Specific Environmental Temperature and Relative Humidity Conditions and Grafting Affect the Persistence and Dissemination of Salmonella enterica subsp. enterica Serotype Typhimurium in Tomato Plant Tissues

Tomatoes are one of the most widely produced vegetables around the world; however, fresh tomatoes have been connected to multiple wide-scale salmonellosis outbreaks over the past decades. Salmonella is commonly found in the environment and can persist in hostile conditions for several weeks before being internalized into plant tissues, where it is protected from conventional sanitation methods. In addition to biotic factors (host, inoculum size, and phytobiome), abiotic factors (environmental conditions) may affect the persistence of Salmonella in crop production. This study demonstrates that specific environmental conditions, the inoculation method, and the inoculum density affect the persistence and dissemination of JSG626 in tomato plant tissues. Our findings enhance the understanding of interactions between Salmonella enterica and fresh produce and may lead to the development of novel management practices on farms.

ABSTRACT

Little is known about the abiotic factors contributing to the preharvest persistence of Salmonella in tomato tissues. Therefore, we investigated the effects of specific environmental conditions and contamination methods on the persistence and dissemination of Salmonella enterica subsp. enterica serotype Typhimurium (JSG626) in tomato plants. When plants were sprayed on the leaves with a JSG626-contaminated solution, JSG626 persistence in the phyllosphere (bacteria located on the surface of the inoculated foliage and stem tissues) was lower at higher temperatures (30°C day/25°C night) than at lower temperatures (20°C day/15°C night). However, wounding cotyledons with contaminated tools improved JSG626 persistence and the internalization rate (2.27%) in planta compared to spray inoculation (0.004%). The systemic dissemination of JSG626 to other tissues increased when contaminated plants were grown under low relative humidity (<40%); however, JSG626 was only detected in the root systems at later sampling times (between 21 and 98 days postinoculation [dpi]). Further, after tomato scions were grafted onto rootstocks using contaminated cutting tools, dissemination of JSG626 was preferentially basipetal and occasionally acropetal in the plants, with higher persistence rates and loads of JSG626 in root systems compared to foliar tissues. JSG626 was detected in the grafting point and root systems up to 242 dpi; however, none of the fruits harvested from contaminated plants between 90 and 137 dpi were positive for JSG626. This study demonstrates that environmental temperature and relative humidity could be good indicators for estimating the persistence of Salmonella enterica in tomato plants. Further, root systems may represent a risk for long-term persistence of Salmonella enterica in tomato plants.

IMPORTANCE Tomatoes are one of the most widely produced vegetables around the world; however, fresh tomatoes have been connected to multiple wide-scale salmonellosis outbreaks over the past decades. Salmonella is commonly found in the environment and can persist in hostile conditions for several weeks before being internalized into plant tissues, where it is protected from conventional sanitation methods. In addition to biotic factors (host, inoculum size, and phytobiome), abiotic factors (environmental conditions) may affect the persistence of Salmonella in crop production. This study demonstrates that specific environmental conditions, the inoculation method, and the inoculum density affect the persistence and dissemination of JSG626 in tomato plant tissues. Our findings enhance the understanding of interactions between Salmonella enterica and fresh produce and may lead to the development of novel management practices on farms.

INTRODUCTION

Salmonella enterica is a major cause of foodborne gastroenteritis cases worldwide. It accounts for millions of illnesses, thousands of hospitalizations, hundreds of deaths, and up to $3.6 billion in public health costs annually to the United States (1, 2). Salmonella enterica subsp. enterica serotypes Enteritidis and Typhimurium are the most common serotypes encountered in outbreaks. Salmonellosis cases generally occur following consumption of contaminated animal products; however, fresh produce has also been associated with salmonellosis outbreaks in the United States over the past several decades (3–5). Approximately 15% of all salmonellosis cases caused by Salmonella enterica are believed to be associated with the consumption of fruits, seeds, and sprouts (2, 6). Salmonella enterica can persist in dozens of crops, including parsley, alfalfa, papaya, cantaloupe, cucumber, and tomato (7). Three salmonellosis outbreaks related to consumption of contaminated tomatoes were reported between 1990 and 2000, while more than twelve outbreaks have been recorded since 2001 (8). Nevertheless, the occurrence of tomato-related outbreaks remains infrequent, suggesting that specific biotic and abiotic conditions are required for the contamination of crops by Salmonella enterica (9–12). Therefore, like plant pathogens, Salmonella enterica might be bound to the concept of the disease triangle, a conceptualization of the interaction between pathogen, host, and environment (9).

The first contact between the plant, plant-associated microorganisms, and foodborne pathogens is a decisive moment that affects the dissemination of foodborne pathogens throughout the crop (13–22). Salmonella enterica can be introduced into tomato fields by applying contaminated agricultural inputs (i.e., manure and water) on the fields, or it can be vectored by animals and invertebrates (17, 23–25). However, the persistence of Salmonella enterica in the environment can be affected by the host (genotype, maturity, physiological status, and plant exudates), the inoculum (genotype, abundance, and physiological status), the route of transmission, and other parameters surrounding Salmonella enterica and the plant host (environmental conditions, nutrient availability, and phytobiome diversity) (9, 13, 22, 26–28). Salmonella enterica can survive for several weeks in the plant rhizosphere and phyllosphere (14, 27). Nevertheless, the plant surface is a hostile environment for most microorganisms due to the limited access to water and nutrients, exposure to environmental extremes, and competition. The persistence of Salmonella enterica on or in tomato plants can be modulated by the presence of specific plant tissues such as trichomes, the formation of biofilm, and abiotic factors such as relative humidity, temperature, and light quality and intensity (27, 29). Specific environmental conditions (temperature and relative humidity) can induce histological modifications of the plant tissues, which might affect the abundance of free nutrients and water content in the phyllosphere (30, 31).

Over time, internalization of Salmonella enterica can occur in leaves, roots, flowers, and fruits using natural openings (stomata, hydathodes, and lenticels) and wounds (15, 32). Salmonella enterica can persist inside plant tissues in the apoplast, where it is protected against conventional sanitation methods, making it difficult to ensure the public safety of fresh vegetables (7, 21, 27). Furthermore, previous studies have provided evidence that Salmonella enterica can be translocated into other plant tissues using phloem vessels (18, 32). This phenomenon is rare and not completely understood. Therefore, the risks of salmonellosis related to the consumption of fresh tomatoes may be reduced by identifying and managing abiotic factors that may increase the persistence of Salmonella enterica in a preharvest setting. Furthermore, it would provide significant information for the development of novel management practices to reduce the Salmonella burden in fresh produce.

Tomatoes are produced seasonally in temperate regions and in specific structures (tunnels and greenhouses) for longer periods and require specific growing conditions and intensive care for optimum crop production (33). Therefore, we studied the impact of specific environmental temperature and relative humidity conditions on the persistence of S. Typhimurium (JSG626) on the surfaces of inoculated plant tissues and its translocation throughout plant tissues after spraying leaves with contaminated water or wounding cotyledons with contaminated tools. Further, grafting a horticulturally desirable scion on a vigor-enhancing and/or disease-resistant rootstock is a common horticultural technique used in tomato production, particularly in hydroponic systems (34). Therefore, we also studied the impact of grafting using contaminated tools on the persistence and dissemination of JSG626 in plant tissues. This study demonstrates that foliar tissues represent the major site for internalization of JSG626 in tomato plant tissues and that specific environmental temperature and relative humidity conditions significantly affect the persistence, internalization, and dissemination of JSG626 in tomato plants. Environmental temperatures were more detrimental to JSG626 in the phyllosphere (the surfaces of inoculated foliage and stem tissues) than internally (inside the foliage, stem, and root system), while the relative humidity affected JSG626 persistence and dissemination inside the plant tissues. Furthermore, our grafting experiment supported previous observations that the dissemination of JSG626 inside plant tissues is preferentially basipetal and that root systems represent a higher risk for long-term persistence of JSG626 in planta than above-ground tissues. JSG626 was not detected in fruits harvested from contaminated plants throughout the experiments performed in this study, suggesting that the contamination of tomato fruits is less likely to occur following a preharvest contamination of the foliage or stem by S. Typhimurium.

RESULTS

Environmental temperature, relative humidity, and inoculation method affected the persistence, internalization, and dissemination of S. Typhimurium in tomato plant tissues.

3-week-old “Tiny Tim” tomato plants were inoculated with JSG626 using leaf spray or cotyledon clip methods. After inoculation, the plants were grown under specific environmental conditions, and JSG626 persistence outside and inside the plants was monitored weekly (Fig. 1).

FIG 1.

Persistence of Salmonella enterica subsp. enterica serotype Typhimurium (JSG626) in “Tiny Tim” tomato plants during the first weeks following leaf spray inoculation with Salmonella-contaminated water. Each column displays the population of Salmonella in the different plant tissues for a determined combination of relative humidity and temperature used to grow the plants. White bars, Salmonella populations recovered from the phyllosphere (bacteria located on the surface of the inoculated plant tissues; i.e., foliage and stem tissues); gray bars, Salmonella populations recovered inside the leaves; *, the Salmonella population in the designated tissue was significantly lower than for the previous time point (P < 0.01); arrow, senescence of inoculated leaves; bars, standard deviations (n = 9 plants per time point per treatment).

The persistence of S. Typhimurium on the surface of inoculated foliage was higher under lower environmental temperatures than under higher temperatures.

Immediately after the foliage was spray inoculated with JSG626, ∼7.7-log CFU per plant were detected in the phyllosphere across the four growing conditions (0 day postinoculation [dpi]; Fig. 1A to D). The abundance of JSG626 in the phyllosphere (the surfaces of the inoculated plant tissues; i.e., foliage and stem tissues) was significantly reduced by ∼3.5-log at 7 dpi compared to 0 dpi (P < 0.05), independent of growing condition. Further, the persistence of JSG626 in the phyllosphere significantly declined by 2.1-log and 1.7-log at 14 dpi compared to 7 dpi when the plants were grown under higher temperatures in both low and high relative humidity, respectively (Fig. 1B and D; P < 0.05). Similar trends were observed at 21 dpi (an ∼1.4-log reduction in the phyllosphere) compared to 14 dpi (Fig. 1B and D). On the other hand, the abundance of JSG626 in the phyllosphere of plants grown under low temperatures at either low or high humidity stabilized at 4.3-log (Fig. 1C) and 3.3-log (Fig. 1A) CFU in the phyllosphere per plant, respectively, at 14 and 21 dpi compared to 7 dpi. However, JSG626 was not detected in the phyllosphere of plants sampled at green mature and ripened fruiting stages due to senescence and detachment of the inoculated leaves between 14 dpi and green mature fruiting stages (black arrow; Fig. 1A to D). The same senescence rate was observed with mock-inoculated plants (inoculated with sterile water; data not shown).

Low relative humidity levels reduced the persistence of S. Typhimurium inside the foliage.

The internalization and persistence of JSG626 inside the foliage were significantly reduced by specific combinations of relative humidity and environmental temperature conditions (P < 0.05; Fig. 1E to H). JSG626 was not detected inside the inoculated leaf tissues at 0 dpi, but it was detected at 7 dpi inside the foliage, independent of the growing condition. However, the abundance of JSG626 inside the foliage was significantly lower at 7 dpi when plants were grown under a higher temperature and a low relative humidity (∼1.1-log CFU inside the foliage per plant; Fig. 1H) compared to plants grown under higher relative humidity levels at either a low (Fig. 1E) or a high (Fig. 1G) environmental temperature (∼3.4-log CFU inside the foliage per plant; P < 0.05). The low abundance of JSG626 observed under a high temperature and a low relative humidity (Fig. 1H) can be explained by the absence of JSG626 inside the foliage in many of the plants processed at 7 dpi (n = 4/9 plants negative for JSG626), 14 dpi (n = 7/9 plants negative for JSG626), and 21 dpi (all plants were negative for JSG626; Fig. 1H). The abundance of JSG626 inside the foliage did not significantly change at 14 and 21 dpi compared to 7 dpi, within the same growing condition (Fig. 1E to H). JSG626 was not detected inside the foliage after senescence of the inoculated leaves.

Low relative humidity increased the probability of dissemination of S. Typhimurium into noninoculated tomato plant tissues.

As mentioned above, specific growing conditions affected the senescence of the foliage. Following wilting, death, and detachment of contaminated leaves from the plant, JSG626 was not detected in the tomato phyllosphere, inside the foliage, or in other tissues when the plants were grown under high relative humidity levels (Fig. 1I and J, Fig. 2A and B). However, JSG626 was detected inside the stem and root system of plants grown under low relative humidity at 21 dpi through the ripe fruit stages, up to 98 dpi (Fig. 1K and L; Fig. 2C, G, and H). JSG626 populations in stem tissues and root systems were low. On average, there was less than 1.1-log CFU/stem at 21 dpi (Fig. 1K and L) at the green mature fruiting stage (Fig. 2C) and less than 2.0-log CFU/root system at the green mature and ripe fruit stages (Fig. 2G and H), and the presence of JSG636 was sporadic in these tissues, with only a few samples showing positivity for JSG626. The soil samples collected between 21 and 98 dpi and the fruits collected at the green mature and ripe fruit stages (between 45 and 98 dpi) from these plants were negative for JSG626 after enrichment in tetrathionate bile (TTB) broth.

FIG 2.

Persistence of Salmonella enterica subsp. enterica serotype Typhimurium (JSG626) in “Tiny Tim” tomato plants during the fruiting stages after leaf spray inoculation of 3-week-old plants with Salmonella-contaminated water. Each column displays the population of Salmonella in the different plant tissues for a determined combination of relative humidity and temperature used to grow the plants. The GM and RP stages occurred between 45 and 98 dpi (n = 6 plants per treatment per fruiting stage). Striped bars, Salmonella populations recovered inside stems; dotted bars, Salmonella populations recovered inside roots; bars, standard deviations; GM, green mature fruit stage; RP, ripe fruit stage.

Wounding plant tissues with contaminated tools increased the persistence and internalization of S. Typhimurium in inoculated tissues.

Different trends were observed when plants were inoculated by clipping the cotyledons using scissors contaminated with S. Typhimurium (JSG626; Fig. 3) than with foliar spray inoculation (Fig. 1). The abundance of JSG626 significantly decreased over time in the phyllosphere independent of the growing condition (Fig. 3A to D; P < 0.05), while being more static inside the cotyledons over time (Fig. 3E to H). After clip inoculation of the cotyledons, ∼5.5-log CFU in the phyllosphere per plant were detected at 0 dpi and significantly decreased to ∼2.8-log CFU in the phyllosphere per plant at 7 dpi across the four environments (Fig. 3 to D; P < 0.05). The relative humidity affected the senescence rate of the inoculated cotyledons. Cotyledons senesced between 14 and 21 dpi when plants were grown under low relative humidity and between 21 dpi and the green mature fruit stage when plants were grown under high relative humidity. Similar senescence rates were observed with the mock-inoculated plants grown under the same conditions (data not shown). Once the inoculated cotyledons senesced and became detached from the plants (arrows in Fig. 3), JSG626 was not detected in the phyllosphere at 21 dpi when plants were grown under low humidity (Fig. 3C and D), whereas 1.0-log CFU in the phyllosphere per plant remained in plants grown under high humidity at both low and high temperatures at 21 dpi (n = 6/9 and 9/9 plants positive for JSG626, respectively; Fig. 3A and B, respectively).

FIG 3.

Persistence of Salmonella enterica subsp. enterica serotype Typhimurium (JSG626) in “Tiny Tim” tomato plants during the first weeks following wounding of cotyledons using Salmonella contaminated scissors. Each column displays the population of Salmonella in the different plant tissues for a determined combination of relative humidity and temperature used to grow the plants. White bars, Salmonella populations recovered from the phyllosphere (bacteria located on the surfaces of the inoculated plant tissues; i.e., cotyledons); gray bars, Salmonella populations recovered inside the pair of cotyledons; *, the Salmonella population in the designated tissue was significantly lower than for the previous time point (P < 0.01); arrow, senescence of inoculated leaves; bars, standard deviations (n = 9 plants per time point per treatment).

JSG626 was detected inside all the inoculated cotyledons at 7 dpi, independent of the growing conditions. However, the abundance of JSG626 inside the cotyledons was significantly lower at 7 dpi when plants were grown under high temperature and low relative humidity conditions (∼2.4-log CFU inside the cotyledons per plant; Fig. 3H) compared to three other growing conditions (∼3.6-log CFU inside the foliage per plant; Fig. 3E to G; P < 0.05). The abundance of JSG626 inside the cotyledons remained the same until senescence of the inoculated cotyledons between 14 and 28 dpi (Fig. 3E to H).

As observed with the leaf spray inoculation, JSG626 was not detected inside stems or roots at any time in plants grown under high humidity, regardless of temperature (Fig. 3I and J; Fig. 4A and B) after cotyledon clip inoculation. However, JSG626 was detected inside the stem or root system of plants grown under low relative humidity at 21 dpi until the ripe fruit stage at 98 dpi (Fig. 3K and 3J; Fig. 4C and D). JSG626 populations in stem tissues and root systems were low. On average, there were <0.3-log CFU/stem at 21 dpi (Fig. 3L) and <1.4-log CFU/root system at 21 dpi (Fig. 3K), the green mature and ripe fruit stages (Fig. 4C and D, respectively), and the presence of JSG636 was sporadic in these tissues, with only few samples showing positivity for JSG626. The soil samples and fruits collected from these plants were negative for JSG626 after enrichment in TTB broth.

FIG 4.

Persistence of Salmonella enterica subsp. enterica serotype Typhimurium (JSG626) in “Tiny Tim” tomato plants during the fruiting stages following wounding of cotyledons in 3-week-old plants using Salmonella-contaminated scissors. Each column displays the population of Salmonella in the different plant tissues for a determined combination of relative humidity and temperature used to grow the plants (n = 6 plants per treatment per fruiting stage). Striped bars, Salmonella populations recovered inside stems; dotted bars, Salmonella populations recovered inside roots; *, the Salmonella population in the designated tissue was significantly lower than for the previous time point (P < 0.01); bars, standard deviations; GM, green mature fruit stage; RP, ripe fruit stage.

Overall, by comparing the JSG626 population detected inside the plant tissues (leaves or cotyledons, depending on the inoculation method) at 7 dpi with the overall JSG626 population detected at 0 dpi (internalization ratio between 0 and 7 dpi), we observed that this ratio was significantly higher after cotyledon clipping inoculation (approximately 2.3% across the four growing conditions; Fig. 3) compared to the leaf spray inoculation (approximately 0.004% across the four growing conditions; Fig. 1; P < 0.01). Nevertheless, the internalization rate between 0 and 7 dpi was significantly reduced for both inoculation methods when plants were grown under low relative humidity and high temperatures (ratios of 0.00003 and 0.07% with leaf spray and cotyledon clipping inoculations, respectively; Fig. 1H and 3H) compared to the ratio obtained with the other growing conditions (approximately 0.005 and 2.96% with leaf spray and cotyledon clip inoculations, respectively; P < 0.01).

Inoculum density and inoculation methods affected the persistence and dissemination of S. Typhimurium in grafted tomato plant tissues.

Tomato plants were grafted using JSG626-contaminated blades soaked in an inoculum of S. Typhimurium (JSG626) ranging between 2- and 8-log CFU/ml before cutting the stem of the scions (“Celebrity”) or rootstocks (“MaxiFort”). The persistence and dissemination of JSG626 in planta were monitored weekly.

Inoculum density affected the amount of JSG626 introduced into the grafted tissues at 0 dpi, as well as its persistence over time (Fig. 5). The quantity of JSG626 introduced during grafting was proportional to the concentration of inoculum used (R² = 0.93; P < 0.0001) for scion- and rootstock-inoculated plants. Between 0.5- and 5.6-log CFU/plant were detected on the day of grafting/inoculation (0 dpi; P < 0.05), using an inoculum of between 2- and 8-log CFU/ml, respectively (Fig. 5A to D). The population of JSG626 recovered from grafted plants at 0 dpi was approximately 0.05% of the original inoculum applied (2-log lower than that of the corresponding inoculum used). Therefore, it would explain why JSG626 could not be detected in half of the plants (n = 4/8) grafted with a 2-log CFU/ml inoculum at 0 dpi, even after enrichment in TTB (Fig. 5A).

FIG 5.

Persistence of Salmonella enterica subsp. enterica serotype Typhimurium (JSG626) in tomato plant tissues following splice grafting using Salmonella-contaminated blades. Each column displays the population of Salmonella in the different plant tissues for a determined inoculum density used for grafting (n = 8 plants per time point per treatment). Gray bars, grafted plants inoculated by cutting the scion stem using a contaminated blade; white bars, grafted plants inoculated by decapitating the rootstock stem using a contaminated blade; USS (upper stem section), stem section from the scion (“Celebrity”), without the leaves; LSS (lower stem section), stem section from the rootstock (“MaxiFort”), without the root system; foliage, all leaves attached to the plant; *, the Salmonella abundance was significantly different between the two inoculation methods (P < 0.01); bars, standard deviations.

The total abundance of JSG626 per plant (combination of the bacterial population in the root system, lower stem section from the rootstock [LSS; just below the graft union], upper stem section from the scion [USS; just above the graft union], and foliage) significantly increased (approximately 1.6-log CFU/plant) from 0 to 7 dpi when an inoculum of ≥4-log CFU/ml was used (P < 0.05; Fig. 5B to D). The population remained the same at 14 and 21 dpi compared to 7 dpi. The JSG626 population was significantly lower (0.9- and 2.1-log CFU/plant with rootstock- and scion-inoculated plants, respectively) at 21 dpi compared to 7 dpi only when plants were grafted with 8-log CFU/ml inoculum (P < 0.05; Fig. 5D). JSG626 was not detected in plants grafted with the 2-log CFU/ml inoculum at 21 dpi (Fig. 5A) and in some of the plants grafted with the 4-log CFU/ml inoculum at 21 dpi (n = 8/8 were positive for JSG626 following rootstock inoculation, and n = 4/8 were positive for JSG626 following scion inoculation; Fig. 5B and Table 1). All the plants inoculated using 6- and 8-log CFU/ml inocula were positive for JSG626 at 21 dpi and displayed similar abundance levels (approximately 4.8- and 5.5-log CFU/plant in 6- and 8-log CFU/ml inocula, respectively; Fig. 5C and D). Interestingly, the abundance of JSG626 from plants scion inoculated using 2- and 4-log CFU/ml inocula (Fig. 5A and B) was significantly lower than that of plants that were scion inoculated using 6- and 8-log CFU/ml inocula at 21 dpi (Fig. 5C and D; P < 0.05), while the abundance observed with a 4-log CFU/ml inoculum (Fig. 5B) was similar to that observed with a 6-log CFU/ml inoculum (Fig. 5C) when plants were rootstock inoculated at 21 dpi.

TABLE 1.

Percentage of tissues from splice-grafted tomato plants positive for Salmonella enterica subsp. enterica serotype Typhimurium (JSG626) after enrichment in tetrathionate bile broth^a

^aValues in the table indicate the percentages of tissues positive for JSG626 after enrichment in tetrathionate bile broth (TTB) at 0, 7, 14, and 21 dpi. The shade of gray is proportional to the percentage value. Rootstock- and scion-inoculated plants, plants grafted using JSG626-contaminated blade presoaked into inoculum ranging between 2- and 8-log CFU/ml; USS (upper stem section), stem section from the scion (“Celebrity”), without the leaves; LSS (lower stem section), stem section from the rootstock (“MaxiFort”), without the root system; foliage, all leaves attached to the plant; n = 8 plants per time point per inoculation method; DPI, days postinoculation. The superscript letters A to D represent statistical differences in tissues positive for JSG626.

The overall abundance and persistence of JSG626 in grafted plants were not affected by the inoculation method—rootstock or scion—at any time point (Fig. 5A to D). The majority of the JSG626 population was localized in the upper (USS) and lower (LSS) stem sections (close to the grafting point) over time for both inoculation methods (Fig. 5 and Table 2). Nevertheless, different dissemination patterns of JSG626 inside the plants were observed between the two inoculation methods (Fig. 5E to T). The general trends were that the abundance of JSG626 in the root system was higher and/or more consistently detected in plants inoculated through the rootstocks than in those inoculated through the scions (Fig. 5Q to T and Table 1). More root systems from rootstock-inoculated plants were positive for JSG626 compared to scion-inoculated plants when a 4-log CFU/ml inoculum was used (P < 0.05). Significantly higher JSG626 populations were detected with 8-log CFU/ml inoculum in rootstock-inoculated plants (approximately 6.0- and 5.7-log CFU per root system at 14 and 21 dpi) than in scion-inoculated plants (approximately 1.4- and 2.6-log CFU per root system at 14 and 21 dpi; Fig. 5T; P < 0.05). The abundance of JSG626 was significantly higher in the rootstock tissues (LSS plus root system combined together; Fig. 5P and T) than the scion tissues (USS plus foliage combined together; Fig. 5H and L) at 14 and 21 dpi in rootstock-inoculated plants when grafted with an 8-log inoculum (P < 0.05).

TABLE 2.

Percentage of stem sections in the proximity of the grafting zone positive for Salmonella enterica subsp. enterica serotype Typhimurium (JSG626) based on stem imprinting data^a

^aValues in the table represent the percentages of stem sections positive for JSG626 after stem imprinting at 7, 14, and 21 dpi. The shade of gray is proportional to the percent positive stems. Values between 0 and 20, distance (mm) of the cut in the stem from the grafting zone, which is labeled as 0 mm; grafting zone, location where both scion (“Celebrity”) and rootstock (“MaxiFort”) stems were connected together; rootstock- and scion-inoculated plants, plants grafted using JSG626-contaminated blade presoaked into the inocula ranging between 2- and 8-log CFU/ml; scion and rootstock, stem sections collected from the scion (upper) or rootstock (lower) parts of the plants; n = 8 plants per time point per inoculation method; DPI, days postinoculation. The letters A to D represent statistical differences in tissues positive for JSG626.

The presence of JSG626 in the foliage was sporadic and positively associated with the inoculum density and the time postinoculation (Fig. 5E to H and Table 1). JSG626 was more often detected in the foliage tissues at 7 dpi compared to 14 dpi and/or 21 dpi for both inoculation methods when an inoculum of ≥4-log CFU/ml was used (n = 8 plants per time point; Table 1; P < 0.05). These results were supported with the imprinting data of stem sections cut at proximity to the grafting point (n = 8 plants per time point; Table 2). More stem sections were positive for JSG626 at 7 dpi compared to 14 and 21 dpi.

In addition, the dissemination and abundance of JSG626 in plant tissues were also influenced by the density of the inoculum used. The frequency of detecting JSG626 in foliage and root systems, as well as its abundance in these plant tissues, was higher when the concentration of the inoculum used increased (Table 1; P < 0.05). Only plants grafted using the 2-log CFU/ml inoculum did not harbor JSG626 in the foliage and root systems (Fig. 5E and Q, respectively).

The inoculum density affected the long-term persistence of S. Typhimurium in grafted tomato plants, but not its transmission to tomato fruits.

None of the green mature and ripened fruits collected between 90 and 137 dpi were positive for S. Typhimurium (JSG626), even after enrichment in TTB. Yet, JSG626 was still detected at 137 dpi in the root system and grafting point of the grafted plants, but not in the stem and foliage (Fig. 6). As mentioned above, the density of the inoculum significantly affected the abundance of JSG626 detected in the plant tissues (P < 0.05); however, the inoculation method had no effect. JSG626 was not detected in grafted plants when 2- and 4-log CFU/ml inocula were used. JSG626 was detected in all the grafting points (n = 8), while only 50% of the root systems were positive for JSG626 when plants were inoculated with a 6-log CFU/ml inoculum. On the other hand, JSG626 was detected in every root system and grafting point (n = 8) when plants were inoculated with an 8-log CFU/ml inoculum. The abundance of JSG626 in the grafting point was significantly higher with the plants inoculated with an 8-log CFU/ml inoculum [approximately (3.7 ± 0.2)-log CFU] compared to the plants inoculated with a 6-log CFU/ml inoculum [approximately (2.6 ± 0.3)-log CFU; Fig. 6A]; However, no differences were detected between the inocula in the root systems (Fig. 6B).

FIG 6.

Long-term persistence of Salmonella enterica subsp. enterica serotype Typhimurium (JSG626) in tomato plants and fruits at 137 dpi following splice grafting using Salmonella-contaminated blades. (A) Salmonella population in the grafting zone at 137 dpi. Grafting zone, stem section 3 cm long containing the zone of grafting in its middle. (B) Salmonella population in the root system at 137 dpi. Gray and white bars, tissues from grafted plants inoculated using Salmonella-contaminated blade presoaked in 6- and 8-log CFU/ml inocula, respectively; scion and rootstock, rootstock (“MaxiFort”)- and scion (“Celebrity”)-inoculated plants using Salmonella-contaminated blade; *, the Salmonella abundance was significantly higher in grafted plants inoculated with an 8-log CFU/ml inoculum compared to a 6-log CFU/ml inoculum (P < 0.01); bars, standard deviations; n = 8 plants per group.

The remaining grafted plants (n = 10 plant per group) were subjected to a drought period of 105 days after termination of the fruit harvest at 137 dpi. JSG626 was detected only after enrichment in TTB in the grafting point plus root system of plants inoculated with an 8-log CFU/ml inoculum. Interestingly, 100% of grafting points plus root systems from scion-inoculated plants were positive for JSG626, while only 40% of them were positive for JSG626 with rootstock-inoculated plants.

DISCUSSION

The control of foodborne pathogens is a challenging task in fresh produce due to the limited efficacy of conventional control methods and the lack of symptoms on contaminated fresh produce. The internalization of Salmonella enterica into plant tissues can occur before harvesting (18); therefore, identifying agronomic parameters affecting the persistence, internalization rate, and dissemination of Salmonella enterica into plant tissues could lead to means of reducing the risk of wide-scale outbreaks related to the consumption of fresh produce (35).

Environmental temperature and relative humidity have significant impacts on the persistence of pathogens in the phyllosphere (28, 36, 37). This study showed that the persistence of S. Typhimurium in the “Tiny Tim” tomato plant phyllosphere was highly affected by environmental temperatures but not by the relative humidity levels. Further, absolute humidity levels estimated based on relative humidity/temperature combinations were not correlated (P > 0.01; multivariate analysis using JMP Pro 14) with bacterial counts (Fig. 1 and 3). As previously observed with Escherichia coli O157 (38), high environmental temperatures (30°C day/25°C night) significantly reduced the persistence of JSG626 in the phyllosphere over time, while the Salmonella population stabilized (∼4.0-log CFU/plant) after 7 dpi until senescence of the inoculated tissues when plants were grown under lower environmental temperatures (20°C day/15°C night). Environmental temperature fluctuations induce physiological and histological alterations of the foliage related to the chemical composition of the cuticular wax layer, the quality and abundance of trichomes and stomata, the transpiration rate, the mineral composition, the photosynthesis rate, and the internal carbon dioxide concentration in planta (39–42). Further, these alterations have consequences on the environment of the phyllosphere, as well as its epiphytic microbial populations, which might have detrimental effects on the persistence of Salmonella enterica on the plant tissues (30, 43, 44). As observed with Botrytis, some plant-associated microorganisms with higher competitiveness in high temperatures might have antagonistic potential toward Salmonella enterica via the production of volatile organic compounds (44).

Inoculation method impacted the capacity of JSG626 to enter and persist inside plant tissues. Without wounding of plant tissues, JSG626 only internalized in foliar tissues when plants were spray inoculated with a contaminated solution, which suggests that S. Typhimurium might use natural openings (i.e., stomata and hydathodes) to enter into plant tissues (15). In fact, plants grown under high-temperature and low relative humidity conditions displayed a drastic reduction in JSG626 internalization ratio (44% of the plants were negative for JSG626 at 7 dpi) and its persistence inside the plant tissues compared to other growing conditions tested, which mimic more moist conditions. Plants grown under low relative humidity are more likely to experience higher transpiration rates, resulting in the closure of the stomata to prevent water losses through transpiration (13). Therefore, these results also suggest indirectly that stomata might be the main entrance for S. Typhimurium into tomato leaf tissues (13). Therefore, growing tomato plants in an environment with low relative humidity might be a good practice to reduce the persistence of S. Typhimurium on tomato plant tissues, as well as shorten the life of contaminated foliage; however, growing tomato plants under these conditions might have a significant impact on the yield and quality of the tomatoes. The use of ventilation systems in greenhouses is common to keep the leaf surfaces dry, which reduces the risk of disease (45, 46). Therefore, studying the impact of ventilation systems on the persistence of Salmonella on tomato foliage could lead to the development of effective management practices to limit Salmonella in greenhouses. It was also observed that, overall, freshly wounded plants displayed a significantly higher internalization rate (2.3%) compared to plants spray inoculated with a contaminated solution (0.004%), suggesting that stomatal closure and plant basal defenses represent an effective barrier against the internalization of S. Typhimurium.

Once JSG626 was inside the foliage (i.e., hypothetically in stomatal chamber and/or apoplastic area between the cells) (13, 18, 47), its population remained stable and was not affected in most cases by environmental conditions. Again, only plants grown under high temperatures and low relative humidity conditions harbored lower or nondetectable populations of JSG626 inside the inoculated leaves, indicating the effectiveness of a dry environment to control internalization of S. Typhimurium in tomato plant tissues. However, these observations did not apply to wounded plants, which accentuated the importance of nutrients released by plant wounds for the persistence of S. Typhimurium outside and inside the plant tissues independently of environmental conditions. Previous studies also showed that the persistence of Salmonella enterica was enhanced when coinoculated with phytopathogens due to the release of water and nutrients freely available for Salmonella enterica (48–50; unpublished data).

Despite the high density of inoculum used (8-log CFU/ml inoculum), JSG626 was restricted to the foliage of spray-inoculated plants when grown under high relative humidity conditions. It is only when spray-inoculated plants were grown under low relative humidity conditions that JSG626 was occasionally translocated inside the stem tissues and root systems after 21 dpi. Further, translocation of JSG626 inside noninoculated stems and roots for foliage inoculated by cotyledon clipping was similar to that for foliage inoculated by leaf spray, suggesting that environmental conditions affect the translocation of JSG626 from the phyllosphere to the root system. Our grafting experiment supported observations that, similar to the case for plant pathogens (Erwinia tracheiphila and Clavibacter michiganensis subsp. michiganensis), the translocation of JSG626 in tomato plant tissues was both basipetal and acropetal (51, 52). In addition, the translocation of JSG626 inside the stem vascular vessels was inoculum density dependent. JSG626 was consistently detected inside the root system when at least 3.7-log CFU/plant were introduced in the grafting point. Interestingly, clip-inoculated plants were inoculated with ∼5.5-log CFU/pair of cotyledons, but the systemic translocation of JSG626 into the root systems was sporadic even though vascular vessels in the cotyledons were injured using contaminated scissors. Therefore, we hypothesize that the translocation of Salmonella Typhimurium from wounded cotyledons or the phyllosphere to the vascular vessels in the stem required specific environmental conditions, while these restrictions did not apply when Salmonella was directly introduced into the stem vascular vessels. Moreover, our stem imprinting data showed that the persistence of JSG626 in vascular tissues away from the grafting point was limited except at high inoculum levels, which suggests that the sieve element properties might not be favorable for S. Typhimurium survival. On the other hand, JSG626 survived for an extended period of time inside the root system (up to 137 dpi in plants grafted with an inoculum of 6- or 8-log CFU/ml and watered daily and up to 242 dpi in plants grafted with an inoculum of 8-log CFU/ml and not watered for 105 days after fruit harvest at 137 dpi). These differences in JSG626 persistence between plant tissues might be explained by the regulation of sugar sources between source-to-sink tissues (53). Salmonella enterica can use glucose and starch as sources of carbon, which might contribute to its persistence inside the foliage, fruits, and root system; however, some S. Typhimurium strains do not possess the enzymatic tools to ferment sucrose, which is the major transported form of carbon in the sieve elements between source (leaves) and sink tissues (roots, young tissues, and fruits) (54, 55).

We demonstrated that S. Typhimurium requires specific conditions to persist and be disseminated throughout tomato tissues. Environmental temperature and relative humidity significantly affected the persistence, internalization, and dissemination of S. Typhimurium on or in tomato plant tissues. Therefore, even if S. Typhimurium was not translocated into the fruits via the plant vessels, the risk of transmission of Salmonella enterica from contaminated plant tissues to the fruits remains high if proper sanitation measures are not followed. The data provided in this study may help to improve management practices and enhance the safety and sustainability of fresh produce against the Salmonella burden. Nevertheless, we acknowledge that these studies have been performed in a controlled environment. Other parameters may have significant contributions leading to the internal contamination of fresh produce before harvest. Further, plant grafting represents a significant risk for the introduction of S. Typhimurium into vegetable crops. Root systems and wounded tissues represent a risk for the long-term survival of S. Typhimurium in tomato plants. Therefore, our study also emphasizes the importance of sanitation during grafting.

MATERIALS AND METHODS

Bacterial strains.

Salmonella enterica subsp. enterica serovar Typhimurium (JSG626) was used to study abiotic factors (environmental temperature and relative humidity) influencing JSG626 persistence in planta. S. Typhimurium JSG626 is an LT2 strain well characterized and commonly used in our laboratory. This strain was provided by John Gunn (Ohio State University, Columbus, OH). Moreover, we selected JSG626 based on preliminary studies (unpublished data) that showed that JSG626 was most persistent over time out of seven S. Typhimurium strains tested when tomato seedlings were either spray or clip inoculated as described below. Use of this genetically defined strain also allows further characterization of bacterial genes needed for Salmonella to persist and interact with plant tissues. JSG626 was grown in Luria-Bertani broth at 37°C and 180 rpm in the dark for 8 h before preparation of the inoculum for each plant experiment. Isolation of JSG626 from plant tissues was performed on xylose lysine tergitol-4 (XLT-4) agar plates.

Tomato varieties.

“Tiny Tim” tomato plants were used to understand the impact of specific environmental parameters and inoculation methods on the persistence of JSG626 and its dissemination in tomato plants. “Tiny Tim” seeds (Johnny’s Selected Seeds, Winslow, ME) were sown individually in 250-cell trays filled with Baccto professional growers’ potting mix (Baccto, Houston, TX) and grown in a greenhouse (minimum and maximum values: approximately 22 to 28°C and a 20 to 80% relative humidity, with a 12-h photoperiod). Prior to sowing, seeds were heat treated as previously described (56). Nineteen-day-old plants were transplanted into 4-in. pots filled with autoclaved soil. “Celebrity” and “MaxiFort” were used as the scion and rootstock, respectively, to assess the role of grafting in the introduction and dissemination of JSG626 in plant tissues. Tomato seeds were provided by Johnny’s Selected Seeds. “Celebrity” and “MaxiFort” heat-treated seeds were planted into 50-cell trays filled with Baccto professional growers’ potting mix and grown under greenhouse conditions (22 to 28°C; 20 to 60% relative humidity; 12-h photoperiod). Watering was performed daily in the trays only after inoculation.

Impact of environmental temperature and relative humidity on S. Typhimurium survival and the colonization of tomato plants.

3-week-old “Tiny Tim” tomato plants (four leaves fully developed) were inoculated with a suspension of JSG626 normalized to an optical density of 0.2 at 600 nm (approximately 1 × 10⁸ CFU/ml) in sterile water (pH 7). Plants were inoculated either (i) by cotyledon clipping using scissors dipped in inoculum or (ii) by spraying leaves with inoculum (∼1 ml/plant). Water-inoculated plants were used as negative controls (mock inoculation). Plants were spray inoculated under a biosafety hood using a commercial hand sprayer located approximately 10 cm above the plants with a 45° angle. For the clip inoculation, scissors were sanitized between each inoculation of a cotyledon by soaking them in 70% alcohol for 1 min and for a few seconds in sterile water (pH 7). Sprayed or clipped inoculated plants were kept at room temperature for approximately 5 to 10 min to allow drying before transferring them into a growth chamber (BDR16 model; Conviron, Winnipeg, Canada) at low (20°C day/15°C night) or high (30°C day/25°C night) temperatures, with a constant relative humidity level (80 or 40%) and a 12-h photoperiod. The relative humidity and temperature stabilized in the growth chambers in less than 1 h between day and night shifts. Three plants were sampled at 0, 7, 14, and 21 dpi and at two specific fruit development stages (green mature and ripening stage; between 45 and 98 dpi, depending on the environmental conditions). The experiments were performed three times for each growing condition tested (n = 4).

Prior to quantification of the JSG626 populations present in the phyllosphere and inside the tissues of each plant, the majority of the soil was manually removed by shaking the root system. The rest of soil attached to the root system was then carefully removed by rinsing the roots in tap water (pH 7; 50 to 100 ml, depending on the root system size). The water used to rinse the root systems was transferred into a centrifugation tube and centrifuged for 7 min at 3,000 × g at room temperature. The pellet was resuspended in 1 ml of sterile water (pH 7), 10-fold serially diluted, plated on XLT-4 agar plates, and incubated at 37°C for 36 h to determine the JSG626 populations. The remaining suspensions were stored at 4°C. If the samples were determined to be negative for JSG626 by direct plating (detection limit of 10 CFU/ml), then the remaining suspensions of the negative samples were enriched in tetrathionate bile (TTB) broth (ratio of 1 volume of sample for 9 volumes of TTB) for 18 h at 150 rpm and 40°C. The product of the enrichment was plated on XLT-4 agar plates and incubated at 37°C for 36 h to detect the low abundance of JSG626 in plant tissues. Approximately 1 g of soil from the pot of each plant was collected, dilution plated, and enriched in TTB broth if required as described above.

To determine the JSG626 population present on the surface of the inoculated plants, each plant was individually transferred into a Whirl-Pack sample bag (Nasco, Fort Atkinson, WI) containing sterile water (pH 7). The size of the bag and the volume of sterile water added were based on the size of the plant (7- to 24-oz. bag; 10 to 50 ml of sterile water). Bags were manually shaken for 1 min without disrupting the plant tissues. The liquid was transferred into a 50-ml centrifugation tube and centrifuged for 7 min at 3,000 × g at room temperature. The pellet was resuspended in 1 ml of sterile water (pH 7), dilution plated, and enriched in TTB broth if required as described above The product of the enrichment was plated on XLT-4 agar plates and incubated at 37°C for 36 h to detect the low abundance of JSG626.

To quantify internal JSG626 populations, the plants were surface sterilized by soaking them separately in 70% ethanol and 1% sodium hypochlorite for 30 s each with manual shaking, followed by three rinses in sterile water (pH 7). This surface disinfection protocol was shown to successfully remove all JSG626 bacteria from the surface of freshly inoculated tomato plants, as described above. The root system, stem, and foliage were separated aseptically, placed in individual sample bags containing sterile water (pH 7), and macerated. The volume of water added was based on the size of the tissues (between 0.5 to 10 ml). The macerated tissues were 10-fold serially diluted, plated on XLT-4 agar plates, and incubated at 37°C for 36 h to determine JSG626 CFU. The remaining macerated tissues were stored at 4°C. If the samples were negative for JSG626 by direct plating, then 1 ml of the remaining macerated tissues of the negative samples was enriched in TTB and dilution plated as described above.

Impact of inoculum density and the type of inoculated tissues on S. Typhimurium persistence in tomato plants after grafting.

The persistence of JSG626 within grafted tomato was assessed using 18-day-old “Celebrity” plants as scion and 21-day-old “MaxiFort” plants as rootstock. Plants were grafted using the splice grafting technique (https://u.osu.edu/vegprolab/grafting-guide/). For this experiment, the scion or the rootstock was cut using a blade presoaked in a JSG626 suspension normalized at 2-, 4-, 6-, or 8-log CFU/ml in sterile water (pH 7). Inoculated plants were grafted with another scion or rootstock cut with another blade presoaked in sterile water (pH 7). Alcohol (70%) was used as hand and surface sanitizer between each grafting step. The rootstock plants were decapitated approximately 1.5 cm above the soil at a 45° angle, while scions were cut at least 5 mm above the cotyledon at a 45° angle. Scion and rootstock were held together with a spring loaded side-grafting clip (Johnny’s Selected Seeds). Grafted plants were kept in a healing chamber set at 20°C day/night with an increasing amount of light as previously described (https://horticulture.ucdavis.edu/information/tomato-grafting-guide) for at least 9 days after grafting. Plants were removed from the healing chamber and grown in a greenhouse at 28°C day/22°C night, 20 to 40% relative humidity, and a 12-h photoperiod. Four plants per group were sampled at 0, 7, 14, and 21 dpi. The grafting experiment was conducted two times.

Roots were washed in sterile water (pH 7) to remove soil and other debris. Roots and leaves were detached from the stem using sterile blades and collected in individual Whirl-Pak bags containing sterile water (pH 7; the volume varied depending on the tissue size). The stems were horizontally cut using sterile blades into 5-mm-long sections in both directions starting from the grafting zone. The blades were sterilized between each cut by removing the plant tissues with a sterile towel, soaking the blades in 70% alcohol for at least 1 min and then in sterile water (pH 7) for few seconds, and drying the blades on a sterile towel. Each cut section was imprinted on XLT-4 agar plates, as previously described (52). Stem sections cut from the scion and rootstock stems were collected in two Whirl-Pak bags containing 1 ml of sterile water (pH 7). Tissues were macerated, 10-fold serially diluted, and plated on XLT-4 agar plates. Plates were incubated at 37°C for 36 h to determine JSG626 abundance in tissues. The remaining macerated tissues were stored at 4°C. If the samples were negative for JSG626 by direct plating, then 1 ml of the remaining macerated tissues of the negative samples was enriched in TTB and dilution plated as described above.

The abundance of JSG626 in the grafted plants was also determined at 137 dpi (n = 4 plants per group). Roots were washed in sterile water (pH 7) to remove soil and other debris. The plant tissues (foliage, stem, grafting point, and root system) were collected in individual bags containing a volume of sterile water fluctuating (pH 7; between 5 and 50 ml) depending on the plant tissue size. The grafting point was collected from the stem by cutting 1.5 cm above and below the grafting zone. Plant tissues were macerated and dilution plated as described above. No imprinting was performed for these plants. The remaining macerated tissues were stored at 4°C. If the samples were negative for JSG626 by direct plating, then 1 ml of the remaining macerated tissues of the negative samples was enriched in TTB and dilution plated as described above.

To estimate the risk of grafting using JSG626-contaminated tools on the long-term persistence of JSG626 in tomato plants and the production of contaminated tomato fruits, at least seven green mature and twelve ripe tomato fruits per inoculum density were collected between 90 and 137 dpi from different grafted plants (n = 8 plants per groups). Fruits were collected in individual Whirl-Pak bags containing 5 ml of sterile water (pH 7). Tissues were macerated and dilution plated as described above. The remaining macerated tissues were stored at 4°C. If the samples were negative for JSG626 by direct plating, then 1 ml of the remaining macerated tissues of the negative samples was enriched in TTB and dilution plated as described above.

Once the fruits were harvested at 137 dpi, ten plants per group were incubated in the same greenhouse for 105 days using the same growing conditions descibed above but with no watering to estimate the long-term persistence of JSG626 in dry tomato plant tissues. At 242 dpi, the dry plant tissues (grafting point plus root system) were collected in individual Whirl-Pak bag containing 50 ml of sterile water (pH 7) and soaked for 3 h at room temperature. Tissues were macerated, and dilutions were plated as described above; a portion of the macerated tissue was stored at 4°C. If the samples were negative for JSG626 by direct plating, then 1 ml of the macerated tissues of the negative samples was enriched in TTB and plated as described above.

Statistical analysis.

For each experiment, plants were randomized between treatment groups. The randomization of the treatment groups was performed on the day of inoculation with JSG626 for the temperature and relative humidity study, while it was performed after 9 days in the healing chamber for the grafting experiment. JSG626 population data were log transformed. Statistical analyses were performed using JMP Pro 12 software (SAS Institute, Cary, NC). Before combining the data from replicated experiments, a one-way analysis of variance (ANOVA) combined with a Tukey test was used to confirm that no significant differences (P > 0.01) in JSG626 abundance were observed between the two sets of data at 0 dpi. A one-way ANOVA combined with a Tukey test was used to assess the differences in JSG626 abundance between the treatment groups for the different plant tissues and time points studied. Internalization ratios were processed using angular transformation before statistical analysis using a one-way ANOVA combined with a Tukey test (P < 0.01). A multivariate analysis was performed to identify correlations between the absolute humidity values obtained using online calculators (http://www.michell.com/us/calculator/ and https://planetcalc.com/2167/) and the bacterial counts. A chi-square test combined with Pearson and likelihood ratio tests was performed to the assess the frequency differences in tissues and plants positive for Salmonella in both the environmental and grafting experiments (P < 0.05). A linear regression analysis combined with a pairwise correlation test was performed to study the correlation between the inoculum density and the abundance of JSG626 detected in the grafting zone at 0 dpi.