Interactions Affecting the Proliferation and Control of Human Pathogens on Edible Plants

Authors

  • D. Aruscavage,

    1. Authors Aruscavage and Lee are with Dept. of Food Science and Technology, The Ohio State Univ., 2015 Fyffe Road, Columbus, OH 43210-1007, U.S.A. Author Aruscavage and LeJeune are with Food Animal Health Research Program. Author Miller is with Dept. of Plant Pathology. Direct inquiries to author LeJeune (E-mail: lejeune.3@osu.edu).
    Search for more papers by this author
  • K. Lee,

    1. Authors Aruscavage and Lee are with Dept. of Food Science and Technology, The Ohio State Univ., 2015 Fyffe Road, Columbus, OH 43210-1007, U.S.A. Author Aruscavage and LeJeune are with Food Animal Health Research Program. Author Miller is with Dept. of Plant Pathology. Direct inquiries to author LeJeune (E-mail: lejeune.3@osu.edu).
    Search for more papers by this author
  • S. Miller,

    1. Authors Aruscavage and Lee are with Dept. of Food Science and Technology, The Ohio State Univ., 2015 Fyffe Road, Columbus, OH 43210-1007, U.S.A. Author Aruscavage and LeJeune are with Food Animal Health Research Program. Author Miller is with Dept. of Plant Pathology. Direct inquiries to author LeJeune (E-mail: lejeune.3@osu.edu).
    Search for more papers by this author
  • J.T. LeJeune

    1. Authors Aruscavage and Lee are with Dept. of Food Science and Technology, The Ohio State Univ., 2015 Fyffe Road, Columbus, OH 43210-1007, U.S.A. Author Aruscavage and LeJeune are with Food Animal Health Research Program. Author Miller is with Dept. of Plant Pathology. Direct inquiries to author LeJeune (E-mail: lejeune.3@osu.edu).
    Search for more papers by this author

Abstract

ABSTRACT:  Pathogens on edible plants present a significant potential source of human illness. From 1991 to 2002, 21% of Escherichia coli O157:H7 outbreaks were from produce-related sources. E. coli O157 and other enteric bacteria can contaminate the surface of edible plants both pre- and postharvest. Some pathogens do not survive on the leaf surface or are removed by washing, but a significant portion of these enteric pathogens can persist on the surface and proliferate. Proliferation of these dangerous pathogens can increase the likelihood of foodborne disease associated with fresh or minimally processed produce. Several intrinsic and extrinsic factors determine the ability of enteric pathogens to attach and proliferate in the phyllosphere of plants. These include motility of the pathogen, leaching of nutrients by the plant, and interaction with epiphytic organisms. The interaction of enteric pathogens with the environment can lead to internalization into tissue, incorporation into biofilms, and genetic transfer. Current produce sanitation practices can reduce the microbial load from 1 log10 to 3 log10, so there are many new treatments possible. Understanding the ecology of enteric pathogens on plants is important to the development of sanitation methods and biocontrol agents. This knowledge can also assist the farmer in preventing contamination. With increasing consumption and importation of produce, its safety is a high priority for processors and U.S. consumers. Food safety may be markedly improved with proper attention to pathogens on edible plants.

Fruit and Vegetable Microbiological Safety Concerns

Increased consumption of minimally processed fruit and vegetables has focused attention on how the microenvironment of these plants affects produce safety. This microenvironment can help or hinder food safety, affecting proliferation and persistence of enteric pathogens on plants. Understanding these effects is important in developing new technologies to improve washing and disinfection of fresh produce. Current and future research leading to improved safety of fresh produce is the focus of this review. Brandl (2006) recently reviewed the fitness of human enteric pathogens on plants. This review advances this literature with focus on control of pathogens with emphasis on washing and antibiotic resistance. Understanding the importance of the phyllosphere in enteric pathogen survival will help ensure safety of fresh fruits and vegetables, a food category with markedly increasing consumption and increasing illness outbreaks in recent years.

Unreported before the 1990s, it is no longer surprising that a significant number of outbreaks of foodborne disease can be linked to fresh produce. By the mid-1990s, illnesses associated with cantaloupe, tomatoes, alfalfa, lettuce, and several other fruits and vegetables were documented (Tauxe 1997). The U.S. Department of Agriculture (USDA) recommends adults consume 5 cups of fruit and vegetable products daily to maintain good health (http://mypyramid.gov). The United States has become a major net importer of fruits and vegetables, with imports more than doubling during the decade 1994 to 2004 to reach a record $12.7 billion (http://tse.export.gov). This huge array of global produce has questionable safety and represents a large potential source of foodborne illness. With increased consumption and importation, the risk of foodborne illness associated with produce has increased (Beuchat 2002). Increased reporting by clinicians also provides a better surveillance system of produce-associated foodborne disease outbreaks (Suslow 2002).

The most common bacterial enteropathogens associated with fruits and vegetables are Salmonella spp. (Thunberg and others 2002). In 1990, a multistate outbreak of salmonellosis in the midwestern United States was caused by consumption of raw tomatoes (Hedberg and others 1999). Cantaloupe and strawberries were also identified as sources of Salmonella outbreaks in the 1990s (Hedberg and others 1999). Rangel and others (2005) showed produce was the transmission vehicle for E. coli O157:H7 in 21% of the foodborne outbreaks from 1982 to 2002 with the 1st recorded incident occurring in 1991. From 2000 to 2004, produce-related outbreaks were usually the 2nd most common identified outbreaks of E. coli O157 (Figure 1). E. coli O157 outbreaks were associated with apple cider, lettuce, radish, alfalfa sprouts, and other mixed salads since 1991 (Beuchat 2002).

Figure 1—.

Vehicles of Escherichia coli outbreaks by year, adapted with permission from Rangel and others (2005)

Soil microorganisms such as Listeria monocytogenes and Clostridium botulinum can also attach to fruits and vegetables (Beuchat and Ryu 1997). Both of these pathogens have been isolated from cabbage (Buck and others 2003). Shigella flexneri was the causative agent in an outbreak associated with scallions (Tauxe 1997). Campylobacter jejuni was isolated from lettuce, peppers, and spinach (Buck and others 2003). Viruses such as hepatitis A and Norwalk viruses have also been implicated in outbreaks associated with fresh produce. A hepatitis A outbreak was caused by consumption of contaminated raspberry mousse, and lettuce was implicated as the source of a Norwalk virus outbreak in a school cafeteria (Beuchat 1996). In 1996, Cyclospora was the causative agent in a raspberry-associated outbreak of disease that affected over 900 people (Tauxe 1997). Although all these microorganisms are capable of surviving on produce (Beuchat 2002), the organisms most likely to cause an outbreak that need to be studied for produce safety are enteric pathogens such as Salmonella and E. coli O157 (Buck and others 2002). These organisms deserve close scrutiny in terms of produce safety.

Enteric pathogens are transient residents of plants that only appear after contamination from the environment (Suslow 2002). Several preharvest and postharvest sources for bacterial contamination of produce are documented. These include contaminated irrigation or wash water, improperly managed manure used as fertilizer, wild animals, and, of course, human handling (Beuchat and Ryu 1997). Chlorine washes are used by fruit and vegetable producers to lower counts of microbes during processing, but washes can only create about a 1 log10 to 2 log10 reduction in total microbial counts (Beuchat 1996).

Enterobacteriacea survival on plants is affected by multiple factors, including nutrient availability, UV radiation, competition, toxic compounds released by the plant and other microorganisms, and desiccation. However, the principal factors limiting the proliferation and survival of microorganisms on plant surfaces are the frequent and large fluctuations that occur in temperature, humidity, and radiation (Beattie and Lindow 1994; Suslow 2002). To better understand how human pathogens persist, and possibly proliferate, in this environment requires a more thorough understanding of the factors governing the colonization of plants (phyllosphere and rhizosphere) with these organisms.

Contamination and Persistence of the Phyllosphere of Edible Plants with E. coli O157 and Salmonella

Johnston and others (2005) found coliforms averaged between about 1.0 log10 and 3.4 log10 cfu/g (colony-forming unit per gram) of vegetable tissue at different times of production and stages of processing (harvest, washing, and packaging). Coliforms are common indicator organisms for animal and human fecal contamination of the environment. Their presence can indicate that Salmonella, E. coli O157, or other enteric pathogens might also be present. Contrary to conventional wisdom, E. coli and other Enterobacteriacea survive quite well outside the animal host, as shown in Table 1. The persistence of E. coli O157 in agricultural and food processing environments indicates that nutrient sources are available for this pathogen outside of the digestive system of mammals. For example, Kudva and others (1998) found E. coli O157 persisted in nonaerated sheep manure for over a year. The organism remained in aerated sheep manure for about 4 mo (Kudva and others 1998). Wang and others (1996) found E. coli O157 survived in dairy cattle manure inoculated with 103 cfu/g for 49 d at 22 °C. Compost inoculated with Salmonella Typhimurium, either directly or through contaminated irrigation water, was applied to soil where parsley and lettuce were being grown. The pathogen persisted on lettuce for 63 d and parsley for 161 d (Islam and others 2004). Another route of transmission to plants could be from pesticide solutions; Guan and others (2005) showed both Salmonella and E. coli O157 survived in different brands of pesticides and fungicides. After 24 h, the concentration of E. coli O157 averaged about 2.3 log10 across the 7 solutions initially contaminated with a 4 log10 inoculum. The 7 solutions containing a 4 log10 concentration of Salmonella Heidelberg averaged 3.8 log10 after 24 h. When fungicides contaminated with E. coli O157:H7 and Salmonella Enteriditis were sprayed on tomato plants, the pathogens survived 2 and 15 d, respectively. Salmonella survived better than E. coli O157:H7, Shigella flexneri, and Listeria monocytogenes in the pesticides used for the study (Guan and others 2005). Solomon and others (2002a) studied the impact of the types of irrigation water on E. coli O157 survival. Ninety percent of lettuce plants spray irrigated with 107 cfu/mL of E. coli O157 became contaminated with E. coli O157, but only 19% of the plants were contaminated with E. coli O157 when surface irrigated with the same concentration. In some instances, the pathogen remained in the phyllosphere for 20 d (Solomon and others 2002a). Solomon and others (2003) found when lettuce was sprayed intermittently over a period of several days with 102 cfu/mL of E. coli O157, the population of the pathogen increased with each watering. Therefore, overhead spray irrigation is a food safety risk for E. coli O157 in fresh produce (Solomon and others 2002a, 2003). As water is necessary for plant growth, the quality of the water supply needs to be considered to develop safe practices for produce production. Since pathogens have been shown to persist in different water sources for extended periods of time (LeJeune and others 2001), treatment of the water supply can be as important as treating the produce.

Table 1—.  Persistence of E. coli O157 and Salmonella in farm environments and produce
PersistenceMediumPathogenReference
1 yNonaerated sheep manureE. coli O157Kudva and others (1998)
49 dDairy cattle manureE. coli O157Wang and others (1996)
63 dLettuceSalmonella typhimuriumIslam and others (2004)
161 dParsleySalmonella typhimuriumIslam and others (2004)
2 dFungicidesE. coli O157:H7Guan and others (2005)
15 dFungicidesSalmonellaGuan and others (2005)
20 dLettuceE. coli O157Solomon and others (2002)
21 d to 45 dFallow soilE. coli O157Gagliardi and Karns (2002)
47 d to 96 dAlfalfaE. coli O157Gagliardi and Karns (2002)
100 dGrasslandE. coli O157Bolton and others (1999)
49 dTomatoSalmonella MontevideoGuo and others (2001)
27 dTomatoSalmonella PoonaGuo and others (2001)

Persistence of enteric pathogens is important because these organisms can develop a niche in the phyllosphere long before sanitation occurs, possibly making them more difficult to remove. Brandl and Mandrell (2002) demonstrated that Salmonella Thompson proliferated in the phyllosphere and sometimes formed aggregates with epiphytic Pantoea agglomerans on the surface of cilantro. This Salmonella serotype did not grow as well as the epiphytes, demonstrating they did not use nutrients as efficiently as the epiphytes on cilantro. On the other hand, E. coli P36 and Salmonella P2 that contaminated bean seeds grew rapidly and became the dominant microflora on the sprouts of these seeds for up to 3 d (Warriner and others 2003). E. coli O157 persisted for 25 d to 41 d in fallow soil and persisted for 47 d to 96 d in the rhizosphere of alfalfa. However, populations were lower than those of native coliforms in the soil or rhizosphere but were also capable of surviving in frozen soil samples (Gagliardi and Karns 2002).

When a nontoxigenic strain of E. coli O157 was applied to grassland, population size decreased from that of the original inoculum, but the pathogen persisted for about 100 d (Bolton and others 1999). Human infections of E. coli O157 can be caused by as few as 100 cells (Jay and others 2005), so persistence even of low numbers of microorganisms on produce intended for human consumption with minimal processing can be just as important as proliferation of this pathogen. Guo and others (2001) studied persistence of Salmonella on tomatoes. The Montevideo serotype was identified in stem tissue 49 d after inoculation, and the Poona serotype was identified 27 d after inoculation of the stem. When the flower was inoculated, Montevideo and Michigan serotypes persisted for 39 d, and these were identified in pulp of the tomato and the stem scar. The Enteriditis serotype persisted for 27 d and was identified only in wash water used on the tomato (Guo and others 2001). These studies verify the persistence of E. coli O157 and Salmonella on produce.

Impact of Structural Damage on Colonization by Human Pathogens

The plant cuticle contains waxes that make it difficult for microorganisms to attach to plant tissues. This layer repels water that might contain immigrating microbes. The cuticle also protects the plant from environmental stresses. For plant and human pathogen attachment, the cuticle usually must be penetrated (Beattie 2002). Damaged fruits and vegetables might harbor human pathogens more than intact fruits and vegetables. Wounded apples had a 1 log10 to 3 log10 increase in E. coli concentration 2 d after infection compared to noninjured apples (Janisiewski and others 1999a). The increase in population was greater with a small inoculum than a large inoculum. It is possible that E. coli adjusted to the microenvironment in the wounds, allowing it to reproduce more readily, especially after the first 48 h postinoculation (Janisiewski and others 1999a). Dingman (2000) also showed about a 2 log10 increase in E. coli O157 within 48 h postinoculation in damaged tissue of apples but used higher starting concentrations. There was a significant increase in pH of damaged apples as compared to similar varieties that were not damaged (Dingman 2000). On damaged honeydew melons and apples, Salmonella Enteriditis replicated at 10 °C and 20 °C. The pathogen had a greater than 5 log10 increase over the 168 h study when incubated at 20 °C on the honeydew slices (Leverentz and others 2001). We hypothesize that it is possible that damaged tissue can enhance microbial proliferation in 2 ways: first, the tissue might be more readily colonized by human pathogenic enterobacteria because of changes in the microenvironment or, alternatively, the microenvironment conditions change independent of bacterial colonization, such as through enzyme or nutrient leakage.

Wounds can also act as sites of coinfection with other microorganisms that can alter the microenvironment and stimulate proliferation of E. coli O157. E. coli O157:H7 had increased in number 3 log10 from the initial dose in wounded apple tissue in the presence of Glomerella cingulata (Riordan and others 2000). This fungal plant pathogen erodes the cuticle of apples, providing a way for E. coli O157 to enter wounds and further alter the environment to allow for proliferation (Riordan and others 2000). Brandl and Mandrell (2002), using confocal laser scanning microscopy, showed that Salmonella Thompson was in lesions on cilantro plants. There is additional evidence that human enterobacterial pathogens are more likely to be found in the presence of fungal or bacterial plant pathogens than in noninfected retail produce (Wells and Butterfield 1997, 1999).

Nutritional factors

Since E. coli O157 and Salmonella, transient species in the plant phyllosphere, proliferate to greater numbers in damaged plants than whole plants, it stands to reason that there must be an ecological disturbance on the damaged plant surface. Epiphytic microorganisms normally present in the harsh conditions of the plant phyllosphere use nutrients leached by the plant for survival and proliferation. Measuring carbon utilization can determine the structure of a microbial community in the plant phyllosphere (Wilson and Lindow 1994). Understanding the interactions in the epiphytic flora might assist in understanding how human pathogens act in the phyllosphere.

The average epiphytic aerobic heterophilic bacterial population is between 105 and 107 cfu/g of leaf (Mercier and Lindow 2000). Early studies showed approximately 2 × 106 cfu/g of leaf tissue were present on unwashed cabbage leaves (Lund 1992). Lettuce had approximately 105 cfu of aerobic bacteria/g of leaf tissue (Lund 1992). Not all microorganisms have adapted to survive in this harsh habitat. Thompson and others (1993) cultured an average of 13 bacterial species on different leaf types as compared to 40 different species in the soil where plants were grown. The most common group of epiphytic microorganisms present in the phyllosphere of edible plants is gram-negative bacteria (Lund 1992). Colonization is not spatially distributed uniformly across the entire leaf surface. Instead, bacteria cluster in specific regions. The most common areas of bacterial aggregations are located at the base of trichomes, around the stomata, and along veins in the leaves. These areas allow the best proliferation conditions for epiphytic bacteria (Beattie and Lindow 1999; Leveau and Lindow 2001). Cooley and others (2003) found E. coli O157:H7 clustered near the veins on the surface of Arabidopsis thaliana. Salmonella Thompson clustered in regions along the veins of cilantro plants. It was concluded that this region has greater wettability than other areas of the plant, thus increasing the likelihood of nutrient leaching and water availability (Brandl and Mandrell 2002). Knowing what nutrients leach and how they are made available to human pathogens can help in the development of biocontrol agents for human pathogens in fresh produce. Janisiewski and others (1999a) showed that a formula of antagonist microbes used to slow decay of stored apples also decreased E. coli O157:H7 present in wounds of apples.

The predominant sugars leached from the surface of plants are glucose, sucrose, and fructose (Mercier and Lindow 2000). Jaegar and others (1999) found rhizosphere microorganisms were in largest numbers near the root tips where the highest concentration of sucrose diffused into the soil. Plants also leach minerals, amino acids, sugar alcohols, pectic substances, proliferation hormones, and vitamins (Tukey 1970). These leachates supply the carbon and nitrogen sources for the microbial flora present on or in the plant tissue. It is possible that after the bacteria attach, they multiply in areas around these nutrient sources (Wachtel and others 2002).

The amount and type of substances leached by different plants vary considerably. Bean leaves release as much as eight times more potassium than sugar beet leaves. This, in turn, can influence how much bacterial colonization occurs on a plant. Dingman (2000) found damaged Macintosh apples inhibited the proliferation of E. coli O157 2 d to 3 d after infection, but damaged Red Delicious apples showed increases in the E. coli O157 population. It is likely that Macintosh apples released a compound that inhibited the proliferation of the human pathogen (Dingman 2000). Reinders and others (2001) showed that caffeic acid, a common phenolic compound in apples, inhibited proliferation of E. coli O157.

Very young, growing leaves release limited amounts of leachate compared to those that are near senescence (Tukey 1970). Thompson and others (1993) showed microbial communities on sugar beets were greater on old and senescent leaves than on young leaves. Jacques and Morris (1995) observed more microbial contamination on older leaves of endive than on the younger inner leaves. Plant pathogens caused more decay on these older leaves than on the young leaves, further contributing to leaching and other microenvironmental changes on the plant surface that would favor the survival and proliferation of human pathogens (Jacques and Morris 1995).

Injury to leaves also increases the amount of leaching from leaves (Tukey and Morgan 1963). There is evidence that suggests damaged tissue is more likely to positively affect the proliferation of E. coli O157 than healthy tissue (Janisiewski and others 1999a; Dingman 2000; Riordan and others 2000). The type of damage can be an important factor. Kenney and others (2001) showed no significant difference in E. coli O157 concentration on the surface of bruised or nonbruised apples, but Dingman (2000) extracted the juice from the bruised tissue and showed E. coli O157 proliferated in the nutrient solution. Janisiewski and others (1999a) found proliferation in the cut surface of an apple, as did Riordan and others (2000). Confocal laser scanning microscopy showed that E. coli O157 attached to lettuce tissue more readily at cut edges than on intact lettuce tissue (Seo and Frank 1999). It is possible that bruising alone might not cause increased proliferation, but the cracking of the surface that allows nutrient release might be needed. This could be associated with the pathogen's inability to penetrate the cuticle that is necessary for attachment (Beattie 2002).

Survival of microorganisms in the phyllosphere is often limited by availability of carbon and nitrogen sources leached from the plant. Across several vegetable samples tested, carbohydrates range from 3.4 to 74 μg/mL and amino acids range from 3.0 to 52 μg/mL in leaf exudate (Morris and Rouse 1985). Different plants do not replenish these pools at the same rate required by the microbes, so competition for nutrients is very important in epiphyte survival (Leveau and Lindow 2001). The availability of sugars is not uniform around the leaf, so when these pools are depleted, rapid proliferation of the epiphytic microorganisms usually stops. Jablasone and others (2005) determined that large counts of E. coli O157:H7 associated with alfalfa sprouts were due to nutrients in the exudate released by emerging seedlings. A decrease in bioluminescence over time was likely associated with changes in the nutrient patterns of the plant exudate (Jablasone and others 2005).

The type and amount of exudates released by plants are limiting factors in the survival and proliferation of epiphytic organisms. The ability of leaf surface organisms to find and use these nutrient sources is a very important determinant of epiphytic fitness. For resident epiphytes, the limiting factor for proliferation is nitrogen (Lindow 1991). It is not known if nitrogen is a limiting factor for enteric pathogens because as a casual species to the phyllosphere, human pathogenic bacteria might use carbon and nitrogen sources not commonly used by epiphytes. It is necessary to know what extracts are being released from the plant that might give these enteric pathogens a competitive advantage when they are present on produce associated with outbreaks.

Water availability

Spatial experiments have shown that epiphytic bacteria grow better in areas where there is high wettability of the tissue. The population size and distribution of phyllosphere microorganisms might be altered by the amount of usable water present on the leaf surface (Bunster and others 1989). There is also evidence that high relative humidity increases bacterial communities on plant surfaces (Leben 1988), so besides nutrients, water is a factor in bacterial proliferation in the phyllosphere. Cooley and others (2003) found at 100% humidity and in the absence of competition, E. coli O157 and Salmonella enterica grew to as much as 107 cfu/g of leaf tissue in the phyllosphere of Arabidopsis thaliana.

Chet and others (1973) observed Pseudomonas lachrymans was attracted by chemotaxis to water droplets from different plant sources. Some plant epiphytes can release biosurfactants that changes leaf wettability in a way that makes it easier for these organisms to use water. It has also been demonstrated that even after these biosurfactant-producing organisms have become less dense on the leaf surface, the biosurfactant activity continues; thus water availability is still increased for remaining bacteria (Bunster and others 1989). This is an important food safety issue because plants that have been previously diseased by pathogens capable of producing biosurfactants have increased leaf wettability. This increased wettability favors proliferation of enteric pathogens that contact these plants.

Competition

A competitive advantage can be gained by an organism that grows quickly, thus establishing dominance when nutrient levels are high or by being able to grow when there are few nutrients remaining. Competitors who have highly efficient modes of nutrient uptake or ability to produce antimicrobial compounds have a competitive advantage (Beattie and Lindow 1994). Schuenzel and Harrison (2002) found 3% of epiphytes isolated from produce contained inhibitory compounds that acted against 1 or more of the following pathogens: Staphylococcus aureus, E. coli O157, Salmonella Montevideo, and Listeria monocytogenes. Isolates from shredded lettuce were most likely to produce inhibitory compounds that were effective against all 4 pathogens. Most of the inhibitory epiphytes were gram-negative, with the highest percentage being pseudomonads (Schuenzel and Harrison 2002).

If organisms use different carbon and energy sources, microbial populations might actually coexist in the phyllosphere and might enhance the epiphytic fitness of other populations. In transgenic plants that released excess opines, opine-catabolizing pseudomonads coexisted with other pseudomonads that used different carbon sources. In the wild-type variety, these 2 microorganisms would normally compete for nutrients (Wilson and others 1995). Wilson and Lindow (1995) used salicylate as the sole carbon source for one of the Pseudomonas strains so that it did not compete with a known competitor and found similar nutrition partitioning between the species. Wilson and Lindow (1994) hypothesized that coexistence could occur between epiphytes that have different nutrient resource patterns. If 2 or more organisms use different carbon sources, they all might be able to survive in the same region of the leaf surface. Most cohabitation of microbes is not complete coexistence (reaching same population sizes with or without the other species) or complete competition (sharing the same nutrients), but a coexistence of microbes between the 2 extremes (Wilson and Lindow 1994). It is evident that finding what nutrients are available and how they are utilized by human pathogens is essential in devising a strategy to control human pathogens in the phyllosphere by use of biocontrol agents.

Plant lesions in which E. coli O157 and other pathogens proliferate might be a source of rich nutrients (Dingman 2000; Riordan and others 2000). High concentrations of Pseudomonas syringae acted as an antagonist to E. coli O157 in these wounds, possibly competing for the same carbon and energy sources. When P. syringae was not present, E. coli O157 had at least a 2 log10 increase in concentration compared to a coinoculation (Janisiewski and others 1999b). Liao and Fett (2001) found 6 out of 120 resident epiphytes inhibited the proliferation of at least 1 of the following: E. coli, L. monocytogenes, Salmonella Chester, or Erwinia carotovora. Two Pseudomonas species reduced the proliferation of both E. coli and L. monocytogenes (Liao and Fett 2001). These organisms were found to be present in the phyllosphere of different types of vegetables. Enterobacter asburiae competed with E. coli O157 and Salmonella enterica (Cooley and others 2003). Janisiewski and others (1999b) found no evidence of compounds produced by Pseudomonas that inhibited E. coli O157. Pectin-degrading epiphytes of bell peppers, romaine lettuce, and baby carrots did not inhibit proliferation of human pathogens (Liao and Fett 2001). Some strains of E. coli might also participate in pectin degradation or enhance the ability of other organisms to degrade pectin. Dongowski and others (2000) showed increased pectin degradation when Bacteroides thetaiotaomicron (a known pectin-degrading microbe) was mixed with E. coli than when B. thetaiotaomicron was alone in culture. While some residents of the phyllosphere act as antagonists of enteric pathogens, some might actually have neutral or symbiotic association with these types of casual residents.

Effects of Damage Caused by Plant Pathogens

Plant surfaces contain several resident species of microorganisms (Carmichael and others 1999). When transient foodborne pathogens enter the phyllosphere, there are several possible interactions between the resident epiphytes and transient enteropathogens. In cases where fresh produce has been implicated in outbreaks (Tauxe 1997; Hedberg and others 1999), cooperation between human pathogens and epiphytes was likely because competition between a commensal epiphytic flora and a transient contaminant would favor the exclusion of the latter. Wells and Butterfield (1997) state that Salmonella contamination increased in vegetables that have bacterial soft rot compared to noninfected produce. This implies that plant pathogens might have a synergistic or commensal relationship with the natural flora. Knowledge of interactions between human pathogens and plant pathogens can be beneficial in the study of foodborne disease associated with fresh and minimally processed fruits and vegetables because this interaction is occurring in a harsh environment where cooperation seems to be a necessity for proliferation of both epiphytic organisms and transient microorganisms.

Epiphytic organisms that can be pathogenic might help create an environment that makes it easier for transient species to proliferate (Beattie and Lindow 1999). These environmental changes include changes altering leakage from plant to increase nutrients and water availability. Atkinson and Baker (1987) found P. syringae pv. syringae disrupted sucrose transport in bean plants without damaging the cell membrane of the plant cell. P. syringae alters the ion gradient between cells, causing more sucrose to be released from the cell and be used by epiphytic organisms (Atkinson and Baker 1987). If there are microenvironmental changes caused by leaf surface residents that allow enteric pathogens to proliferate on the surface, there are major food safety concerns. Salmonella Typhimurium coinoculated with soft rot bacteria had a 1 log10 greater increase in population size on potatoes, carrots, and peppers than when inoculated alone (Beuchat 2002). In soft rotting plant diseases, the fleshy part of plant tissue is destroyed and the debris can ooze from cracks. Soft rot-causing Erwinia and some Pseudomonas can actively destroy the tissue of living cells of the plant. These plant pathogens enter wounds and multiply using nutrients from broken cells. They use pectolytic and cellulytic enzymes to damage the middle lamella and cell walls of intact cells. Water is released from these cells and the pectic substances will be liquefied, thus softening the infected tissue (Agrios 2005). Besides the water that is released, nutrients are released either directly from the plant cell or by enzymatic activity of the plant pathogen.

Wells and Butterfield (1997) showed an association between bacterial soft rot and the incidence of presumptive Salmonella in retail market produce. Nearly twice as many suspect Salmonella colonies were observed when bacterial soft rot was present as compared to healthy plant tissue. There were similar results with rots caused by fungi. There was a significant increase in presumptive Salmonella organisms when fungal rot was present compared to the healthy tissue. Coinfection with Alternaria or Cladosporium enhanced proliferation of Salmonella enterica in tomatoes (Wade and Beuchat 2003). It is speculated that these fungal pathogens cause an increase in the pH of the tomato, possibly altering the microenvironment to enhance survivability and reproduction of Salmonella (Wade and Beuchat 2003). Riordan and others (2000) showed coinfection with E. coli O157:H7 and Glomerella cingulata increased proliferation of the human pathogen. There was a noticeable increase in pH in the wounds where there was coinfection (Riordan and others 2000). Similar findings by Conway and others (2000) showed L. monocytogenes grew in apple tissue decayed by G. cingulata, which increased the pH of the tissue to 7.0, but did not grow in tissue decayed by Penicillium expansum, which decreased the pH of the tissue to 3.7. These data have led many researchers to speculate that pH change due to tissue damage is an important factor in allowing proliferation of E. coli O157:H7 in plant lesions.

Besides altering pH, plant pathogens can disturb the microenvironment in other ways, with ramifications on the proliferation of human pathogens. A nutritive advantage for epiphytic microorganisms could be attributed to substances toxic to plants being released by plant pathogens. Toxins alter the microenvironment by causing the release of sugars from plant cells and having antimicrobial capabilities. Syringomycin produced by P. syringae forms pores in the membrane of the plant cell, releasing the contents (Alfano and Collmer 1996). Tabtoxin, produced by 3 strain of P. syringae, is a β-lactam that could act against other epiphytes (Gross 1991).

Most plant pathogens use enzymes to degrade cell wall components. The byproducts of these catabolic reactions can create new carbon sources for the pathogen or a cohabitant in the phyllosphere. Cutinases produced by fungi and Streptomyces scabies are esterases that act on cutin, a major protein of the cuticle. Cellulases include a broad range of enzymes that degrade cellulose completely to glucose. Pectolytic enzymes degrade pectic substances in the cell wall. There are different pectic enzymes that act on the pectin molecule. Pectin is composed mostly of polygalacturonic acid but also contains rhamnose, arabinose, and galactose. There are also small amounts of other carbon-based molecules (Agrios 2005). Many of these compounds can be used by bacteria as carbon and energy sources. Pectolytic activity is common in many plant pathogens (Agrios 2005).

Plant pathogens can assist in the fitness of enteric pathogens in the phyllosphere by weakening the plant to enhance survival of microorganisms. Some plants can resist against infiltration by enteric pathogens by salicylic-independent and salicylic-dependent pathways. Iniguez and others (2005) showed that both flagella and elements of type III secretion system of Salmonella Typhimurium are recognized by plant defense mechanisms when inoculated inside the root tissue of alfalfa seedlings. Mutant strains lacking flagellin or type III secretion genes infiltrated the root epidermis of alfalfa sprouts more successfully than wild-type strains (Iniguez and others 2005). Jakobek and others (1993) demonstrated a suppression of plant defenses when Pseudomonas syringae not detected by the plant was inoculated before a Pseudomonas sp. that normally would be suppressed by the plant. The response of the bean defenses to E. coli was also lessened when previously infected with the Pseudomonas sp. not detected by the plant (Jakobek and others 1993). Plant pathogens that weaken plants' defense mechanisms can alter the microenvironment in a way that increase both proliferation and persistence of human pathogens in the phyllosphere, making removal of human pathogens by sanitizing less likely to be successful.

More research is needed to determine how plant pathogens influence human pathogenic proliferation on plants. Besides nutrition and competition, plant pathogens might aggregate with human pathogens and protect them from desiccation, UV radiation, and washing. These plant pathogens have also been shown to enter protected sites, such as the substomatal chamber, which make them less likely to be affected by sterilization methods (Wilson and others 1999). These hypothetical human pathogen–plant pathogen aggregates in protected sites might be resistant to chlorine washing. Thus they are a postharvest concern for produce safety.

Internalization and Dissemination of Bacteria

The potential internalization of pathogens is a concern in food safety because these pathogens are less likely to be removed during the washing steps after harvest than surface contaminants. Meneley and Stanghellini (1974) found bacteria inside each of 100 cucumber samples tested. They noted that most internalized bacteria belonged to Enterobacteriaceae, Pseudomonaceae, Corynebacteriaceae, Bacillaceae, and Micrococcaceae (Meneley and Stanghellini 1974). However, not all enteric bacteria are internalized equally well within plants; Klebsiella pneumonia had higher counts inside alfalfa seedlings than E. coli K-12 (Dong and others 2003). Pathogenic E. coli O157 was also more efficient at colonizing the interior than E. coli K-12 (Dong and others 2003). Dong and others (2003) also noticed differences in internalization capabilities between different serovars of Salmonella.

Solomon and others (2002b) showed that E. coli O157:H7 when inoculated in manure and added to planting soil contaminated and survived on lettuce plants grown in that soil. Using confocal microscopy, E. coli O157:H7 was observed inside of the tissue at depths as low as 45 μm (Solomon and others 2002b). It has also been shown that E. coli O157:H7 internalized in cress, lettuce, radish, and spinach seedlings that had been contaminated as seeds but did not remain internalized as mature plants (Jablasone and others 2005). In the absence of competitors, E. coli O157 grew inside root tissue and within the vasculature (Cooley and others 2003). Johannessen and others (2005) found different results from plants grown in contaminated soil. They did not detect E. coli O157 on lettuce seedlings from soil contaminated with E. coli O157:H7. They suggested that the concentration of bacteria in the soil and time of infection might be important in the attachment and internalization of E. coli O157:H7 (Johannessen and others 2005).

Zhuang and others (1995) found internalization of Salmonella Montevideo into core tissue of tomatoes was higher when inoculated at 25 °C than when inoculated at 10 °C. The inoculum dose required for Salmonella to internalize can be as low as 1 cfu (Dong and others 2003). Dong and others (2003) noted that the internalization of Salmonella does not seem to be a passive process, but one that requires energy either from the organism or the environment. Likewise, motility seems to be an important factor for infiltration of E. coli O157 in plants. E. coli O157 and S. enterica are capable of movement along the surface of the plant (Cooley and others 2003). This ability to move along the surface can also assist in allowing these pathogens to enter into wounds, stomata, or other openings such as the calyx. Motility has also been shown to be an important factor in the ability of P. syringae subsp. glycinea to cause infection of plants (Hatterman and Ries 1989). Infiltration of E. coli O157 into internal regions of apples occurred through the open calyx with subsequent travel into the core region. These cells also attached to structures inside the core (Burnett and others 2000).

Internalization of pathogens in produce is of concern to the food industry. If these pathogens can internalize, then washing procedures are ineffective at removing them. Future sanitation practices that can disinfect protected sites and internal plant tissue may be more effective than washing. It is important to note that because of internalization, preharvest contamination control is potentially more critical than the best postharvest washing strategies.

Aggregates or Biofilms

Biofilms or aggregates on plant surfaces are collections of cells in an exopolymer matrix that can protect members from environmental stresses, desiccation, and bactericidal agents and act as pools for exchange of genetic material (Morris and Monier 2003). Epiphytic biofilms have been shown in the phyllosphere of several different types of vegetables, including endive, spinach, celery, parsley, and lettuce (Morris and others 1997). Fett (2000) found biofilms composed of rod-shaped bacteria on young alfalfa sprouts. The incorporation of human pathogens into these biofilms might protect them from produce washes and serve as a possible entry site for genetic resistance elements into human pathogens (Fett 2000). In the phyllosphere of cilantro, Salmonella Thompson formed aggregates with Pantoea agglomerans along the veins of the plant (Brandl and Mandrell 2002). Using confocal microscopy, Seo and Frank (1999) found E. coli O157 attached to cut edges of lettuce leaves and not to biofilms of Pseudomonas fluorescens, but some E. coli O157 were present in the pseudomonad biofilm.

Epiphytic biofilms have been observed in several locations on the plant surface. These are most commonly found at the leaf base with many attached to trichomes. The communities associated with the biofilms contained diverse populations of bacteria and fungi. Both gram-positive and gram-negative bacteria were seen in the phyllosphere of vegetables (Morris and others 1997). Of the epiphytic community in the phyllosphere, Morris and others (1998) showed that the percentage of bacteria in a biofilm ranges from 13% to 38% on endive compared to free-living epiphytes.

The exopolysaccharide produced by microbes in the biofilm aids in protecting members from the natural environment and also maintains water and nutrients for reproduction. Pectolytic phytopathogens of the genera Pseudomonas and Erwinia are capable of producing biofilms (Carmichael and others 1999). Carmichael and others (1999) observed that biofilm formation of minimally processed vegetables was associated with damaged tissue when stored. Since biofilm formation and human pathogen proliferation are associated with damage (Janisiewski and others 1999b; Wells and Butterfield 1999), heterogeneous biofilms containing human pathogen might occur on the plant surface. There is a need to understand the ecology of human pathogen heterogeneous biofilms to determine their effect on fruit and vegetable safety.

Bacterial cells within a biofilm differ from those cells that remain isolated on the plant surface. There are physiological changes in cells within a biofilm. Exopolysaccharide layers of biofilms can create abundant nutrients available to the biofilm members. They can also alter pH to enhance survival (Morris and others 2002). Other qualities of the biofilm noted by Morris and others (2002) include the ability to help with attachment to the surface of the plant and absorbance of volatile compounds created by the plant. The bacteria within the biofilms also work together by degrading xenobiotic compounds and having other members of the biofilm use these byproducts as a carbon source (Morris and others 2002). More research needs to be conducted to see if nutritional changes in biofilms help the proliferation and survival of human pathogens.

Antibiotic resistance is also increased in microbes within a biofilm when compared to those existing as individual cells. Sensitivity of Klebsiella pneumoniae cells to both ampicillin and ciprofloxacin was reduced in a biofilm compared to planktonic cells. Ampicillin did not penetrate the biofilm created by wild-type K. pneumoniae. Ciprofloxacin penetrated the biofilm but did not eliminate cells present in the biofilm after prolonged exposure (Anderl and others 2000). Anderl and others (2000) suggested that slow growing and stressed cells in specific areas of the biofilm are less susceptible to antibiotics so it is likely that these cells are most responsible for resistance observed. Brooun and others (2000) showed that it is most likely not multidrug resistant pumps that cause this increased resistance, but a small collection of superresistant cells that increase the overall resistance of the biofilm. These cells were resistant to increasing levels of ciprofloxacin (Brooun and others 2000). The mechanisms of these superresistant cells and the ability of transfer of these traits within the biofilm are unknown.

Biofilms are a potential site for horizontal gene transfer (Morris and Monier 2003). It was suggested that genes on mobile elements might readily transfer to new microbes introduced into an established biofilm. Transfer seems to be associated with proliferation, thus increasing the amount of plasmids present (Christensen and others 1998). Normander and others (1998) showed conjugation in the phyllosphere of bean plants. The most common places for these transfers to occur were in the stomata and interstitial spaces (Normander and others 1998). Bjorklof and others (1995) showed a positive correlation between increased conjugation and nutrient concentration. Because of the proximity, high nutrient concentrations, and high cell density created by biofilms, it is possible that antibiotic resistance can be passed to human pathogens that have attached to or become incorporated into an epiphytic biofilm.

Since heterogeneous biofilms have been shown to occur in the phyllosphere, the possibility of human pathogens incorporating into these aggregates is an issue of concern for fruit and vegetable safety. The ability of human pathogens to form homogeneous biofilms also needs to be studied. Adams and McLean (1999) showed that rpoS is an important gene in the formation of E. coli biofilms; its deletion decreases the formation of biofilms by E. coli. Ryu and others (2004) showed a spontaneous mutant of E. coli O157 that overproduced extracellular polymeric substances (EPS) formed biofilms on stainless steel chips. Scanning electron microscopic photos document the polysaccharide network that could help protect these cells from bactericidal agents (Ryu and others 2004).

Genetic Transfer of Antimicrobial Agents

The transfer of antibiotic resistant elements is a major concern for the meat industry since antibiotics are often used as growth and health adjuvants for livestock, but there is very little concern about antibiotics used in the produce industry. Because of the large populations of bacteria present on the surfaces of fruits and vegetables, exchange of genetic material is likely. If resistant genetic elements are transferred to human pathogens in the plant phyllosphere, this becomes a food safety concern, just like with livestock. One hundred thirty-seven vegetable samples were studied for antibiotic resistance of Enterobacteriaceae in Finland. Cefuroxime resistance was identified in 14% of the samples, chloramphenicol resistance was identified in 12%, and tetracycline resistance was identified in 5.5% of the samples. Resistance to other antibiotics was negligible (Osterblad and others 1999). Osterblad and others (1999), however, concluded that antibiotic-resistant microorganisms from vegetable sources did not influence the fecal flora of humans. Hamilton-Miller and Shah (2001) found multidrug-resistant epiphytes from carrots and lettuce. The most common epiphytes identified were Pseudomonas fluorescens and Pantoea agglomerans, both opportunistic pathogens. It was concluded from this work that produce can harbor multidrug-resistant organisms that could be a potential risk to humans (Hamilton-Miller and Shah 2001).

Multidrug-resistant bacteria have been found in tomato, lettuce mix, chicory, and sprouts. Each multidrug-resistant strain was a member of Enterobacteriaceae, and each isolate was resistant to at least 6 different antibiotics. Some of the antibiotics that these bacteria were found to be resistant to are not used in horticulture (Boehme and others 2004). Boehme and others (2004) also showed sprouts are likely to be more contaminated than whole produce and have more resistance associated with them. The samples collected were from retail markets, and it was found that much of the microbial load were common transient members of the phyllosphere that could have attached during processing (Boehme and others 2004).

Since the transient species are sometimes members of Enterobacteriaceae, there is the chance for resistance to pass from these bacteria to potential pathogens in the phyllosphere during processing. Under selective pressure, resistant elements could be passed more readily to other members of a community (Atlas and Bartha 1998). The phyllosphere of a plant is a stressful environment due to the rapid change in conditions that occur (Alfano and Collmer 1996). Mercury-resistant plasmids were transferred by conjugation from epiphytic organisms of sugar beet plants to introduced Pseudomonas fluorescens at a frequency of 10−2 to 1 for transconjugants to recipients (Lilley and Bailey 1997). HI plasmids of Enterobacteriaceae have been shown to transfer to Vibrio cholerae and S. typhi at 24 °C and 14 °C with high frequency (Maher and Taylor 1993). The IncHI1 plasmids conjugate better at temperatures between 20 °C and 30 °C, but are not very successful at 37 °C (Taylor and Levine 1980). This suggests that transfer to human pathogens is possible at ambient temperatures.

Transfer of genetic material differs for both inter- and intraspecies transfer. Dionisio and others (2002) demonstrated that the frequency of interspecies transfer by conjugation from Erwinia chrysanthemi (a plant pathogen) to E. coli K12 was similar for intraspecies transfer by E. coli M4, both showing transfer frequencies of 10−1.7 (Dionisio and others 2002). Transfer between species in a community by conjugation might also occur more quickly in the presence of an amplifier species. These are populations of organisms that do not transfer genes but their presence increases the rate of transfer between other organisms (Dionisio and others 2002). Since the phyllosphere of plants has many different members, it is possible that when a transient species such as a human pathogen enters, horizontal gene transfer could occur. This human pathogen could be a recipient to epiphytic enteric bacteria with another epiphyte acting as an amplifier.

Extended spectrum beta-lactamases (ESBL) are able to hydrolyze 3rd generation cephalosporins (Tzouvelekis and others 2000). These cephalosporins are not likely to be used in horticulture bacterial control, but Vimont and others (2002) found evidence of an ESBL from a natural contaminant of produce. It is possible that EPR-1 can be a progenitor for unknown beta-lactamases. Different human pathogens have been reported in the literature that have ESBL type resistance. Yan and others (2000) found 18 ESBLs out of 1210 E. coli isolates in China. Ceftizamide-resistant Klebsiella pneumoniae was identified in hospitals in Boston, Chicago, and New York (Bradford and others 1995).

Streptomycin-resistant elements have been identified in numerous plant pathogens, including Pseudomonas syringae and Xanthomonas campestris. Transposon 5393 (Tn5393) carries a streptomycin-resistant gene in several plant and soil gram-negative bacteria (Sundin and Bender 1995). For Xanthomonas campestris, there is an insertion sequence (IS6100) in Tn5393 that increases expression of the streptomycin-resistant genes (strA-strB) (Sundin 2002). Salmonella enterica isolates also contain Tn5393, but with an insertion sequence (IS1133) that acts in the same way as IS6100 (Sundin 2002). This sequence had only been previously identified in Erwinia amylovora, a plant pathogen (Pezzella and others 2004). Both IS6100 and IS1133 are common for plant pathogenic bacteria, but Pezzella and others (2004) found that the IS1133 sequence was also in a human pathogen. There is some indication from these studies that transfer between human and plant pathogens might have occurred.

Antibiotic-resistant forms of Erwinia amylovora, the causative agent of fire blight in pears and apples, have become a problem in the United States (Nuclo and others 1998). These pathogens became less problematic with the addition of secondary colonizers on the plant. Both Pseudomonas fluorescens and Erwinia herbicola limited the disease of the pears and apples. Thus, competitive bacteria might be used instead of antibiotics to reduce the effects of these plant pathogens (Nuclo and others 1998).

Washing of Edible Plants

Processing of fruits and vegetables results in cut surfaces and a prepackaged nonsterile food (Nguyen-the and Carlin 1994). Packaging implies a level of safety to some consumers that may or may not be achieved. Processing includes washing, trimming, possibly peeling or slicing, and a sanitation step (Carmichael and others 1999). Washing and sanitation steps are critical with produce because this type of food is often eaten raw. Besides the esthetic problems caused by heat treatment of vegetables (texture and color), Madden (1992) suggests that heat treatment of produce disrupts the physical barrier that keeps bacteria from entering, so heat could be detrimental. It is necessary to limit microbial proliferation in these unheated products to enhance shelf life and to ensure safety.

Johnston and others (2005) observed an increase in microbial proliferation from harvest to processing. They found a 10-fold increase in microbial counts for cilantro from harvest samples to boxed and processed samples. The strength of attachment to plant surfaces is a factor in removal of enteric pathogens from a surface. Compared to Salmonella and other E. coli strains, E. coli O157 does not attach as well to the surface of alfalfa sprouts (Barak and others 2002). These authors observed most E. coli O157 were easily removed from the surface of the sprout with 2 washes of 10 mL sterile water. The OmpA protein is an essential component needed for E. coli O157 to bind to alfalfa sprouts. There are genes for fimbriae and adhesion elements that are redundant in the genome of E. coli O157 that makes these less important than the ompA gene (Torres 2005). Barak and others (2005) found that some virulence factors of Salmonella enterica were needed for attachment to alfalfa sprouts. One such virulence factor, RpoS, helps in the initial attachment to the sprout surface and regulates other adhesion proteins.

Cut surfaces and other wounds are sites for proliferation of bacteria (Janisiewski and others 1999b; Dingman 2000; Riordan and others 2000; Brandl and Mandrell 2002). These are also potential sites for biofilm formation that could reduce the effectiveness of sanitation washes because bacteria in biofilms are less affected by disinfecting solutions than free-living bacteria. Another area where sanitation practices will be less successful is if human pathogens internalize or colonize in protected sites. Using a 3-D reconstruction, Beuchat (1999) found E. coli O157:H7 20 μm to 100 μm below the surface. Itoh and others (1998) used contaminated radish seeds and found E. coli O157:H7 inside the tissue of the young leaves. They were not successful at removing the pathogen with HgCl2 for 10 min. They identified E. coli O157 inside hypocotyls and stomata of radishes that were contaminated as seeds. The stomata and trichomes of lettuce are common sites of attachment for E. coli O157. At cut surfaces, E. coli O157 had better attachment than on the surface of intact lettuce tissue. It was also observed that E. coli O157 entered into the tissue at these cut areas of the lettuce (Seo and Frank 1999). Buchanan and others (1999) showed that E. coli O157:H7 can move inside intact apples if they are immersed in a bacterial solution that is colder than the fruit. Bartz and Showalter (1981) saw a similar phenomenon in tomatoes and said internalization was associated with internal air pockets that, when cooled, absorb fluid from the outside.

A produce sanitation step typically involves a solution of 50 ppm to 200 ppm chlorine that may or may not be necessary (Table 2). Chlorine is as successful at removing bacteria as nonchlorinated water, but it is necessary to keep the wash water free of contaminants (Zagory 1999). Chlorinated water does not sterilize produce, but Zhuang and others (1995) suggested using 200 ppm free chlorine in wash water to ensure some reduction of Salmonella Montevideo from tomatoes. Some producers feel that high concentrations of chlorine in the wash water cause esthetic problems with the product (Hurst and Schuler 1992). Beuchat (1999) found deionized water compared with a 200 ppm chlorine solution was equally successful at removing pathogens from the surface of lettuce. Solomon and others (2002a) used a 200 ppm chlorine solution to clean lettuce and found that it did not remove E. coli O157. Among potential reasons for the failure of chlorine washes are protected sites on the plant where these bacteria hide, avoiding contact with the wash (Leben 1988). Raiden and others (2003) determined detergents were as successful as water at removing Salmonella from strawberries and tomatoes. Most research indicates that enteric pathogens are removed from produce surfaces effectively with water, with chlorine, and with common detergents, yet chlorine is often considered mandatory for produce sanitation. Chlorine may help keep the rinse water clean but keep in mind there are other ways to do this.

Table 2—.  Effectiveness of sanitation methods used on fruits and vegetables infected with enteric pathogens
MethodDoseEffectReference
WaterSpray for 1 or 5 minApprox. 2.5 log10 reduction of E. coli O157:H7Beuchat (1999)
Chlorinated water200 ppm spray for 1 or 5 minApprox. 2.5 log10 reduction of E. coli O157:H7Beauchat (1999)
Chlorine dioxide0.62 and 1.24 g/L3.0 to 6.0 log10 reduction of E. coli O157:H7Han and others (2000)
Produce wash5 sprays of 4.4 to 4.8 mL and 30 s rubbingReduced Salmonella approx. 4.6 log10 cfu/mLHarris and others (1999)
Bacteriophage cocktail25 μL to contaminated woundsReduced Salmonella approx. 3 log10 cfu/mLLeverentz and others (2001)

Chlorinated water may only be as effective as water at disinfecting produce, but Han and others (2000) found chlorine dioxide (ClO2) was much more effective than water. The authors found a 3.0 log10 reduction of E. coli O157:H7 when disinfected with 0.62 g/L ClO2 and a 6.5 log10 reduction of E. coli O157:H7 when 1.24 g/L ClO2. New technologies show promise for produce sanitation to ensure product safety. Harris and others (1999) tested an alkaline produce wash with generally regarded as safe (GRAS) components. The produce wash reduced Salmonella 2 to 4 logs more than the sterile water on the surface of tomatoes. Leverentz and others (2001) used Salmonella-specific phage cocktail to reduce the counts of Salmonella Enteriditis on fresh-cut fruit. A 3.5 log reduction was found on honeydew slices when the phage cocktail was used, but no such reduction was noted on apple slices. The ineffectiveness on apples was likely due to a rapid decline in the phage cocktail caused by the low pH of the apple tissue (Leverentz and others 2001).

With current technology and packaging of fresh produce, microorganisms are present at the time of purchase in grocery stores. Farmers and processors must be vigilant in food safety, but so must consumers. Proper storage and cleaning before eating will help prevent foodborne disease associated with fresh produce. The U.S. FDA recommends buying produce that is not bruised or damaged and buying cut produce that is refrigerated or packed in ice. It is also recommended that all produce be washed with water before consumption, even if it is prepackaged. The U.S. FDA suggests scrubbing produce capable of handling it, but does not recommend use of detergents. It is important to realize that produce safety is now a partnership that requires attention from farmers, processors, retailers, and consumers.

Future Research

Since internalization of human pathogens has been identified (Solomon and others 2002a; Cooley and others 2003), the next important step is to determine if the internalized pathogen can travel to other parts of the plant. Kamoun and Kado (1990) report some plant pathogens lose motility once inside plant tissue. Internalized human pathogens have been identified as deep as 20 μm inside lettuce plants (Solomon 2002b). It is not known how these bacteria get to this depth nor if the pathogen moved outside of the plant or inside vascular bundles. Internalization is a very important factor to study in the safety of fresh fruits and vegetables because current washing methods are ineffective at removing internal bacteria. Some effort is needed to reveal the extent of internalization and dissemination of human pathogens in edible plants.

A greater understanding of nutrient use by human pathogens when in the phyllosphere will lead to effective biocontrol agents. Nutrient use by human pathogens in their preferred environment has been well studied, but research is needed to determine proliferation of human pathogens on leaf surfaces. Understanding the limiting nutrients can help determine how well these organisms grow once attached to the plant. When the preferred nutrient source of human pathogens is detected, effective biocontrol methods could be readily developed. Yang and others (2001) found microbes in the phyllosphere that had previously not been identified as epiphytes. If these organisms use the same nutrients as human pathogens, these epiphytes could be used as biocontrol agents. This is why there is potential benefit in understanding nutrient requirements of human pathogens in the phyllosphere.

Direct gene transfer from epiphytic microorganisms to human pathogens in the phyllosphere should not be overlooked. Similar antibiotic resistance elements were found in both human and plant pathogens. Since both are members of Enterobacteriaceae, then transfer becomes more likely. As explained earlier, transfer between and Erwinia and E. coli occurred as frequently as transfer of resistance elements between E. coli species (Dionisio and others 2002). If natural transfer is possible, this acts as another potential pool of resistance elements that affect safety. Human and livestock antibiotic use might still be the major areas of concern for transfer of resistance, but with the large number of microorganisms present in the phyllosphere, the study of the transfer of resistance elements makes sense.

Nutrients and antibiotics need to be studied to assist in the prevention of produce-associated outbreaks, but changes in the growing, harvesting, and processing of fresh fruits and vegetables can have a larger impact on the fruit and vegetable industry. Attention to esthetic quality and safety factors is essential in processing of produce. The current use of a 200 ppm chlorine rinse has been questioned by producers due to esthetic reasons, but food safety experts believe this level is necessary for produce safety. Finding wash procedures that can eliminate bacteria and keep the produce looking healthy is essential in this industry. Attempts should also be made to find biocontrol methods that eliminate human pathogens and prevent any internalization.

Produce-related foodborne illness and safety concerns have grown in the last 2 decades, so research is now beginning to focus on this emerging problem. Future research endeavors have to consider effects of pathogens on both the plants and the humans who eat the plants. A key to safer processing technology is to understand how these foodborne pathogens act on the plant surface. Once the ecology is better understood, effective methods can be designed to either eliminate the hazard preharvest or remove them from the plant, postharvest. The ultimate goal that may be obtainable within the current generation of science is farm-fresh fruits and vegetables that are free of any harm.

Conclusion

In 2004, the FDA instituted an action plan to reduce foodborne disease associated with fresh produce. The action plan calls for improvement in every step of produce production from the farm to the retail store (cfsan.fda.gov/∼dms/prodpla2.html). To help in the improvement in these areas, we must fully understand conditions that affect the proliferation of enteric pathogens (Table 3). Microenvironmental changes can enhance or adversely affect survival and proliferation of these transient organisms. Surface characteristics can determine if enteric pathogens adhere to food; for example, protected sites on a leaf surface are important for survival of an organism. Such natural sites and created sites such as wounds may enhance survival and proliferation because additional nutrients and protection from drastic environmental shifts will help any microorganism not normally present in the phyllosphere. These drastic shifts in the environment are important to all epiphytes. Temperature and UV radiation are factors that affect enteric pathogen survival. Competition between organisms is important in all microbial communities. For example, Cooley and others (2003) found E. coli O157 could grow to as much as 107 cfu/g on Arabidopsis thaliana when no other organisms were present. Some organisms are antagonists of enteric pathogens, others have no effect on survival, and some may enhance proliferation of the enteric pathogens. Many questions remain. Research that provides answers to the many questions about epiphytic fitness of human pathogens on fresh produce will ultimately achieve the goal of improved food safety.

Table 3—.  Factors that affect proliferation and gene transfer in the phyllosphere of edible plants
Positive effects on enteric pathogen proliferation:
 Damaged leaf tissue—mechanical and biological
 Large amounts of leaf exudates
 Neutral pH E. coli O157 can survive in acidic conditions
 Presence of usable water
 High humidity
 Fungal disease and bacterial soft rot
 Motility
 Protected sites
 Aggregates
 Cutting of produce during processing
Negative effect on enteric pathogen proliferation:
 Intact leaf tissue
 Desiccation of the leaf tissue
 Low humidity
 Presence of antagonist organisms
 Lack of flagella
 Chlorine washes with 200 ppm solution
 Refrigeration temperature
Positive effect on gene transfer between epiphytes
 and enteric pathogens:
 Selective pressure
 Aggregate or biofilm incorporation
 Temperature—Some plasmids will transfer better at 25 °C while others will transfer better at higher temperatures

Ancillary