Survival of Escherichia coli O157:H7 in Soils from Jiangsu Province, China

Escherichia coli O157:H7 (E. coli O157:H7) is recognized as a hazardous microorganism in the environment and for public health. The E. coli O157:H7 survival dynamics were investigated in 12 representative soils from Jiangsu Province, where the largest E. coli O157:H7 infection in China occurred. It was observed that E. coli O157:H7 declined rapidly in acidic soils (pH, 4.57 – 5.14) but slowly in neutral soils (pH, 6.51 – 7.39). The survival dynamics were well described by the Weibull model, with the calculated td value (survival time of the culturable E. coli O157:H7 needed to reach the detection limit of 100 CFU g−1) from 4.57 days in an acidic soil (pH, 4.57) to 34.34 days in a neutral soil (pH, 6.77). Stepwise multiple regression analysis indicated that soil pH and soil organic carbon favored E. coli O157:H7 survival, while a high initial ratio of Gram-negative bacteria phospholipid fatty acids (PLFAs) to Gram-positive bacteria PLFAs, and high content of exchangeable potassium inhibited E. coli O157:H7 survival. Principal component analysis clearly showed that the survival profiles in soils with high pH were different from those with low pH.


Introduction
Applications of animal manure as fertilizers or soil amendments to agricultural soils are routine, world-wide. In the UK, the annual amount of animal manure applied to land was recently estimated at 4.3610 5 tons dry weight [1]. Though animal manure can provide nutrients, a variety of pathogenic bacteria may survive in the manure, which in turn may serve as a primary hazardous material for environmental contamination and as a public health threat [2]. For example, Escherichia coli O157:H7 (E. coli O157:H7), which can cause severe hemorrhagic colitis and haemolytic uraemia in humans, can persist in soil for days to more than 1 year following manure application to land [3]. It was reported that 20 people were infected with E. coli O157:H7 through manurecontaminated soil after camping on a field in Scotland that was previously grazed by sheep [4]. Increasing evidence shows that soil and animal manures are the main transport agents of E. coli O157:H7 to contaminate fresh vegetables, fruits and drinking water [1], [5], [6]. Therefore, it is important to understand the nature of E. coli O157:H7 survival and its infective risk in soil or soil-related (manure) environments.
(2012) proved E. coli O157:H7 survival was affected by indigenous microorganisms in soil [9], [15]. Different E. coli O157:H7 survival rates indicate the different potential risks of the pathogen contamination under various soil environments. Consequently, a better understanding of E. coli O157:H7 survival in soils will help in reducing the potential risk of pathogen contamination and avoiding human infection from the pathogen.
In the present study, experiments were carried out to investigate E. coli O157:H7 survival in 12 soils taken from Jiangsu Province, China. In 1987, researchers firstly detected E. coli O157:H7 from the fecal samples of patients in Jiangsu Province, where the largest E. coli O157:H7 outbreak in China occurred [16], [17]. Later, Xu et al.(1990) found that the biochemical reactions of five strains of E. coli O157:H7, which were isolated from 486 stool specimens of patients with diarrhea in Xuzhou City, Jiangsu Province, were almost identical with those of the well-known E. colt O157:H7 (strain EDL933) [18]. Recently, numerous researchers reported that E. coli O157:H7 (strain EDL933) had been detected in excrement, sewage, foods, and soils from many provinces of China, including Jiangsu Province [16], [19][20][21]. Thus, E. coli O157:H7 (strain EDL933) was selected as a representative strain in this study.
Most E. coli O157:H7 outbreaks occur in summer [17][18][19]. The summer mean temperature in Jiangsu Province is about 26uC [22]. Furthermore, the water content under -33 kPa, indicating the water holding capacity of the soil and representing the highest available water contents in soil, was generally used to simulate field conditions [23]. Hence, our simulation experiments also used incubation conditions of 2561uC and water content under -33 kPa. The aims of this study were to (1) investigate E. coli O157:H7 survival dynamics in soils, (2) identify the relationships between E. coli O157:H7 survival time and soil physicochemical properties and microbial community structure, and (3) understand the possible risks of pathogen contamination to prevent further disease outbreaks from E. coli O157:H7.

Ethics Statement
The samples were not collected from national parks, protected areas or private land. Hence, no specific permission was required to obtain these samples. The sampling did not cause any disturbance to the environment or to protected species at the sampling sites.

Soils
The 12 soils (S1-S12) used in this study were taken from Jiangsu Province, China (32.05 o N -34.70 o N). Three replicates were sampled at each soil site from the surface horizon (0-20 cm) and each sample was a composite of several individual soil cores taken at 5-m interval. After sampling, the 36 individually bulked fresh soil samples were immediately taken to the laboratory in coolers containing ice packs. The samples were then hand-picked to remove discrete plant residues, sieved ,2 mm, homogenized thoroughly, and then stored at 4 uC. According to the protocols of the Agricultural Chemistry Committee of China [24], a subsample of the sieved soil from each sample was air-dried for physical and chemical analyses, including pH, total organic carbon (OC), exchangeable potassium (K), humic acid, fulvic acid, sand, clay and silt. Total nitrogen (TN) was determined following digestion using a Kieldahl apparatus (BÜ CHI Labortechnik AG, Flawil, Switzerland). Total dissolved organic carbon (DOC) and total dissolved nitrogen (TDN) were measured by a Multi N/C TOC analyzer (Analytic Jena AG, Jena, Germany). Exchangeable potassium (K) was extracted with 1 M ammonium acetate and measured by Flame Atomic Absorption Spectrophotometry (Analytik Jena AG, Jena, Germany). The soil water content at -33 kPa was determined using a pressure membrane apparatus (Soil Moisture Equipment Corp, CA,USA) as described by Richards [23].

Incubation experiments
The preparation of E. coli O157:H7 (strain EDL933) and inoculums were described previously by Wang et al. (2013) [32]. E. coli O157:H7 cells in sterile deionized water were inoculated into soils to achieve a cell density of about 10 7 colony-forming units per gram soil oven-dry weight (CFU g 21 ). The incubation experiments were conducted on each of the three soil replicates taken from each of the twelve sampling sites (a total of 36 soil samples). All the inoculated and uninoculated control soil samples were incubated in the dark at 2561uC and kept at soil moisture of -33 kPa.
The inoculated soils were sampled at 0, 0.25, 1, 2, 3, 5, 7, 10, 15, 20, 25, 30 and 35 days after inoculation (DAI), and E. coli O157:H7 was extracted by 0.1% peptone buffer (Lab M, Lancashire, UK). The resulting soil suspension was then subjected to 10-fold serial dilutions and E. coli O157:H7 enumerated at the last three of the serial dilutions. The detection limit of the plating technique is about 100 CFU g 21 . The sampling was stopped after plate counts of zero appeared twice in succession during the incubation.

Statistical analysis
The survival data was converted to log 10 (CFU g 21 ) and then analyzed by the Weibull survival model (Equation 1), as described by Wang et al. (2013) [32] and Ma et al. (2013) [33].
N t represents the number of surviving cells remaining at time t, N 0 is the initial cells of the inoculum population; p is the shape parameter; d is the scale parameter that represents the time needed for the first decimal reduction. The time when N t reaches the detection limit (100 CFU g 21 ) of the culturable E. coli O157:H7, t d , can also be calculated from Equation (1).
In addition, principal component analysis (PCA) of the parameters (p and d) and t d values were performed using the R software vegan package v 2.0-5 [34] to visualize survival patterns of E. coli O157:H7 in the test soils. Stepwise multiple-linear regression analysis was carried out by using SPSS 13.0 for Windows (SPSS Inc., IL, USA) to better understand how soil properties affected the survival of E. coli O157: H7 in soils. Analysis of variance (ANOVA) was also carried out to test the differences at the 5% significant level in the survival parameters (p and d) and t d values among soils by SPSS statistic software.

Survival of E. coli O157:H7 in soils
The E. coli O157:H7 population declined by 1.14 log 10 (CFU g 21 ) within the first day after inoculation in all the test soils. After 1 day, two different survival dynamics of E. coli O157:H7 were observed (Fig. 1). E. coli O157:H7 declined rapidly in the acidic Table 1. Physical, chemical, and biological properties of the soils used in this study.   soils (S1, 4,7,8,9,11) and was not detectable after 7 days. In the neutral soils (S2, 3, 5, 6, 10, 12), E. coli O157:H7 declined much more slowly during 1 -3 days, and then entered a relatively rapid decline period, reaching the detection limit after 30 days.

Modeling of survival data
The statistical measures and parameter values of the fitted Weibull model describing the survival of E. coli O157:H7 in the test soils are presented in Table 2. The E. coli O157:H7 survival dynamics in all soils were accurately described by the model, with R 2 ranging from 0.84 to 0.99 (P , 0.001), while the first decimal reduction time (d) and the shape parameter (p) varied in the test soils (Table 2). Survival times (t d ) of E. coli O157:H7 in the 12 soils were calculated from the Weibull model (Equation 1). The survival times (t d ) varied considerably among the soils, with t d of 6.5061.34 days in acidic soils with pH of 4.57 to 5.14 and significantly longer (32.7961.16 days, P , 0.05) in neutral soils (pH, 6.57-7.39). A long survival time of 32.07 days was detected in soil from Xuzhou (S2), where the biggest infection ever reported in China.

The relationship between soil properties and the survival of E. coli O157:H7
For all soils, stepwise regression analysis showed that soil pH, the ratio of G 2 bacteria PLFAs to G + bacteria PLFAs (G 2 /G + ); exchangeable K and OC were the key factors affecting the survival of E. coli O157:H7. Soil pH (P , 0.001) and OC (P , 0.05) had positive effects on E. coli O157:H7 survival time (t d ). In contrast, a high G 2 /G + ratio, and high exchangeable potassium concentration decreased the survival time. The results also suggested that soil pH and G 2 /G + were the most important factors determining the survival of E. coli O157:H7 in the test soils (Table 3).

Discussion
The results indicated that E. coli O157:H7 survival times (t d ) mainly depended on the initial soil pH. Principal component analysis (PCA) clearly showed a substantial difference in E. coli O157:H7 survival times between acidic and neutral soils (Fig. 2).
In the acidic soils, E. coli O157:H7 could not be detected by 7 days after inoculation (Fig.1). However, in the neutral soils, E. coli O157:H7 declined much slowly than in the acidic soils and survived for up to 38 days. There have been previous reports which linked E. coli O157:H7 survival times (t d ) with soil pH [9], [12], [32], [35], However, our data showed an exceptionally highly statistically significant correlation between these two parameters. It therefore appears that the different physical, chemical and biological properties of acidic and neutral soils resulted in the remarkable differences in E. coli O157:H7 survival.
Stepwise multiple regression analysis indicated that E. coli O157:H7 survival time was significantly and positively correlated with soil pH and OC (Table 3). In addition, the initial G 2 /G + and exchangeable K were negatively correlated with E. coli O157:H7 survival time (t d ).
There may be several reasons why soil pH had the most significant positive correlation with E. coli O157:H7 survival time (t d ). Firstly, most soil microbial species are generally adapted to neutral or slightly alkaline environments and the quantity of soil bacteria increases from pH 4 to 7, and they can readily adapt to changing pH within this range [36], [37]. Secondly, pH can affect the adsorption of bacteria to soil minerals. The adsorption decreased gradually with increasing pH [38], [39]. For example, Zhao et al. (2013) found that more E. coli was absorbed by soil colloids when solution pH decreased from 9 to 4 [38]. Also,  showed that large decreases in the viability of E. coli O157:H7 can be caused by the sorption of E. coli O157:H7 to soil minerals [40]. Therefore, a large amount of E. coli O157:H7 adsorbed to soil minerals would result in a high loss of viability of E. coli O157:H7 in low pH soils. Also, the low biological availability of phosphorus and organic nitrogen and the high toxicity of Al and Mn in the soil with low pH [1], [41][42][43], might indirectly affect E. coli O157:H7 survival and activity in the acidic soils.
Because of competition for nutrients and niche space, as well as predation, the survival of introduced pathogens can be affected by the diverse coexisting populations of the indigenous microorganisms [9], [15]. The stepwise multiple regression analysis showed  that the initial G 2 /G + ratio was negatively correlated with E. coli O157:H7 survival time (t d ) ( Table 3). van Elsas et al. (2012) and Ma et al. (2013) also indicated that survival of the invading E. coli O157:H7 was negatively correlated with soil microbial diversity, due to the competition between native bacterial communities and the introduced species for nutrients and niche space [15], [44]. The G 2 bacteria are known to out-compete G + bacteria for nutrients in soil [45]. Furthermore, the production of bacteriocins to kill closely related species drives the negative interspecies interactions in bacterial systems, which plays an important role in determining the fate of invading bacteria [46]. E. coli O157:H7 belongs to G 2 bacteria, therefore we surmised that G 2 bacterial species in soil would have a stronger antagonism to E. coli O157:H7 than G + bacteria. Another study confirmed that Bacteroidetes, Gammaproteobacteria and Firmicutes, mainly G 2 bacteria, could inhibit E. coli O157:H7 survival due to antagonism [47]. Therefore, the greater direct or indirect antagonism of G 2 bacteria than G + bacteria to E. coli O157:H7 might account for the statistically significant negative correlation between E. coli O157:H7 survival time (t d ) and the initial G 2 /G + ratio (Table 3).
E. coli O157:H7 survival times in the test soils were decreased by increasing exchangeable K. A recent study revealed that t d values decreased significantly (P, 0.05) with increasing electrical conductivity and concentrations of individual soil cations, e.g. K + , Na + , Ca 2+ , and Mg 2+ . This could interfere with ion transport, enzyme activity, and crucial protein synthesis in E. coli O157:H7, and finally result in reduced E. coli O157:H7 survival in soils [33]. In addition, our results showed that soil OC was significantly correlated with E. coli O157:H7 survival time (t d ). Soil organic matter is a major energy source for microorganisms, and can provide carbon sources for the growth and survival of E. coli in soil and water [48]. Furthermore, abundant organic carbon can decrease the competitive pressure between organisms through providing easily available energy sources in soil, and thus possibly enhance the persistence of E. coli O157:H7 [15]. Soil organic carbon also helps to improve soil structure by forming multi-pored aggregates which serves as microbial habitats [49]. Therefore, the abundance of organic carbon in soil can provide more nutrients, water, air and biological niches for E. coli O157:H7 and decrease the competition with indigenous microorganisms, slowing the decline of E. coli O157:H7 [3], [15].

Conclusions
This research has enhanced our understanding of the survival of E. coli O157:H7 in soils. E. coli O157:H7 could survive for 32.7961.16 days in neutral soils, and only 6.561.34 days in acidic soils. Special attention should be paid to the different survival times of E. coli O157:H7 in acidic and neutral soils when evaluating the environmental risk associated with it. The survival of E. coli O157:H7 in soils might relate to the interactions between numerous physical, chemical and biological factors. Our findings suggested that high soil pH and organic carbon could prolong E. coli O157:H7 survival times (t d ), while the initial G 2 /G + ratio and exchangeable K were negatively correlated with survival times. Soil pH governs the fate of E. coli O157:H7 directly and indirectly through the adsorption/desorption of soil minerals, nutrition availability, and metal toxicity. However, the experiment was carried out under laboratory conditions. In order to better understand the survival dynamics of E. coli O157:H7 and provide precise information to assess the possible risk of contamination by this pathogen, further research should be done under natural conditions.  # Survival time to reach detection limit (t d ), the ratio of Gram-negative bacteria phospholipid fatty acids (PLFAs) to Gram-positive bacteria PLFAs (G 2 /G + ); exchangeable potassium (K); total soil organic carbon (OC); correlation is significant at the 0.001 probability level(***). doi:10.1371/journal.pone.0081178.t003  Table 2. A-S: acidic soils; N-S: neutral and slight alkaline soils. doi:10.1371/journal.pone.0081178.g002