Internalization of Salmonella in Leafy Greens and Impact on Acid Tolerance

ABSTRACT

Salmonella colonizes the surface or the inner part of leafy greens, while the ability of internalized bacteria to evade common disinfection practices may pose a considerable risk. Hereby, we aimed to assess how the colonization and internalization of Salmonella spp. (i) vary with the type of leafy green, the storage conditions (temperature, time), and Salmonella serovar at phenotypic and gene transcriptional level (regarding stress- and virulence- or type III secretion system [T3SS]-associated genes) and (ii) potentially impact the survival of the pathogen against subsequent exposure at lethal pH (2.7), mimicking the gastric acidity. Internalized Salmonella reached 3.0 to 5.0 log CFU/g depending on storage conditions and vegetable, with spinach and chicory allowing the highest (P < 0.05) internalization. Prolonged storage (48 h) at 20°C increased the recovery of internalized Salmonella in spinach and green amaranth by 1.0 to 1.5 log units. Colonization of Salmonella on/in leafy vegetables induced the transcription (maximum fold change [FCmax], ∼2,000) of T3SS-related genes. Interserovar variation regarding the internalization ability of Salmonella was observed only in lettuce and green amaranth in a time- and temperature-dependent manner. Attached cells exhibited higher survival rates against low pH than the internalized subpopulation; however, habituation at 20°C in lettuce and amaranth induced acid tolerance to internalized cells, manifested by the 1.5 to 2.0 log CFU/g survivors after 75 min at pH 2.7. Habituation of Salmonella in vegetable extracts sensitized it toward acid, while indigenous microbiota had limited impact on acid resistance of the organism. These findings reveal physiological aspects of Salmonella colonizing leafy vegetables that could be useful in fresh produce microbial risk assessment.

IMPORTANCE Consumption of leafy greens has been increasingly associated with foodborne illnesses, and their contamination could occur at pre- and/or postharvest level. Human pathogens may become passively or actively internalized in plant tissues, thereby escaping decontamination procedures. Plant colonization may impact bacterial physiology such as stress resistance and virulence. In this study, it was demonstrated that internalization of Salmonella spp., at the postharvest level, varied with type of vegetable, serovar, and storage conditions. Attached and internalized subpopulations of Salmonella on/in leafy greens showed distinct physiological responses regarding transcriptional changes of stress- and virulence-associated genes, as well as survival capacity against subsequent exposure to lethal pH (2.7). These findings could contribute to a better understanding and potential (re)definition of the risk of enteric pathogens colonizing leafy greens, as well as to the design of intervention strategies aiming to improve the microbiological safety of fresh produce.

INTRODUCTION

Consumption of fresh produce has been increasingly associated with foodborne illnesses, with Salmonella spp. and norovirus in leafy salads being ranked among the five food-pathogen combinations most often linked to foodborne human infections in the European Union (1). In both the European Union and the United States, most of the fresh produce-associated outbreaks have been linked to Salmonella spp. In the United States from 2010 to 2017, Salmonella enterica has been reported as the etiological agent for 56 multistate outbreaks associated with fresh produce (2). Contamination of fresh produce with human pathogens may occur at preharvest or postharvest level by colonizing plant surfaces or penetrating plant tissue through a process known as internalization. Localization of pathogens on plant surface niches and internalization assist bacteria in surviving decontamination procedures applied in the fresh produce industry (3, 4). At preharvest level, the major contamination sources are irrigation water, soil, animals, physical damage, and handling procedures, while postharvest contamination is linked to washing with contaminated water, hydrocooling, vacuum cooling, (5), improper handling by workers with poor personal hygiene, and potential cross-contamination during shredding, cutting, and packaging (6).

Salmonella is known to be capable of penetrating the epidermis of vegetables through open stomata and colonizing within stomata or substomatal cavities of plant leaves (3, 7, 8). The degree of internalization depends on numerous factors, such as the route and mechanism of entry, the plant type and age, the plant morphology and exudates, the soil type, indigenous microbiota, and the bacterial strain or serovar (9–15). Salmonella serovars use strain-specific mechanisms to attach to different vegetable leaves, such as lettuce, rocket, and spinach (16), while upregulation of genes that improve carbon source utilization potentially offer a competitive advantage by enhancing the survival and persistence of these strains. Numerous studies have shown that indigenous bacteria may either impair the growth of introduced enteric pathogens (17–21), enhance their colonization by supplying carbon sources, or intercept antimicrobials (22–24).

Plants possess mechanisms that assist them in sensing, recognizing, responding to, and controlling undesirable bacterial intruders, inducing mechanisms to impede their establishment on/in plants (25). Human pathogens should be capable of avoiding or subverting plant defense systems in order to successfully colonize plants and become part of the phyllosphere microbial community (26). The Salmonella and Escherichia coli O157:H7 type III secretion system (T3SS) is also associated with bacterial adhesion and motility on the leaf surface, a fact that renders it critical to the initiation of the internalization process (4, 27). Several studies support the hypothesis that the suppression of the innate immune response of plants is a result of the secretion of effector proteins through the T3SS, with Salmonella T3SS mutants inducing a stronger immune response of the plant and lower proliferation rates than the wild type (28–30). While previous studies have proposed that E. coli O157:H7 was unable to overcome stomatal closure and resulted in an extended stomatal immune response, there is strong evidence that S. enterica serovar Typhimurium is able to bypass the plant stomatal immune response, keeping the stomata open and entering the plant tissue (31, 32).

Foodborne pathogens colonizing leafy vegetables potentially encounter multiple stresses within the plant environment (vegetable), in the food processing plant, during storage, and along the infection route inside the human host. Exposure of pathogens to sublethal stresses may result in the development of tolerance against homologous stresses of higher intensity (adaptation) or against different stresses (cross-tolerance) (33). In fact, several studies have shown that Salmonella associated with fresh produce (apple, tomato, cucumber, lettuce) may exhibit increased acid tolerance (31, 34, 35). The acid tolerance response (ATR) may play an important role in invasion and colonization, leading to enhanced virulence of the pathogen (36). The low infectious dose of Salmonella (37) recovered from clinical samples of salmonellosis outbreaks may be attributed partly to the latter phenomena.

During the last decade, the colonization of Salmonella on (attachment) or in (internalization) leafy greens, as well as the risk of internalized cells evading common disinfection practices, has drawn increasing attention. Nevertheless, comparative studies assessing variability between plant species (inter-plant species) and between Salmonella serovars (interserovar) regarding the internalization ability of the pathogen are limited. In addition, there is insufficient knowledge regarding the impact of internalization on pathogen physiology with relevant implications for survival in the gastrointestinal (GI) tract. Considering the above, the objectives of the present study were to investigate how colonization and internalization of Salmonella spp. (i) vary with the type of leafy green, the storage conditions (temperature and time), and Salmonella serovar at phenotypic and gene transcriptional level and (ii) potentially impact the survival of the pathogen against subsequent exposure at lethal pH (2.7), mimicking the gastric acidity.

RESULTS

Assessment of endogenous microbiota of vegetables.

Total endogenous microbiota (i.e., colonizing the surface and/or interior of vegetable leaves) as well as internalized subpopulations were determined for all tested vegetables (n = 2 × 2). The total viable counts (TVC) in lettuce, spinach, green amaranth, arugula, and chicory ranged from 5.3 to 7.4 log CFU/g, while the pH values of the vegetables were 5.8 for lettuce, 6.1 for arugula, 6.2 for spinach and chicory, and 6.4 for green amaranth. An endogenous Salmonella species population was not detected (38) (data not shown) in any of the batches of vegetables tested. Pseudomonas sp. was the dominant microorganism in all vegetables, while Enterobacteriaceae, lactic acid bacteria, and yeasts/molds were present in almost all vegetables. The internalized TVC ranged from 4.5 to 6.0 log CFU/g for all vegetables tested, except for lettuce (3.5 log CFU/g). Pseudomonas sp. was also the dominant internalized microorganism, followed by enterobacteria and yeasts, while lactic acid bacteria were below the enumeration limit (10 CFU/g) in all tested vegetables (Fig. 1). Storage at 5°C for 48 h resulted in no statistically significant changes (P > 0.05) both for the microbial population (TVC) and the pH, regardless of vegetable. On the other hand, storage at 20°C resulted in a 100-fold TVC increase, while pH change (by 0.2 U) was vegetable dependent (data not shown).

FIG 1.

Total and internalized populations (log CFU/g) of indigenous microbiota (total viable counts, Pseudomonas sp., yeasts and molds, enterobacteria, lactic acid bacteria) colonizing spinach (A), arugula (B), lettuce (C), chicory (D), and green amaranth (E). The enumeration limit for TVC, enterobacteria, and lactic acid was 10 CFU/g, while for Pseudomonas sp., yeasts, and molds, the enumeration limit was 100 CFU/g. Error bars represent the standard deviation of results between four replicates (n = 2 × 2).

Internalization of Salmonella in different vegetables.

The total Salmonella population colonizing fresh produce 2 h post inoculation (Fig. 2A and C) by immersion (in a bath containing 7.0 log CFU/mL Salmonella spp.) was 5.0 to 6.0 log CFU/g, regardless of vegetable and storage temperature, while internalized Salmonella ranged between 3.0 and 5.0 log CFU/g depending on the vegetable and storage conditions (n = 3 × 2). Under refrigerated conditions (5°C) (Fig. 2A and B), spinach and chicory allowed the highest internalization (4.5 to 5.0 log CFU/g) of the pathogen, being significantly different (P < 0.05) from that of green amaranth (3.5 to 4.0 log CFU/g), which allowed the lowest internalization. Storage for 48 h at 5°C did not affect the level of internalized Salmonella, except for arugula, where a small decrease in total Salmonella was also reflected in the internalized Salmonella. Short storage (2 h) of spinach, arugula, and chicory at 20°C (Fig. 2C) allowed the highest internalization (4.5 to 5.0 log CFU/g), while in green amaranth, the lowest levels of internalized Salmonella (3.5 log CFU/g) were observed (P < 0.05). Storage for 48 h at 20°C (Fig. 2D) resulted in similar internalization levels of Salmonella among all vegetables. Compared to short storage (2 h) at 20°C, an increase in total Salmonella colonizing lettuce by ∼1.0 log units (P < 0.05) was observed, and this increase was also reflected in the internalized population (P < 0.05). A similar increase in internalized Salmonella by approximately 1.0 to 1.5 log units (P < 0.05), compared to that of 2 h storage at 20°C, was also observed in green amaranth and spinach, with this increase not being reflected in the total Salmonella population (Fig. 2C and D). Under the same storage temperature in arugula and chicory, both total and internalized Salmonella populations exhibited no statistically significant changes with storage time (P > 0.05).

FIG 2.

Total and internalized populations (log CFU/g) of Salmonella Enteritidis colonizing lettuce, green amaranth, spinach, arugula, and chicory after storage at 5°C for 2 h (A), 5°C for 48 h (B), 20°C for 2 h (C), and 20°C for 48 h (D). Different letters indicate significant differences (P < 0.05) in bacterial population among vegetables within each population group (total, internalized) and time-temperature condition. Asterisks indicate significant differences (P < 0.05) in bacterial population between storage time (2 h and 48 h) within each population group, vegetable, and storage temperature. The enumeration limit was 10 CFU/g (1.0 log CFU/g). Error bars represent the standard deviation of results between six replicates (n = 3 × 2).

Interstrain variability in internalization of Salmonella in different vegetables.

No statistically significant differences were observed among the three tested serovars colonizing spinach, chicory, and arugula (P > 0.05) regardless of storage time and temperature (n = 2 × 2) (Fig. 3). Conversely, in green amaranth and lettuce leaves, the internalized population of S. Typhimurium was significantly different from that of S. Infantis and S. Enteritidis (P < 0.05). Even though short storage of inoculated green amaranth at 5°C resulted in similar internalized Salmonella populations for all serovars within each vegetable (P > 0.05), longer storage (48 h) at 5°C led to a higher internalized population for S. Typhimurium (P < 0.05) than for the other two serotypes on lettuce and green amaranth. Nevertheless, part of the observed differences was likely associated with a decrease in internalized populations of S. Infantis and S. Enteritidis. On the other hand, short storage (2 h) at 20°C resulted in a lower internalized population (P < 0.05) of S. Typhimurium in green amaranth by ca. 1.0 log unit compared to S. Enteritidis and S. Infantis. Long storage (48 h) at 20°C resulted in an increase of internalized S. Enteritidis and S. Infantis in green amaranth, while internalized S. Typhimurium decreased and remained at lower levels than the other two serovars (P < 0.05).

FIG 3.

Internalized subpopulations (log CFU/g) of three different Salmonella serovars (Enteritidis, Infantis, Typhimurium) colonizing lettuce, green amaranth, spinach, arugula, and chicory after storage at 5°C for 2 h (A), 5°C for 48 h (B), 20°C for 2 h (C), and 20°C for 48 h (D). Significant differences (P < 0.05) among serotypes (for each vegetable) are indicated with different letters. The enumeration limit was 10 CFU/g (1.0 log CFU/g). Error bars represent the standard deviation of results between four replicates (n = 2 × 2).

Acid tolerance of total Salmonella population colonized on/in different vegetables.

Survival of total Salmonella (including attached and internalized subpopulations) against lethal pH 2.7, following habituation on vegetable leaves for 2 and 48 h at 5 and 20°C (Fig. 4), was dependent on habituation temperature, time, and vegetable (P < 0.05) (n = 3 × 2) (see Table S1 in the supplemental material). Salmonella cells colonizing spinach at 20°C for 2 h resulted in higher survival after 75 min at low pH (ca. 3.5 log CFU/g) than longer habituation at the same temperature or storage at 5°C, which resulted in 3.0 to 3.2 log CFU/g survivors after 75 min of acid challenge. In the case of lettuce and green amaranth, storage temperature appeared to be significant for the survival of Salmonella against subsequent acid stress. In lettuce, habituation of Salmonella at 20°C allowed higher survival (P < 0.05) of the organism after 75 min at low pH, than at 5°C, while in green amaranth, habituation at 5°C and 20°C resulted in the absence of survivors and a surviving population of 3.0 to 4.0 log CFU/g after 45 min of exposure to low pH, respectively (P < 0.05). Prolonged habituation (48 h) of the pathogen on/in chicory at 20°C resulted in acid sensitization of the pathogen, as indicated by the 2.5 log CFU/g survivors at the end of acid challenge (75 min), compared to the ca. 4.0 log CFU/g survivors of Salmonella following habituation at refrigeration temperature (P < 0.05). Salmonella colonizing arugula demonstrated higher survival toward lethal pH after storage at 20°C for 2 h and 48 h than at 5°C (P < 0.05), reaching ca. 3.0 and 2.5 log CFU/g at the end of acid challenge (75 min), respectively. Conversely, habituation of Salmonella on/in arugula leaves at 5°C led to no survivors after 60 min of acid challenge.

FIG 4.

Survival of total S. Enteritidis against acid stress (pH 2.7) colonizing different vegetables (lettuce [A], green amaranth [B], chicory [C], spinach [D], and arugula [E]) after storage at 5°C (black) and 20°C (gray) for 2 h (solid) and 48 h (dashed). Initial microbial populations (log CFU/g) are indicated above each graph. Survival is expressed as the ratio of the bacterial population at each time point (CFU/g) to the initial bacterial population (CFU/g). The enumeration limit was 10 CFU/g (1 log CFU/g). Error bars represent the standard deviation of results between six replicates (n = 3 × 2).

Acid tolerance of internalized versus attached Salmonella population for different vegetables.

Subpopulations (attached and internalized) of Salmonella inhabiting leafy greens demonstrated different (P < 0.05) (n = 3 × 2) acid tolerance responses against lethal pH (pH 2.7) at 37°C simulating the acidity of the human gastric environment (Fig. 5–9), and their responses varied with vegetable type and storage conditions (P < 0.05) (Table S2). In general, internalized Salmonella was found to be less acid resistant than the surface-attached subpopulation (P < 0.05). Notably, storage for 48 h at 20°C resulted in strengthened internalized Salmonella subpopulation in lettuce and green amaranth against subsequent exposure to lethal pH. Internalized Salmonella inhabiting lettuce leaves for 2 h could not be recovered after 30 min of exposure at low pH, while the attached subpopulation decreased by only 1.5 log units at the end (75 min) of acid challenge (Fig. 5). On the other hand, longer habituation (48 h) of the pathogen on/in lettuce leaves at 20°C resulted in almost the same response (P > 0.05) of the pathogen toward acid stress, reaching a reduction of approximately 3.0 log units, with the survivors of the internalized and attached Salmonella at the end of the acid challenge being 1.5 and 2.8 log CFU/g, respectively. Attached and internalized Salmonella inhabiting spinach leaves for 2 h demonstrated comparable acid responses (P > 0.05), resulting in approximately 2.5 to 3.5 log CFU/g survivors, at the end of acid challenge (75 min), while longer habituation in plant leaves resulted in sensitization of internalized subpopulations (P < 0.05), with no countable survivors at the end of acid challenge (Fig. 6). In the case of green amaranth, acid resistance of Salmonella spp. was found to be temperature dependent (P < 0.05) (Fig. 7). In particular, after habituation at 5°C, no survivors were observed after 75 min of exposure to low pH. Notably, internalized Salmonella was not countable from the first 15 min of the acid challenge. A similar response was observed for internalized Salmonella inhabiting green amaranth at 20°C for 2 h, decreasing below the enumeration limit within the first 15 min of acid challenge, while attached cells showed higher survival, remaining countable until 60 min of exposure to lethal acid stress. Longer habituation (48 h) at 20°C, though, enhanced the survival of internalized Salmonella, resulting in approximately 1.5 log CFU/g survivors at the end (75 min) of acid challenge. Acid tolerance of Salmonella inhabiting arugula was found to be temperature dependent (P < 0.05) too (Fig. 8). However, prolonged habituation at 20°C resulted in sensitization of both attached and internalized subpopulations. In particular, storage of inoculated arugula at refrigeration temperature resulted in negligible survival of both internalized and attached subpopulations during 60 min of exposure to low pH. Conversely, attached Salmonella inhabiting arugula at 20°C (2 h) demonstrated higher acid tolerance (ca. 2.0 log CFU/g survivors at the end of acid challenge) than the corresponding internalized subpopulation that fell below the enumeration limit upon 45 min of exposure to lethal acid (P < 0.05). Prolonged (48 h) habituation of Salmonella on/in arugula leaves at 20°C sensitized both bacterial subpopulations, with internalized Salmonella resulting in no survivors from the first 15 min of acid challenge and the respective attached Salmonella remaining at ca. 2.0 log CFU/g until the end of acid challenge. In the case of Salmonella inhabiting chicory, the internalized population was found to be acid sensitive, resulting in no survivors after 30 to 45 min of acid challenge (Fig. 9). On the other hand, the attached subpopulation resulted in approximately 2.0 to 2.5 log CFU/g survivors at the end of exposure to lethal pH regardless of storage time and temperature. Nevertheless, considering that the initial populations of attached Salmonella were ca. 4.0 log CFU/g and 5.2 log CFU/g for short (2 h) and prolonged (48 h) habituation of Salmonella on chicory at 20°C, respectively, it seems that prolonged storage sensitized the pathogen, resulting in similar numbers of survivors (ca. 2.0 log CFU/g) at the end of acid challenge.

FIG 5.

Survival of attached (black) and internalized (gray) subpopulations of S. Enteritidis against subsequent acid stress (pH 2.7) following habituation on/in lettuce at 5°C for 2 h (A), 5°C for 48 h (B), 20°C for 2 h (C), and 20°C for 48 h (D). Initial microbial populations (log CFU/g) are indicated above each graph. Survival is expressed as the ratio of the bacterial population at each time point (CFU/g) to the initial bacterial population (CFU/g). The enumeration limits were 10 CFU/g and 1 CFU/g for internalized and attached subpopulations, respectively. Error bars represent the standard deviation of results between six replicates (n = 3 × 2).

FIG 9.

Survival of attached (black) and internalized (gray) subpopulations of S. Enteritidis against subsequent acid stress (pH 2.7) following habituation on/in chicory at 5°C for 2 h (A), 5°C for 48 h (B), 20°C for 2 h (C), and 20°C for 48 h (D). Initial microbial populations (log CFU/g) are indicated above each graph. Survival is expressed as the ratio of the bacterial population at each time point (CFU/g) to the initial bacterial population (CFU/g). The enumeration limits were 10 CFU/g and 1 CFU/g for internalized and attached subpopulations, respectively. Error bars represent the standard deviation of results between six replicates (n = 3 × 2).

FIG 6.

Survival of attached (black) and internalized (gray) subpopulations of S. Enteritidis against subsequent acid stress (pH 2.7) following habituation on/in spinach at 5°C for 2 h (A), 5°C for 48 h (B), 20°C for 2 h (C), and 20°C for 48 h (D). Initial microbial populations (log CFU/g) are indicated above each graph. Survival is expressed as the ratio of the bacterial population at each time point (CFU/g) to the initial bacterial population (CFU/g). The enumeration limits were 10 CFU/g and 1 CFU/g for internalized and attached subpopulations, respectively. Error bars represent the standard deviation of results between six replicates (n = 3 × 2).

FIG 7.

Survival of attached (black) and internalized (gray) subpopulations of S. Enteritidis against subsequent acid stress (pH 2.7) following habituation on/in green amaranth at 5°C for 2 h (A), 5°C for 48 h (B), 20°C for 2 h (C), and 20°C for 48 h (D). Initial microbial populations (log CFU/g) are indicated above each graph. Survival is expressed as the ratio of the bacterial population at each time point (CFU/g) to the initial bacterial population (CFU/g). The enumeration limits were 10 CFU/g and 1 CFU/g for internalized and attached subpopulations, respectively. Error bars represent the standard deviation of results between six replicates (n = 3 × 2).

FIG 8.

Survival of attached (black) and internalized (gray) subpopulations of S. Enteritidis against subsequent acid stress (pH 2.7) following habituation on/in arugula at 5°C for 2 h (A), 5°C for 48 h (B), 20°C for 2 h (C), and 20°C for 48 h (D). Initial microbial populations (log CFU/g) are indicated above each graph. Survival is expressed as the ratio of the bacterial population at each time point (CFU/g) to the initial bacterial population (CFU/g). The enumeration limits were 10 CFU/g and 1 CFU/g for internalized and attached subpopulations, respectively. Error bars represent the standard deviation of results between six replicates (n = 3 × 2).

Effect of plant tissue presence on the acid tolerance of Salmonella spp.

The acid response of Salmonella Enteritidis, following habituation on/in leafy greens, was assessed in the presence of macerated plant tissue simulating chewed vegetable following human consumption. Intact tissue with internalized Salmonella was homogenized in acidified tryptone soy broth (TSB), while attached Salmonella (recovered from leaf surface by swabbing) was resuspended in homogenized uninoculated plant tissue in acidified TSB. In order to evaluate the effect of the presence of plant tissue in the acid challenge medium on the survival of the pathogen, stationary-phase Salmonella cells without prior habituation were exposed to acidified TSB in the presence of different plant tissues (Fig. S1) (n = 2 × 2). Acidified TSB without the presence of any plant tissue (TSB−) resulted in inactivation of Salmonella similar to that of TSB with lettuce (TSB-L), green amaranth (TSB-GA), and chicory (TSB-C), leading to approximately 2.3 to 2.8 log CFU/mL survivors (P > 0.05) by the end of acid challenge (75 min). The presence of arugula (TSB-A) and spinach (TSB-S) in the acid challenge medium resulted in a higher inactivation (P < 0.05) of the bacterial population than that of TSB alone (TSB-), resulting in 0.7 to 1.2 log CFU/mL survivors after 75 min at pH 2.7.

In order to evaluate the effect of habituation of the pathogen on its subsequent survival against low pH, Salmonella habituated in TSB at 5°C and 20°C for 2 and 48 h was exposed to acidified TSB (pH 2.7) in the presence or absence of different plant tissues (Fig. S2) (n = 2 × 2). Habituation of Salmonella in TSB reduced its acid resistance, with TSB-S and TSB-A resulting in the lowest survival of Salmonella, while the highest survival of the pathogen was observed in TSB-GA. This response deviates from that of Salmonella habituated in intact leaves, depending on both habituation conditions and type of vegetable, highlighting the impact of the plant microenvironment as well as the impact of the complexity of the host-pathogen interaction on the stress response of Salmonella.

Acid tolerance of Salmonella habituating plant extracts.

Survival of the pathogen against lethal pH varied with the leafy vegetable and habituation conditions, while the effect of indigenous microbiota on the acid response of Salmonella, albeit limited, was still dependent on the vegetable (Fig. S3 and S4) (n = 2 × 2). All tested plant extracts supported the growth of Salmonella at 20°C (data not shown). Habituation of the pathogen in vegetable extracts compared to habituation in intact leaves sensitized the pathogen toward lethal pH, while the observed differences in acid resistance of cells habituated in different leafy greens were limited. The acid response of Salmonella habituating plant extracts better approximated the response of internalized than attached Salmonella in leafy greens, pinpointing the potential stress adaptation-inducing microenvironment encountered during bacterial colonization on the plant surface. Even though habituation of Salmonella in plant extracts (Fig. S3 and S4) and on the surface of intact leaves (Fig. 5 and 9) in most cases resulted in similar maximum numbers of survivors (2.5 to 3.0 log CFU/mL) after exposure to low pH, the initial bacterial populations were 6.0 to 9.0 log CFU/mL and 3.5 to 5.9 log CFU/mL, respectively, highlighting the enhanced survival potential of Salmonella attached to plant leaves. Moreover, the presence of indigenous microbiota had diverse effects on acid survival of the pathogen. Salmonella habituation at low temperature (5°C) in sterile chicory extract compared to nonsterile extract resulted in higher acid tolerance toward lethal pH, while this phenomenon was not observed at 20°C. Conversely, Salmonella habituation at 20°C in green amaranth extract in the presence of indigenous microbiota enhanced the survival of the pathogen against low pH compared to habituation in sterile extracts.

Relative gene transcription of internalized and attached Salmonella.

In order to investigate the physiology of attached and internalized subpopulations of Salmonella colonizing leafy vegetables at transcriptional level, the normalized relative quantification (NRQ) of genes associated with virulence (hilA, invA, spvR, ssrB, avrA), stress adaptation (cadB, proV), and attachment to biotic surfaces (prgH) was assessed. Spinach was chosen because it showed the highest internalization of the pathogen, while internalized bacteria were sensitive toward subsequent exposure to low acid (pH 2.7). On the other hand, green amaranth was among the vegetables that allowed the lowest internalization of Salmonella, resulting in increased acid tolerance of Salmonella following habituation on this tissue in a temperature-dependent manner. Transcriptional changes (fold change [FC]) were expressed relatively to the reference condition (Salmonella grown in TSB at 20°C for 2 h) as a mean of three biological replicates. To our knowledge, there is limited information regarding the transcriptional changes of internalized Salmonella. Potential limitations of this particular approach emerged due to a low initial bacterial population (3.0 to 6.0 log CFU/g), leading in many cases to gene transcription levels below the detection limit (threshold cycle [C_T], >35) and subsequently to nonconclusive results. These cases are indicated by asterisks in the figures.

Gene transcription of Salmonella subpopulations colonizing leafy greens varied with vegetable and temperature (Fig. 10 and 11). Upregulation of T3SS-associated genes (invA, prgH, avrA, ssrB) was observed mainly for cells colonizing spinach rather than green amaranth. In particular, prgH was positively regulated in internalized (FC, >60 to 230) and attached (FC, <10) bacteria inhabiting spinach at 5°C, while at higher temperature, prgH was downregulated or not detected in internalized bacteria. Regarding Salmonella colonizing green amaranth, the largest change in prgH transcription level was observed for attached bacteria after habituation at 20°C for 48 h (FC, ∼2,000). Attachment of the pathogen on both spinach and green amaranth leaf surfaces resulted in activation (FC, 6 to 2,700) of virulence genes (hilA, avrA, invA). The highest upregulation of spvR (FC, ∼1,200) was observed for internalized Salmonella in spinach following habituation at 20°C for 2 h. In the present study, the relative transcription levels of ssrB remained unchanged in Salmonella inhabiting spinach, where higher internalization of the pathogen was observed, while colonization of green amaranth resulted in downregulation of ssrB by up to 100 FC under the majority of tested conditions. ssrB was positively regulated (FC, ∼1,000) in the attached bacterial subpopulation inhabiting green amaranth for 48 h at 20°C. Habituation of Salmonella on/in both vegetables resulted in downregulation of proV, which encodes an ATP-binding protein, a member of the osmoregulated glycine/betaine transport system (39, 40), except for attached Salmonella on green amaranth, where positive regulation (FC, ∼100) of this gene was observed. On the other hand, cadB, which encodes the lysine/cadaverine transport protein known to contribute to pH homeostasis (41), was induced (FC, ∼18) only in the internalized bacterial population in spinach at 5°C, while in attached bacteria, the NRQ levels of cadB remained almost unchanged. Habituation of Salmonella on green amaranth led to downregulation of cadB by up to 100 FC, except for attached bacteria at 20°C for 48 h, where induction of cadB by ca. 300 FC was observed.

FIG 10.

Transcriptional changes of stress (cadB, proV)- and virulence/T3SS (hilA, avrA, invA, spvR, prgH, ssrB)-associated genes during habituation of Salmonella Enteritidis on spinach at 5°C and 20°C for 2 h and 48 h relative to the control (TSB, 20°C, 2 h). Each figure corresponds to a different set of genes: hilA, avrA, invA, and spvR (A); cadB, proV, prgH, and ssrB (B). In cases of a gene lacking data (asterisk), the population of Salmonella was inadequate for reliable estimation (no product in qPCR or cycle threshold value [C_T] of >35) of transcriptional changes. Error bars represent the standard deviation of results between three biological replicates (n = 3).

FIG 11.

Transcriptional changes of stress (cadB, proV)- and virulence/T3SS (hilA, avrA, invA, spvR, prgH, ssrB)-associated genes during habituation of Salmonella Enteritidis on green amaranth at 5°C and 20°C for 2 h and 48 h relative to the control (TSB, 20°C, 2 h). Each figure corresponds to a different set of genes: hilA, avrA, invA, and spvR (A); cadB, proV, prgH, and ssrB (B). In cases of a gene lacking data (asterisk), the population of Salmonella was inadequate for reliable estimation (no product in qPCR or cycle threshold value [C_T] of >35) of transcriptional changes. Error bars represent the standard deviation of results between three biological replicates (n = 3).

DISCUSSION

During the last decade, bacterial internalization of Salmonella spp. in leafy vegetables has gained increasing interest. Internalization refers to the entry and establishment of the pathogen beneath the plant surface. Plants, traditionally, are not considered hosts for human pathogens (42). However, among the endophytes that have been identified inhabiting various plant species, human pathogens such as Salmonella enterica, Escherichia coli, Staphylococcus saprophyticus, and others (43) may be found, indicating a potential change in their ecological niches. Colonization, including both attachment and/or internalization, of fresh produce depends on numerous factors regarding vegetable characteristics, including leaf surface structure, thickness of the leaf, chemical and natural microbiota composition, and metabolic activity of the plant (44–47). Attachment is considered the first step in plant colonization by foodborne pathogens. While numerous reports suggest that internalization varies widely among different vegetables (13, 48, 49), there are limited studies that quantitatively approach bacterial internalization. Hereby, in a postharvest scenario of leafy vegetables contamination with Salmonella, it was demonstrated that internalization of the pathogen occurred in all vegetables. Although PAMPs (pathogen-associated molecular patterns) of Salmonella are recognizable by plant innate immune response mechanisms, e.g., extracellular receptors called pattern recognition receptors (PRRs) (50, 51), and induce plant pattern-triggered immunity (PTI), there is strong evidence that Salmonella also possesses powerful means to manipulate the plant immune system, avoiding or circumventing it (29). In the present study, the internalization capacity of Salmonella varied with vegetable and storage conditions (time and temperature). Spinach and chicory allowed the highest internalization of the pathogen following short (2 h) storage at either 5 or 20°C, while green amaranth was the vegetable with the lowest internalization capacity for Salmonella. Prolonged storage (48 h) at 20°C resulted in similar internalized Salmonella levels across all tested vegetables. Several studies have demonstrated the ability of Salmonella spp. and E. coli O157:H7 to replicate on or inside the plant by taking advantage of the nutrient content of the plant (20, 51–54), while other studies claim that Salmonella does not proliferate within spinach or lettuce leaves (55). In the present study, though, it is unclear whether the observed increase in internalized Salmonella, after 48 h at 20°C, could be the outcome of the growth of already internalized Salmonella or of the increase in the internalization phenomenon due to either the ability of the pathogen to enter into the apoplastic space of the leaves or the potential dissipation of leaf microstructure following prolonged storage.

It is well known that bacterial strains belonging to different Salmonella serovars exhibit variable internalization abilities (48). Herein, it was shown that the interserovar variability of Salmonella regarding its ability to internalize leafy greens was vegetable dependent. Salmonella serovars Enteritidis, Infantis, and Typhimurium exhibited similar internalization capacities in spinach, arugula, and chicory, i.e., in the vegetables that allowed the highest internalization of the pathogen. In contrast, S. Typhimurium demonstrated internalization capacity in lettuce and green amaranth different from those of S. Enteritidis and S. Infantis, depending on storage conditions. There are numerous studies elucidating the role of plant type as well as of bacterial strain variability in plant colonization. Klerks et al. (48) described the interserovar variability of S. enterica with respect to colonization of lettuce seedlings, while Dong et al. (56) showed that Salmonella serovars Cubana, Infantis, and Typhimurium exhibited various internalization rates in alfalfa sprouts. Later studies, conducted on tomato plants, highlighted that plant-bacterium interactions could vary depending on plant species and bacterial strain (57), with serovars Montevideo and Michigan being most prevalent in contaminated hydroponically grown tomatoes, while Enteritidis, Hartford, and Poona were not detected in the tomato tissue samples (58).

Plant colonization by Salmonella spp. may potentially lead to physiological changes that could impact the responses of the pathogen against subsequent (lethal) stresses. A limited number of studies exist on the acquired acid resistance of foodborne pathogens when colonizing leafy vegetables. In fact, according to Kroupitski et al. (49), Salmonella strains demonstrated enhanced tolerance against low pH (3.0) following attachment on lettuce leaves compared to planktonic cells. In the present study, the survival of Salmonella against subsequent exposure to low pH (2.7) did not follow a consistent pattern, varying significantly with the vegetable and the time and temperature of habituation in/on plant tissues, highlighting the importance of the substrate and storage conditions in the acid-adaptive response mechanisms of Salmonella, with likely implications about enhanced survival during gastrointestinal passage. Habituation on/in lettuce, spinach, and chicory resulted in the highest survival of Salmonella after 75 min of exposure to lethal acid pH (2.7). Prolonged storage (48 h) at 20°C, where the Salmonella total populations (including surface-attached and internalized subpopulations) were similar among vegetables, resulted in the highest survival against lethal pH of those pathogen populations that inhabited green amaranth. Previous studies have shown that Salmonella inhabiting soil or lettuce may exhibit increased acid tolerance in a simulated gastric environment in a strain-dependent manner, with 40 to 50% of the initial bacterial population capable of entering the gastrointestinal tract and invading the Caco-2 layer (59). This is also supported by the findings of Gawande et al. (34, 60), who demonstrated that attachment of Salmonella spp. on solid abiotic or biotic surfaces, such as fresh-cut produce surfaces, may increase the survival potential of the organism against lethal acid, a fact that could possibly further explain the low infectious dose implicated in some foodborne salmonellosis outbreaks.

Salmonella colonizes fresh produce either by attaching to the plant surface or by internalizing into plant tissues. These bacterial subpopulations (attached, internalized) potentially encounter different macro- and microenvironments that may impact their physiological response to subsequent challenges. In the present study, the distinct performances of the two subpopulations (attached and internalized) of Salmonella inhabiting fresh produce toward subsequent lethal pH were assessed, providing phenotypic data and insights into a rather scarcely documented area. Under the majority of tested conditions and vegetables, attached Salmonella exhibited higher acid tolerance than the internalized subpopulation (P < 0.05), confirming and enriching the findings of existing studies. It should be noted, however, that attached cells subjected to acid stress constituted only a fraction of the total attached subpopulation due to technical restrictions. Consequently, from a quantitative microbial safety aspect, the underlying risk from the attached subpopulation surviving subsequent exposure toward lethal pH, such as the gastric environment, might be underestimated. Refrigeration temperature (5°C) resulted in acid sensitization of both subpopulations (P < 0.05) of the pathogen colonizing leafy greens. This is in line with existing studies reporting that the acid resistance of stationary-phase Salmonella cells was higher than that of exponentially growing cells (61) and that low (10°C) temperature resulted in decreased acid tolerance toward simulated gastric fluid (pH 2.0) (62). Prolonged habituation (48 h) of Salmonella spp. in lettuce and green amaranth leaves at 20°C enhanced their survival against lethal pH. Lettuce and green amaranth were the vegetables, along with spinach, where storage at 48 h at 20°C resulted in an increase in internalized Salmonella, suggesting that either the environment was favorable for growth of Salmonella or the local plant macro- and microenvironment enhanced the internalization process. In fact, the majority of metabolites in lettuce extracts have been shown to be polysaccharides, compared to those in spinach extracts that consist mainly of organic acids and especially oxalic acid, which is highly prevalent in spinach leaf lysates (63). Nevertheless, there are contradictory data regarding the growth capacity of Salmonella in plant extracts. In particular, Segura et al. (64) reported that spinach leaf extracts were inhibitory to the growth of Gram-negative bacteria, due to the presence of defensin-like antimicrobial peptides, while other studies (65, 66) reported growth of Salmonella in spinach leaf juices. Plants are complex substrates, and acid tolerance of Salmonella colonizing fresh produce could be affected by numerous intrinsic factors, e.g., plant surface, plant composition, and indigenous microbiota, and extrinsic factors, e.g., temperature and package atmosphere, as well as the specific plant-pathogen interactions.

The acid response of Salmonella seems to also be affected by the presence of plant tissue in the acid challenge medium. Stationary-phase Salmonella cells without prior habituation that were exposed to acidified TSB in the presence of spinach (TSB-S) or arugula (TSB-A), resulted in higher pathogen inactivation than cells exposed to acidified medium without plant tissue (TSB-). TSB-S and TSB-A also resulted in lower survival of Salmonella at lethal pH, following habituation of Salmonella in TSB at 5 and 20°C for 2 and 48 h, while the presence of green amaranth in the acid challenge medium (TSB-GA) resulted in the highest acid tolerance of the pathogen. Leafy greens are rich in phenolic compounds and other bioactive substances such as nitrates, and therefore, their extracts have been shown to be effective against numerous foodborne pathogens (67). Moreover, the adverse effect of the presence of spinach in the challenge medium on inactivation of the pathogen could also be linked to the high flavonoid content of approximately 1,000 mg/kg (68).

According to previous studies (13, 31), internalized Salmonella might be localized in the apoplastic area of parenchymal tissue of plant leaves. Within plant tissue, physiological responses of Salmonella may be affected by the local microenvironment as described by the microstructure of leaf, composition of intercellular space, and indigenous internalized microbiota. The crucial role of plant structure and microenvironment on the physiology of Salmonella was also evident from the fact that habituation of the pathogen in vegetable extracts in comparison to intact leaves seemed to sensitize Salmonella toward lethal pH, while the presence of indigenous microbiota in the vegetable extracts had diverse effects on the acid tolerance of the pathogen, varying with vegetable and storage conditions. This is in line with previous studies suggesting that Salmonella inhabiting biotic surfaces led to increased resistance toward lethal acid conditions (60). With regard to indigenous microbiota, Pseudomonas sp. was the dominant internalized microorganism in the intact vegetables, followed by enterobacteria and yeasts, while lactic acid bacteria were below the enumeration limit in all tested vegetables. Even though there are limited studies quantitatively assessing bacterial internalization, it has been shown that the genera Pantoea and Pseudomonas have been suggested as the bacteria inhabiting the internal part of lettuce leaves (69). Endophytes are known to contribute to plant growth and pathogen suppression (20), assisting in removal of contaminants, solubilization of phosphate, and providing assimilable nitrogen to plants (43). Moreover, it is assumed that endophytes may even regulate the plant immune system, thus affecting their colonization status inside plants (48, 50). Salmonella has been shown to potentially benefit from the immune-suppressing effects of plant-pathogenic bacteria, as manifested by the higher growth of Salmonella observed when the pathogen was coinoculated with Pseudomonas syringae pv. tomato with an active P. syringae T3SS rather than a T3SS-deficient mutant (70).

T3SS-related genes are necessary for the invasion of epithelial cells as well as survival and proliferation in host cells. Nevertheless, T3SS-associated genes are also known to play a crucial role in the internalization of Salmonella in plants, since the absence of particular genes resulted in decreased internalization of the pathogen in Arabidopsis thaliana (30, 71). This is in agreement with the present study showing that T3SS/virulence-associated genes (invA, prgH, avrA, ssrB) were upregulated in Salmonella colonizing spinach rather than green amaranth, corresponding to the plants with high and low internalization capacities, respectively. Virulence-associated genes have been previously shown to be highly expressed in biofilm bacterial cells (36), and these genes are also required for the attachment of S. enterica to plant tissue (72). spvR, which hereby demonstrated the highest transcriptional levels in internalized Salmonella colonizing spinach, is known to regulate along with the rpoS the expression of the spv (Salmonella plasmid virulence) operon that is induced in intracellular bacteria (73) and encodes for effector proteins necessary for effective virulence expression in animal host cells (74, 75) and in plant hosts (76).

Transcriptional levels of stress-associated genes cadB and proV varied with the leafy green, indicating the impact of variable microenvironments that the pathogen encounters in each vegetable. Although cadB has been reported previously to be induced in acid-adapted bacteria (77) and its upregulation could be correlated with survival at low pH, in the present study, this was not observed. The acid tolerance response in Salmonella includes pH homeostatic systems, modifications of the membrane composition, and synthesis of acid shock proteins (ASPs) (78). This suggests that the bacterial stress responses are rather complex and are regulated by multiple pathways rather than single genes, so that the pathogen is able to survive the potentially inimical environment of plant and animal hosts (79).

It is known that plant metabolic responses are different between preharvest and postharvest. In assays involving detached leaves, the relationship between the root system (mineral and organic nutrition, water uptake, soil microbiota, etc.) and the foliar system (photosynthesis, transpiration, foliar emission, etc.) is broken up. There are many indications supporting the importance of this relationship in host-pathogen interactions. Many studies have reported that plants at postharvest level are susceptible to certain phytopathogenic microorganisms. At preharvest level, the multiplication and establishment of these pathogens would not be feasible (80). It has also been shown that plants exhibit an immune response system at postharvest level as well (81), yet it is less effective than at preharvest, while the differences in plant microenvironment (photoperiod, air, humidity, microbiota, water, etc.) between the pre- and postharvest levels may have a significant impact on the plant-pathogen interaction. With regard to food safety, it remains to be elucidated how the host-pathogen interaction, as shaped at pre- and postharvest levels, may affect the risk of human foodborne infection associated with consumption of fresh produce. Mapping the physiological responses of the pathogen at molecular level (transcriptional or/and proteomic) could shed light on the impact of the environment that the pathogen encounters when colonizing plants, as well as on the underlying risk of induction of virulence or stress adaptation mechanisms that could potentially compromise the safety of these products.

Conclusions.

It is well known that Salmonella is able to colonize leafy greens by attaching to plant surfaces or being internalized in plant tissues. Plant colonization and internalization depend on multiple intrinsic and extrinsic factors and require the engagement of certain mechanisms in order for the pathogen to avoid or subvert the plant immune system. Here, it was shown that postharvest internalization of Salmonella varied with the vegetable and storage conditions, with spinach and chicory allowing the highest internalization, while attachment and internalization of Salmonella was characterized by upregulation of T3SS- and virulence-associated genes. To our knowledge, there are limited comparative studies reporting the acquired acid tolerance or sensitization of attached and internalized subpopulations of Salmonella habituated in different vegetables. In general, attached cells exhibited higher survival rates against subsequent exposure of the pathogen to low pH than internalized cells. However, habituation of Salmonella in lettuce and green amaranth resulted in the induction of acid tolerance and increased survival of the internalized subpopulation during exposure to low pH, emphasizing the impact of plant-associated factors such as complexity and structure of plant surface, as well as plant composition, on bacterial physiology. The acid sensitization of Salmonella following habituation in plant extracts approximated better the response of the internalized pathogen, while the presence of indigenous microbiota was diverse and vegetable dependent. The findings of the present study could contribute to elucidation of the physiology of Salmonella colonizing plants at postharvest level and to improvement in quantitative assessment of the underlying risk of the pathogen in fresh produce. Similar approaches at preharvest level could definitely shed light on the physiological profile of Salmonella in response to plant colonization, while further studies at pre-and postharvest levels approximating realistic scenarios (e.g., low initial bacterial population, stress-adapted bacteria, etc.) would contribute to a better assessment of intervention strategies to ensure the microbiological safety of fresh produce.

MATERIALS AND METHODS

Bacterial strains and inoculum preparation.

Salmonella enterica subsp. enterica serovar Enteritidis PT4 P167807, isolated from animal feed, was the main strain used throughout the study. In order to investigate interstrain variability of the internalization capacity of the pathogen, two additional Salmonella serovars, Typhimurium and Infantis (Salmonella enterica subsp. enterica serovar Typhimurium 4/74 and Salmonella enterica subsp. enterica serovar Infantis 167) were also used, isolated from calf bowel and animal feed, respectively. All three strains were maintained on tryptone soy agar (TSA) (LAB011; Lab M, Lancashire, UK) at 4°C. Bacterial strains were activated by transferring a single colony in 10 mL tryptone soy broth (TSB) (LAB004; Lab M) and incubation at 37°C for 24 h, followed by a subculture in 10 mL of the same medium and temperature for 18 h. Activated bacterial cultures were harvested by centrifugation (3,600 rpm for 15 min at 4°C) (Megafuge 1.0R; Heraus, Buckinghamshire, England), washed twice, and finally resuspended in 10 mL of one-quarter-strength Ringer’s solution (Lab M, Lancashire, UK) to obtain an approximately 9.0 log CFU/mL working culture.

Preparation and inoculation of vegetables.

Fresh lettuce (Lactuca sativa), spinach (Spinacea oleracea), green amaranth (Amaranthus viridis), arugula (Eruca sativa), and chicory (Chicorium intybus) were purchased from a local retail store. Injured leaves and damaged tissues of the vegetables were discarded, while the remaining leaves were thoroughly washed with tap drinking water (Athens city water supply network) to remove organic soil. Tap water was used in order to mimic consumers’ practices and procedures applied to the fresh produce industry. The average chlorine residual in the Athens water network for year 2020 was 0.37 ppm (Athens Water Supply and Sewerage Company; https://www.eydap.gr/), which is lower than the minimum bactericidal concentration of 1 to 20 ppm. Considering that the rinsing time of vegetables was ca. 10 s, it is unlikely that tap drinking water would have any bactericidal impact on the bacterial communities of vegetables. The absence of an endogenous Salmonella species population was verified (38) (data not shown) in all batches of vegetables tested. To simulate a potential postharvest contamination of fresh produce associated with handling in the processing environment or by the consumers, intact leaves of each vegetable were individually inoculated by immersion in a Salmonella-containing (7.0 log CFU/mL) one-quarter-strength Ringer’s solution for 2 min. The same inoculum level and procedure were also used in order to assess interstrain variability in the internalization capacity of different serovars (Enteritidis, Typhimurium, and Infantis). Following inoculation, the vegetables were allowed to dry from inoculum excess, transferred to sterile stomacher bags, and subsequently stored at 5 and 20°C for 2 h and 48 h. The temperature of 5°C corresponded to a typical domestic or retail refrigeration, while 20°C mimicked the temperature that fresh produce may encounter on a household bench or at a farmers’ market. According to a recent study (82), the 82.39% of the sellers/traders at farmers’ markets reported room temperature as the most likely mean of storage of fresh produce, while in the same study, fresh produce was most commonly stored for 1 (36.60%) or 2 (38.56%) days before sale at farmers’ markets. A storage temperature of 0°C is recommended for maximizing the shelf life of leafy greens (83), while temperatures above 15°C result in a shelf life of ca. 2 to 3 days (84).

Recovery of total, attached, and internalized Salmonella.

Microorganisms colonize the surface of the vegetables, while some microbial cells may become internalized and established in the area beneath the plant surface (apoplastic area of plant tissue). Colonization of Salmonella on/in vegetables includes both bacterial attachment (attached subpopulation) and internalization (internalized subpopulation) processes. In the present study, both the total colonizing microbial population (consisting of attached and internalized bacterial subpopulations) and the internalized subpopulation were quantitatively determined. Following incubation of inoculated vegetables at 5°C and 20°C for 2 h and 48 h, leaves were immersed in tap water in order to remove loosely attached cells. The total microbial population colonizing the plant leaves was recovered by homogenizing 10 g of vegetable tissue (hand mixer) in 90 mL one-quarter-strength Ringer’s solution. Assessment of bacterial internalization may be performed by confocal microscopy or plate count-based approaches (classical microbiology) (85). Both approaches have advantages and disadvantages. The plate count approach, though, is fit for quantifying internalized bacteria and subsequently evaluating the physiology of this particular subpopulation. Recovery and/or quantification of the internalized bacteria requires the prior removal of surface contamination by disinfection of the leaf surface. Although it cannot be excluded that disinfectants may contact internalized cells, this is an inevitable compromise compared to the benefit of their selective quantification. In the present study, disinfection of leaf surfaces was performed using ethanol and AgNO₃ by following a modified protocol of Franz et al. (86), with the consideration that this method has minimal or no effect on leaf structure and integrity while exhibiting high disinfection efficiency in terms of no countable surface survivors. Briefly, to remove and/or inactivate bacteria colonizing leaf surfaces, the leaves were immersed in 80% (vol/vol) ethanol for 20 s and then in 1% (wt/vol) AgNO₃ for 5 min and finally were washed in a deionized water bath to remove AgNO₃ residues. In order to assess the effectiveness of surface disinfection, treated leaves were tested for any attached Salmonella survivors by (i) swabbing the surface (abaxial and adaxial sides) of 10 g of leaves by using two cotton swabs, transferring the swabs to 5 mL Ringer’s solution, and plating 1 mL on selective (xylose lysine decarboxylase [XLD] agar) and nonselective (tryptic soy agar [TSA]) media, followed by incubation at 37°C for up to 48 h, and (ii) the leaf print evaluation method modified from Zhang et al. (87). Briefly, leaves were placed directly on the surface of selective (XLD) and nonselective (TSA) media, with half of leaves placed with the adaxial side facing down and the other half with the abaxial side facing down, and gently pressed against the agar surface followed by preincubation at 37°C for 1 h. Thereafter, the leaves were discarded, and the agar plates were incubated at 37°C for up to 48 h. The presence of one or more bacterial colonies was indicative of a positive sample, corresponding to inadequate disinfection. The leaf print method was used only at the preexperimental level as preliminary testing of the disinfection efficacy, while swabbing of surfaces postdisinfection was systematically performed in the experimental process.

The attached bacterial subpopulation was also recovered and enumerated from inoculated leaves. The attached microbial subpopulation was recovered from the surface of 10 g of inoculated plant leaves by using two cotton swabs (from both sides of the leaves, abaxial and adaxial) and subsequently immersed in 10 mL of one-quarter-strength Ringer’s solution. Considering, though, that the swab method only partially recovers the attached microbial population colonizing plant leaves, the recovered fraction of the attached microbial population was enumerated, only in the framework of assessing its survival toward subsequent exposure to low pH (acid challenge).

Determination of microbial load in vegetables.

Throughout this study, Salmonella spp. as well as indigenous microbiota were enumerated. Assessment of indigenous microbiota was performed by spread plating (Pseudomonas sp., yeasts, and molds) or pour plating (total viable counts) or by pour plating with overlay (enterobacteria, lactic acid bacteria) in the appropriate selective or nonselective media. For Pseudomonas species (CFC [Cephalothin, Fucidin, Cetrimide] counts; Pseudomonas agar base, LAB108 [Lab M]; 25°C for 48 h) and yeasts and molds (Rose-Bengal chloramphenicol agar base [RBC]; LAB036 [Lab M]; 25°C for 120 h), the enumeration limit was 100 CFU/g, while for enterobacteria (violet red bile glucose agar [VRBG]; LAB088 [Lab M]; 37°C for 24 h), lactic acid bacteria (LAB; DeMan Rogosa Sharpe agar [MRS]; LAB223 [Lab M], pH 5.7; double layer; 30°C for 72 h; confirmation under microscope), and total viable counts (TSA; LAB011 [Lab M]; 30°C for 48 h), the enumeration limit was 10 CFU/g. The population of Salmonella spp. was determined by plating on xylose lysine decarboxylase agar (XLD; LAB032 [Lab M]; 37°C for 24 h) or by applying the thin-layer agar method (88) in order to maximize the recovery of potentially injured cells. The bacterial population (100 μL, or 1 mL dispensed in three petri dishes) was surface plated onto nonselective nutrient TSA and incubated at 37°C for 2 h, to enhance recovery of injured cells. Then, an 8 mL overlay of XLD was poured on the surface of preincubated TSA plates, followed by incubation at 37°C for 22 h.

The pH values of leafy green homogenates (1:10 in one-quarter-strength Ringer’s solution) were assessed with a digital pH meter (WTW pH meter 526; Metrohm Ltd., Switzerland).

Preparation and inoculation of leafy green extracts and TSB.

To evaluate the hypothesis that the physiological response (survival against lethal pH) of internalized Salmonella in leafy greens could be a result of the plant local microenvironment (i.e., composition of and presence of indigenous microbiota), Salmonella was inoculated into plant extracts with (nonsterile) or without (sterile) the plant indigenous microbiota, followed by storage at similar conditions as the intact vegetables. Cells were then also exposed to lethal pH (2.7).

Fresh lettuce, spinach, green amaranth, arugula, and chicory were prepared as described in “Preparation and inoculation of vegetables.” The leafy greens were homogenized 1:10 in one-quarter-strength Ringer’s solution using a hand mixer, followed by thermal treatment at 80°C for 30 min to denature proteins and enzymes and form two distinct phases (solid and liquid). Extracts were collected by filtration using Whatman paper and subsequently autoclaved at 121°C for 20 min in order to inactivate indigenous microbiota (89). The absence of an endogenous Salmonella species population in plant extracts was verified (38) (data not shown). Nonsterile (i.e., including endogenous microbiota) and sterile extracts were dispensed in Falcon tubes (10 mL), inoculated with Salmonella Enteritidis to obtain a final concentration of 10⁷ CFU/mL and subsequently stored under similar conditions with the intact vegetables (5 and 20°C for 2 h and 48 h).

In order to investigate the effect of habituation time and temperature on the acid tolerance of the pathogen, basal medium TSB was inoculated with Salmonella Enteritidis to obtain a final population of approximately 10⁶ CFU/mL. Inoculated TSB was then incubated at 5 and 20°C for 2 h and 48 h. Acid challenge was performed as described below.

Acid challenge of Salmonella cells colonizing fresh produce.

In an effort to investigate the impact of habituation of Salmonella on and in different types of vegetables on the acid-adaptive response of the bacterium, the acid resistance of total, attached, and internalized populations of Salmonella against low pH was evaluated. In particular, fresh lettuce, spinach, green amaranth, arugula, and chicory were inoculated as described in “Preparation and inoculation of vegetables” and subsequently stored at 5 and 20°C for 2 h and 48 h. Following habituation on/in leafy greens, the total Salmonella population (consisting of both attached and internalized subpopulations) and the attached and internalized subpopulations of the pathogen were individually subjected to lethal pH. Acid challenge took place at pH 2.7 (adjusted using HCl) at 37°C, and survival of Salmonella spp. was determined after 0, 15, 30, 45, 60, and 75 min of exposure (n = 3 × 2). The initial population corresponded to that of Salmonella prior to exposure to acid challenge and was determined according to “Recovery of total, attached, and internalized Salmonella,” while time zero (t₀) corresponded to the time right after exposure to acid challenge medium. The conditions of acid challenge (pH, temperature) were chosen to approximate the temperature that the pathogen likely experiences in the human gastric environment, while the duration of the challenge (75 min) corresponded to short-term gastric digestion (90, 91). In order to better simulate human digestion, in which case plant tissue maceration would have occurred previously in the human mouth, acid challenge was performed by direct homogenization of inoculated plant tissue in acid challenge medium (total and internalized Salmonella) or, in case of the attached Salmonella subpopulation, by recovering bacterial cells by swabbing and subsequently transferring them into acid challenge medium containing uninoculated macerated plant tissue. More specifically, in order to assess the acid resistance of the total Salmonella population colonizing vegetables (consisting of attached and internalized subpopulations), 10 g of inoculated leaves was directly homogenized (hand mixer) in 90 mL of prewarmed acidified TSB so that the final pH was 2.7. Regarding the internalized subpopulation, inoculated plant leaves were first surface disinfected, as described in “Recovery of total and internalized microbial population,” in order to eliminate the attached bacterial subpopulation, and subsequently, 10 g of surface disinfected leaves was homogenized in 90 mL of prewarmed acidified TSB. The attached Salmonella subpopulation was recovered from the surface of 10 g of inoculated plant leaves by using two cotton swabs (from both sides of the leaves, abaxial and adaxial). The acid challenge medium, to which the attached Salmonella was subjected, was prepared by homogenizing 10 g of uninoculated plant tissue in 90 mL of prewarmed acidified TSB to achieve a final pH of 2.7 and thus approximate the acid challenge environment that total Salmonella and the internalized subpopulation were exposed to. Uninoculated plant tissue had been stored under similar conditions as inoculated plant tissue (5 and 20°C for 2 h and 48 h). Cotton swabs, hosting the attached Salmonella subpopulation, were subsequently placed in 10 mL of acid challenge medium containing prewarmed acidified TSB and macerated uninoculated plant tissue with a final pH 2.7. Survival of Salmonella during acid challenge was determined by applying the thin-layer agar method for maximizing recovery of injured cells, as described above (88). Survival of the pathogen is expressed as the ratio of the bacterial population at each time point (CFU/g) to the initial bacterial population (CFU/g). The enumeration limit was 10 CFU/g for total and internalized Salmonella and 1 CFU/g for the attached subpopulation.

Acid challenge of Salmonella cells habituated in TSB or vegetable extracts.

In order to investigate the impact of the habituation medium of Salmonella on its subsequent acid tolerance, 10 mL of TSB or vegetable extract containing Salmonella from the second subculture or habituated as described above was centrifuged (3,600 rpm for 15 min) at the temperature that the habituation took place (5°C and 20°C) in order to harvest Salmonella cells. The supernatant was discarded, and the remaining pellet (ca. 1.0 mL) was resuspended in 19 mL of prewarmed acidified TSB with or without uninoculated plant tissue at final pH 2.7 (at a 1:10 ratio of vegetable tissue to TSB) and incubated at 37°C. Survival of Salmonella spp. was determined at 0, 15, 30, 45, 60, and 75 min of exposure by the thin-layer agar method. Applying 1.0 mL in three plates resulted in an enumeration limit of 2 CFU/mL.

Assessment of gene transcription of Salmonella colonizing leafy greens.

Changes in gene transcription were assessed only for spinach and green amaranth. Spinach was chosen as the vegetable that allowed the highest internalization of the pathogen, although internalized bacteria were sensitive toward subsequent exposure at low acid (pH 2.7). Green amaranth was among the vegetables that allowed the lowest internalization of Salmonella, with temperature-dependent acid resistance of the internalized subpopulation. Gene transcriptional changes of attached, and internalized Salmonella colonizing leafy vegetables were assessed after 2 h and 48 h at 5°C and 20°C. To recover internalized microbial populations, surface disinfection of plant leaves (as described in “Recovery of total and internalized microbial population”) preceded. Then, 10 g of surface-disinfected plant leaves was homogenized (hand mixer) in 90 mL one-quarter-strength Ringer’s solution. The attached Salmonella subpopulation was collected from the surface of 10 g of plant leaves by using two cotton swabs (from both sides of the leaves, abaxial and adaxial). Cotton swabs were subsequently placed in 10 mL one-quarter-strength Ringer’s solution. Ten-milliliter volumes of bacterial suspensions containing attached, or internalized Salmonella populations were collected in sterile Falcon tubes containing 1.0 mL RNA stabilization solution (5% [vol/vol] water-saturated phenol in ethanol). Bacterial cells were then harvested by centrifugation (4,000 rpm for 2 min at 4°C) and stored at −80°C. RNA was extracted using a modified hot phenol-chloroform purification protocol (92). Briefly, collected samples in stabilization solution were initially washed with 1 mL one-quarter-strength Ringer’s solution, followed by centrifugation at 12,000 rpm for 1 min. The supernatant was discarded, and lysis was performed in 500 μl lysis buffer (Tris HCl [10 mM], 1 mM EDTA, pH 8.0, and 1 mg/mL lysozyme; Applichem, Germany) at room temperature. The hot phenol/SDS protocol was used for the purification of RNA, followed by ethanol precipitation at −80°C overnight, by adding 1/10 volume CH₃COONa (3 M; pH 4.8) and 2.5 volume cold absolute ethanol. Nucleic acids were then collected by centrifugation at 13,000 rpm for 30 min (at 4°C), followed by a wash step with 500 μl ethanol (70%, vol/vol) and centrifugation at 13,000 rpm for 5 min (at 4°C). RNA pellets were allowed to dry and finally were resuspended in double-distilled water (ddH₂O). Total RNA quality and quantity were determined spectrophotometrically (NanoPhotometer; Implen). Removal of DNA from RNA samples was performed using RQ1 RNase-free DNase (Promega) according to the manufacturer’s instructions, following incubation for 1 h at 37°C. DNase-treated RNA was then purified using phenol-chloroform, followed by ethanol precipitation as described above. Effectiveness of DNA removal was assessed by real-time PCR, with samples yielding threshold cycle values greater than 35 considered acceptable.

cDNA was synthesized from approximately 500 ng of total RNA using a First Strand synthesis kit (Invitrogen) in accordance with the manufacturer's protocol. The primers used in this study correspond to T3SS-, virulence-, and stress-associated genes and were synthesized by Eurofins MWG Operon. Their sequences as well as their sources are described in Table 1. Gene transcription was quantified by fluorometric RT-PCR (SYBR green-based real-time PCR) and normalized using the 16S rRNA gene as the reference gene. In particular, 10-μL reaction mixtures were prepared using KAPA SYBR Fast ABI Prism 2× qPCR master mix (Kappa Biosystems, USA), 200 nM forward and reverse primers and 1 μl of cDNA, and the final volume was adjusted using water. Reactions were performed in 96-well microplates using a StepOnePlus real-time PCR platform (Applied Biosystems). A negative control (reaction without cDNA as the template) as well as a positive control (Salmonella Enteritidis DNA) was included for each assay. The PCR thermal protocol consisted of a 1-min initial denaturation step at 95°C, followed by 40 cycles of 95°C for 3 s, 59°C for 30 s, 72°C for 30 s, and, finally, melting curve analysis (95°C for 15 s, 60°C for 1 min, and progressive heating up to 95°C with a ramp rate of 0.3°C/s, with continuous fluorescence measurement). The amplification efficiency for each primer set was determined through decimal dilutions of cDNA and genomic DNA and calculated as E = 10^(−1/slope) (93, 94) (Table 1). No amplification product was detected for the 16S rRNA gene, as well as all tested genes, in uninoculated plant samples for all tested storage conditions. Relative gene expression was calculated using the corrected for amplification efficiency Pfaffl's model (94) and expressed relatively to the reference condition (Salmonella grown in TSB at 20°C for 2 h). Gene transcription data were expressed as the mean of results of three biological replicates.

TABLE 1.

Primers used in the present study

Target gene(s)	Forward primer (5′−3′)	Reverse primer (5′−3′)	Product length (bp)	PCR efficiency (E)	Gene function	Reference
spvR	GAGAGCCGTTAACAGCCAAA	ATATCAGGTTTGCCGCAGAG	171	2.00	spv	95
hilA	ATTAAGGCGACAGAGCTGGA	GCAGAAATGGGCGAAAGTAA	134	1.91	SPI-1–T3SS	95
invA	CAACGTTTCCTGCGGTACTGT	CCCGAACGTGGCGATAATT	116	1.95	SPI-1– T3SS	96
prgH	GCTCTTTCTTGCTCATCGT	ATCTCTATCTGGCTGGATACCT	121	1.92	SPI-1–T3SS	97
avrA	GAGCTGCTTTGGTCCTCAAC	AATGGAAGGCGTTGAATCTG	173	1.93	SPI-2–T3SS	95
ssrB	GCGTTGGCCAGCAATGAATA	TTGCAATGCCGCTAACAGAA	136	1.93	SPI-2– T3SS	98
proV	CCACAATGGTACGCCTTCTCA	GCATGAGCGCAAATGACTGGA	153	1.90	Osmotic stress	99
cadB	TCCAGACCGGCGTTCTTTAT	ATGCCGGGAAGAACGTAGAA	96	1.95	Acid stress	77
16S rRNA gene	CAGAAGAAGCACCGGCTAAC	GACTCAAGCCTGCCAGTTTC	167	1.92	Reference gene	95

Statistical analysis.

Data were collected from at least two independent experimental reproductions and two biological replicates within each independent reproduction (n = 2 × 2). Experimental reproductions corresponded to experiments independently reproduced on different days with different batches of vegetables (from the same retailer and the same plant variety) and different bacterial cultures. Biological replicates corresponded to different inoculated samples within an experimental reproduction that were processed independently in order to evaluate the measurable parameter (e.g., CFU/g). Statistical analysis was performed using an SPSS statistical software program (SPSS for Windows version 255 16; SPSS, Inc., USA) and XLSTAT (version 2016.02.27444; Addinsoft). One-way analysis of variance (ANOVA) with Tukey’s honestly significant different (HSD) post hoc multiple comparisons and Student's t test were applied. Differences were considered significant at a P of <0.05.