Correspondence: Adam Schikora, Institute for Phytopathology and Applied Zoology (IPAZ), Research Center for BioSystems, Land Use and Nutrition, Justus-Liebig University Giessen, Heinrich-Buff-Ring 26-32, 35392 Giessen, Germany. Tel.: +49 641 9937497; fax: +49 641 9937499; e-mail: firstname.lastname@example.org
Infections with non-typhoidal Salmonella strains are constant and are a non-negligible threat to the human population. In the last two decades, salmonellosis outbreaks have increasingly been associated with infected fruits and vegetables. For a long time, Salmonellae were assumed to survive on plants after a more or less accidental infection. However, this notion has recently been challenged. Studies on the infection mechanism in vegetal hosts, as well as on plant immune systems, revealed an active infection process resembling in certain features the infection in animals. On one hand, Salmonella requires the type III secretion systems to effectively infect plants and to suppress their resistance mechanisms. On the other hand, plants recognize these bacteria and react to the infection with an induced defense mechanism similar to the reaction to other plant pathogens. In this review, we present the newest reports on the interaction between Salmonellae and plants. We discuss the possible ways used by these bacteria to infect plants as well as the plant responses to the infection. The recent findings indicate that plants play a central role in the dissemination of Salmonella within the ecosystem.
Numerous pathogenic bacteria appear to have a fairly broad spectrum of host organisms. Among them, Salmonella spp. efficiently infect animal and plant organisms. In this review, we focus on Salmonella enterica, a Gram-negative enteropathogenic bacteria and one of the main causes of food-borne poisoning today. Salmonellosis is unfortunately a constant threat to human health in developed as well as developing countries. A large study conducted in 2007 revealed that in the UK, the Netherlands, Germany and Ireland, 0.1–2.3% of pre-cut products were contaminated with Salmonella bacteria (Westrell et al., 2009). In the USA, an estimated one of six citizens will become infected with food-borne pathogens (Centers for Disease Control & Prevention, 2011). In cases related to domestic food poisoning, salmonellosis was responsible for 11% of infections in the USA, 35% of these resulting in hospitalization and 28% in death (Centers for Disease Control & Prevention, 2011). The incidence of Salmonella infections has not declined in the USA in the last 15 years, making the non-typhoidal strains the second most common cause (after Norovirus) of food poisoning, with around 1million infections per year (Centers for Disease Control & Prevention, 2011). Poultry and eggs are commonly associated with Salmonella outbreaks; however, 33% of infections in 2004–2008 were linked to other sources including: vine vegetables, fruits, nuts, sprouts, leafy greens, roots, and beans (Centers for Disease Control & Prevention, 2011). The assumption that Salmonellae passively survive on plants after occasional contamination has changed in the last few years. Research on the interaction between plants and these bacteria revealed an active infection process (for review see Brandl, 2006; Holden et al., 2009; Schikora et al., 2012a). Additionally, numerous reports point to suppression of the plant immune system prior to infection (Barak et al., 2005, 2008, 2011; Iniguez et al., 2005; Klerks et al., 2007; Schikora et al., 2008, 2011; Kroupitski et al., 2009; Golberg et al., 2011; Shirron & Yaron, 2011). Consequently, the consumption of Salmonella-contaminated raw vegetables and fruits has been identified as the source of infection in many recent outbreaks. Table 1 lists examples of Salmonella outbreaks related to plants in recent years.
Table 1. Salmonellosis outbreaks related to the consumption of contaminated fruits or vegetables during the last 14 years. The table is not an exhaustive list of outbreaks; we considered only those that were traced back to a vegetal contamination source
Salmonella bacteria are able to persist in soil. Studies on native Conzattia multiflora, a legume tree indigenous to Mexico, revealed the existence of Salmonella plant-borne lines isolated from nodule-like structures (Wang et al., 2006). Salmonella enterica ssp. enterica serovar Typhimurium (S. Typhimurium) was detected in the rhizosphere of several crop plants including: wheat (Triticum sativum), oilseed rape (Brassica napus) and strawberry (Fragaria x ananassa) (Berg et al., 2005). An additional infection route could be water, which was in contact with human or animal waste. Contaminated water is a well-known dissemination path for numerous pathogens, including Salmonella (Islam et al., 2004; Kisluk & Yaron, 2012). In this way, bacteria can move throughout the water supply system and subsequently reach crop fields where the contaminated water is used for irrigation (see Fig. 1 for different infection routes). Salmonella enterica populations, for instance, were bigger in the phyllosphere of tomato plants irrigated with contaminated water than in plants grown from seeds in pre-infested soil (Barak et al., 2011). Infection of plants with Salmonellae may occur also with the help of other organisms. Co-inoculation with the nematode Caenorhabditis elegans and Salmonella Newport of lettuce, strawberries or carrots placed on the top of animal or composted manure resulted in plant infection with these bacteria; conversely, the plant tissues were not infected with the pathogen in the absence of the nematode (Kenney et al., 2006). In addition, the free-living C. elegans has been found to ingest S. enterica (Aballay et al., 2000). Salmonellae can also be transported via insects, e.g. by the pharaoh ant (Monomorium pharaonis), moving from indoor facilities to the environment, and therefore coming in contact with a range of possible hosts, including plants (Beatson, 1972). A recently published report on the interaction between Salmonellae and arbuscular mycorrhzal fungi (AMF) has demonstrated a higher persistence of Salmonella in plants colonized by AMF, which indicates another layer of interactions in the rhizosphere (Gurtler et al., 2013).
Susceptibility to contamination by S. enterica via soil differs among agricultural crops. Members of the Brassicaceae family presented higher bacterial populations than tomato and lettuce (Barak et al., 2008). However, lettuce was shown to have a higher contaminated phyllosphere, suggesting that other contamination routes, such as irrigation water, are also effective (Barak et al., 2008). Important to note is the possibility that the plant itself might contribute to infection. Study performed on alfalfa sprouts demonstrated that during seed germination, the endosperm breakdown causes the release of reducing sugars and other organic molecules, which can be found in the irrigation water and, in consequence, produce a growth medium for S. enterica. Since bacteria metabolize these molecules as a source of nutrients, plant exudates seem to be a suitable nutrient source. The high multiplication rates of some of the Salmonella strains in effluxes of germinating seeds, prompted the suggestion of saprophytic growth (Howard & Hutcheson, 2003). Furthermore, other plant pathogenic bacteria may contribute to the infection with Salmonella. Pectinolytic bacterial pathogens, which cause soft rot and in this way mobilize nutrients, are frequently associated with infection of fruits and vegetables by S. enterica (Barak & Schroeder, 2012).
Adhesion and attachment to plant host
Successful attachment of bacteria to plant surfaces is the first step in the infection process. Studies on Salmonella attachment to diverse crop plants revealed that both the host plant and the bacterial genetic equipment influence the efficiency with which bacteria adhere to plants. Lettuce leaves, for example, show distinct attachment properties between older leaf parts and leaf regions near the petiole (Kroupitski et al., 2011). Bacterial attachment to basil, lettuce or spinach leaves can differ among S. enterica serovars. Whereas S. Typhimurium, Enteritidis and Senftenberg are efficient in attaching to those plants, other serovars including Agona, Heidelberg or Salmonella arizonae are less effective (Berger et al., 2011).
Biofilm formation could also influence the success of attachment to plant tissues. Biofilms are formed of a matrix of exopolymers, thin aggregative fimbriae (Tafi or curli), cellulose and lipopolysaccharide O-polysaccharide (also known as O-antigen) capsules that are involved in multicellular behavior and persistence under harsh environmental conditions (for review see Barnhart & Chapman, 2006; Latasa et al., 2006). AgfD, a transcription regulator from the Tafi operon, positively regulates expression of those structural elements and confers two distinct Salmonella morphotypes: red dry and rough (rdar) or smooth and white (saw). Thus, the Tafi operon supports the persistence of the biofilm-associated bacteria in response to environmental conditions (Gibson et al., 2006). In comparison with the saw morphotype, the rdar morphotype, which was isolated during tomato-originated outbreaks, adhered and attached better to tomato leaflets (Cevallos-Cevallos et al., 2012). Similarly, on parsley plants, attachment and resistance to disinfection treatments are improved in biofilm-associated bacteria. The efficiency with which a biofilm protects bacteria from disinfecting agents seems to increase during storage and food processing because no significant difference was observed between the wild-type and mutants failing to produce biofilm when treated with chlorine at the early stages of infection (Lapidot et al., 2006). A large screen of 6000 S. Newport mutants identified 20 mutants with lower attachment ability to alfalfa sprouts (Barak et al., 2005). Interestingly, genes identified in this study code for the surface-exposed aggregative fimbria nucleator curli (agfB) and for the global stress regulator rpoS. Both proteins regulate the production of curli, cellulose and adhesins, important for biofilm formation. AgfD was also identified in this study. In addition, Barak et al. (2007) showed that yihO (involved in O-antigen capsule formation) and bcsA (coding for a cellulose synthase) are important for adhesion to alfalfa sprouts (Barak et al., 2007), whereas cellulose and curli are involved in the transmission of S. Typhimurium from water to parsley leaves (Lapidot & Yaron, 2009). Another report characterized two additional genes (STM0278 and STM0650) as important factors for the infection of alfalfa sprouts, due to their essential role in formation of biofilm and swarming (Barak et al., 2009).
The disease-like symptoms appearing on leaves in response to the attachment and recognition of bacteria (wilting and chlorosis) seem to depend on the structure of the bacterial capsule. Strains such as S. Senftenberg, S. Cannstatt, S. Krefeld and S. Liverpool, all of which belong to the serogroup E4 (O: 1, 3, 19), possess the O-antigen and induce rapid wilting and chlorosis in Arabidopsis plant. In contrast, infiltration with serovars lacking the O-antigen does not provoke such symptoms (Berger et al., 2011). In contrast to Arabidopsis, infection with S. Typhimurium seems to cause no disease-like symptoms in tobacco plants (Shirron & Yaron, 2011). Whether this difference depends on the different immune responses remains to be verified.
Salmonella gain access to plant interior
Several natural entry points used by Salmonellae to infect plant organisms were described in recent reports. Stomata are natural openings responsible for gas exchange. Salmonella enterica was shown to aggregate near the open stomata of iceberg lettuce leaves and subsequently invade the inner leaf tissues under light conditions, suggesting that the pathogen is attracted to nutrients produced in photosynthetic active cells (Kroupitski et al., 2009), and even competes for carbon sources with the natural endophytic microflora (Klerks et al., 2007). Internalization via stomata was also documented by electron microscopy in arugula, basil and parsley (Golberg et al., 2011). However, differences in Salmonella internalization were observed in different seasons; summer had the highest frequency of internalization (Golberg et al., 2011). In addition, the type I glandular trichomes and hydratodes were shown to be a potential colonization site in tomato leaves, as confirmed by fluorescence and confocal laser microscopy (Barak et al., 2011; Gu et al., 2011, 2013).
The movement of bacteria within plants grown in fields contaminated with Salmonellae and the colonization of fruits is a very important agronomical concern. If present in the soil, S. Typhimurium systemically infects tomato plants including the fruits, without inducing any disease symptoms except for a slight reduction in plant growth (Gu et al., 2011). Since plant-originated bacteria retain the ability to infect animals (Schikora et al., 2011), plants infected before the harvest and storage/processing procedures are likely to be responsible for at least a part of the salmonellosis outbreaks associated with raw fruits and vegetables.
A very intriguing question is whether, as in the case of animals, Salmonella gains access to the intracellular region of plant host. Observations on Arabidopsis roots inoculated with S. Typhimurium strain 14028s expressing the green fluorescent protein (GFP) showed that bacteria are present within root hair cells already 3 h after infection (Schikora et al., 2008); 20 h later, bacteria were also visible in other rhizodermal cells. The internalization rate observed in Arabidopsis protoplasts and tobacco cells was relatively low (Schikora et al., 2008; Shirron & Yaron, 2011). Other reports that focused on mesophyll cells, indicated only the extracellular presence of Salmonella (Kroupitski et al., 2009, 2011). Thus, bacterial internalization in root cells, or in leaf or fruit cells, needs to be examined further.
Bacteria require T3SS for efficient infection
Recent findings regarding the mechanism of plant infection indicate that the mechanisms used for plant and animal infections are similar (Schikora et al., 2012b). Infiltration of Arabidopsis plants with the wild type S. Typhimurium and type III secretion system (T3SS) mutants demonstrated that both T3SSs encoded on Salmonella genome are important for a successful colonization (Schikora et al., 2011, 2012b; Shirron & Yaron, 2011). The isogenic T3SS mutants of S. Typhimurium 14028s prgH and invA (encoded by Salmonella pathogenicity island (SP-1), and ssaV and ssaJ (encoded by SPI-2) had lower proliferation rates and enhanced hypersensitive response (HR)-related symptoms in Arabidopsis plants, suggesting that these mutants are unable to suppress HR-like symptoms (Schikora et al., 2012b). Moreover, a transcriptome comparison of responses to infection with Salmonella wild-type and prgH mutant revealed that highly conserved Arabidopsis genes involved in defense are up-regulated upon infection with the prgH mutant (Schikora et al., 2011). Another study analyzed the oxidative burst after exposure to S. Typhimurium in Nicotiana tabacum plants and cell suspensions. Living Salmonella did not trigger an oxidative burst, whereas heat-killed or chloramphenicol-treated bacteria were effective elicitors, indicating that the pathogen actively suppresses this immune response. Furthermore, deletion of invA reduced the ability of Salmonella to suppress the oxidative burst (Shirron & Yaron, 2011). In summary, all three reports imply that a functional secretion system is required for the suppression of the plant defense mechanism.
Plant responses to Salmonella infection
The involvement of the plant immune system in the response to a Salmonella attack was demonstrated in several independent publications. The previously mentioned oxidative burst in response to treatment with killed bacteria or the invA mutant (Shirron & Yaron, 2011), or the transcriptional reprogramming in response to the prgH mutant (Schikora et al., 2011) are good examples of immune responses suppressed by wild-type bacteria. Another recently published example is the recognition of the Salmonella SseF effector protein (Ustün et al., 2012). SseF elicits HR-like symptoms in Nicotiana benthamiana, and a loss of symptoms upon silencing of the SGT1 suppressor indicates that SseF is recognized in an R protein-mediated mechanism. This recognition activates the effector-triggered immunity of the host plant (Ustün et al., 2012) otherwise induced by a disarmed pathogen (Schikora et al., 2012b). Additionally, plant hormones usually associated with defense responses (ethylene and jasmonic acid) were found to be of major importance; the JA-mediating pathway seems to be crucial for restricting Salmonella proliferation in Arabidopsis (Schikora et al., 2008). However, salicylic acid and ethylene are also important in the resistance to Salmonella (Iniguez et al., 2005). Furthermore, the mitogen-activated protein kinase (MAPK) cascades are involved in the response to Salmonella infection. Three MAP kinases commonly associated with immune response (MPK3, MPK4 and MPK6) are activated 15 min after contact with Salmonella bacteria (Schikora et al., 2008). The role of MAPK cascades in the defense against these bacteria is further supported by the fact that the mpk3 and mpk6 mutants are highly susceptible to S. Typhimurium infection (Schikora et al., 2008).
The infection potential of Salmonella-infected Arabidopsis leaves towards mice (Schikora et al., 2011) and the increasing number of outbreaks related to raw fruit and vegetables (Table 1 and references within) are clear indications of the role of plants as dissemination vectors. In addition, recent findings support the notion that Salmonellae infect plants in an active process, which resembles the infection of animal hosts. The characterization of invasion mechanism(s), immune suppression and cellular reprogramming during infection of plant hosts is a very intriguing question, still requiring answers. The knowledge gained will surely increase our understanding of the bacterial infection mechanism and also provide suggestions for future prevention and food protection strategies.
The authors would like to acknowledge Dr. Heriber Hirt, Unité de Recherche en Genomique Végétale Plan Genomics, and Marek Schikora, Fraunhofer-Institut für Kommunikation, Informationsverarbeitung und Ergonomie FKIE Wachtberg, for helpful discussions on this paper. The work of C.H.R. is supported by the Mexican National Council for Science and Technology fellowship (CONACYT).