Interaction of a Free-Living Soil Nematode, Caenorhabditis elegans, with Surrogates of Foodborne Pathogenic Bacteria

Transcription

1 1543 Journal of Food Protection, Vol. 66, No. 9, 2003, Pages Copyright q, International Association for Food Protection Interaction of a Free-Living Soil Nematode, Caenorhabditis elegans, with Surrogates of Foodborne Pathogenic Bacteria GARY L. ANDERSON, 1 KRISHAUN N. CALDWELL, 2 LARRY R. BEUCHAT, 2 AND PHILLIP L. WILLIAMS 1 * 1 Department of Environmental Health Science, University of Georgia, Athens, Georgia ; and 2 Center for Food Safety and Department of Food Science and Technology, University of Georgia, 1109 Experiment Street, Grif n, Georgia , USA MS : Received 13 December 2002/Accepted 28 March 2003 ABSTRACT Free-living nematodes may harbor, protect, and disperse bacteria, including those ingested and passed in viable form in feces. These nematodes are potential vectors for human pathogens and may play a role in foodborne diseases associated with fruits and vegetables eaten raw. In this study, we evaluated the associations between a free-living soil nematode, Caenorhabditis elegans, and Escherichia coli, an avirulent strain of Salmonella Typhimurium, Listeria welshimeri, and Bacillus cereus. On an agar medium, young adult worms quickly moved toward colonies of all four bacteria; over 90% of 3-day-old adult worms entered colonies within 16 min after inoculation. After 48 h, worms moved in and out of colonies of L. welshimeri and B. cereus but remained associated with E. coli and Salmonella Typhimurium colonies for at least 96 h. Young adult worms fed on cells of the four bacteria suspended in K medium. Worms survived and reproduced with the use of nutrients derived from all test bacteria, as determined for eggs laid by second-generation worms after culturing for 96 h. Development was slightly slower for worms fed gram-positive bacteria than for worms fed gram-negative bacteria. Worms that fed for 24 h on bacterial lawns formed on tryptic soy agar dispersed bacteria over a 3-h period when they were transferred to a bacteria-free agar surface. The results of this study suggest that C. elegans and perhaps other free-living nematodes are potential vectors for both gram-positive and gram-negative bacteria, including foodborne pathogens in soil. The agricultural impacts of plant and animal parasitic nematodes have long been recognized and, by virtue of their direct effects on fruit and vegetable production, have been extensively studied. Comparatively little is known regarding the impact of free-living microbivorous nematodes on produce production and safety, although these nematodes are the most abundant and widespread soil mesofauna. They play an important role in microbial degradation processes and nutrient ow in soil (20), which includes the inoculation of new substrates by phoretic transport or through the excretion of viable microorganisms. The colonization of new substrates by microorganisms is favored by the presence of free-living bacterivorous nematodes (10, 18). Free-living nematodes are reportedly attracted to soils receiving agricultural inputs, including organic amendments and fertilizer (16, 29), and thus have the potential to play a role in colonizing agricultural soils with foodborne pathogens originating from animals. Bacterivorous nematodes have been shown to ingest and transmit a broad range of bacterial species, including plant and human pathogenic species. Chantanao and Jensen (12) compiled a list of bacteria ingested and transmitted in viable form by Pristionchus iheritieri (Nematoda: Diplogasterinae), a free-living saprozoic nematode. The list included several species of plant pathogenic bacteria. Jensen (24) demonstrated that Rhabditis spp. are able to ingest and void viable spores of various bacteria that are pathogenic * Author for correspondence. Tel: ; Fax: ; pwilliam@arches.uga.edu. to plants. In addition, rhabditid nematodes fed a diet of Escherichia coli defecated viable cells (2). Nematodes are also able to protect bacteria present in their digestive tracts from chlorination (28). Salmonella ingested by P. iheritieri is reported to have survived the chlorination of wastewater at a treatment facility (33). Thus, nematodes may harbor, protect, and disperse bacteria, including those ingested and passed in viable form in feces. The association of free-living nematodes and various genera of bacteria has been studied. While it is recognized that free-living nematodes avoid certain bacteria (4, 31), it is clear that they do not uniformly avoid foodborne pathogens. Wasilewska and Webster (35) reported that a few species of bacterivorous nematodes can disseminate plant or human pathogenic bacteria. Two human enteric pathogens, Salmonella and Shigella, are reportedly ingested and defecated by free-living saprozoic nematodes (11). Salmonella Typhimurium is known to infect the free-living soil nematode Caenorhabditis elegans. Infected nematodes survive for up to 11 days, thus potentially serving as reservoirs for this foodborne pathogen. Darby and Falkow (13) reported that C. elegans was unaffected by 9 of 10 pathogens and fed on them as they do on nonpathogenic E. coli. From these reports, it appears that free-living nematodes may be important vectors of pathogenic bacteria, including some bacteria capable of causing human disease. Soil is a source of microbial contamination for fruits and vegetables, as evidenced by the isolation of soil-residing pathogenic bacteria from produce (6, 17). Jay (23) stated that the microbiota of soil-grown fruits and vegetables

2 1544 ANDERSON ET AL. may be expected to re ect the microbiota of soils in which they grow. Microbivorous nematodes are among the primary grazers of bacteria in soil (34). The possibility that nematodes may transmit foodborne pathogens to raw produce and thus play a role in the epidemiology of foodborne disease has been given only meager research attention. In a survey undertaken to determine the presence of amoebae and Salmonella in vegetables, Rude et al. (32) recovered nematode eggs and larvae by a nacconol-ether method. The recovery of nematodes from uncooked vegetables indicates that agronomic conditions and marketing practices may be conducive to the survival of nematodes on fresh produce (32). This result also indicates that if free-living nematodes are present on raw produce, they may serve as vehicles for contamination with pathogenic bacteria, either by surface contact or via eggs or voided material from their gastrointestinal tracts. We undertook a study to evaluate the interaction of C. elegans with bacterial surrogates for foodborne pathogens occasionally occurring or persisting in soil. Nematode-bacterium interactions were characterized to determine the propensity of young adult worms to be attracted to bacterial colonies, to compare the feeding and development of young adult worms cultured on this diverse group of bacteria, and to examine the dispersal of bacteria by C. elegans after feeding on monoxenic cultures. MATERIALS AND METHODS Nematode culture. Cultures of N2 (wild-type) strain C. elegans were used. A stock suspension of dauer-larval-stage worms in M9 buffer was kept at 208C and renewed monthly according to procedures described by Donkin and Williams (15). The dauers were used to generate age-synchronous adult worms. Dauers were placed on the surface of K agar (36) in petri plates with an established lawn of E. coli OP50 (7) and incubated at 208C. After 3 days, eggs and worms were collected from the plates and treated for 15 min in Clorox solution (1% NaClO and M NaOH) to isolate eggs (22). Eggs were then placed on K agar with an E. coli OP50 lawn and incubated for 3 days to obtain age-synchronous adults. Selection and culture of bacteria. The bacteria chosen for the investigation were E. coli OP50, an avirulent strain of Salmonella Typhimurium, Listeria welshimeri, and Bacillus cereus. E. coli OP50 is a uracil-de cient strain commonly used as a food source in the culturing of C. elegans. The virulence of Salmonella Typhimurium is attenuated owing to an alteration of the rpos gene caused by natural mutation, rendering it incapable of infecting mice according to standard pathogenicity assays (38). A laboratory stock strain of L. welshimeri and an atoxigenic strain (F3812/ 84) of B. cereus isolated from pasteurized milk were used. These four test bacteria served as surrogates for pathogenic E. coli, Salmonella, L. monocytogenes, and B. cereus. Attraction of C. elegans to bacterial colonies. C. elegans may be either attracted to or repulsed by bacteria, and a number of methods are available to assess its behavior in this regard. We developed a simple method to rapidly screen for the attraction of C. elegans to bacterial colonies. Brie y, 12 replicate colonies of a single bacterial species were established on the periphery of the surface of tryptic soy agar (TSA, ph 7.3; BBL/Difco, Sparks, Md.) in 6-cm-diameter petri plates. Each colony was ca cm in diameter (area, ca cm 2 ), and the total area covered by colonies was ca cm 2. Approximately 22.1 cm 2 of the agar surface was free of bacterial colonies. Nematodes were placed on a colony-free area in the center of the plate. Their movement to bacterial colonies was monitored and documented with video capture and image analysis (NIH Image v1.59, for 16 min. Following tracking, nematodes were monitored for 24 h to determine whether they remained associated with bacteria or tended to disperse into the colony-free area of the agar surface. Ingestion of bacteria by C. elegans. The feeding of C. elegans on bacteria can be quanti ed by monitoring the change in optical density at 570 nm (OD 570 ) of bacterial suspensions inoculated with known numbers of worms (25, 26). A modi cation of a method described by Anderson et al. (3) was used to measure the ingestion of four test bacteria by C. elegans. C. elegans was placed in suspensions of bacteria (ca CFU/ml of K medium, 1 ml per well) in 12-well tissue culture plates. Four wells containing bacterial suspension were inoculated with young adult worms. Two wells containing bacterial suspension not inoculated with worms served as controls. Duplicate determinations of OD 570 values for bacterial suspensions were made with the use of a spectrophotometric plate reader after feeding intervals of 4 and 20 h. The consumption of bacteria by C. elegans was expressed as DOD 570 per 100 worms over de ned periods (net DOD DOD 570 for suspension with worms 2 DOD 570 for suspension without worms). The number of worms in each well was determined from digital images by image analysis (3). The quantitative relationship between OD 570 and the number of bacteria in a suspension varied among test bacteria. For each test bacterium, a calibration curve establishing the relationship between the number of cells in a suspension and OD 570 was constructed. Populations of bacteria were determined by serially diluting suspensions in tryptic soy broth (TSB), surface plating the dilutions on TSA, incubating the plates at 378C for 24 h, and counting colonies. Populations (CFU/ml) were plotted against the OD 570 values for suspensions to construct calibration curves. The dependence of feeding by C. elegans upon the density of bacterial cell suspensions was determined. Young adult worms were fed bacteria in suspensions of various cell densities as determined by the initial OD 570 values. For each trial, four suspensions of bacterial cells were prepared by diluting 24-h TSB cultures in K medium (at culture/k medium ratios of 4:1 to 0.5:1). Two or three trials were carried out for each of the four test bacteria; thus, feeding was determined at 8 to 12 cell densities for each of the test bacteria. For each trial, the feeding of C. elegans (DOD 570 per 100 worms) in suspensions containing each population of bacteria was calculated relative to the average across the test populations (relative feeding 5 feeding at initial OD 570 [c-n]/ average feeding for initial OD 570 values [c-1, c-2, c-3, and c-4]). The initial OD 570 values (c-1, c-2, c-3, and c-4) were determined for four dilutions ranging between 4:1 to 0.5:1 stock solution to K medium; thus, a 2.4-fold range in cell density was tested for each feeding trial. These procedures make it possible to explore the relationship between cell density and feeding even when changes in OD 570 vary between trials. Dispersal of bacteria by C. elegans. Three-day-old adult C. elegans fed for 24 h on 24-h colonies of E. coli OP50, Salmonella Typhimurium, L. welshimeri, or B. cereus on TSA. Worms 1 to 2 cm away from colonies were picked with a sterile wire and transferred to the surface of an uninoculated TSA plate. Nematodes were allowed to move about freely on plates. After 3 h, worms were killed with a hot wire, and plates were incubated for 24 or 48 h at 378C to allow the growth of bacteria. Digital image

3 INTERACTIONS OF C. ELEGANS WITH SOIL MICROFAUNA 1545 analysis was used to quantify the extent of bacterial dispersal on the basis of the number of foci of bacterial growth. A second series of bacterium dispersal experiments was carried out with worms that had fed on bacteria and were then surface sanitized by washing with sodium hypochlorite before being placed on uninoculated TSA. Worms were collected in 10 ml of K medium (37) in a 15-ml test tube and allowed to settle to the bottom. The K medium was removed, and worms were suspended in 0.4 mm sodium hypochlorite at 208C for 5 to 6 min. The hypochlorite solution was removed, and the worms were rapidly washed in K medium (twice, 10 ml per wash) before they were deposited on the surface of TSA. The dispersal of ingested cells, as well as the presumably small number of cells that may have survived hypochlorite treatment on the external surfaces of worms, was determined as described for worms not treated with sodium hypochlorite. RESULTS AND DISCUSSION E. coli OP50 served as a standard against which the responses of C. elegans to other test bacteria were compared. The relationship between C. elegans and Salmonella Typhimurium is unique in that Salmonella is among the few pathogens known to infect nematodes (11). Salmonella Typhimurium has occasionally been isolated from fruits and vegetables and can persist in soil for several months (6). B. cereus, a gram-positive sporeformer, is commonly found in soil. L. welshimeri, another gram-positive bacterium, is not pathogenic to humans, but Listeria species, including the pathogenic L. monocytogenes, are known to reside in soil as saprophytes (5). Attraction of C. elegans to surrogates. Young adult worms placed at the center of a TSA plate on which 24-h colonies of test bacteria had formed migrated into the colonies. Digital images documenting the movement of nematodes into colonies were captured at 1-min intervals for 6 min and at 16 min (Fig. 1). Worms were rapidly and strongly attracted to colonies of all four test bacteria. Over 80% of the worms migrated to colonies of Salmonella Typhimurium, L. welshimeri, and B. cereus within the rst 4 min. After 6 min,.90% of the worms were located in colonies of these bacteria. Eighty-nine percent of the worms were associated with colonies of E. coli OP50 within 6 min. At the end of the 16-min tracking period, 95, 95, 95, and 99% of the nematodes were located in colonies of E. coli, L. welshimeri, Salmonella Typhimurium, and B. cereus, respectively. Andrew and Nicholas (4) reported that C. elegans is strongly attracted to E. coli, Pseudomonas uorescens, and Pseudomonas aeruginosa. These gram-negative bacteria were found to support the growth and reproduction of the nematode. Bacillus megaterium, a gram-positive sporeformer, was found to repel C. elegans. We observed that C. elegans was strongly attracted to 24-h colonies of gram-negative as well as gram-positive bacteria. Behavior and development of C. elegans on surrogates. The association between nematodes and 24-h colonies of test bacteria growing on TSA plates was monitored over a 96-h period following inoculation with 3-day-old C. elegans adults. As observed in the attraction assays, worms became strongly associated with bacterial colonies within FIGURE 1. Attraction of C. elegans to 24-h colonies of bacteria growing on TSA. Twenty to 30 young adult worms were placed on the surface of TSA ca. 0.5 mm away from colonies. The percentage of worms found within colonies was calculated from digital images captured at de ned intervals up to 16 min after inoculation. 16 min after they were deposited on the surface of TSA. The percentage of worms migrating to bacterial colonies was determined at 1-h intervals for 4 h. More than 90% of the worms were found within colonies of E. coli throughout this 4-h period. The association was also strong for the three other test bacteria, with $95% of worms migrating to bacterial colonies. After 20 h, the association between young adult worms and colonies of L. welshimeri and B. cereus changed. For L. welshimeri, 38% of the worms migrated to areas outside the colonies, and over half of the worms were located at the outer edges of colonies. Bacterial colony trails left on TSA by worms leaving colonies were also clearly evident. After 20 h, worms were embedded in colonies of B. cereus, making their enumeration dif cult. No worms were found outside colonies, and approximately half of those within colonies were located at the edges of those colonies. This behavior may re ect worms avoidance of concentrated areas of B. cereus cells. Some wild isolates of C. elegans have been shown to exhibit similar behavior, which led Davis and Avery (14) to suggest that the worms were seeking the lowest concentration of volatile compounds while remaining in the bacterial lawn. After 48 h, C. elegans in the adult and early larval stages of were found on TSA inoculated with test bacteria. On TSA with E. coli OP50 colonies, adults and larvae were

4 ø 1546 ANDERSON ET AL. detected inside bacterial colonies. Most of the adult worms (.80%) were located within Salmonella Typhimurium colonies, whereas worms in the early larval stages were located both inside and outside colonies. For L. welshimeri, both worms in the early larval stages and adult worms were found primarily outside the colonies. Trails of bacterial colonies were formed on TSA as worms left the original colonies. It was unclear whether worms outside the original colonies were associated with less dense new colonies formed by bacteria left on TSA by C. elegans moving over the agar surface. At 48 h, early-larval-stage worms and adult worms remained associated with B. cereus colonies, although they preferred the colony edges over the dense growth in the centers of colonies. Thus, on the basis of observations on the association of worms inside and outside colonies of test bacteria, worms were not as strongly attracted to L. welshimeri and B. cereus as they were to E. coli OP50 and Salmonella Typhimurium. Seventy-two hours after inoculation, viable adult worms and worms in all larval stages were present on TSA plates containing E. coli or Salmonella Typhimurium colonies. Worms were located both inside and outside the original colonies and within secondary colonies resulting from the movement of C. elegans over the 72-h period. On TSA with colonies of L. welshimeri or B. cereus, both adult worms and larvae were located outside the original colonies. Because of the density of colonies, the presence of worms within B. cereus colonies was dif cult to assess. The development of C. elegans in L. welshimeri and B. cereus colonies was slower than that in E. coli or Salmonella Typhimurium colonies. Few late-stage larvae were found after 72 h when C. elegans fed on L. welshimeri or B. cereus. By 96 h, it was not possible to discern the association of worms with original colonies, which were not spatially distinct from secondary colonies formed by cells dispersed on TSA as a result of the movement of worms. Secondgeneration adult worms, eggs, and larvae at various stages were present on plates inoculated with all test bacteria. On TSA inoculated with L. welshimeri or B. cereus, adults tended to be smaller and late-stage larvae were less numerous than those on plates inoculated with E. coli or Salmonella Typhimurium. Second-generation adults were nonetheless as numerous on L. welshimeri or B. cereus plates as on those inoculated with E. coli or Salmonella Typhimurium. For example, in one trial, numbers of second-generation adults per rst-generation adult inoculated were 30, 51, 53, and 40 for E. coli, Salmonella Typhimurium, L. welshimeri, and B. cereus, respectively. The behavioral pattern of attraction followed by altered association with bacterial colonies observed for B. cereus and L. welshimeri may be relevant to the associations of nematodes with pathogenic bacteria. Pujol et al. (31) reported that C. elegans is initially strongly attracted to virulent and attenuated strains of the gram-negative bacterium Serratia marcescens. Over time, however, the virulent strain had a strong tendency to repel C. elegans. The percentage of worms associated with bacteria decreased from 100% at 1 h to.75% at 24 h to,30% at 48 h. This behavior is strikingly similar to that of C. elegans in the presence of colonies of L. welshimeri. Over time, C. elegans also tended to locate more at the periphery of B. cereus colonies, although it did not dissociate from them. The slow development of C. elegans on B. cereus colonies is consistent with reportedly lower reproduction on Bacillus spp., including B. cereus, than on several other bacteria known to attract C. elegans (18). C. elegans remained associated with colonies of E. coli and Salmonella Typhimurium for at least 96 h, a period longer than the worm s life cycle (ca. 3.5 days at 208C when feeding on E. coli). Pujol et al. (31) similarly demonstrated that C. elegans remains associated with E. coli for up to 48 h. The association of C. elegans with Salmonella Typhimurium for 96 h is consistent with a report that 50% of 1-day-old adults survive for 17 days when feeding on Salmonella Typhimurium SL 1344 (1). Differences in the associations of C. elegans with the four test bacteria are consistent with patterns reported for other bacteria, as is the apparent slow development of C. elegans on B. cereus. While C. elegans does not respond similarly to all bacterial species, it is attracted to and can reproduce in association with a diverse group of bacteria known to be present, or even to persist, in soil. Feeding of C. elegans on bacteria in suspension. Previous research suggests that bacterivorous nematodes can use several microbial taxa to sustain growth (34). The fact that we found viable adult worms 48 h after worms were deposited on bacterial lawns of E. coli, Salmonella Typhimurium, L. welshimeri, and B. cereus suggests that C. elegans can use these bacteria as nutrient sources. Rhabditidae, including C. elegans, feed nonselectively (30), and in bacterial suspensions consumption is dependent on nutrient availability (26). We quanti ed the effect of food (bacteria) availability (initial level of bacteria [log 10 CFU/ml]) on feeding (DOD 570 per 100 worms) for E. coli, Salmonella Typhimurium, L. welshimeri, and B. cereus. Feeding in these experiments is expressed on a relative scale. A value of.1 indicates a level of feeding above the average level across all initial cell densities for each bacterium, and a value of,1 indicates a level of feeding below the average level (Fig. 2). Feeding generally increased with increasing populations of E. coli, Salmonella Typhimurium, and L. welshimeri up to an initial OD 570 of ca Above this level, feeding decreased. A similar pattern was observed for B. cereus, but peak feeding (relative value, ca. 1.40) occurred for a smaller bacterial population (OD ). The biphasic nature of the relationship between feeding and initial cell density (initial OD 570 ) was con- rmed by regression analysis. Quadratic equations tted to the data presented in Figure 2 with the use of Microsoft Excel exhibit a maximum feeding level for each bacterial species within the range of cell densities tested (correlation coef cients ranged from 0.67 to 0.78). The calculated cell densities (initial OD 570 ) at which maxima occurred were 0.104, 0.131, 0.167, and for B. cereus, Salmonella Typhimurium, E. coli, and L. welshimeri, respectively. Thus, the feeding behavior of C. elegans tends to be more sensitive to high densities of B. cereus than to concentrated

5 INTERACTIONS OF C. ELEGANS WITH SOIL MICROFAUNA FIGURE 2. Dependence of number of bacteria ingested by C. elegans on level of bacteria in suspension. Relative feeding values (RF) for any initial bacterial population (x) are calculated as RF[x] 5 net DOD570(ODintialx) 4 [net DOD570(OD intialx1) 1 net DOD570(OD intialx2) 1 net DOD570(ODintialx3) 1 net DOD570(ODintialx4)/4], where x1 through x4 are each of four initial concentrations ranging from 0.06 to suspensions of the other test bacteria. Bacillus thuringiensis, a species closely related to B. cereus, produces toxins that are effective against nematodes (21). The observation that C. elegans exhibits a distinct feeding pattern in suspension may re ect reactions caused by toxins similar to or the same as those produced by B. thuringiensis. Relative numbers of bacteria consumed within 3 or 20 h of feeding in suspensions with an initial OD570 of ca are shown in Figure 3. At 3 h, C. elegans fed more heavily on E. coli than on other test bacteria. After 20 h, values for DOD570 per 100 worms were , , , and for E. coli, Salmonella Typhimurium, L. welshimeri, and B. cereus, respectively, as estimated with the use of a standard curve (OD570 versus CFU/ml). Thus, C. elegans continues to consume bacteria for at least 20 h. Venette and Ferris (34) reported that CFU of bacteria per worm per day is suf cient to support the growth of C. elegans unconstrained by food availability. In our study, at 20 h, changes in OD570 (corrected for modest changes observed for bacterial suspensions in the absence of worms) were equivalent to , , and CFU per worm for E. coli, Salmonella Typhimurium, and L. welshimeri, respectively. Worms fed on these bacteria at a level that supported rapid population growth. In contrast, C. elegans consumed and CFU of B. cereus at initial OD570 values 1547 FIGURE 3. Ingestion of E. coli OP50, Salmonella Typhimurium, L. welshimeri, and B. cereus by C. elegans in suspensions of bacteria. Feeding is calculated as net DOD570 5 DOD 570 for suspension with worms 2 DOD570 for suspension without worms. Values for net DOD570 are means 6 standard errors; n $ 6 for each bacterium. of 0.18 and 0.12 (peak of feeding; see Fig. 2), respectively. Thus, the growth of C. elegans would be food-limited on B. cereus, assuming the threshold of CFU per worm given by Venette and Ferris (34) is applicable. Gramnegative bacteria consistently support the growth of Caenorhabditis briggsae, a close relative of C. elegans, whereas gram-positive bacteria do not (8). In addition, Grewal and Wright (19) reported that Bacillus spp. are less attractive to C. elegans than are many other genera, and B. megaterium (4) is toxic to C. elegans. Thus, C. elegans s reduced level of feeding on B. cereus is consistent with other observations regarding this nematode s association with Bacillus spp., including the slow development described above. Dispersal of bacteria by C. elegans. Free-living nematodes disperse bacteria by phoretic transport or through the excretion of viable cells (20). C. elegans was attracted to colonies of E. coli, Salmonella Typhimurium, L. welshimeri, and B. cereus. In dispersal assays, young adult worms fed for 24 h on lawns of bacteria and were then transferred to sterile TSA. Worms removed from bacterial lawns left trails of colonies that developed on TSA from bacteria carried on the surfaces of worms or excreted during movement. Dispersal was quanti ed by counting the number of bacterial foci that developed after 3 h on uninoculated TSA plates. The average number of bacterial foci that developed

6 1548 ANDERSON ET AL. FIGURE 4. Dispersal of viable bacteria by C. elegans after exposure to lawns of E. coli OP50, Salmonella Typhimurium, L. welshimeri, and B. cereus for 24 h at 208C. Dispersal values are means 6 standard errors; n $ 4 for each bacterium. following the dispersal period for each test bacterium is shown in Figure 4. Worms dispersed Salmonella Typhimurium foci per worm over the 3-h dispersal period, a level similar to that determined for E. coli ( foci/ worm). The dispersal level for L. welshimeri ( foci per worm) exceeded those for other bacteria, while the dispersal level for B. cereus ( foci per worm) was lower than those for the other test bacteria. In spite of differences in the feeding and development of C. elegans on different species of bacteria, this nematode is capable of dispersing all four species of bacteria. Exposure of worms that had fed on lawns of bacteria to 0.4 mm sodium hypochlorite for 5 to 6 min followed by two washes in K medium reduced the dispersal of cells on TSA. Dispersal rates were 36, 28, 39, and 9 foci per worm for E. coli, Salmonella Typhimurium, L. welshimeri, and B. cereus, respectively. Reduced phoretic transport following hypochlorite treatment may account for the reduced levels of dispersal. In addition, bacteria may have been excreted through defecation during the 5- to 6-min hypochlorite treatment and during ca. 10 min of washing following the treatment. The saprozoic nematode P. iheritieri defecates viable plant pathogenic bacteria over a 27-h period following feeding (12). For C. elegans, defecation is limited in the rst hour after feeding (27); however, excretion prior to deposition of worms on TSA cannot be precluded as a factor in the reduced level of dispersal following hypochlorite treatment. The persistence of dispersal following hypochlorite treatment is consistent with, although it does not unequivocally prove, the excretion of viable bacterial cells in feces. The dispersal of viable E. coli by C. elegans after hypochlorite treatment was not unexpected, since Adamo and Gealt (2) have reported the defecation of viable E. coli by rhabditid nematodes. Viable Salmonella are also reportedly excreted in the feces of free-living nematodes (11). For terrestrial bacterivorous nematodes, the dispersal of ingested bacteria depends on an attraction to and the consumption of bacteria naturally occurring in or introduced to soil. Literature suggests that bacterivorous nematodes are attracted to and feed on a diverse spectrum of bacterial species, including some bacterial species that are potentially pathogenic to humans. Rhabditid nematodes are nonselective feeders, thus having the greatest potential to disperse pathogens in soils. Observations reported here demonstrate that C. elegans, a rhabditid form, is strongly attracted to and consumes L. welshimeri and B. cereus, both soil inhabitants, as well as E. coli and Salmonella Typhimurium, which can be introduced into soils through the deposition of animal feces, untreated irrigation water, or runoff water from livestock feeding lots. C. elegans was acutely attracted to all of the test bacteria but over time appeared to limit its exposure to L. welshimeri and B. cereus. Behavioral avoidance, even after an initial attraction, could decrease the potential for the vectoring of bacteria by rhabditid nematodes. We also observed that C. elegans consumed less B. cereus, particularly if populations of the bacterium were large, than other test bacteria, including an avirulent strain of Salmonella Typhimurium. Reduced feeding, like the delayed avoidance of bacteria, may also reduce the potential for vectoring. C. elegans dispersed B. cereus, but in lower numbers (fewer foci per worm per unit of time) than it dispersed the other test bacteria. Thus, a reduced level of feeding appears to limit but not preclude the dispersal of B. cereus under laboratory conditions. In contrast to B. cereus, the level of feeding for Salmonella Typhimurium was equal to or above that observed for E. coli. This observation, along with the fact that some Salmonella serotypes may infect the gut of C. elegans (1), suggests the need for studies to determine the role of C. elegans as a vector for the transmission of Salmonella in soil to the surfaces of fruits and vegetables. We have explored the association between C. elegans and Salmonella Poona (9). Worms that had fed on Salmonella Poona were able to transport the pathogen in soil to cantaloupe rind placed on the soil surface. Further studies, particularly those involving soil media amended with manure, are needed to explore factors affecting this process. In addition, foodborne pathogenic microorganisms that are less favored but nonetheless consumed by bacterivorous nematodes should also be investigated. The consumption and dispersal of these microorganisms could be ampli ed by factors imposed by the soil environment. ACKNOWLEDGMENT The nematode strain used in this work was provided by the Caenorhabditis Genetics Center, which is funded by the NIH National Center for Research Resources. REFERENCES Aballay, A., P. Yorgey, and F. M. Ausubel Salmonella Typhimurium proliferates and establishes a persistent infection in the intestine of Caenorhabditis elegans. Curr. Biol. 10: Adamo, J. A., and M. A. Gealt A demonstration of bacterial conjugation within the alimentary canal of Rhabditis nematodes. FEMS Microbiol. Ecol. 20: Anderson, G. L., W. A. Boyd, and P. L. Williams Assessment of sublethal endpoints for toxicity testing with the nematode Caenorhabditis elegans. Environ. Toxicol. Chem. 20:

7 INTERACTIONS OF C. ELEGANS WITH SOIL MICROFAUNA Andrew, P., and W. Nicholas Effect of bacteria on dispersal of Caenorhabditis elegans (Rhabditidae). Nematologica 22: Beuchat, L. R Listeria monocytogenes: incidence on vegetables. Food Control 7: Beuchat, L. R Ecological factors in uencing survival growth of human pathogens on raw fruits and vegetables. Microbes Infect. 4: Brenner, S The genetics of Caenorhabditis elegans. Genetics 77: Briggs, M. P Culture methods for a free-living soil nematode. M.A. thesis. Stanford University, Stanford. Caldwell, K. N., G. L. Anderson, P. L. Williams, and L. R. Beuchat. Attraction of a free-living nematode, Caenorhabditis elegans, to foodborne pathogenic bacteria, and its potential as a vector of Salmonella Poona for preharvest contamination of cantaloupe. Appl. Environ. Microbiol., in press. Cayrol, J. C., J. P. Frankowski, and C. Quiles Recherches sur les possibilities d tilisation des nematodes libres pour inoculer les terres de gobetage au moyen de bacteries inductricesde la fructi cation. Bull. Fed. Natl. Synd. Agric. Cult. Champignons 11: Chang, S., G. Berg, N. A. Clark, and P. W. Kabler Survival, and protection against chlorination, of human enteric pathogens in free-living nematodes isolated from water supplies. Am. J. Trop. Med. Hyg. 9: Chantanao, A., and H. Jensen Saprozoic nematodes as carriers and disseminators of plant pathogenic bacteria. J. Nematol. 1: Darby, C., and S. Falkow. 1 October Personal communication (Worm Breeder s Gazette 16(1):39). Davis, M. W., and L. Avery Personal communication, (Worm Breeder s Gazette 11(5):60). Donkin, S. G., and P. L. Williams In uence of developmental stage, salts and food presence on various end points using Caenorhabditis elegans for aquatic toxicity testing. Environ. Toxicol. Chem. 14: Fu, S., D. C. Coleman, P. F. Hendrix, and D. A. Crossley Responses of trophic groups of soil nematodes to residue application under conventional tillage and no-till regimes. Soil Biol. Biochem. 32: Geldreich, E. E., and R. H. Bornder Fecal contamination of fruits and vegetables during cultivation and processing for market: a review. J. Milk Food Technol. 34: Grewal, P. S In uence of bacteria and temperature on the reproduction of Caenorhabditis elegans (Nematoda: Rhabditidae) infesting mushrooms (Agaricus bisporus). Nematologica 37: Grewal, P. S., and D. J. Wright Migration of Caenorhabditis elegans larvae towards bacteria and the nature of the bacterial stimulus. Fundam. Appl. Nematol. 15: Grif ths, B Microbial-feeding nematodes and protozoa in soil their effects on microbial activity and nitrogen mineralization in decomposition hotspots and the rhizosphere. Plant Soil 164: Hale, K., J. Wei, R. Aroian, and L. Carta Bt toxins are effec tive against phylogenetically diverse nematodes, abstr. 1074, p Abstr. Thirteenth International C. elegans Meeting, University of California, Los Angeles. Hitchcock, D. R., M. C. Black, and P. L. Williams Investigations into using the nematode Caenorhabditis elegans for municipal and industrial wastewater toxicity testing. Arch. Environ. Contam. Toxicol. 33: Jay, J. M Fruit and vegetable products: whole, fresh-cut, and fermented, p In Modern food microbiology, 6th ed. Aspen Publishers, Gaithersburg, Md. Jensen, H Do saprozoic nematodes have a signi cant role in epidemiology of plant diseases. Plant Dis. Rep. 51: Jones, D., and E. P. M.. Candido Feeding is inhibited by sublethal concentrations of toxicants and by stress in the nematode Caenorhabditis elegans: relationship to cellular stress response. Exp. Zool. 284: Jones, D., E. G. Stringham, S. L. Babich, and E. P. M. Candido Transgenic strains of the nematode C. elegans in biomonitoring and toxicology: effects of captan and related compounds of the stress response. Toxicology 109: Liu, D. W. C., and J. H. Thomas. 1 October Personal communication (Worm Breeder s Gazette). Lupi, E., V. Ricci, and D. Burrini Recovery of bacteria in nematodes isolated from a drinking water supply. J. Water Supply Res. Technol. AQUA 44: McSorley, R., and J. Frederick Nematode population uctuations during decomposition of organic amendments. J. Nematol. 31: Nicholas, W. L A study of a species of Acrobeloides [cephalobidae] in laboratory culture. Nematologica 8: Pujol, N., E. Link, L. X. Liu et al A reverse genetic analysis of components of the Toll signaling pathway in Caenorhabditis elegans. Curr. Biol. 11: Rude, R. A., G. J. Jackson, J. W. Bier, T. K. Sawyer, and N. G. Risty Survey of vegetables for nematodes, amoebae, and Salmonella. J. Assoc. Off. Anal. Chem. 67: Smerda, S. M., H. J. Jensen, and A. W. Anderson Escape of salmonellae from chlorination during the ingestion by Pristionchusiheriteri (Nematoda Diplogasterinae). J. Nematol. 3: Venette, R. C., and H. Ferris In uence of bacterial type and density on population growth of bacterial-feeding nematodes. Soil Biol. Biochem. 30: Wasilewska, L., and J. Webster Free-living nematodes as disease factors of man and his crops. Int. J. Environ. Stud. 7: Williams, P. L., and D. B. Dusenbery Using Caenorhabditis elegans to predict mammalian acute lethality to metallic salts. Toxicol. Ind. Health 4: Williams, P. L., and D. B. Dusenbery Aquatic toxicity testing using the nematode Caenorhabditis elegans. Environ. Toxicol. Chem. 9: Wilmes-Riesenberg, M. R., J. W. Foster, and R. Curtis An altered rpos allele contributes to the avirulence of Salmonella Typhimurium LT2. Infect. Immun. 65: