Intra-protozoal growth of bacterial pathogens has been associated with increased environmental survival, virulence and resistance to biocides and antibiotics. Using laboratory microcosms we have shown that Escherichia coli 0157 survives and replicates in a common environmental protozoan, Acanthamoeba polyphaga. As protozoa are widely distributed in soils and effluents, they may constitute an important environmental reservoir for transmission of E. coli 0157 and other pathogens.
The importance of protozoa in soil, sewage and water ecosystems has been recognised for several decades and the relevance of predatory protozoa in the control of bacterial populations is widely acknowledged . Yet, the potential role of protozoa as reservoirs of human/animal pathogens has only recently received attention [2–4]. Bacterial pathogens that multiply/survive in amoebae and animal/human cells include: Legionella spp., Listeria, opportunist mycobacteria, and coliforms [2–7]. Intra-protozoal growth has been associated not only with enhanced environmental survival  but also with increased virulence  and greatly increased resistance to biocides  and antibiotics . Legionella spp. have been found in all phases of sewage treatment and population numbers do not significantly decline through the treatment process . Survival could be assisted by internalisation within protozoan hosts.
Vero-cytotoxin-producing Escherichia coli 0157:H7 strains have caused major outbreaks of haemolytic uraemic syndrome in Britain and North America . The infection is highly transmissible and may be acquired after ingestion of less than 100 bacterial cells . The organism readily infects cattle herds and it can be found in soil . Mud contaminated with E. coli 0157 may have been responsible for nine cases of E. coli 0157 infection that occurred at the Glastonbury Festival in Somerset, UK in 1997 . An environmental source may be an important reason for the continuing E. coli 0157 re-infection of cattle herds. The role of protozoa in the environmental survival of E. coli 0157 has not been studied although the organism can grow in cattle manure slurry for several weeks . There is growing concern about the survival of pathogens in sewage sludges disposed of on land now that green laws limit sea dumping. Soils contaminated with organic matter and sewage waste contain greatly increased numbers of protozoa such as acanthamoebae [1, 18]. Thus it is highly likely that E. coli 0157 in soil and slurry will be preyed on by free-living amoebae which could be potential vectors for the spread of this pathogen. Using laboratory microcosms we have studied the potential for a ubiquitous free-living protozoan, Acanthamoeba polyphaga to support the growth of E. coli 0157:H7 (vero-cytotoxin-producing).
2Materials and methods
A well characterised strain of A. polyphaga was obtained from T. Rowbotham, Leeds Public Health Laboratory, UK . It was grown axenically at 35°C in peptone-yeast-extract-glucose broth as shallow monolayers in 75 cm2 tissue culture flasks as described previously [10, 11]. After 3 days incubation at 35°C the exponentially growing amoebae were harvested by centrifugation (400×g, 5 min), washed twice and resuspended in sterile amoeba-saline  to give densities of ca. 105 amoebae ml−1. For some assays, suspensions of washed trophozoites (105 ml−1) were lysed by ultrasonic disintegration and sterilised by passing through a 0.2 μm membrane filter.
E. coli 0157:H7 (vero-cytotoxin-producing) associated with an outbreak of haemolytic uraemic syndrome was obtained from the Centre for Applied Microbiology, Porton Down, UK. Broth-grown (Oxoid No 2, Oxoid Ltd, UK) stationary phase cells of E. coli 0157 were used throughout. The bacteria were washed (3×) and resuspended in amoeba-saline.
Co-cultures with final densities of A. polyphaga trophozoites (ca. 105 cells ml−1) and E. coli (ca. 106 cfu ml−1) were prepared in tissue culture flasks. E. coli microcosms in amoeba-lysate and amoeba-saline were also prepared. Static incubation was carried out in the dark at 4 and 25°C for 35 days. Bacterial survival was assessed weekly by plating ten-fold dilutions onto MacConkey agar (Oxoid Ltd, UK). For some experiments internalisation and survival of E. coli was determined after prior staining of the bacterial cells with a BacLight Live/Dead stain (Molecular Probes, Leiden, The Netherlands) which stains live cells green and cells with damaged membranes, red. Internalisation and replication was unaffected by the staining procedure.
3Results and discussion
Stationary phase E. coli 0157 multiplied in the co-cultures containing Acanthamoeba trophozoites increased in number by >1.0 log cycle after 4 days at 25°C (Fig. 1). Thereafter the bacterial count decreased by <0.5 log cycle and then remained constant. Lysed Acanthamoeba trophozoites provided a substrate for rapid growth of the E. coli 0157: after 4 days incubation at 25°C the colony count increased by ca. 1 log cycle. Over the next 3 days the count decreased by <0.5 log cycle but then remained constant for the remaining 28 days. E. coli 0157 survived in the amoeba-saline microcosms, without added nutrients, at 25°C, showing about a 1 to 2 log cycle reduction in the viable count over the 35 day incubation period (Fig. 1). As an additional control, exponential phase E. coli 0157 was inoculated into amoeba-saline microcosms, without added nutrients, at 25°C. These cells rapidly lost viability and could not be recovered after 10 days incubation (data not shown). Co-cultures were also incubated at 4°C. At this temperature, E. coli 0157 survived and multiplied within Acanthamoeba trophozoites and after 3 days incubation the count had increased by 1 log cycle (data not shown). Bacteria suspended in amoeba-saline without trophozoites decreased by 1 log cycle over the same period.
Phase contrast microscopy of co-cultures incubated at 25°C showed that some trophozoites contained ten or more E. coli cells, within membrane bound vacuoles (Fig. 2). Staining with the BacLight kit revealed internalised, green fluorescent, live bacteria (data not shown). As the cells multiplied the green fluorescence was lost because of dilution of the stain, as was observed in numerous cases. In some vacuoles both green and red fluorescent bacterial cells were observed, although bacteria initially internalised within the amoebae had demonstrated green fluorescence. Red fluorescence indicates cells with damaged membranes which may be the result of the stress induced by the intracellular environment. On several occasions the trophozoites were directly observed expelling internalised live bacteria from food vacuoles into the environment. Digestion of internalised E. coli occurred in some of the acanthamoebae at both 4 and 25°C. When the bacteria were hydrolysed, the STYO 9 component of the BacLight stain was released, staining the trophozoites bright green and no intact bacteria were observed. The trophozoites which permitted growth of E. coli did not absorb the BacLight stain and remained colourless. The BacLight stain also revealed E. coli located in the outer wall of A. polyphaga cysts, very similar to the report for Mycobacterium avium, which has shown to be associated with the outer cyst wall in A. polyphaga.
Our results show that there is a dynamic and mutually beneficial interaction between E. coli 0157 and A. polyphaga trophozoites in laboratory microcosms. In some instances the trophozoites digested the internalised bacteria as a food source but in others the bacteria multiplied within membrane bound vacuoles. Growth of E. coli was enhanced when lysed amoebae were used as the suspending medium. It has been noted that Acanthamoeba lysate can have an inhibitory effect on bacterial growth through the presence of free radicals , but E. coli grew rapidly in our amoeba-lysate. Acanthamoebae are an abundant source of amino acids, enzymes, fatty acids and lipids [21, 22] and clearly provided E. coli with nutrients for growth.
The mechanisms that determine the internalisation and digestion of bacteria by amoebae are complex and not fully understood . However, these mechanisms are likely to be influenced by the physiological characteristics of both the protozoa and the bacterial prey. In E. coli, the rpoS-encoded sigma factor σs is the master regulator in a complex network of stationary-phase-responsive genes involved in the general response (GSR) . The GSR has a major role in resistance to environmental stress conditions and we speculate that this may include the ability of E. coli to evade digestion when internalised within protozoa. Recently, Steinert et al.  have shown that E. coli can survive in A. polyphaga at 33°C, although they did not study an 0157 vero-cytotoxigenic strain. We also found that a non-pathogenic strain of E. coli (K12 W3110) could survive within A. polyphaga (unpublished data). Thus the ability of vero-cytotoxin-producing E. coli to survive in amoebae is not a unique phenomenon. Vero-cytotoxin-producing E. coli can also survive and replicate in bovine mammary cells indicating that the organism is able to resist lysosomal attack . It has been suggested that this intracellular location may provide a reservoir of bacteria for the contamination of workers, equipment and carcasses at the time of slaughter. Further, it has been noted by King et al.  that internalised coliform bacteria can resist digestion by protozoa and survive the effects of chlorination. Our results confirm bacterial survival within protozoa and indicate that protozoa have important potential to act as vehicles for the dissemination of E. coli (or other pathogens) in the environment. Amoeba trophozoites could provide a protective niche for E. coli against adverse conditions, especially so if the organism is able to survive in amoeba cysts as has been shown for Legionella pneumophila and Vibrio cholerae. Bacteria trapped within amoeba cysts could be blown through the air and enhance distribution. Grazing cattle almost certainly ingest protozoa in silage and grass and if the ingested protozoa contain pathogens, this may be a significant route of transmission between and within cattle herds.
A possible link between bacterial evolution in the natural environment and animal pathogenicity is the general stress response, involving RpoS, and its dual influence on environmental survival and bacterial virulence in animals [24, 28, 29]. It is possible that the co-evolution of bacteria and protozoa has equipped some species of bacteria both for environmental survival and for invasion of and survival in animal cells/tissues [2–4]. We suggest that this area of medical ecology requires further investigation to enhance our understanding of environmental pathogens such as E. coli, where protozoa have a potential role in transmission.
This work was supported in part by grants from BBSRC (GR/J96086) and The Wellcome Trust (041505/Z/94/Z).