Aims: To investigate interactions between rumen protozoa and Shiga toxin-producing Escherichia coli (STEC) and to ascertain whether it is likely that rumen protozoa act as ruminant hosts for STEC.
Methods and Results: The presence of stx genes in different microbial fractions recovered from cattle and sheep rumen contents and faeces was examined using PCR. In animals shedding faecal STEC, stx genes were not detected in the rumen bacterial or rumen protozoal fractions. Direct interactions between ruminal protozoa and STEC were investigated by in vitro co-incubation. Rumen protozoa did not appear to ingest STEC, a STEC lysogen or non-STEC E. coli populations when co-incubated.
Conclusions: The ruminal environment is unlikely to be a preferred habitat for STEC. Bacterial grazing by rumen protozoa appears to have little, if any, effect on STEC populations.
Significance and Impact of the Study: This study indicates that ruminal protozoa are unlikely to be a major factor in the survival of STEC in ruminants. They appear as neither a host that protects STEC from the ruminal environment nor a predator that might reduce STEC numbers.
Shiga-toxin producing Escherichia coli (STEC), causative organisms in outbreaks of food-borne illness, are widely distributed in domestic ruminants, which are considered an important route for transmission to humans (Hancock et al. 2001). There is a need to understand the maintenance of STEC populations in ruminants and their survival in the rumen following ingestion and prior to colonizing the intestines. This knowledge may contribute to understanding STEC ecology and to possible strategies to limit their presence in ruminants.
Free-living protozoa are found in a diverse range of habitats, from water and soils, as well as within the gastrointestinal tract (GIT) of a range of vertebrate hosts. These protozoa can act as hosts for other micro-organisms, especially bacteria (Barker and Brown 1994). It has been shown that increased survival of E. coli O157 occurs in association with the soil protist Acanthamoeba polyphaga (Barker et al. 1999). Ciliate protozoa are common inhabitants of the rumen of domestic livestock that may constitute a potential reservoir for the survival of E. coli O157 and other STEC within the rumen and provide a mechanism for survival and transit through the rumen to the lower digestive tract in ruminants. Rumen protozoa are known to engulf a range of bacteria, kill and digest some of them and excrete their digestion products (Coleman 1989). However, other bacteria maintain commensal relationships with ruminal protozoans (Lloyd et al. 1996). To date, little is known about the predatory interactions between ruminal protozoa and bacterial pathogens (Williams and Coleman 1992) and a symbiotic association of STEC with rumen protozoa has not been reported.
The current study was designed to investigate the interactions between rumen protozoa and STEC. A particular aim was to determine the presence or absence of STEC in different microbial fractions from cattle and sheep rumen contents and to assess the in vitro impact of mixed rumen protozoa on STEC.
Materials and methods
Collection of animal faeces and rumen fluid
Samples of faeces and ruminal contents were taken in accordance with the Australian code of practice, from ruminally cannulated sheep and cattle in ethically approved projects, within the Department of Primary Industries and Fisheries (DPI & F, Qld) and the University of Queensland (UQ).
Sheep were maintained in an open pasture situation feeding on subtropical grasses, predominantly Rhodes grass (Chloris gayana) at the DPI&F's Animal Research Institute (Yeerongpilly). High-grade Brahman steers (Bos indicus) were housed at the University of Queensland Mt Cotton Research Station and fed tropical pangola (Digitaria eriantha) hay. Faecal samples were collected from recently voided material or by rectal palpation. Rumen fluid samples were collected by aspiration of ruminal contents through nylon gauze.
Media and reagents
Nutrient agar, nutrient broth and Luria–Bertani (LB) broth (Oxoid) were used as general purpose, nonselective media. Modified E. coli broth (mEC) (Okrend et al. 1990) was used to enrich for STEC (see below). Media were prepared according to the manufacturer's instructions and sterilized by autoclaving at 121°C for 15 min. To select antibiotic-resistant strains of bacteria, agars and broths were supplemented with either nalidixic acid (Nal) (20 mg ml−1 in 0·1 mol l−1 NaOH; Sigma) or streptomycin (Sm) (10 mg ml−1; Sigma). Protozoal media were prepared under anaerobic conditions (Hungate 1969). Salt solutions for the in vitro cultivation of ruminal protozoa were prepared as previously described (Coleman 1958).
Bacterial strains and DNA standards
One known STEC strain (EC 2257 E. coli O111:H- Nalr [stx1 and stx2)], a STEC lysogen [Q358 Smr::φ608 (stx1)] and E. coli K12 [Q358 Smr (no stx)] were maintained and utilized in co-incubations with mixed rumen protozoa. Escherichia coli strains were prepared for inoculation into protozoal cultures by growth at 37°C in LB or nutrient broth, supplemented with nalidixic acid or streptomycin, to a density of 108 CFU ml−1. Boiled cell lysate containing DNA from an E. coli O157:H7 isolate carrying stx1 and stx2 operons was used as a positive control for PCR detection of stx1 and stx2 amplicons.
Rumen ciliate protozoa were cultured in vitro using the methods described by Coleman (1987). Briefly, strained rumen fluid from a rumen cannulated steer was collected by aspiration through a nylon stocking and inoculated, using a needle and syringe, into sealed Hungate tubes, containing 15 ml of anaerobic salt solution. Rumen ciliate cultures were fed daily with dry ground lucerne (10–50 mg) and 0·05–0·25 ml of a wheat starch suspension (1·5% w/v; Sigma). As the protozoal density increased, cultures were transferred to 150 ml serum bottles containing 100 ml of salt solution and the amount of feed supplied daily was increased accordingly. Following the daily feeding, cultures were overlaid with a gas phase of anaerobic CO2 : H2 (95 : 5), sealed tightly with a butyl rubber stopper and incubated at 39°C. Once per week the volume of the cultures was halved and replaced with an equal volume of anaerobic salt solution. This procedure was carried out under a stream of CO2 : H2 (95 : 5). The harvested protozoans were then used in the co-incubation experiments.
Enumeration and identification of protozoa
A 4-ml aliquot was taken from cultures and mixed with 16 ml formal saline for enumeration of protozoa. Concentrations of protozoa were determined using a counting chamber (Hawksley) as described previously (Klieve et al. 1998). The protozoa were classified to the level of genera based on morphological characteristics described by Dehority (1993) which was facilitated by staining with Gram's iodine. Protozoa were examined and enumerated by a combination of light and phase contrast microscopy at a magnification of 100× (Olympus BH-2, Tokyo, Japan), as reported previously (Klieve et al. 1998).
Microbial fractionation of rumen contents from sheep and cattle and enrichment for STEC in faeces
The presence of stx genes in rumen bacterial and protozoal fractions from five sheep and four beef cattle was examined by PCR. Prior to this examination the status of these animals in terms of whether they were actively shedding STEC was determined. Samples of faecal material (ca 10 g) were diluted with sterile mEC (1 : 10), homogenized (Colworth; Stomacher 400 A.J. Seward, London, UK) in stomacher bags for 45 s and enriched for STEC by incubation for 18–24 h at 37°C. Subsequently, DNA was extracted from 1 ml of the enrichment broth and used as a template for PCR to detect stx genes (see below).
From animals shown to be positive for STEC in faeces, the rumen bacterial and protozoal fractions were separated using a modified sucrose cushion method (Ogimoto and Imai 1981). Briefly, rumen fluid samples (100–200 ml) were allowed to settle for 3 h at room temperature when a milky white layer could be seen (putative protozoal layer). One millilitre of this layer was placed in a 15-ml tube containing 4 ml of sucrose solution (30% w/v) and centrifuged at 300 g for 10 min. The supernatant contained the bacterial fraction and the pellet the protozoal fraction. These fractions were washed with 0·9% NaCl solution and centrifuged for 10 min (17 000 g, bacterial fraction or 300 g, protozoal fraction). The supernatants were removed and pelleted cells were stored frozen (−20°C) prior to DNA extraction.
DNA extraction and PCR
DNA was extracted from microbial fractions by physical disruption in a bead-beater based on the procedure described by Hensiek et al. (1992). The extracted DNA was then used as a template for PCR. PCR reactions to amplify stx genes were performed in 25 μl volumes, following the method of Karch and Meyer (1989) and Paton and Paton (1998). DNA fragments were separated by electrophoresis through agarose gels (2%) and visualized by u.v. light following staining with ethidium bromide.
Co-incubation of rumen protozoa and STEC
Cultures of mixed rumen protozoa were harvested from a stock culture for co-incubation experiments. For each co-incubation experiment, 15 μl of STEC at 1 × 108 CFU ml−1 were inoculated into a 15-ml ciliate culture to give a final concentration of 1 × 105 STEC ml−1. Prior to STEC inoculation, a 1-ml aliquot of protozoa culture (the T = 0 sample) was removed and placed in a centrifuge tube on ice. This was performed to arrest the metabolism of the microbes. The mixed cultures were then incubated for a total of 4 h at 37°C. Aliquots (1 ml) of the cultures were removed at 30, 60, 120 and 240 min. Stereomicroscopy (AIS–SZ–M; AIS-OPTICAL, Melbourne, Australia) confirmed that aliquots of the ciliates taken before, during and after the incubation period were healthy and motile.
Three strains of E. coli, O111:H- Nalr (stx1 and stx2), Smr::φ608 (stx1) and Smr (no stx) were incubated with mixed rumen protozoa cultures in separate experiments. Control incubations containing either STEC or protozoa were set up in an identical manner to the co-incubation. The STEC-only control was incubated in a culture supernatant of mixed rumen ciliate culture after the ciliates were removed by centrifugation (1000 g). No ciliates were observed by stereomicroscopy in 100 μl samples of the supernatant.
From samples taken from the mixed cultures, the number of viable E. coli cells was determined by serial dilution and the spread plate technique on nutrient agar. For detection of STEC by PCR, washed fractions of protozoa and bacteria were prepared from aliquots taken from co-incubation experiments. Aliquots were centrifuged at 100–150 g for 3 min and the bacterial fraction (supernatant) was drawn off into a fresh Eppendorf tube, the pellet contained the protozoal fraction. Bacteria were subsequently pelleted by centrifugation at 17 000 g. Both the protozoal fraction and the bacterial fraction were washed (three times each) with caudatum-type salt solution or 0·9% (w/v) NaCl respectively. Following centrifugation the supernatant was discarded and the pelleted protozoa and bacteria were stored frozen (−20°C) prior to DNA extraction and PCR.
Cultivation of rumen ciliates
Mixed rumen ciliate cultures were maintained continuously for 3 months. The cultures consisted predominantly of mixed Entodinium spp., with Epidinium spp. also present (ca 1–5%). Other genera present in the original rumen fluid did not establish in the mixed culture. The number of protozoa was generally maintained at >10 000 ml−1.
In vivo investigations of stx gene presence
Both stx1 and stx2 genes were detected in all five sheep faecal samples. However, stx genes were not detected in either the rumen bacterial or protozoal fractions recovered from these sheep.
stx was detected in all cattle (steer nos 66, 224, 229 and 232) faecal samples (Fig. 1, lanes 1–4). Further characterization demonstrated stx2 in faecal samples from steer nos 224 and 229 while steer nos 66 and 232 were positive for both stx1 and stx2. However, no stx genes were detected in the cattle ruminal microbial fractions (Fig. 1, lanes 5–12).
In vitro investigations
The interactions of STEC and E. coli with mixed rumen protozoa were investigated by co-incubation experiments. Escherichia coli O111:H- Nalr decreased in number by <0·2 log cycle after 30 min when co-incubated with protozoa (Fig. 2). Thereafter the bacterial count remained constant. Escherichia coli O111:H- Nalr; incubated under the same conditions but without the mixed rumen protozoa, also decreased by <0·2 log cycle after 30 min (Fig. 2) and remained constant thereafter. The E. coli K12 stx-phage lysogen, Q358 Smr::φ608 and its isogenic parent strain E. coli K12 Q358 Smr exhibited similar results.
The results of STEC gene amplification using DNA from the bacterial and protozoal fractions sampled from co-incubation of E. coli O111:H- Nalr with mixed rumen protozoa are presented in Fig. 3. Bacterial fractions at all time periods were positive for stx genes (Fig. 3, lanes 6–10). Bacterial fractions of the STEC only incubation were also positive (Fig. 3, lanes 16–21). All protozoal fractions were negative for the presence of stx genes (Fig. 3, lanes 1–5 and 11–15). The results following co-incubation of E. coli K12 Q358 Smr::φ608 and its isogenic parent strain Q35 Smr with ruminal protozoa were essentially the same as those presented for E. coli O111:H- Nalr (with the exception that Q358 Smr did not show stx amplicons).
The results of this study do not support a relationship between STEC and ruminal protozoa.
Using an enrichment technique, sheep and cattle were shown to be actively shedding STEC in faeces. Samples from the ruminal protozoal and bacterial fractions from the same animals were PCR-negative for stx genes. Enrichment was used with faecal samples to ensure that the animals that were used did harbour STEC, even if populations were small. It was not thought necessary to enrich ruminal fractions as these had been concentrated from the original material and for a significant involvement with protozoa their numbers would need to be reasonably high, at least in the protozoal fraction. With all DNA samples that were to be used in stx PCR reactions, a control 16S rRNA gene PCR was used to ensure that amplifiable DNA was present and that inhibitors were not present in the samples. The findings of some researchers differ from those reported by us and it has been suggested that the upper GIT is a primary site of STEC localization and proliferation and that STEC can be consistently isolated from rumen fluid (Brown et al. 1997). However, this conclusion was based on observations from a weaned calf model inoculated with 1010 CFU ml−1 STEC and not from healthy animals with naturally occurring microbial populations. In addition, studies demonstrating STEC localization within the rumen were with young animals (Rasmussen and Casey 2001) and the results may not extrapolate to adult animals, as used in our work. Other reports (Grauke et al. 2002) have indicated that the culture of STEC from the rumen of animals experimentally inoculated with STEC is infrequent. Although STEC pass through the rumen, they may only be transiently localized in this part of the GIT and may only persist in the lower GIT/colon. Laven et al. (2003) surveyed the distribution of E. coli O157 in the digestive tract of cattle under 30 months of age and detected E. coli O157 more frequently in the colon than the rumen. In this study the majority of rumen locations (seven of eight) harbouring E. coli O157 were associated with the rumen wall.
STEC did not appear to be ingested or become attached to cultured rumen protozoa of the genera Entodinium or Epidinium. The initial decrease in number by <0·2 log cycle of E. coli O111:H- Nalr was observed across both co-incubations and control incubations and did not appear to be due to predation by protozoa as cells initially decreased by the same extent with or without the presence of protozoa. In addition to this strain, the STEC lysogen, E. coli Q358 Smr::φ608 and a non-STEC strain of E. coli exhibited similar results. The reduction in viable E. coli cells in the initial 30 min may be a result of rapid addition into an anaerobic environment. After this initial drop in viability, numbers of E. coli remained constant across the diversity of strains used and in both the presence and absence of protozoa.
Although stx was detected in the bacterial fractions of the in vitro co-incubation experiments, stx genes were not detected in the protozoal fraction. These same findings were observed with all E. coli strains co-incubated with rumen protozoa. The stx PCR results confirmed the results of the viable count experiment, both experiments indicating that ingestion of STEC by rumen protozoa was not occurring.
In contrast to our results, Coleman (1964) reported uptake of E. coli by rumen protozoa in vitro, using 14C-labelled bacteria, at rates as high as 200 cells min−1. A major difference with the present study was the density of E. coli cells in the protozoal growth medium. Coleman (1964) used very dense cultures (1 × 108–1010E. coli CFU ml−1) whereas the density of STEC were at least a 1000-fold fewer, at 1 × 105 CFU ml−1, in the current study. As the density of E. coli in ruminal contents is typically between 103 and 105 cells g−1 of rumen contents (Russell et al. 2000), the current study more closely approximates to the lower densities of E. coli likely to be encountered in vivo, than used in the earlier work.
We speculate that rumen protozoa may preferentially ingest rumen bacteria when E. coli are present at densities ≤105 CFU ml−1. When E. coli are at high concentrations, selective grazing of only rumen bacteria by protozoa may not be possible. Further research would be required to elucidate whether this is the case.
Contrasting with the rumen ciliate protozoans, Barker et al. (1999) found that the free-living soil protozoan, Acanthamoeba polyphaga supported the survival and multiplication of E. coli O157:H7 in vitro. The reason for this may be the phylogenetic distance between rumen ciliate protozoa and soil amoebae and the different environments that they inhabit.
The results from the current study tend to suggest that rumen protozoa are unlikely to constitute a reservoir for STEC in the rumen and therefore probably do not contribute to survival of STEC. Rumen protozoa do not appear to ingest STEC either and bacterial grazing by rumen protozoa would not be a factor in reducing STEC numbers in the rumen. The objectives of this study were to determine the association and interactions between rumen protozoa and STEC. Our data provides evidence that there appears to be little, if any, interaction between ruminal protozoa and STEC and we conclude that rumen protozoa are unlikely to play a role in hosting STEC in ruminant animals.
The authors would like to thank Brett Knight and Dr Maree Bowen, DPI&F Qld, Animal Research Institute, Yeerongpilly for all their assistance and donation of time in collection of samples from animals. We thank Dr Patricia (Trish) Desmarchelier of Food Science Australia for fruitful discussions and the STEC cultures used in this study. The financial assistance of the Cooperative Research Centre for Cattle and Beef Quality is gratefully acknowledged.