S. enterica serovar Baildon strain 05x-02123 [2] and serovar Enteritidis strain 99A-23 (California Health Department, July 2005 outbreak) are clinical isolates from tomato-related outbreaks and were used as a 1:1 mixture. These strains were chosen due to their involvement in tomato Salmonellosis outbreaks. Although the strains were not differentiated during population enumeration, both strains were used in this study to ensure inclusion of biological variability which may exist among S. enterica serovars during soil survival or plant contamination and colonization. The plant pathogen, Xanthomonas campestris pathovar vesicatoria was isolated from the cultivated tomato (Lycopersicon esculentum Mill.), diseased with bacterial spot, field grown in 2004 (Davis, CA). All bacteria were grown on Luria-Bertani (LB) media and S. enterica populations were enumerated on Salmonella Shigella (SS) media (Difco/BBL; Sparks, MD). Kanamycin (Sigma, St. Louis, MO) was incorporated into all media at 40 mg/liter. Plasmid pKT-kan, in which a 131 bp nptII promoter fragment from Tn5 was fused to the gfp gene of plasmid pPROBE-KT, is a stable, broad-host-range vector that confers kanamycin resistance and green fluorescent protein (gfp) expression [10]. Plasmid pKT-kan was transformed into both strains of S. enterica; this plasmid has been shown to have no affect on survival and growth of S. enterica [11].

S. enterica was inoculated directly into Supersoil (Rod McLellan Co., San Mateo, CA), an enriched potting soil (total nitrogen 0.14%, available P₂O₅ 0.09%, soluble potash K₂O 0.02%, total iron 0.25%, Canadian sphagnum peat moss, ground fir bark, compost, and sand in a proprietary blend) at a pH of 5.5–6.5. Overnight bacterial streak cultures grown at 37°C were suspended in sterile water with a sterile swab to an OD_{600 nm} of 0.2 (∼10⁸ CFU/ml) and bacteria were mixed 1:1 and diluted to the necessary concentration. Soil (approximately 215 g) was either autoclaved or not (non-sterile), placed in 10.5 cm² pots, and irrigated once with 25 ml of S. enterica suspensions (10³, 10⁵, or 10⁷ CFU/ml). Pots were kept in a controlled-environment growth chamber under a day and night cycle of 12 h, during which the day temperature was 26°C and the night temperature was 18°C. Humidity was constant at 75%. Controls were soil irrigated with sterile water. Tomato seeds (cultivar Moneymaker; Tomato Bob, Hilliard, OH) were surface sanitized with 3% calcium hypochlorite as described previously [12] and soaked in sterile water in Petri plates for 1 h to remove any remaining sanitizer. Seeds were sown in S. enterica contaminated soil 24 h post soil inoculation. Pots were returned to the growth chamber immediately following seed sowing. Pots were irrigated (approximately 25 ml sterile water) every 48 h or 24 h, once tomato plants were four weeks old.

Soil was assayed for S. enterica 24 h post-inoculation and once weekly for long term survival assays. Soil samples were placed in tared 15 ml conical tubes (approximately 3 g samples), weighed, and 10 ml of sterile water was added. The suspension was vortexed on high for 1 min. Serial dilutions of the suspension were plated on SS agar with kanamycin, plates were incubated at 42°C (to select for the growth of S. enterica over indigenous soil bacteria) for 24 h, and S. enterica populations were enumerated. Black colonies were confirmed as the inoculated S. enterica strains by confirmation of gfp expression under UV illumination of the plates.

Plants were removed from the soil whole and soil was gently removed from the roots by shaking. Using a sterile razor blade, tomato plants were cut in two; separating the above the soil (phyllosphere) plant parts from that which was below the soil line (rhizoplane). At the seedling stage (approximately 10 to 13 days-old), plant parts were put into separate tared microfuge tubes, weighed, and 1 ml of sterile water was added. The tubes were vortexed for 1 min, serial dilutions of the suspension were made and aliquots were plated on SS agar. LB broth (with kanamycin) was added to the plant samples and incubated overnight at 37°C, with shaking at 150 rpm. If no colonies grew on the original SS agar-Kan, one microliter loop of the enrichment was streaked on SS agar-Kan and incubated at 42°C for 24 h to confirm the presence of S. enterica in the plant samples.

Soil was inoculated with S. enterica and seeds sown as described above. Plants were grown for 30 days and then the entire plant was cut into approximately 2.5 cm pieces, including the roots. Soil was removed from the roots by shaking. Plant debris was mixed with non-sterile soil at a ratio of 1:4 by weight, placed in pots, and kept in a growth chamber, same conditions as described above. Twenty-four hours or one week later, seeds for the second crop were treated with calcium hypochlorite and sown as described above. The second crop was assayed for S. enterica populations at the seedling and three to five leaf stages. Controls were seeds sown in non-sterile soil with debris from plants grown in soil irrigated with sterile water instead of S. enterica. Pots were irrigated (approximately 25 ml sterile water) every 48 h.

To investigate whether the presence of a bacterial tomato plant pathogen effects the S. enterica population, soil was autoclaved and inoculated with the S. enterica cocktail as described above. An overnight streak culture of X. campestris pv. vesicatoria was suspended in sterile water with a sterile swab to an OD_{600 nm} of 0.2 (∼10⁸ CFU/ml) in 20 ml. Following calcium hypochlorite treatment and 1 h soaking in sterile water, tomato seeds were soaked in the X. campestris pv. vesicatoria suspension for 1 h, continuously shaking at 40 rpm. Seeds were then sown in the S. enterica contaminated soil as described above. Populations of X. campestris pv. vesicatoria on seed were determined immediately proceeding planting and 24 and 48 h post sowing. To enumerate the bacteria, individual seeds were placed in 1 ml of sterile water and vortexed for 1 min, serial dilutions of the suspension were made, aliquots were plated on LB agar, and plates were incubated overnight at 28C. Seeds not inoculated with X. campestris pv. vesicatoria were soaked in sterile water and otherwise treated similarly to the others.

Tomato plants were assayed for S. enterica populations at the seedling stage, three to five leaf stage (approximately 25 to 28 days-old), and pre-bloom (approximately 35 to 42 days-old). Plants were sampled as described above with the exception of three to five leaf and pre-bloom stage plant parts were placed into tared 50 ml conical tubes, weighed, and 10 ml of sterile water was added.

For each study, pots were seeded with six to eight seeds per pot. For the long term survival assays, two pots were seeded for each week (6 time periods) starting one week after soil contamination. At the seedling stage, all plants were sampled per time period. The experiment was repeated twice. For the plant debris studies, five pots were seeded for each fallow period. Approximately, half of the plants were sampled at each plant growth stage, seedling and 3–5 leaf stage. The experiment was repeated twice. For the plant pathogen studies, three pots were used for each S. enterica inoculum level and with or without plant pathogen seed inoculation. Approximately half of the plants were sampled at each plant growth stage. The experiment was repeated twice. To determine whether the average populations or incidence of S. enterica differed between treatments or over time, populations were log transformed and unpaired t-tests with a Welch correction were performed using GraphPad Instat (version 3.06, GraphPad Software, Inc., San Diego, CA).

To test the longevity of the S. enterica population in soil and whether these cells could contaminate plants, soil was irrigated with S. enterica contaminated water, seeds were sown at weekly intervals, and plants were examined for S. enterica. S. enterica was able to survive in fallow soil and remained capable of plant attachment and contamination at least six weeks following introduction to the soil by irrigation (Fig. 1). S. enterica soil populations declined for three weeks, stabilized for week three to five, and declined again for the sixth week. At six weeks fallow, S. enterica soil populations were approximately three logs lower than their initial levels. Overall, S. enterica populations in the rhizoplane were significantly smaller with each weekly sampling, except between the fourth and fifth week. S. enterica populations in the phyllosphere were stable for five weeks and smaller each week thereafter. S. enterica was not recovered from control plants.

To determine the role crop debris may play in the persistence of S. enterica in fields used for continuous tomato cropping, a contaminated tomato crop was produced by irrigating soil with S. enterica, direct seeding tomato into the S. enterica contaminated soil, and allowing the crop to develop for 30 days. The crop was mulched and mixed with soil, and tomato seeds (second crop) were sown in the crop debris soil mixture either 24 h or 7 d later. S. enterica was recovered from the phyllosphere and rhizoplane of the second crop when seeds were sown 24 h following plant debris incorporation (Table 1). At the three to five leaf stage, the S. enterica population in the phyllosphere was below the level of enumeration and required enrichment for detection. S. enterica was not recovered from the phyllosphere of the second crop whose seeds were sown in soil which had lain fallow for seven days. S. enterica rhizoplane populations remained below 100 CFU/g for the duration of the study. S. enterica was not recovered from control plants.

To investigate whether the presence of a bacterial tomato plant pathogen affects S. enterica in association with tomato, seeds were inoculated with X. campestris pv. vesicatoria, causal agent of bacterial spot of tomato, and sown in S. enterica contaminated autoclaved soil. Immediately proceeding sowing, X. campestris pv. vesicatoria treated seeds contained 4.8×10⁴±4.2×10³ CFU/seed and decreased to 1.2×10⁴±1.4×10³ CFU/seed after 24 h in soil. At the seedling, 3–5 leaves, and pre-bloom stages, all rhizoplane samples, regardless of the presence of X. campestris pv. vesicatoria, were colonized by S. enterica. There was no statistical difference between the incidence of S. enterica contaminated tomato plants, rhizoplane or phyllosphere, with and without X. campestris pv. vesicatoria plant colonization.

To determine whether S. enterica soil contamination levels could affect the subsequent contamination of the tomato plant, tomato seeds with and without X. campestris pv. vesicatoria were seeded in either soil with an initial high (∼10⁵ CFU/ml) or low (∼10³ CFU/ml) S. enterica inoculum level. There was no difference between the S. enterica populations on plants with or without X. campestris pv. vesicatoria seeded in the high inoculum soil. However, the S. enterica populations on plants seeded in the low inoculum soil differed at each plant development stage between those with or without X. campestris pv. vesicatoria plant colonization (Fig. 2; p≤0.005). At the seedling stage, S. enterica populations on plants not co-colonized by X. campestris pv. vesicatoria, were significantly higher than those of the co-colonized plants. At the three to five leaves and pre-bloom stages, S. enterica populations were significantly higher on those plants which were co-colonized by X. campestris pv. vesicatoria.