colony numbers on the CHROM+NaCl (or CHROM+PEA) and on CHROM plates were found in samples inoculated at 6 or 8 log CFU/inoculated part, suggesting that most L. monocytogenes cells survived on the surface of tomato fruit but were injured.
Changes in viable counts of L. monocytogenes on the surface of tomato fruits during cultivation. ●, LM counts on CHROM; △, LM counts on CHROM+PEA; ▲, LM counts on CHROM+NaCl. Data are means ± SD (n=3).
In the present paper, we investigated the contamination pathway of L. monocytogenes from soil to tomato fruit, a representative type of fresh produce, during cultivation. For this purpose, we prepared mixtures of cell suspensions of L. monocytogenes of three clinical isolates with the 1/2a, 1/2b, or 4b serotype, because these three serotypes reportedly account for over 90% of pathogens causing listeriosis (Tappero et al., 1995). First, we cultivated tomato plants in soil with cells of L monocytogenes, and investigated viable counts of L. monocytogenes in the soil for 16 weeks (Fig. 1). Van Renterghem et al. (1991) showed that L. monocytogenes, though it had been artificially inoculated into soil or liquid manure samples, was not detected 8 weeks after inoculation. Furthermore, in the same paper, cultivation of radishes in soil artificially contaminated with L. monocytogenes led to contamination of radishes even after 3 months. Thus, Van Renterghem et al. (1991) suggested that the plant-soil rhizosphere might be the natural reservoir, leading to contamination of radish by L. monocytogenes. If soil for cultivation of any vegetable were contaminated with L. monocytogenes by some routes such as contaminated water and insufficiently fermented compost, and the bacteria accessed the roots, the bacteria might be able to survive in soil for a while. In the present study, viable counts of L. monocytogenes that had been inoculated into soil decreased immediately after the start of cultivation, and the counts of L. monocytogenes fell below the detection limit by 6 – 8 weeks (Fig. 1). Formerly, we showed that Salmonella survives over 90 days in soil during cultivation of tomato (Mishima et al., 2012). Compared to Salmonella, viable counts of L. monocytogenes decreased rapidly, suggesting that the soil conditions, such as the pH and nutrient components, might not be favorable for survival of L. monocytogenes. Prazak et al. (2002) detected L. monocytogenes from water samples in furrows of a field used for cultivation of cabbage. In the present paper, L. monocytogenes in contaminated soils at concentrations of 6 and 8 log CFU/g survived for 16 weeks, and surviving cells were detected by a qualitative test (Table 1). Thus, if the residual L. monocytogenes in soil adheres to vegetables, inappropriate processing and storage might improve the nutritional conditions for the bacteria and induce proliferation, leading to food poisoning. Furthermore, Chitarra et al. (2014) reported survival of L. monocytogenes in soil where lettuce was cultivated over 80 days at 24°C but not at 30°C. They also showed that L. monocytogenes was not detected in soil where basil plants were cultivated, even 20 days after inoculation at 24°C, and suggested that basil might release antibacterial compounds in root exudates (Chitarra et al., 2014). It is not clear whether roots of tomato plants release such compounds. However, in the present paper, tomato plants were cultivated at 25°C, and temperature might affect the survival period of L. monocytogenes in soil. Thus, the type of cultivated plants and cultivation temperature should affect the viability of L. monocytogenes in soil.
Occurrence of injured bacteria is another important problem for detection of pathogenic bacteria. To determine the frequency of occurrence of injury of L. monocytogenes during cultivation of tomato, we used PEA and sodium chloride for discrimination of normal cells from injured cells. PEA seems to damage cytosolic membranes (Silva et al., 1976) and inhibit an enzyme involved in fatty acid biosynthesis, sn-glycerol-3-phosphate acyltransferase (Heath et al., 2001). Furthermore, the substance is lipophilic, and thus might disorder membranes. Of course, if PEA is used at high concentrations, even the growth of normal bacteria might be affected by it. However, the counts of L. monocytogenes that had been sampled from soils just after inoculation were almost the same on CHROM and CHROM+PEA (Fig. 1), so a PEA concentration of 0.25% was not considered to influence the viability of normal cells. On the other hand, 2 or 4 weeks later, differences in counts between CHROM and CHROM+PEA appeared. These results suggest that membrane-injured cells probably appeared over longer cultivation time and were affected by PEA, leading to lower colony counts on CHROM+PEA than on CHROM. There have been few research papers using PEA for discrimination of bacterial injury. However, based on the present results, inhibition of bacterial growth by PEA suggests the occurrence of injury to the bacterial membrane.
Sodium chloride has often been used for discrimination of normal cells from injured cells by its presence or absence in agar medium (Jofré et al., 2010; Jantzen et al., 2006; Smith and Archer, 1988). A criterion of inability to grow in the presence of 4% sodium chloride in media has often been used for detection of injured cells of L. monocytogenes (Jofré et al., 2010; Jantzen et al., 2006; Smith and Archer, 1988; Jasson et al., 2007; Golden et al., 1988). L. monocytogenes is generally reported to be able to grow under conditions of 10% sodium chloride at pH 5 – 8 (McClure et al., 1989). Normal cells of L. monocytogenes easily grow in 4% sodium chloride, but some injured cells are not able to grow at this concentration. Thus, the sodium chloride concentration at which the cells grow has been used as a criterion for discrimination between normal cells and injured cells in the present paper. Sodium chloride at higher concentrations might impose osmotic stress on L. monocytogenes and limit some functions of the cells, leading to inactivation or death of injured cells. We used PEA and sodium chloride for the numeration of L. monocytogenes injury. In Figs. 1 and 2, differences in counts between CHROM+PEA and CHROM+NaCl were found at several time points. On the other hand, in Fig. 3, counts between CHROM+PEA and CHROM+NaCl were similar. The counts in Figs. 1 and 2 reflect normal cells of L. monocytogenes in soil and the counts in Fig. 3 the normal cells on the surface of tomato fruit. Although we could not determine the mechanisms causing injury, they might be different in soil and on tomato surfaces.
On the other hand, as shown in Fig. 1, total viable bacterial counts in all soil samples increased to 7 – 8 log CFU/g and remained similar for 16 weeks. The existence of indigenous bacteria including spore-forming bacteria in soil is well-known; thus, autoclaving for over 1 h is needed for complete sterilization of soil (Trevors 1996). Although soil was autoclaved for 20 min before inoculation with L. monocytogenes, the indigenous spore-forming bacteria survived autoclaving and seemed to grow during cultivation of tomato. Since countable numbers of L. monocytogenes on CHROM decreased during cultivation (Fig. 1), it was judged that the increase in number of total viable bacteria was dependent on proliferation of indigenous bacteria in soil. However, because we used CHROM as a selective medium for L. monocytogenes, and colonies of L. monocytogenes were discriminated from the other indigenous bacteria, the influence of the indigenous bacteria on the results were quite low.
Several kinds of vegetables have reportedly become contaminated with L. monocytogenes during cultivation. From October 1987 through August 1988, Heisick et al. (1989) investigated the contamination status of Listeria spp. in fresh produce sold at supermarkets in Minneapolis (USA) and isolated L. monocytogenes from cabbage, cucumbers, potatoes, and radishes. Prazak et al. (2002) analyzed cabbage for L. monocytogenes contamination during cultivation and postharvest processing. They isolated the bacteria from cabbage in the field, showing that the cabbage was contaminated by the bacteria during cultivation. Soni et al. (2014) surveyed 10 vegetables and their respective rhizospheric soils, and detected L. monocytogenes from brinjal, cauliflower, dolichos bean, tomato, chappan-kaddu, and chili as well as from the soils. Contamination of vegetables by bacteria during cultivation seems to occur sometimes.
On the other hand, in order to clarify the contamination route of L. monocytogenes to vegetables, a few researchers have investigated the mechanism of contamination in some vegetables cultivated in contaminated soil or irrigated with contaminated water. Van Renterghem et al. (1991) showed that cultivation of radish in soil contaminated with L. monocytogenes led to survival of the bacteria in the rhizosphere and contamination of some of edible parts of radish. Settanni et al. (2012) investigated counts of Listeria spp. in soil and in aromatic plants such as basil and rocket. These plants were grown in soil contaminated with L. monocytogenes before sowing. Seedlings and plants at harvest on day 42 were used for counts of Listeria spp. Seedlings were likely to be contaminated with Listeria spp.; however, plants at harvest were not contaminated. In that paper, internalization of the bacteria into the aromatic plants was not observed. Thus, even if the bacteria attach to the plants, survivability of the bacteria on the surface of these aromatic plants would be quite low.
In order to investigate whether internalization of L. monocytogenes via roots into tomato plants occurs, two trials were done. One was cultivation in highly contaminated soils (Tables 4 and 5) and the other was cultivation of plants with artificially injured roots in contaminated soil (Table 6). From the results of both experiments, internalization of L. monocytogenes into plants was not confirmed at harvest. Listeria spp. have flagella and this characteristic gives the bacteria motility.
Salmonella also have flagella and are one of the bacterial species suspected to contaminate fresh produce. Zheng et al. (2013) investigated the internalization of five serotypes of Salmonella into tomato plants during cultivation, excluding the serotype Typhimurium. Although internalization of the bacteria from soil to plants were not clearly observed, the possibility of internalization of four of the serotypes into edible parts of tomato was shown via blossom inoculation. Previously, we also investigated the internalization of Salmonella Enteritidis (Mishima et al., 2012); however, we could not confirm internalization of the bacteria. Because we investigated only three serotypes, we could not disprove the existence of any strains of L. monocytogenes that can enter tomato plants. However, the possibility of internalization of L. monocytognes in tomato plants seems low based on our results (Tables 5 and 6).
We investigated the survival of L. monocytogenes on the surface of tomato fruit (Fig. 3). Survival counts of the bacteria on the fruit surface gradually decreased to the lower limit of detection. Furthermore, residual samples qualitatively tested for detection of L. monocytogenes (data not shown). At any inoculation level, L. monocytogenes survived on the fruit surface even after three weeks. Thus, most alive cells were considered as being in an injured state. However, if the surrounding nutritional conditions improve, the injured bacterial cells would again proliferate.
Regarding the adhesion of L. monocytogenes on vegetables, Van Renterghem et al. (1991) showed that three out of six edible parts of radishes were contaminated with L. monocytogenes after 3 months of cultivation in soils contaminated with the bacteria. However, after washing the radishes, L. monocytogenes was not detected. Thus, even if contamination of the bacteria from soil to radish occurred, the contamination would be based on the attachment of the bacteria on the radish surface, and internalization of the bacteria into radishes is not likely Dreux et al. (2007) reported that, under low relative humidity (47 – 69%), L. monocytogenes on parsley leaves tended to remain viable, but in a non-culturable state (VBNC); however, the countable L. monocytogenes population increased after the cells were exposed to high relative humidity (100%). Dreux et al. also suggested that the increase in culturable cell number was independent of recovery of VBNC cells and dependent on the growth of uninjured normal cells. Thus, humidity, one of the environmental conditions during cultivation, is likely to be involved in bacterial viability and survival on the surface of vegetables. In the present paper, because we did not control the humidity during cultivation of tomato plants, cells of L. monocytogenes inoculated onto the surface of tomato fruit were not considered to be exposed to high relative humidity (100%). Thus, most cells of L. monocytogenes on the surface of tomato fruit were exposed to relatively dry conditions, and were consequently injured or dead, resulting in a decrease in survival rate 3 weeks after inoculation.
Acknowledgements This study was financially supported by the Ministry of Agriculture, Forestry and Fisheries of Japan under the title for improvement of food safety and animal hygiene [Elucidation of the emergence mechanism of injured bacteria, and development of technology for detection and control of the bacteria; Project No. 11]. The Micro-Tom seeds (TOMJPF00001) used in this research were provided by the University of Tsukuba, Gene Research Center, through the National Bio-Resource Project (NBRP) of MEXT, Japan.