Differential E. coli Die-Off Patterns Associated with Agricultural

Copyright © 2006 American Chemical Society ..... Within the slurry microcosms, E. coli declined from 8.5 log CFU g-1 on day 0 to cell levels below de...
0 downloads 0 Views 155KB Size
Environ. Sci. Technol. 2006, 40, 5710-5716

Differential E. coli Die-Off Patterns Associated with Agricultural Matrices D A V I D M . O L I V E R , * ,† PHILIP M. HAYGARTH,‡ CHRISTOPHER D. CLEGG,‡ AND A. LOUISE HEATHWAITE† Centre for Sustainable Water Management, The Lancaster Environment Centre, Lancaster University, Lancaster LA1 4YQ, United Kingdom, and Institute of Grassland and Environmental Research, North Wyke Research Station, Okehampton, Devon EX20 2SB, United Kingdom

The investigation of fecal bacterial die-off in various agricultural and catchment related matrices remains important because of the growing concern over pathogens in agricultural environments and watercourses. The aim of this research was to investigate the die-off of Escherichia coli within cattle manure (both slurry [liquid mix of excrement and urine produced by housed livestock] and feces), soil, and runoff water and to determine if cell numbers would be influenced by the presence of cattle manure within soil and runoff water. E. coli survived better within feces than in slurry; cells within feces declined from 7.5 to 3.3 log CFU g-1 in 76 days. Within slurry, cells fell from 8.5 log CFU g-1 to below levels of detection by day 42. E. coli died off more quickly within manure and slurry than in soil amended with the same fecal material, and declined significantly faster within microcosms when introduced to the soil via sterile water rather than cattle manure. E. coli was found to decline more rapidly within wet (50% moisture w/w), rather than dry (25% moisture w/w), soil. Conversely, in runoff water, die-off of E. coli was increased in the presence of feces. Overall, E. coli die-off was most rapid in water incorporated with cattle manure > unincorporated cattle manure > soil incorporated with cattle manure. The derived die-off characteristics including halflife and decimal reduction times can now provide (i) input for predictive models and (ii) information upon which to consider mitigation strategies associated with both manure and land management.

Introduction The die-off of fecal bacteria derived from agricultural sources is a key issue when considering the risk of contamination of surrounding land and water (1-4). Grassland farming practices in the UK routinely apply animal manure to pasture to provide replenishment of nutrients for plant growth. Additionally, feces are deposited directly on pasture by grazing livestock and this constitutes a potentially direct route by which E. coli, and other fecal bacteria, may enter soil (2). * Corresponding author phone: +44 (0)1524 595808; fax: +44 (0)1524 510217; e-mail: [email protected]. † Centre for Sustainable Water Management, The Lancaster Environment Centre, Lancaster University. ‡ Institute of Grassland and Environmental Research. 5710

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 18, 2006

Some research has already shown that fecal bacteria such as E. coli are not well-adapted to survive in soil (6-8). However, when introduced into the soil and protected within a manure medium such as feces, their rate of die-off may be reduced through provision of favorable micro-sites and available nutrients. The mechanisms by which microorganisms are transported through the environment are poorly understood (9) but, under typical grassland farming regimes, we can assume that E. coli may enter the soil system in association with feces or slurry (liquid mix of excrement and urine produced by housed livestock), or as cells freely suspended in water (following disassociation from livestock manure) (see ref 10). Derivation of E. coli die-off characteristics linked to these different scenarios will provide valuable information for predictive modeling purposes which can help provide a basic understanding of the risk of bacterial contamination of surface waters when coupled with cell transfer models. We know that fecal bacteria delivered to grasslands within cattle manures are unlikely to all remain at the soil surface. This is because (i) cattle may tread feces into the soil, particularly in areas of congregation, such as around water troughs, where feces accumulate in greater abundance (11); (ii) liquid slurry applied to grassland may infiltrate through the soil profile (12), assisted by precipitation; (iii) applications of slurry via shallow injection (shallow, vertical slots, typically about 50 mm deep, cut into the soil by a tine or disc) may deliver cells directly to the soil environment; and (iv) cells may percolate vertically into the soil profile following rainfall or overland flow (see ref 10). The purpose of this study was to determine the effect of the presence of cattle manure (slurry and feces) on the dieoff of E. coli in soil and runoff water and thus obtain a series of die-off rates to assist the parametrization of field models (e.g., refs 13, 14). Fecal bacteria are indicative of fecal contamination but are also used to highlight the possible presence of pathogenic bacteria in environmental samples; both are valid reasons for determining fecal bacteria die-off in the environment. The results presented here describe a set of experiments that examined a range of E. coli die-off curves likely to be associated with conditions typical of grassland farming in the UK. It was hypothesized that incorporating cattle manures with soil would impact the dieoff of E. coli when compared with E. coli introduced to soil free of a fecal substrate. This study also examined the potential to reduce die-off rates of E. coli within collected runoff water following the addition of feces, a nutrient source and natural habitat of E. coli, to water microcosms, when compared with the control condition of cells inoculated into runoff water free of fecal material. This study contributes a series of E. coli decimal reduction times (D-values), which may be directly used in field scale models and identifies practical implications for manure management on the farm.

Materials and Methods The work reported here describes two sets of microcosm experiments that were conducted to establish E. coli die-off within (i) cattle manure (feces and slurry) and soil contaminated with cattle manure, and (ii) runoff water contaminated with cattle feces. Bacterial Strain and Maintenance. In both experiments, an E. coli strain originally isolated from feces collected from cattle (Simmental × Friesian steers) at the Institute of Grassland and Environmental Research (IGER), North Wyke, was used. An overnight culture (100 mL) of E. coli (originally isolated from feces) in Nutrient Broth (Oxoid) was harvested 10.1021/es0603249 CCC: $33.50

 2006 American Chemical Society Published on Web 08/17/2006

TABLE 1. Summary of Microcosm Treatments treatment

soil moisturea %

soil matric potentialb (kPa)

1 slurry only 2 feces only 3 intact cores + E. coli 4 intact cores + E. coli 5 repacked cores + E. coli 6 repacked cores + E. coli 7 repacked cores + E. coli via slurry 8 repacked cores + E. coli via slurry 9 repacked cores + E. coli via feces 10 repacked cores + E. coli via feces

N/A N/A 25 50 25 50 25

N/A N/A -100 0 -100 0 -100

50

0

25

-100

50

0

a Gravimetric soil moisture on a fresh weight basis. b A water potential component resulting from capillary and adsorptive forces. Value of 0 relates to saturated soil. Values derived from water retention curve for a clay soil.

by centrifugation at 2900 g for 30 min using a MSE centaur 2 centrifuge. The pelleted cells were re-suspended in sterile water and washed twice by centrifugation at 2900 g for 30 min then re-suspended once more in sterile water. Sterile water was used because it was the least likely carrier to impact on the soil chemistry for the survival experiments. The culture was then maintained at 20 °C overnight ( feces. Overall, a significant (P e 0.001) difference was observed between E. coli die-off within repacked soil of high and low moisture content after 111 days. Numbers of E. coli declined significantly faster within wet soil cores (10-fold drop over 29 days) than within dry soil cores (10-fold decline over 39 days). However, comparison of individual treatments revealed that soil moisture only made a real difference under some conditions (repacked soil with E. coli inoculated and E. coli incorporated via slurry) and not when E. coli was added to soil via feces. Escherichia coli introduced to the soil cores without a fecal substrate and in the presence of high moisture

FIGURE 2. Linear trend-line associated with E. coli die-off within microcosm treatments. A, intact soil cores + freely suspended cells; B, repacked soil cores + freely suspended cells; C, repacked soil cores with slurry incorporated; D, repacked soil cores with feces incorporated. Solid line data represents dry (25% moisture) treatments, dashed line data represents wet (50%) moisture soil treatments. Values of the four replicate microcosms are shown for each sampling time.

TABLE 2. Linear Decline Rates and Decimal Reduction Times for E. coli within Microcosm Treatments Incubated at 15 °C treatment 3 intact, freely suspended, dry 4 intact, freely suspended, wet 5 repacked, freely suspended, dry 6 repacked, freely suspended, wet 7 repacked, slurry, dry 8 repacked, slurry, wet 9 repacked, feces, dry 10 repacked, feces, wet a

D-values modeled linear decline rate (day-1)a (days)

R2

0.088

26

0.974

0.069

33

0.805

0.076

30

0.994

0.096

24

0.950

0.054 0.094 0.054 0.058

43 25 43 39

0.939 0.987 0.985 0.942

Linear decline rate constant ) (2.303 × Figure 2 slope gradient).

were most susceptible to a rapid decline in population numbers (10-fold decline over 24 days), whereas for those cells associated with a manure carrier and introduced to drier soil the decline in numbers (10-fold decline over 43 days for both slurry and feces) was the slowest. Overall, the results showed a significant (P e 0.001) interaction between soil moisture content and carrier medium. The modeled log-linear decline of E. coli showed the weakest fit with the wet intact soil core treatment (r2 ) 0.81; see Table 2). The decline followed a shallow exponential drop and tended toward a reduction in die-off rate with time. Comparison of the two intact treatments suggested that E. coli died off less quickly within the wet intact cores in the longer term because of the stabilization of die-off in later stages of the experiment. However, this intact wet treatment underwent a rapid cell decline in the initial period after inoculation. The difference in decline of E. coli numbers introduced as a cell suspension within intact dry and repacked dry cores was insignificant (P > 0.05).

Die-Off of E. coli in Drainage Water. Table 3 (Supporting Information) gives background information for the two runoff treatments investigated. Nonlinear regression analysis was used to fit an exponential model to each of the eight die-off curves derived from the four replicated flasks of the two treatments. The exponential model fitted is described by eq 1:

y ) A + BRx ) A + Be-kx

(1)

where k is the exponential rate of decline (d-1), B is the drop in cell numbers between experiment start and finish (log CFU mL-1), A is the final level of bacterial population stability (log CFU mL-1) and R is a curve fitting parameter. The halflife of E. coli (time for initial inoculum to decrease by 50%) within each treatment was also calculated and analyzed statistically. Values of all of these properties are presented in Table 4 (Supporting Information). Figure 3 shows the measured decline of E. coli within the two runoff water treatments. The k value associated with runoff water amended with feces was 0.33 d-1 compared with 0.19 d-1 for the unaltered runoff water. One-way ANOVA confirmed a significant difference (P e 0.001) in k and half-life between treatments. Escherichia coli within the unaltered runoff water had a significantly greater half-life than that associated with fecally contaminated runoff. The average half-life for the unaltered runoff was 3.75 days in comparison with 2.15 days for the water + feces treatment. However, the modeled response (see Table 4, Supporting Information), and inspection of Figure 3, suggested that E. coli within the unaltered runoff reached the lowest level of bacterial population stability (A). Interestingly, water + feces demonstrated the most rapid rate of decline and shortest half-life, yet maintained an E. coli population at a higher concentration once it stabilized (see Table 4, Supporting Information). One-way ANOVA confirmed that there was a significant difference (P < 0.05) in the total bacterial population VOL. 40, NO. 18, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5713

FIGURE 3. Escherichia coli decline within runoff water microcosms incubated at 15 °C. Error bars show 1 SE of logarithmic means. concentrations between treatments. Greater bacterial numbers were recorded in the treatment amended with feces. The concentration of total background bacterial cells (excluding inoculated E. coli) was 5.56 ( 0.05 and 5.88 ( 0.08 log CFU mL-1 for the water and water + feces treatment, respectively.

Discussion The investigation of fecal bacterial die-off remains important because of the growing concern over livestock-derived pathogens in agricultural environments reaching watercourses. Using generic E. coli as a surrogate for pathogenic strains of this bacteria allowed for a determination of a worstcase scenario of pathogenic E. coli persistence in the environment. This is because pathogenic strains of E. coli have been shown to survive less well in soil than nonpathogenic strains (4, 20). Bacteria die-off parameters such as half-life and D-values provide: (i) input for predictive models, and (ii) information upon which to design mitigation strategies associated with both manure and land management. The combination of die-off parameters with experimentally derived cell transfer rates should then provide a basic understanding of the potential risk of contamination of receiving waters within farmed catchments (21). More specifically, such parameters can be used as inputs to algorithms within cells of distributed fecal indicator organism (FIO) models; as components of FIO field scale mechanistic models to constrain uncertainty associated with assumed die-off rate parameters in model predictions (e.g., ref 13); and for the development of predictive and forecasting tools used for catchment and bathing water quality management (e.g., ref 22). While it was found that E. coli survived relatively well in soil, particularly when incorporated with cattle manure, the results showed that E. coli died off faster when combined with runoff water; thus highlighting the differential die-off rates of a fecally derived bacterium in a range of substrate mixes akin to those experienced in the field in the “die-off transect” from animal, through soil and into stream/river water. (Figure 4, Supporting Information). The die-off process is accelerated when E. coli undergoes hydrological mobilization and cell numbers can be reduced rapidly if delivered to receiving waters or stagnant water on soil surfaces. The farm environment may therefore be “engineered” in a way to increase the rate of cell decline. Die-off studies in soil, water, and slurries consistently suggest that outside the host animal, fecal bacteria do not survive well (2, 4, 5, 7, 23) and so farm and land management options may be a viable approach to 5714

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 18, 2006

accelerate cell death by avoiding those scenarios likely to prolong survival (e.g., incorporation of manures with soil) and by increasing non-conducive environments in fields (e.g., allowing cell delivery to farm water, i.e., farm ponds that do not represent an immediate threat to public health). The die-off of E. coli was reduced once cattle manures were combined with soil. Some work has suggested that slurries should be incorporated with soil because such application methods can reduce the potential for bacterial transfer via overland flow pathways (24). While such incorporation may limit the risk of surface water pollution and reduce ammonia emissions, the research presented here suggests that it has the potential to reduce bacterial die-off rates in soil. Similarly, in a field-based study, where UV radiation would also play a role in microbe decline, Avery et al. (25) concluded that E. coli O157 introduced to the soil via subsurface injections of slurry persisted longer than within slurry applied to the soil surface. Combining soil particulates with the cattle manure component may have reduced E. coli die-off by inhibiting the mobility of E. coli and microbial predators, particularly with respect to liquid slurry. Also, soil particles may have provided more protective sites for the attachment of E. coli, perhaps reducing the risk of potential predation (26). The most rapid decline of cells in soil was observed for those introduced via sterile water. This may be because the cells had no fecal substrate from which to obtain nutrients nor colloids with which to associate for protection. Based on the solids content of the two manure treatments it is apparent that the amount of fecal matter matter added to the slurry-modified cores was around 4× lower. This more dilute manure type may explain the more rapid die-off observed in slurry. Logically, E. coli in feces, the natural habitat for E. coli, underwent a less accelerated die-off. The contribution of soil moisture to the die-off of E. coli is unclear in the literature. Some authors have suggested that low soil moisture promotes E. coli die-off (27-28), other studies have found no significant effect of soil moisture content on E. coli die-off for defined moisture content ranges (6, 29). The overall results reported here tend to suggest that moisture content had little effect on E. coli die-off. In some instances the soil of lower moisture content was found to significantly reduce E. coli die-off and one possible explanation is that, for these treatments, soils at lower matric potentials will hold cells within smaller rather than in larger pores and therefore may have provided E. coli with greater protection from larger grazing protozoa (30). However, since the soil + feces treatment may be considered the best model for real conditions (since indigenous E. coli in the fecal matrix were analyzed), and die-off rates were the same in the two moisture conditions, it may be construed that soil moisture had no real ecological significance. Repacked cores offer an insight into processes operating within soil but have limitations when transferred to field situations. In this experiment, no difference in E. coli die-off was observed for dry soil conditions within repacked and intact soil cores but for higher moisture contents the decline of E. coli was noticeably different between intact and repacked soil. Therefore, the extrapolation of bacterial die-off results for repacked soil studies to intact soil systems should be considered carefully as should extrapolation to larger scales (31) as other environmental factors such as temperature will also play a role in E. coli die-off (7, 28). The addition of fresh cattle feces to runoff water increased the rate of E. coli decline within the microcosms and was a reversal of that observed for E. coli die-off in soil. While statistically, the fitted exponential model produced significantly different half-lives for the two treatments, scientific judgment may argue that the two die-off patterns are remarkably similar. The difference in half-life was only 1.6 days and from an ecological perspective, the significance of

this difference is, perhaps, open to debate. Feces are a natural habitat for E. coli and so it could be expected that a fecal source in water would facilitate a reduction in die-off rate of the bacteria either through available nutrient supplies or surface areas. A potential explanation is that when feces were incorporated with soil, the E. coli and associated manure remained aggregated in the microcosm and thus reduced E. coli die-off. In contrast, the addition of feces to water would have resulted in a dilution of the manure material throughout the microcosm. This may have promoted the antagonism of competing microorganisms to the disadvantage of the introduced E. coli. The observed results may also be linked to a more mobile background bacterial population within the water microcosms following a fecal addition. It may be that competition for the limited available nutrients and surface area is likely to be greater within the microcosms with higher cell counts. Similarly, Korhonen and Martikainen (32) have reported the substantial impact that bacterial competition can have on the die-off of introduced E. coli in water. However, the addition of feces to runoff water appeared to support the longer term survival of E. coli despite the more rapid die-off effects suspected to result from high localized total bacterial populations competing with E. coli for nutrients. Thus, in situations where there is increased likelihood for manureassociated E. coli to collect in stagnant “pockets” of surface water, there may exist the potential for prolonged cell survival despite the bacteria being at a relatively low concentration. The results of these experiments suggest that by increasing the storage time of cattle manures, particularly slurry, it is possible to reduce E. coli numbers. Storage is a relatively simple means of reducing the environmental risk of potential pathogens in farm manures provided stores are not reinoculated with fresh slurry. However, upon land application, there is an apparent paradox; while incorporation of manure with soil may limit the potential for bacterial transfer via a number of biological and hydrological vectors (such as surface runoff), it can also allow for the prolonged persistence of cells within agricultural land by reducing die-off rates. While laboratory-based studies cannot “model” environmental patterns, they are vital in understanding what factors are involved in deciding die-off patterns in the field. Our study has helped constrain the uncertainties associated with a dieoff parameter and thus has potentially improved modelpredictive capacity where die-off times are a critical element of field modeling.

Acknowledgments We thank Dan Dhanoa for advice and guidance on statistical matters and the constructive comments provided by the associate editor Dr J. Suflita and three anonymous reviewers that helped improve the quality of this manuscript. This research was funded by a University of Sheffield research studentship to DMO. The Institute of Grassland and Environmental Research (IGER) is supported by the Biotechnology and Biological Sciences Research Council (BBSRC).

Supporting Information Available Table 3 shows the average values for pH, suspended solids and turbidity within the two runoff water microcosm treatments. Table 4 shows the modeled parameters and values associated with runoff water microcosm treatments. Figure 4 shows conceptualized patterns of die-off associated with E. coli during transfer through the farm environment; the “die-off transect”. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Hutchison, M. L.; Walters, L. D.; Avery, S. M.; Moore, A. Decline of zoontoic agents in livestock waste and bedding heaps. J. Appl. Microbiol. 2005, 99, 354-362.

(2) Avery, S. M.; Moore, A.; Hutchison, M. L. Fate of Escherichia coli originating from livestock faeces deposited directly onto pasture. Lett. Appl. Microbiol. 2004, 38, 355-359. (3) Avery, L. M.; Killham, K.; Jones, D. L. Survival of E. coli O157:H7 in organic wastes destined for land application. J. Appl. Microbiol. 2005, 98, 814-822. (4) Mubiru, D. N.; Coyne, M. S.; Grove, J. H. Mortality of Escherichia coli O157:H7 in two soils with different physical and chemical properties. J. Environ. Qual. 2000, 29, 1821-1825. (5) Maule, A. Survival of verocytotoxigenic Escherichia coli O157 in soil, water and on surfaces. J. Appl. Microbiol Symp. Suppl. 2000, 88, 71S-78S. (6) Vinten, A. J. A.; Lewis, D. R.; Fenlon, D. R.; Leach, K. A.; Howard, R.; Svoboda, I.; Ogden, I. Fate of Escherichia coli and Escherichia coli O157 in soils and drainage water following cattle slurry application at 3 sites in southern Scotland. Soil Use Manage. 2002, 18, 223-231. (7) Cools, D.; Merckx, R.; Vlassak, K.; Verhaegen, J. Survival of E. coli and enterococcus spp. derived from pig slurry in soils of different texture. Appl. Soil Ecol. 2001, 17, 53-62. (8) Mawdsley, J. L.; Bardgett, R. D.; Merry, R. J.; Pain, B. F.; Theodorou, M. K. Pathogens in livestock waste, their potential for movement through soil and environmental pollution. Appl. Soil Ecol. 1995, 2, 1-15. (9) Ramos, M. C.; Quinton, J. N.; Tyrell, S. F. Effects of cattle manure on erosion rates and runoff water pollution by fecal coliforms. J. Environ. Manage. 2006, 78, 97-101. (10) Tyrell, S. F.; Quinton, J. N. Overland flow transport of pathogens from agricultural land receiving fecal wastes. J. Appl. Microbiol. 2003, 94, 87S-93S. (11) White, S. L.; Sheffield, R. E.; Washburn, S. P.; King, L. D.; Green, J. T. Spatial and time distribution of dairy cattle excreta in an intensive pasture system. J. Environ. Qual. 2001, 30, 21802187. (12) Sommer, S. G.; Hansen, M. N.; Sogaard, H. T. Infiltration of slurry and ammonia volatilisation. Biosystems Eng. 2004, 88, 359-367. (13) McGechan, M. B; Vinten, A. J. A. Simulating transport of E. coli derived from feces of grazing livestock using the MACRO model. Soil Use Manage. 2004, 20, 195-202. (14) McGechan, M. B.; Vinten, A. J. A. Simulation of transport through soil of E. coli derived from livestock slurry using the MACRO model. Soil Use Manage. 2003, 19, 321-330. (15) USDA. Soil Conservation Service Soil Taxonomy: A Basic System for Soil Classification for Making and Interpreting Soil Surveys; USDA-SCS: New York, 1975. (16) Harrod, T. R. The soils of North Wyke and Rowden; Harpenden: 1981. (17) MET office: UK climate and weather statistics. http://www.metoffice.com/climate/uk/. Access date 1 February 2006. (18) Oliver, D. M.; Heathwaite, A. L.; Haygarth, P. M.; Clegg, C. D. Transfer of Escherichia coli to water from drained and undrained grassland after grazing. J Environ. Qual. 2005, 34, 918-925. (19) Herbert, R. A. Methods for enumerating microorganisms and determining biomass in natural environments. Methods Microbiol. 1990, 22, 1-39. (20) Ogden, I. D.; Fenlon, D. R.; Vinten, A. J. A.; Lewis, D. The fate of Escherichia coli O157 in soil and its potential to contaminate drinking water. Int. J. Food Microbiol. 2001, 66, 111-117. (21) Oliver, D. M.; Clegg, C. D.; Haygarth, P. M.; Heathwaite, A. L. Assessing the potential for pathogen transfer from grassland soils to surface waters. Adv. Agron. 2005, 85, 125-180. (22) Vinten, A. J. A.; Lewis, D. R.; McGechan, M.; Duncan, A.; Aitken, M.; Hill, C.; Crawford, C. Predicting the effect of livestock inputs of E. coli on microbiological compliance of bathing waters. Water Res. 2004, 38, 3215-3224. (23) Winfield, M. D.; Groisman, E. A. Role of host environments in the lifestyles of Salmonella and Escherichia coli. Appl. Environ. Microbiol. 2003, 69, 3687-3694. (24) Quinton, J. N.; Tyrrel, S. F.; Ramos, M. C. The effect of incorporating slurries on the transport of fecal coliforms in overland flow. Soil Use Manage. 2003, 19, 185-186. (25) Avery, L. M.; Hill, P.; Killham, K.; Jones, D. L. Escherichia coli O157 survival following the surface and sub-surface application of human pathogen contaminated organic waste to soil. Soil Biol. Biochem. 2004, 36, 2101-2103. (26) England, L. S.; Lee, H.; Trevors, J. T. Bacterial survival in soil effect of clays and protozoa. Soil Biol. Biochem. 1993, 25, 525531. VOL. 40, NO. 18, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5715

(27) Entry, J. A.; Hubbard, R. K.; Thies, J. E.; Fuhrmann, J. J. The influence of vegetation in riparian filterstrips on coliform bacteria: II. Survival in soils. J. Environ. Qual. 2000, 29, 12151224. (28) Sjogren, R. E. Prolonged survival of an environmental Escherichia coli in laboratory soil microcosms. Water Air Soil Poll. 1994, 75, 389-403. (29) Ritchie, J. M.,; Campbell, G. R.; Shepherd, J.; Beaton, Y.; Jones, D.; Killham, K.; Artz, R. R. E. A stable bioluminescent construct of Escherichia coli O157:H7 for hazard assessments of longterm survival in the environment. Appl. Environ. Microbiol. 2003, 69, 3359-3367. (30) Postma, J.; Hokahin, C. H.; VanVeen, J. A. Role of microniches in protecting introduced Rhizobium leguminosarum Biovar

5716

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 18, 2006

trifolii against competition and predation in soil. Appl. Environ. Microbiol. 1990, 56, 495-502. (31) John, D. E; Rose, J. B. Review of factors affecting microbial survival in groundwater. Env. Sci. Tech. 2005, 39, 7345-7356. (32) Korhonen, L. K.; Martikainen, P. J. Survival of Escherichia coli and Campylobacter jejuni in untreated and filtered lake water. J. Appl. Bacteriol. 1991, 71, 379-382.

Received for review February 13, 2006. Revised manuscript received July 6, 2006. Accepted July 12, 2006. ES0603249