Many studies have examined the survival of Escherichia coli and foodborne pathogens in agricultural soils. The results of these studies can be influenced by various growth conditions and growth media used when preparing cultures for an experiment. The objectives of this study were to (i) determine the growth curves of rifampin (R)–resistant E. coli in three types of growth media containing R: tryptic soy agar (TSA-R); tryptic soy broth (TSB-R); and poultry pellet extract (PPE-R) and (ii) evaluate the influence of growth media on the survival of E. coli in agricultural soil. Poultry pellet extract (PPE) was prepared by filter sterilizing a 1:10 suspension of heat-treated poultry pellets in sterile water. Generic E. coli (TVS 353) acclimated to 80 μg/mL of R was grown in TSA-R, TSB-R, and PPE-R at 3.0 to 3.5 log CFU/mL and incubated at 37°C. Growth curves were determined by quantifying E. coli populations at 0, 4, 8, 16, 24, and 32 h. Soil microcosms were inoculated with E. coli (6.0 log CFU/g) previously cultured in one of the three media types and stored at 25°C, and soil samples were quantified for E. coli on days 0, 1, 3, 7, 14, 28, and 42. Growth curves and survival models were generated by using DMFit and GInaFiT, respectively. E. coli growth rates were 0.88, 0.77, and 0.69 log CFU/mL/h in TSA-R, TSB-R, and PPE-R, respectively. E. coli populations in the stationary phase were greater for cultures grown in TSA-R (9.4 log CFU/mL) and TSB-R (9.1 log CFU/mL) compared with PPE-R (7.9 log CFU/mL). The E. coli populations in the soil remained stable up to 3 days before declining. An approximate 2 log CFU/g decline of E. coli in soil was observed for each culture type between days 3 and 7, after which E. coli populations declined more slowly from days 7 to 42. A biphasic shoulder model was used to evaluate E. coli survival in soils on the basis of growth media. Using standardized culture growth preparation may aid in determining the complex interactions of enteric pathogen survival in soils.
The growth media used influenced lag, logarithmic, and stationary phases of E. coli.
E. coli populations remained above the limit of detection up to 42 days in the soil.
A biphasic shoulder survival model best fit E. coli in soils for all media types.
Growth media did not affect E. coli survival dynamics or duration in soils.
The survival of microorganisms, specifically foodborne pathogens such as Escherichia coli O157:H7 and Salmonella in agricultural soil, has been investigated extensively to better understand the potential for contamination of fresh produce (3, 5, 14, 19–21, 24). Factors such as temperature (18, 24, 26), soil type (3, 11,15, 24, 27), moisture (3, 18, 21, 24), and existing microbial community (1, 11, 13, 25, 26) have all been shown to significantly influence the survival dynamics of E. coli and Salmonella in soil. These factors contribute to the understanding of their survival in soil and aid in developing predictive models to understand the risk of contamination. Understanding the influences these factors have on pathogen survival is extremely valuable. However, few research studies have reported the effect of specific bacterial growth media used to culture cells on the survival of enteric pathogens in soil. For most soil survival studies, cultures are grown for experiments in bacterial nutrient broth media (1, 3, 19, 24), although other experiments have used extracts from soil and/or animal manure to grow bacteria cultures for experiments (19, 20). Some researchers have recommended that when conducting studies evaluating the survival of enteric bacteria in soils amended with biological amendments (manure or litter), the most appropriate media or matrix for culture preparation is growing cells in extracts of soil amendments (10).
Previous research has been conducted to determine the influence of pregrowth conditions on the growth and survival of foodborne bacteria on food surfaces and in low-moisture foods (7, 9, 12). In a study conducted by Harrand et al. (9), pregrowth conditions significantly influenced the growth and survival of Salmonella, E. coli, and Listeria monocytogenes on fresh produce surfaces. Keller et al. (12) found that Salmonella survived significantly longer in peanut butter when grown on solid media compared with broth. Conversely, Salmonella cultures grown on solid and broth mediums and inoculated into talc powder presented no differences in survival when observed over 30 days (7). Although there is information on the influence growth media has on the survival of bacteria on the food surfaces and in low-moisture products, little is known regarding the influence of growth medium on enteric bacterial survival in soil.
The objectives of this study were to (i) determine the growth rate of rifampin (R)–resistant E. coli in three different types of growth media supplemented with R: tryptic soy agar (TSA-R); tryptic soy broth (TSB-R); and poultry pellet extract (PPE-R) and (ii) evaluate the influence of growth media on the survival of R-resistant E. coli in agricultural soil.
MATERIALS AND METHODS
A nonpathogenic, R-resistant E. coli strain (TVS 353) commonly used in field studies as a model for pathogenic E. coli was used in this study (23). Initial cultures were grown by streaking onto TSA (BD, Franklin Lakes, NJ) supplemented with 80 μg/mL R (TSA-R; Acros Organics, Geel, Belgium) and incubated at 37°C for 24 h. Following incubation, three to five colonies were transferred into a 200-mL aliquot of TSB (BD) supplemented with 80 μg/mL R (TSB-R) and incubated at 37°C for 24 h. The overnight culture was then transferred to one of three growth mediums: TSA-R on petri dishes (150 by 15 mm; Thermo Fisher Scientific, Fair Lawn, NJ); 100 mL of TSB-R; or 100 mL of poultry pellet extract supplemented with R (PPE-R). The PPE-R was prepared following a modified method developed by Sharma et al. (20), which formulated a 1:10 suspension of heat-treated poultry pellets (Everlizer, Live Oak, FL) in sterile deionized water. Fifty grams of the heat-treated poultry pellets (Supplemental Table S1) were added into a Whirl-Pak filter bag (Nasco Sampling/Whirl-Pak, Madison, WI) with 450 mL of sterile deionized water. The poultry pellets were soaked in the water for 5 min. After 5 min, the poultry pellets were pressed through a filter, and the liquid removed. The liquid was then vacuum filter sterilized through a 20-μm-pore-size filter (Fisher Scientific).
Growth rate determinations in different media
To adequately determine the growth rate of E. coli TVS 353 in each type of growth media, overnight cultures were diluted in 0.1% buffered peptone water (BPW; Oxoid, Basingstoke, UK) to 107 CFU/mL for the TSB-R and PPE-R and to 106 CFU/mL for the TSA-R. For the TSA-R growth media, 1 mL of the diluted culture was spread onto a TSA-R petri dish (150 by 15 mm; Fisher Scientific) and incubated at 37°C. For the inoculated TSB-R and PPE-R, a single 10 μL loop was transferred into 100 mL of the respective media and incubated at 37°C. Growth rates were determined by analyzing each growth media at 0, 4, 8, 16, 24, and 32 h for E. coli TVS 353 levels following inoculation. E. coli TVS 353 in TSB-R and PPE-R were quantified by removing 1 mL of the culture, serially diluting in BPW, and plating 0.1 mL in duplicate onto TSA-R. For the TSA-R plates (150 by 15 mm), a plate-scraping method previously described by Moussavi et al. (16) was performed. Briefly, 10 mL of BPW was added to the TSA-R plate (150 by 15 mm), the lawn was loosened by using an L-shaped spreader, and the 10 mL was removed from the plate, added to 90 mL of BPW, and shaken to distribute the culture. The culture was serially diluted and plated on TSA-R. Plates were incubated at 37°C for 24 h and enumerated the following day.
E. coli TVS 353 soil inoculation
E. coli TVS 353 was isolated onto TSA-R and incubated for 24 h at 37°C. A single colony was transferred into 10 mL of TSB-R and incubated for 24 h at 37°C. Following incubation, a 10-μL loop of the culture was transferred to 25 mL of TSB-R or PPE-R and incubated at 37°C for 24 h. Likewise, 1 mL of the culture was spread onto a TSA-R petri dish (150 by 15 mm; Fisher Scientific) and incubated at 37°C for 24 h. After incubation, the TSB-R and PPE-R cultures were centrifuged (MR23i; Jouan SA, Saint-Herblain, France) at 5,400 × g for 10 min. The supernatant was discarded, and cells were washed with 50 mL of 0.1% BPW. The plate-scraping method previously described was performed for the TSA-R lawn plates, and the 10 mL was added to 40 mL of 0.1% BPW. To achieve uniform inoculation levels between the treatments (TSA-R, TSB-R, and PPE-R), the TSA-R and TSB-R inocula were diluted in 0.1% BPW to achieve an approximate 8-log CFU/mL concentration.
Loamy sand soil (Table S2) obtained from the University of Florida, North Florida Research and Education Center, was sifted to remove debris, air dried for approximately a week, and covered for consistency between samples and replications. Following drying, 1 kg of soil was weighed into a sample bag. The soil was inoculated with E. coli TVS 353 by using an adapted method previously described by Bardsley et al. (3). Inoculum (50 mL) of E. coli TVS 353 from each of the three culture preparation methods was added to the 1-kg soil sample. The inoculum was homogenized into the soil by massaging, shaking, and massaging again in sample bags in 30-s increments. Bags containing inoculated soil were left open and stored in an incubator at 25°C.
Analysis of soil for E. coli TVS 353 populations
Following inoculation, soil samples were analyzed for E. coli TVS 353 populations on days 0, 1, 3, 7, 14, 28, and 42. Prior to sampling, each bag of inoculated soil was agitated to provide even distribution across the samples, and 25 g was removed by using a sterile scoop. Inoculated soil samples (25 g) were diluted in 225 mL of 0.1% BPW. Samples were homogenized via hand massaging and shaking for 30 s. Samples were serially diluted, and 0.1 mL were plated in duplicate onto TSA-R, incubated at 37°C for 24 h, and E. coli TVS 353 colonies were counted. To achieve the lowest limit of detection (<1 log CFU/g), 1 mL of the 1:10 soil and BPW homogenate was plated across four TSA-R plates in 250-μL aliquots. Control (uninoculated) soil samples were plated during each sampling period to ensure no cross-contamination had occurred.
Statistical analysis and modeling
The experiments were conducted in triplicate, and three separate replications were performed for each growth media and soil combination (n = 9). Growth curve models for E. coli TVS 353 in different media were generated by using DMFit (2). Decline (die-off) rates of E. coli TVS 353 in inoculated soils were calculated with GInaFiT (8). E. coli TVS 353 levels in each growth medium and inoculated soil treatment were compared at each time point and were evaluated by using an analysis of variance table and Tukey's honestly significant difference test at a significance level P ≤ 0.05. All statistical analysis was performed by using JMP Pro 16 statistical software (SAS Institute Inc., Cary, NC).
Initial (0 h) E. coli populations were approximately 3 log CFU/mL in TSB-R and PPE-R media types. In contrast, populations in the TSA-R (3.8 ± 0.3 log CFU/mL) were significantly higher (P < 0.05) than the other two media types (Fig. 1). For all media types, significant increases (P < 0.05) from the initial starting E. coli population levels were observed until the 16-h sampling period. No significant (P > 0.05) increases were observed after the 16-h period. E. coli TVS 353 populations in TSA-R continued to be significantly (P < 0.05) higher than in TSB-R and PPE-R until the 32-h sampling time. No significant (P > 0.05) differences were observed between E. coli TVS 353 populations in TSB-R and PPE-R at the 0- and 4-h sampling periods, but for the rest of the sampling periods, populations in TSB-R were significantly (P < 0.05) higher than those in PPE-R. Growth rates for E. coli were 0.88 ± 0.05, 0.77 ± 0.08, and 0.69 ± 0.05 log CFU/mL/h for TSA-R, TSB-R, and PPE-R, respectively, although no significant difference between the growth rates of the cultures grown in each media type (Table 1) was noted. Cultures grown in TSA-R had a significantly lower lag phase (1.94 ± 0.3 h) compared with cultures grown in TSB-R and PPE-R (2.86 ± 0.53 and 3.04 ± 0.38 h, respectively). The stationary phases were all significantly different between the cultures grown in different media types, with TSA-R cultures having the highest (9.38 ± 0.05 log CFU/mL), followed by TSB-R (9.09 ± 0.06 log CFU/mL) and PPE-R (7.94 ± 0.04 log CFU/mL).
Survival of E. coli in soil
In inoculated soils, E. coli TVS 353 populations were introduced at a level of ca. 6 log CFU/g, regardless of media type used to prepare the inoculum (Fig. 2). The survival of E. coli populations followed a similar trend regardless of the growth media used to prepare the culture. E. coli populations in soils were not significantly (P > 0.05) different on the basis of media type between days 0 and 3. Compared with day 3, a significant (P < 0.05) decrease in E. coli populations was observed by day 7 of approximately 2 log CFU/g for all three media treatments. Compared with day 7, another significant (P < 0.05) decrease in E. coli population was observed by day 14 for both the TSA-R and PPE-R treatments, although a corresponding significant (P < 0.05) decrease for populations in the TSB-R treatment was not observed until day 42. By day 14 to 28, E. coli populations had an approximate 4 log CFU/g decrease across all treatments (Fig. 2). No significant (P > 0.05) differences in E. coli TVS 353 population levels were observed between the growth media treatments for all sampling periods, except for days 1 and 3, when populations in the PPE-R treatment were significantly (P < 0.05) higher than populations in the TSA-R treatment on day 1. The populations in the PPE-R treatment were significantly (P < 0.05) higher than the populations in the TSB-R treatment on day 3.
A total of 10 different survival curve models were run on each of the growth media–dependent E. coli soil survival data sets to identify the model with the best fit across the three different treatments. Of the 10 models, the biphasic shoulder model had the best fit between the three treatments. The resulting models had shoulder lengths that varied slightly between the E. coli populations from varying growth media types, with PPE-R having the highest (3.78 ± 1.67 days) followed by TSA-R (2.99 ± 1.01 days) and TSB-R (2.78 ± 0.62 days; Table 2). The initial or first-order inactivation rate (Kmax1), when population declines were observed between days 3 and 14, and the second-order inactivation rates (Kmax2, tail), after day 14, were similar across each model but had slight variations. The Kmax1 for TSB-R, PPE-R, and TSA-R E. coli populations were 1.74 ± 0.27, 1.51 ± 0.72, and 1.41 ± 0.34 log CFU/g/day, respectively. The Kmax2 for TSA-R, TSB-R, and PPE-R was 0.04 ± 0.03, 0.03 ± 0.02, and 0.02 ± 0.03 log CFU/g/day, respectively (Table 2).
Growth media significantly affected the E. coli TVS 353 growth rate. TSA-R and TSB-R both supported higher E. coli growth rates in media, carrying capacity, and a shorter lag phase compared with PPE-R. Although the lag phases are different between the media types, only one data point was collected between 0 and 8 h, which may not adeqetaely capture the organism's lag phase in the different growth mediums. These similarities may be attributed to available nutrients between the growth mediums, as TSA and TSB are developed for optimal growth conditions of E. coli. In a study conducted by Shah et al. (19), sterile (filtered) heat-treated poultry pellet–amended soil extracts supported the growth of Salmonella Newport incubated at 25°C and had a carrying capacity of ca. 9 log CFU/mL, compared with the carrying capacity of ca. 8 log CFU/mL for E. coli in PPE incubated at 37°C in this study. These differences could be attributed to the organism (Salmonella versus E. coli) or that the addition of soil as a component of the pellet extract allowed Salmonella Newport to achieve a higher population level compared with E. coli TVS 353 in PPE.
The results of this study and other studies using agricultural extracts as a growth medium show the capability of organisms such as E. coli to grow in commonly used soil amendments such as agricultural teas, which may provide them the ability to survive in soils for durations long enough to transfer to fruits and vegetables. The aim of this study was to determine if culturing E. coli TVS 353 in PPE-R allowed it to survive for either longer durations or at higher levels because the nutrients in PPE more closely resemble those found in soils compared with traditional bacteriological media, similar to an agricultural or compost tea. Agricultural teas present the risk of prolonged persistence or even growth of enteric pathogens. The addition of components, such as molasses, has been shown to increase the populations of E. coli O157:H7 and Salmonella in compost or agricultural teas, most likely by providing a readily usable carbohydrate source for gram-negative enteric pathogens to use for growth (6).
E. coli populations in soil did not decline until day 3 and then showed significant declines between days 3 and 14 compared with levels seen between days 0 and 3. A tail was observed between 14 and 42 days when E. coli populations were stable. These results are similar to other studies that examined the survival of E. coli in soil (5, 24). Some studies have found that E. coli can persist in soils for as long as 364 days in soil held under laboratory conditions (24). The current study was performed in a laboratory setting that was not influenced by dynamic environmental factors such as rainfall, wind, or relative humidity.
Although results of the current study did find that growth media slightly impacted the early (days 0 to 3) survival patterns of E. coli TVS 353 in soils, the data do not support that growth media type affected the survival duration of E. coli TVS 353 in soils. The incubation conditions in this experiment and nutrients available for bacterial growth and survival in soils may have overwhelmed any effect that growth medium may have provided to E. coli TVS 353 cells to survive. Previous work has shown that specific soil nutrients (nitrogen) or properties (moisture content) can impact E. coli survival durations in soils (12, 17, 20). In this study, factors such as soil moisture and humidity were not controlled, which could have an influence on the survival and adjust the model presented. Growth media has affected the survival of other foodborne pathogens on or in other food commodities. Salmonella grown on solid media was found to have populations of up to 2 log CFU/g higher in peanut butter 2 weeks following inoculation compared with cultures grown in broth (12). Likewise, Salmonella had as much as a 3 log CFU/g reduction difference between TSA-grown cultures compared with TSB-grown cultures on the surface of peppercorn after 28 days (4). Results similar to the current study were found in Streufert et al. (22). Salmonella cultures grown on nonselective agar plates were significantly more resistant to desiccation compared with cultures grown in broth. Harrand et al. (9) found that conditions for preparing cultures had a significant effect on the growth and survival of Salmonella, L. monocytogenes, and E. coli on food surfaces; strain differences were not a significant factor.
Although this study found that growth media type slightly influenced E. coli survival immediately after introduction to agricultural soils, environmental and soil-specific factors may have influence the survival of E. coli in soils, making this specific phenomenon challenging to quantify. Although a generic E. coli strain was used for this study, the results presented may not reflect the behavior of other E. coli strains such as O157:H7, and further research is needed to understand the role culture preparation plays on the survival of pathogenic strains in soil. The use of a protocol that uses standardized bacterial strains, culture preparation, and application method may aid in determining the complex interactions of enteric pathogen survival in soils to improve produce safety standards in the preharvest environment.
This research was funded by the U.S. Department of Agriculture, National Institute of Food and Agriculture, Specialty Crop Research Initiative (award 2020-51181-32157).
Supplemental material associated with this article can be found online at: https://doi.org/10.4315/JFP-22-082.s1