CSIRO Livestock Industries, Long Pocket Laboratories, 120 Meiers Rd, Indooroopilly, QLD 4068, Australia
Correspondence
Denis O. Krause
denis.krause{at}csiro.au
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Abbreviations: ANN, artificial neural network; AR, acid resistant; EHEC, enterohaemorrhagic Escherichia coli; LAB, lactic acid bacteria
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Some researchers (Diez-Gonzalez et al., 1998; Scott et al., 2000
; Keen et al., 1999
) have observed an increase in E. coli when cattle consume high-grain diets, while others have demonstrated that high-grain diets result in decreased shedding (Hovde et al., 1999
). In addition, Diez-Gonzalez et al. (1998)
suggested that this results in acidic conditions in the animal, which allows a population of acid resistant (AR) E. coli to proliferate. AR E. coli would in turn survive the acidic conditions in the human gastric stomach and result in an increased risk of human infection. This hypothesis has been challenged on the basis that pathogenic strains of E. coli are no more AR than generic E. coli (Hovde et al., 1999
).
Elder et al. (2000) analysed a large number of cattle for E. coli serotype O157 : H7 pre- and post-evisceration and found that faecal and carcass EHEC were positively correlated, indicating that a reduction in the pre-slaughter abundance of EHEC would have a positive effect on reducing contamination of processed beef. More recently, a draft risk assessment has indicated that reducing feed-lot prevalence (number of lots infected) by 50 % will reduce processed meat contamination by 43 %, while reducing within feed-lot prevalence will reduce contamination by 25 % (Anonymous, 2001
). These data would suggest that a reduction of E. coli in animal faeces would have a beneficial effect on the sanitary condition of the meat manufacturing process.
LAB are important members of the gastrointestinal tract and are a rich source of bacteriocin-like compounds that show inhibitory activity against a variety of microorganisms, including pathogens (Riley & Gordon, 1999). As a group, LAB are not precisely defined, but are generally considered to be bacteria that make lactate as their major fermentation end product, and in the rumen would include Streptococcus, Lactobacillus and Selenomonas as some of the more dominant species. The ecology of LAB is well understood in relation to ruminal acidosis but few studies have defined the ecology of LAB on the lumen and mucosal surfaces of the small and large intestines of ruminants. A more complete understanding of the ecological interrelationships between LAB and E. coli would give insight into the appropriate design of diets and selection of probiotic cultures. It would also provide information on the ecological validity of using LAB as a means of controlling E. coli; ideally, there should be a negative correlation between the abundance of LAB and E. coli in the gut.
In this manuscript we wished to (1) confirm that high-grain diets result in a high abundance of E. coli along the length of the digestive tract of cattle, as opposed to high-fibre diets, (2) determine the proportion of the E. coli population that could resist an acid shock, (3) examine the abundance and composition of the LAB containing population along the digestive tract and (4) develop a method based on artificial neural networks (ANN) that would allow for the rapid identification of isolates enriched on Rogossa MRS agar. Isolates were evaluated with a combination of 16S rDNA analysis and colony characteristics so that an ANN could be developed to identify LAB and Rogossa MRS isolates (Noble et al., 1997, 2000
; Blackburn et al., 1998
). It has previously been demonstrated that a combination of ANN and morphological characteristics can be used to differentiate bacteria (Blackburn et al., 1998
).
![]() |
METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Animals.
Six adult crossbred HerefordxBrahman cattle were fed a maintenance diet consisting of rhodes grass (Chloris gayana) plus a urea supplement, vitamins and minerals. Animals were randomly allocated to one of two dietary treatments and received a 100 % forage diet (rhodes grass) or were adapted step-wise to a high-grain diet (70 % rolled sorghum; plus 30 % rhodes grass, urea, vitamins and minerals) over a period of 1014 days. No rumen-modifying agents were included in the diets and treatment groups were housed separately. Animals receiving the high-grain diet remained on this ration for at least 10 days before sampling.
Sampling procedure.
To enable efficient processing of samples, two animals were processed at a time (one each from the high-forage and high-grain diets). Feed was removed 12 h before slaughter to reduce the chance of contamination between gut compartments and cattle were killed with a dose of lethobarb (Roche). Lumenal contents (50 g) were taken from the rumen, jejunum, ileum, caecum and faeces, and placed into sterile bottles, kept on ice, and taken to the laboratory for further analysis. Ligation at both ends of a section of the ileum and upper colon prevented contamination.
Lumenal samples were weighed out (10 g) and diluted 10-fold with cold anaerobically prepared diluent. These samples were then homogenized (Bamix) for 1 min and serially diluted (10-fold) in anaerobically prepared diluent. To obtain bacteria adherent to the epithelial tissue a 15 cm length of the ileum or upper colon was washed three times in 10 vols anaerobic diluent so that there were no digesta sticking to the tissue. A blunt knife was then used to scrape away the epithelial tissue, which was subsequently diluted 10-fold with anaerobic diluent, homogenized and serially diluted as described for the lumenal samples.
Ten microlitre droplets (10) were pipetted (Maxipettor, Eppendorf) onto Rogossa MRS agar (10-6, 10-7, 10-8, 10-9 dilutions) or SMAC agar (10-1, 10-2, 10-3, 10-4 dilutions), allowed to dry and then inverted and incubated at 39 °C (Rogossa MRS) or 35 °C (SMAC). After 18 h both the red and white colonies on the SMAC plates were counted but no attempt was made to differentiate between sorbitol fermenters and non-fermenters. Colonies appearing on the MRS agar were counted on two consecutive days; at 18 h, the fast-growing small white colonies were counted and marked; at 36 h, the slower growing colonies were counted and care was taken not to count the colonies counted on the previous day. Representative colonies from the highest dilutions from each sampling site were picked into MRS broth, grown up and stored for later identification.
Testing acid shock.
To test for acid resistance, digesta homogenates were serially diluted in anaerobically prepared diluent (pH 2). The samples were allowed to incubate at room temperature for 1 h before being plated out (see above) onto SMAC. Plates were incubated aerobically at 35 °C for 18 h at which time the red and white colonies were counted.
Characterization of Rogossa MRS isolates.
Presumptive identification of Rogossa MRS isolates was largely based on an accurate description of colony morphology. Each colony was carefully noted for size, colour and morphology. A complete explanation of colony description and how it related to each species is given in the footnotes of Table 1. Rogossa MRS picks were amplified with universal primers 27f and 342r for 16S rDNA (Lane, 1991
). Each 20 µl PCR reaction included 16·3 µl H2O, 0·4 µl dNTP (10 mM), 0·2 µl MgCl2 (250 mM), 2·0 µl reaction buffer (10x), 0·2 µl each PCR primer (10 pmol), 1 U Taq polymerase (Promega) and 0·5 µl sample (1/100 dilution of culture). Thermal cycling conditions were: one cycle of denaturation (95 °C, 5 min), annealing (1 min; 50 °C) and extension (72 °C, 30 s); then 30 cycles of denaturation (95 °C, 1 min), annealing (1 min; 50 °C) and extension (72 °C, 30 s); and one final cycle of denaturation (95 °C, 1 min), annealing (1 min; 50 °C) and extension (72 °C, 10 min). Ten microlitres of the PCR product was doubly digested with DdeI and AluI at 39 °C and run on a 15 % acrylamide gel containing 1 ng ethidium bromide ml-1. A representative of each RFLP pattern type was sequenced to confirm identity. All reactions were carried out with the ABI Prism cycle sequencing kit (Applied Biosystems) according to the manufacturer's instructions. Sequencing was done on an ABI automated sequencer (Applied Biosystems).
|
ANN analysis.
An ANN is a statistical technique modelled on the neural networks found in the brain. The brain consists of neurones that transmit information (in this experiment treatments and data) which are interconnected to each other by synapses to form a network (Rao & Rao, 1983). The brain can recognize patterns within data, learns and remembers these patterns, and compares the remembered template with new data. One can teach an ANN to recognize associations in complex datasets so that a mathematical association between inputs and outputs can be developed that is independent of non-linearity and noise associated with complex biological datasets. We used ANNs to develop a relationship between a dataset that consisted of colony size, shape, colour, time incubated and 16S rDNA RFLP. The purpose was to predict the species isolated on Rogossa MRS agar from simple colony morphology. 16S rDNA RFLPs were conducted on 124 out of 563 colonies described and a representative of each RFLP type was sequenced. Cross-validation of the ANN was by the following procedure: the order of the data was randomized, 90 % of the data were used to train the ANN and 10 % of the data were used as the validation dataset. The validation scheme was repeated 10 times to optimize the ANN. All analyses were conducted with Statistica software for neural networks (StatSoft).
Statistical ecological analysis.
Hill's first diversity number, N1, is weighted towards the most dominant species and is calculated as:
![]() |
where pi is the proportional abundance of the ith taxon. N2 is an index of less abundant species and is calculated as: N2=1/S, where S is Simpson's index (Boehm et al., 1993). Simpson's index is calculated as:
![]() |
Proportional similarity between digestive sites was calculated with a multivariate procedure which was performed by calculating Euclidean distances and the dendrogram was constructed with the unweighted paired means method (Bollet & de Micco, 1992).
Other statistical analyses.
Conventional statistical analyses were performed by employing analysis of variance. Bacterial count data did not have homogeneous variances and values were log transformed. The total sums of squares was partitioned between the sums of squares for gut compartment and diet so that the experiment followed a factorial arrangement of treatments with two factors (diet) and six levels (gut compartment). All analyses were conducted with Statistica software for general linear models (StatSoft).
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Numbers of E. coli isolated from the high-grain diet were higher (P<0·05) than for the high-forage diet (mean along the tract: 1·2x106 ml-1 vs 1·05x102 ml-1) (Fig. 1a). AR E. coli (Fig. 1b
) were more numerous (P<0·05) in grain-fed than forage-fed cattle but were biased towards the colon contents (1x101 ml-1 vs 0·13x101 ml-1), upper colon wall (3·9x105 ml-1 vs 0·14x101 ml-1) and faeces (3·2x105 ml-1 vs 0·13x101 ml-1). Other sites (rumen, ileum contents and ileum wall) did not show statistically significant (P>0·05) differences in AR E. coli due to diet. Rogossa MRS isolates were more numerous in grain-fed cattle, which had a mean population of 1·8x107 ml-1, which was approximately 2 log units higher than the mean for forage-fed animals (7·9x105 ml-1) (Fig. 1c
). The greatest differences between diets were observed in the faeces (Fig. 1c
). The pH of the digestive tract was always lowest in the grain diet irrespective of gut compartment, and the pH in the colon and faeces tended to be higher (Fig. 2
).
|
|
|
The major species were Megasphaera elsdenii, Str. bovis, Selenomonas ruminantium, A. fermentans and Lactobacillus species (Fig. 3). Over 90 % of lactobacilli were either Lactobacillus fermentum or Lactobacillus ruminus. A number of isolates with unique RFLP patterns made up a small proportion of the population (
6 %). These isolates with their closest NCBI matches were as follows: Butyrivibrio fibrisolvens (9598 %), unidentified ruminal bacterium (94 %), Mitsuokella multacida (93 %), Enterococcus casseliflavus (100 %), E. coli (99 %) and Clostridium spp. (92 %). ANN analysis predicted species from colony morphology correctly 87·4 % of the time and sensitivity analysis showed that colony colour and form was the most informative while colony size, time incubated and digestive site were not significant (P>0·05). A. fermentans predictions were correct 75 % of the time, Str. bovis 96·1 %, Sel. ruminantium 77·7 %, Lactobacillus spp. 85·7 % and M. elsdenii 81·3 %.
In the high-forage diet, Str. bovis was the dominant species irrespective of digestive site (Fig. 4a), but in the high-grain diet its numbers declined (Fig. 4b
). Sel. ruminantium was a small proportion of the ruminal population (<5 %) in the high-forage diets, but tended to increase in abundance in the lower gut (Fig. 4a
). In contrast, Selenomonas was far more abundant in the rumen of the high-grain diet and declined in the lower gut (Fig. 4b
). Lactobacillus was present at all digestive sites of both (Fig. 4
) diets but was always higher in the high-grain diet (Fig. 4b
). M. elsdenii was associated in significant numbers with the high-grain diet (Fig. 4b
) but this was not the case with the high-fibre diet (Fig. 4a
). A. fermentans only occurred in the colon of the high-fibre diet (Fig. 4a
) but was isolated at all digestive sites from grain-fed animals (Fig. 4b
).
|
|
|
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
To resolve this problem Keen et al. (1999) assessed the effects of diet on 200 cattle that had been adapted to a high-grain diet. The prevalence of O157 in this group was 53 %, and cattle were subsequently divided into two groups of 54 cattle each. One group continued on the high-grain diet and the other group was switched to a high-fibre diet. Within 7 days only 14 hay-fed animals were positive for E. coli, while the prevalence in the grain-fed group remained almost unchanged. These data suggest that experimental dosing results in responses to diet that are not solely dependent on diet. It must however be noted that this does not automatically mean that detection of specific serotypes should not be conducted as the physiology of different pathogenic serotypes may differ from that of O157. This is supported by the fact that subpopulations of O157 have fundamentally different genetic compositions and depending on what continent they are isolated from, may be more or less pathogenic (Kim et al., 2001a
).
In our studies, animals on the high-grain diet had a higher number of E. coli that were viable after acid shock with mineral acid at pH 2 in comparison to the high-fibre diet. Diez-Gonzalez et al. (1998) did not specifically indicate what proportion of the acid-resistant population were generic E. coli but they did confirm that all, or most, of the E. coli isolates which were picked were acid-resistant. In contrast, Hovde et al. (1999)
found no significant differences in the proportion of acid-resistant O157 : H7 present in high-grain or high-fibre diets fed to cattle when animals were experimentally infected. Diez-Gonzalez et al. (1998)
indicated that acid resistance induced by the high-grain diets would select for an acid-resistant E. coli population that would increase the number of bacteria that could potentially survive the acidic conditions of the human gastric stomach. An alternative view is that of Jordan et al. (1999)
who demonstrated that killing of the most acid tolerant E. coli O157 : H7 strains could be increased by as much as 4 log units by the addition of lactate or ethanol, or a combination of the two. Animals consuming high-grain diets produce copious amounts of lactic acid (Brown et al., 2000
; Gill et al., 2000
) and based on the ability of lactate to increase killing of O157 : H7 the potential problem with the acid-resistant phenotype described by Diez-Gonzalez et al. (1998)
would be less likely to be a concern.
Animals in our experiments were fasted for 12 h before slaughter. This may not be optimal from a microbiological point of view because a number of physiological changes occur in the digestive tract during fasting that influence microbial populations (Gate et al., 1999; Kumada et al., 1985
; Dehority & Orpin, 1988
). However, feed is commonly withheld from animals during transport to slaughterhouses to reduce faecal contamination. In Australia it is standard practice to withhold feed from animals for approximately 12 h before processing at slaughter plants. Recent investigations demonstrated no increase in numbers of O157 when examined in experimentally infected adult sheep (Kudva et al., 1997
) and calves (Harmon et al., 1999
).
It is clear that diet influences E. coli numbers in the faeces of cattle but in most situations feeding high-fibre diets is not economically feasible. In some geographical regions, such as the sugar cane growing regions of Northern Australia, molasses can be substituted for grain in the diet. Molasses is made up primarily of sucrose and results in a significant reduction in generic E. coli in the faeces (Krause & McSweeney, 2002). In production systems where grain is the only concentrated energy source available, dietary supplements such as sodium chlorate can be used to reduce faecal E. coli but the level of inclusion is 30 % of the LD50 for cattle (Callaway et al., 2001
, 2002
). A non-toxic alternative that has been proposed is probiotic bacteria that inhibit E. coli and are typically LAB (Naidu et al., 2002
; Kim et al., 2001b
; Ogawa et al., 2001
; Gopal et al., 2001
). Given the popularity for LAB as probiotics (Riley & Gordon, 1999
) we investigated the in vivo ecological relationships between E. coli and LAB. To facilitate this process we developed a method based on ANN that allows for rapid evaluation of large numbers of bacteria isolated on Rogossa MRS agar.
Colony morphology alone does not differentiate between species. However, by including more data, such as information on colony diameter, animal diet, time of incubation, 16S rDNA RFLP, etc., the ability to predict species becomes more robust. In essence, the problem we had to solve was to group like data that were non-linearly related. The classical approach is to use either principle component analysis (Bollet & de Micco, 1992) or cluster analysis (Bollet & de Micco, 1992
) but neither robustly accommodate non-linear relationships and high levels of inherent variability. In a recent study (Noble et al., 2000
), it could be demonstrated that ANN was superior to other statistical techniques when grouping data with high levels of biological variation. These researchers used phospholipid fatty acid profiles of surface and subsurface marine sediments to predict the composition of microbial communities. It was demonstrated that ANN resulted in lower levels (2·7 %) of incorrect classifications than did conventional linear discriminate analysis (8·4 %).
We picked 124 isolates of diverse colony morphology from Rogossa MRS agar plates. The 16S rDNA RFLP pattern was determined and a representative of each pattern was sequenced. Data on diet type, gut compartment, colony colour, morphology and size, as well as time of incubation and 16S rDNA profile were used to optimize an ANN. Sensitivity analysis showed that colony description was the single most important factor for identification of species and colonies on plates could be correctly classified 87 % of the time. The reasons for misclassification were closely examined. It appears that there are a group of isolates within each species population that vary slightly in colour and size from the norm of that species. For example, Sel. ruminantium was frequently misclassified (Table 1). They were typically cream to white translucent colonies greater than 3 mm. Incorrect predictions arose from white translucent colonies of Sel. ruminantium that were smaller than the norm and were either Str. bovis or M. elsdenii. These data suggest that providing that colony morphology can be described in a standardized manner a high level of accuracy can be obtained in identifying an isolate.
Cultivation on Rogossa MRS agar demonstrated that counts were consistently higher from the animals consuming the high-grain diets compared to high fibre diets (Fig. 1c). Examination of the isolates using ANN demonstrated that the bacteria cultured in the hind-gut of cattle on high-grain diets is typical of that found in the rumen (Mackie & Heath, 1979
; Mackie & Gilchrist, 1979
; Mackie et al., 1978
). The predominant lactic acid producing bacterium in the hind-gut was Str. bovis and is classically associated with lactic acid acidosis in the rumen (Fig. 4
). This occurs when animals on forage-based diets are rapidly switched to high-grain diets (Slyter, 1976
). Lactic acid is rapidly produced which, if accumulated in sufficient quantities, can lead to ruminal dysfunction. If however, animals are slowly adapted to high-grain diets, lactic acid consuming bacteria like Sel. ruminantium and M. elsdenii increase in abundance and the concentration of lactic acid declines (Russell & Hino, 1985
). Lactic acidosis is usually ascribed to lactate accumulation in the rumen but hind-gut lactate can be a major contributor (Zust et al., 2000
). This is supported by the fact that LAB were in high abundance in the caecum, colon and faeces (Fig. 1c
).
In relation to lactic acid acidosis it is sometimes assumed that Str. bovis is at low levels in the rumen on high-fibre diets but at high levels on high-grain diets. This is true in that there are less Str. bovis on high-forage diets, but the magnitude of this difference is usually less than one log unit. Latham et al. (1971, 1972
) demonstrated that Str. bovis were approximately 50 % lower in high-fibre than high-grain diets but these numbers were all within one log unit. These authors (Latham et al., 1971
, 1972
) also demonstrated an increase in Lactobacillus numbers on high-grain diets. In contrast we observed a relative decline in Str. bovis with high-grain diets but the change in numbers was within one log unit. Earlier studies did not have the luxury of having access to molecular techniques and were limited by the number of isolates that could be examined in detail. An advantage of our approach was that we not only examined a substantial subset of isolates with molecular techniques but could extend species identification to a much larger population because of the use of ANNs.
Hill's first diversity number (Table 2), N1, indicated that diversity for the most abundant species was always higher in the high-forage diet and this value was a consequence of the dominance of Str. bovis (Fig. 4
). In contrast, N2, which is an index of the less abundant species, was higher in the high-grain diets (Table 2
) and is a consequence of the increase in less abundant species such as M. elsdenii. An interesting feature was that diversity in the ileum was not typical of the rest of the digestive tract. At this site, Str. bovis was more dominant than at other sites; this observation is supported by the fact that the less abundant non-Str. bovis species were lower. For this reason, the evenness index is far lower in the ileum than at other sites, strongly supporting the notion that the ileum has physiological features different to that of the rest of the digestive tract. Although there are site related differences, the physiological differences between diets are the dominant feature influencing diversity. Multivariate analysis supports this hypothesis demonstrating that, in general, digestive sites cluster with the diet (Fig. 5
). An illustration is the isolation of M. elsdenii at all digestive sites from the high-grain diet, while no isolates were obtained in the high-fibre diet. M. elsdenii is a lactic acid consuming organism that increases in abundance in the rumen when animals are adapted to high-grain diets. Our data suggest that the microbial response to high-grain adaptation is similar throughout the digestive tract, irrespective of digestive site.
The data in this manuscript demonstrate that E. coli and Rogossa MRS counts (containing lactic acid bacteria) followed the same population dynamics; when E. coli was high so were MRS counts and visa versa. Lema et al. (2001) fed a probiotic containing Lactobacillus acidophilus, Lactobacillus casei and Streptococcus faecium to lambs and demonstrated that there was a significant decrease in dosed O1572 : H7. Ohya et al. (2000)
fed Str. bovis and Lactobacillus gallinarum to adult cattle inoculated with O157 : H7 and reported a significant reduction in shedding. In these reports, experimentally infected O157 : H7 were used and appropriate controls were not included. For example, the natural rate of decline of O157 : H7 must be assessed. In a similar experiment, Zhao et al. (1998)
was able to demonstrate a significant reduction in O157 : H7 shedding when animals were dosed with a probiotic consisting of naturally occurring E. coli inhibitory to O157 : H7. Alternatively, bacteriocin-producing E. coli can be used and has the advantage that the bacteriocin producer and the pathogen will be competing in the same niche. Jordi et al. (2001)
screened 20 E. coli strains producing well-characterized colicins for their ability to inhibit Shiga toxin-producing E. coli serotypes O26, O111, O128, O145 and O157 : H7. To simulate the rumen of cattle, overlay assays were performed under anaerobic conditions in the presence of 30 % rumen fluid. Colicins E1, E4, E8-J, K and S4 were most active against Shiga-toxin-producing E. coli strains under anaerobic conditions in the absence or presence of rumen fluid.
In conclusion, our data support the view that high-grain diets result in increased levels of E. coli populations throughout the digestive tract of ruminants in comparison to high-fibre diets. The variations in numbers within diet are significant because it points to the fact that nutrient metabolism and physiology in the ruminant hind-gut and rumen differ at the microbial level. We used ANNs to assess Rogossa MRS isolates that were largely LAB. ANN was found to be a powerful tool for analysing large sets of isolates which would be very difficult to process using traditional determinative microbiology. This approach should significantly enhance the ability to screen large numbers of isolates that may be used in secondary screening for probiotic activity. Lastly, it is also important to note that management of E. coli in cattle faeces is not simply a question of understanding dietary interactions. There are other factors that influence prevalence in cattle such as bacteriophages (Koch et al., 2001; O'Brien et al., 2001
) that may be related to the bursts of pathogen often seen in feedlot cattle (Midgley et al., 1999
). In addition, environmental contamination from sources such as drinking water (LeJeune et al., 2001
) and cross contamination between animals (Midgley & Desmarchelier, 2001
) should be taken into account. Similarly, probiotics should be seen as one tool in a suite of tools that can potentially be used to control pathogenic serotypes of E. coli in cattle.
![]() |
ACKNOWLEDGEMENTS |
---|
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Blackburn, N., Hagström, Å., Wikner, J., Cuadros-Hansson, R. & Bjørnsen, P. K. (1998). Rapid determination of bacterial abundance, biovolume, morphology, and growth by neural network-based image analysis. Appl Environ Microbiol 64, 32463255.
Boehm, M. J., Madden, L. V. & Hoitink, H. A. J. (1993). Effect of organic matter decomposition level on bacterial species diversity and composition in relationship to pythium damping-off severity. Appl Environ Microbiol 59, 41714179.[Abstract]
Bollet, C. & de Micco, P. (1992). Taxonomic methods. In Encyclopaedia of Microbiology, pp. 179200. Edited by J. Lederberg. New York: Academic Press.
Brown, M. S., Krehbiel, C. R., Galyean, M. L., Remmenga, M. D., Peters, J. P., Hibbard, B., Robinson, J. & Moseley, W. M. (2000). Evaluation of models of acute and subacute acidosis on dry matter intake, ruminal fermentation, blood chemistry, and endocrine profiles of beef steers. J Anim Sci 78, 31553168.
Bryant, M. P. (1972). Commentary on the Hungate technique for culture of anaerobic bacteria. Am J Clin Nutr 25, 13241328.[Medline]
Callaway, T. R., Anderson, R. C., Anderson, T. J., Poole, T. L., Bischoff, K. M., Kubena, L. F. & Nisbet, D. J. (2001). Escherichia coli O157 : H7 becomes resistant to sodium chlorate in pure culture, but not in mixed culture or in vivo. J Appl Microbiol 91, 427434.[CrossRef][Medline]
Callaway, T. R., Anderson, R. C., Genovese, K. J., Poole, T. L., Anderson, T. J., Byrd, J. A., Kubena, L. F. & Nisbet, D. J. (2002). Sodium chlorate supplementation reduces E. coli O157 : H7 populations in cattle. J Anim Sci 80, 16831689.
Cray, W. C., Casey, T. A., Bosworth, B. T. & Rasmussen, M. A. (1998). Effect of dietary stress on fecal shedding of Escherichia coli O157 : H7 in calves. Appl Environ Microbiol 64, 19751979.
Dehority, B. A. & Orpin, C. G. (1988). Development of, and natural fluctuations in, rumen microbial populations. In The Rumen and its Microbes, pp. 151183. Edited by P. N. Hobson. New York: Elsevier.
Diez-Gonzalez, F., Calloway, T. R., Kizoulis, M. G. & Russell, J. B. (1998). Grain feeding and dissemination of acid-resistant Escherichia coli from cattle. Science 281, 16661668.
Elder, R. O., Keen, J. E., Siragusa, G. R., Barkocy-Gallagher, G., Koohmaraie, M. & Laegreid, W. W. (2000). Correlation of enterohemorrhagic Escherichia coli O157 prevalence in feces, hides, and carcasses of beef cattle during processing. Proc Natl Acad Sci U S A 97, 29993003.
Fegan, N. & Desmarchelier, P. (2002). Comparison between human and animal isolates of Shiga toxin-producing Escherichia coli O157 from Australia. Epidemiol Infect 128, 357362.[CrossRef][Medline]
Gate, J. J., Parker, D. S. & Lobley, G. E. (1999). The metabolic fate of the amido-N group of glutamine in the tissues of the gastrointestinal tract in 24 h-fasted sheep. Br J Nutr 81, 297306.[Medline]
Gill, H. S., Shu, Q. & Leng, R. A. (2000). Immunization with Streptococcus bovis protects against lactic acidosis in sheep. Vaccine 18, 25412548.[CrossRef][Medline]
Gopal, P. K., Prasad, J., Smart, J. & Gill, H. S. (2001). In vitro adherence properties of Lactobacillus rhamnosus DR20 and Bifidobacterium lactis DR10 strains and their antagonistic activity against an enterotoxigenic Escherichia coli. Int J Food Microbiol 67, 207216.[CrossRef][Medline]
Hancock, D. D., Besser, T. E., Gill, C. & Bohach, C. H. (1999). Cattle, hay and E. coli. Science 284, 5153.[CrossRef][Medline]
Harmon, B. G., Brown, C. A., Tkalcic, S., Mueller, P. O. E., Parks, A., Jain, A. V., Zhao, T. & Doyle, M. P. (1999). Fecal shedding and rumen growth of Escherichia coli O157 : H7 in fasted calves. J Food Prot 62, 574579.[Medline]
Hovde, C. J., Austin, P. R., Cloud, K. A., Williams, C. J. & Hunt, C. W. (1999). Effect of cattle diet on Escherichia coli O157 : H7 acid resistance. Appl Environ Microbiol 65, 32333235.
Hungate, R. E. (1950). The anaerobic mesophilic cellulolytic bacteria. Bacteriol Rev 14, 149.
Jordan, S. L., Glover, J., Malcolm, L., Thomson-Carter, F. M., Booth, I. R. & Park, S. F. (1999). Augmentation of killing of Escherichia coli O157 by combinations of lactate, ethanol, and low-pH conditions. Appl Environ Microbiol 65, 13081311.
Jordi, B. J., Boutaga, K., van Heeswijk, C. M., van Knapen, F. & Lipman, L. J. (2001). Sensitivity of Shiga toxin-producing Escherichia coli (STEC) strains for colicins under different experimental conditions. FEMS Microbiol Lett 204, 329334.[CrossRef][Medline]
Keen, J. E., Urlich, G. A. & Elder, R. O. (1999). Effects of hay- and grain-based diets on the fecal shedding of naturally-acquired enterohemorrhagic E. coli (EHEC) O157 : H7 in beef feedlot cattle. In Abstracts of the 80th Conference of Research Workers in Animal Diseases, Chicago. Abstract 86. Ames, IA: Iowa State Press.
Kim, J., Nietfeldt, J., Ju, J., Wise, J., Fegan, N., Desmarchelier, P. & Benson, A. K. (2001a). Ancestral divergence, genome diversification, and phylogeographic variation in subpopulations of sorbitol-negative, ß-glucuronidase-negative Escherichia coli O157. J Bacteriol 183, 68856897.
Kim, S. H., Yang, S. J., Koo, H. C., Bae, W. K., Kim, J. Y., Park, J. H., Baek, Y. J. & Park, Y. H. (2001b). Inhibitory activity of Bifidobacterium longum HY8001 against Vero cytotoxin of Escherichia coli O157 : H7. J Food Prot 64, 16671673.[Medline]
Koch, C., Hertwig, S., Lurz, R., Appel, B. & Beutin, L. (2001). Isolation of a lysogenic bacteriophage carrying the stx(1(OX3)) gene, which is closely associated with Shiga toxin-producing Escherichia coli strains from sheep and humans. J Clin Microbiol 39, 39923998.
Krause, D. O. & McSweeney, C. S. (2002). Use of feed supplements to manage generic and pathogenic Escherichia coli in tropical animal production systems. In Proceedings of the International Association for Food Protection, pp 148149. Des Moines, IA: International Association for Food Protection.
Kudva, I. T., Hunt, C. W., Williams, C. J., Nance, U. M. & Hovde, C. J. (1997). Evaluation of dietary influences of Escherichia coli O157 : H7 shedding by sheep. Appl Environ Microbiol 63, 38783886.[Abstract]
Kumada, K., Koike, K. & Fujiwara, K. (1985). The survival of bacteria under starvation conditions: a mathematical expression of microbial death. J Gen Microbiol 131, 23092312.
Lane, D. J. (1991). 16S/23S rRNA sequencing. In Nucleic Acid Techniques in Bacterial Systematics, pp. 115176. Edited by M. Goodfellow. Chichester: Wiley.
Latham, M. J., Sharpe, E. & Sutton, J. D. (1971). The microbial flora of the rumen of cows fed hay and high cereal rations and its relationship to the rumen fermentation. J Appl Bacteriol 34, 425434.[Medline]
Latham, M. J., Storry, J. E. & Sharpe, M. E. (1972). Effect of low-roughage diets on the microflora and lipid metabolism in the rumen. Appl Microbiol 24, 871877.
LeJeune, J. T., Besser, T. E., Merrill, N. L., Rice, D. H. & Hancock, D. D. (2001). Livestock drinking water microbiology and the factors influencing the quality of drinking water offered to cattle. J Dairy Sci 84, 18561862.
Lema, M., Williams, L. & Rao, D. R. (2001). Reduction of fecal shedding of enterohemorrhagic Escherichia coli O157 : H7 in lamb by feeding microbial feed supplement. Small Rumin Res 39, 3139.[CrossRef][Medline]
Mackie, R. I. & Gilchrist, F. M. C. (1979). Changes in lactate-producing and lactate-utilizing bacteria in relation to pH in the rumen of sheep during stepwise adaptation to a high concentration diet. Appl Environ Microbiol 38, 422430.
Mackie, R. I. & Heath, S. (1979). Enumeration and isolation of lactate-utilizing bacteria from the rumen of sheep. Appl Environ Microbiol 38, 416421.
Mackie, R. I., Gilchrist, F. M. C., Robberts, A. M., Hannah, P. E. & Schwartz, H. M. (1978). Microbiological and chemical changes in the rumen during stepwise adaptation of sheep to high concentrate diets. J Agric Sci 90, 241254.
Meng, J. & Doyle, M. P. (1997). Emerging issues in microbiological food safety. Annu Rev Nutr 17, 255275.[CrossRef][Medline]
Midgley, J. & Desmarchelier, P. (2001). Pre-slaughter handling of cattle and Shiga toxin-producing Escherichia coli (STEC). Lett Appl Microbiol 32, 307311.[CrossRef][Medline]
Midgley, J., Fegan, N. & Desmarchelier, P. (1999). Dynamics of Shiga toxin-producing Escherichia coli (STEC) in feedlot cattle. Lett Appl Microbiol 29, 8589.[CrossRef][Medline]
Naidu, A. S., Xie, X., Leumer, D. A., Harrison, S., Burrill, M. J. & Fonda, E. A. (2002). Reduction of sulfide, ammonia compounds, and adhesion properties of Lactobacillus casei strain KE99 in vitro. Curr Microbiol 44, 196205.[Medline]
Noble, P. A., Bidle, K. D. & Fletcher, M. (1997). Natural microbial community composition compared by a back-propagation neural network and cluster analysis of 5S rRNA. Appl Environ Microbiol 63, 17621770.[Abstract]
Noble, P. A., Almeida, J. S. & Lovell, C. R. (2000). Application of neural computing methods for interpreting phospholipid fatty acid profiles of natural microbial communities. Appl Environ Microbiol 66, 694699.
O'Brien, S. J., Murdoch, P. S., Riley, A. H. & 7 other authors (2001). A foodborne outbreak of Vero cytotoxin-producing Escherichia coli O157 : H-phage type 8 in hospital. J Hosp Infect 49, 167172.[CrossRef][Medline]
Ogawa, M., Shimizu, K., Nomoto, K., Tanaka, R., Hamabata, T., Yamasaki, S., Takeda, T. & Takeda, Y. (2001). Inhibition of in vitro growth of Shiga toxin-producing Escherichia coli O157 : H7 by probiotic Lactobacillus strains due to production of lactic acid. Int J Food Microbiol 68, 135140.[CrossRef][Medline]
Ohya, T., Marubashi, T. & Ito, H. (2000). Significance of fecal volatile fatty acids in shedding of Escherichia coli O157 from calves: experimental infection and preliminary use of a probiotic product. J Vet Med Sci 62, 11511155.[CrossRef][Medline]
Rao, V. B. & Rao, H. V. (1993). C++ Neural Networks and Fuzzy Logic. New York: MIS Press.
Riley, M. A. & Gordon, D. M. (1999). The ecological role of bacteriocins in bacterial competition. Trends Microbiol 7, 129133.[CrossRef][Medline]
Russell, J. B. & Hino, T. (1985). Regulation of lactate production in Streptococcus bovis: a spiraling effect that leads to rumen acidosis. J Dairy Sci 68, 17121721.
Scott, T., Wilson, C., Bailey, D., Klopfenstein, T., Milton, T., Moxley, R., Smith, D., Gray, J. & Hungerford, L. (2000). Influence of diet on total and acid resistant E. coli and colonic pH. In 2000 Nebraska Beef Report pp. 3941. Lincoln, NE: University of Nebraska-Lincoln.
Slyter, L. L. (1976). Influence of acidosis on rumen function. J Anim Sci 43, 910929.[Medline]
Zhao, T., Doyle, M. P., Harmon, B. G., Brown, C. A., Mueller, P. O. E. & Parks, A. H. (1998). Reduction of carriage of enterohemorrhagic Escherichia coli O157 : H7 in cattle by inoculation with probiotic bacteria. J Clin Microbiol 36, 641647.
Zust, J., Pestevsek, U. & Vengust, A. (2000). Impact of lactic acid fermentation in the large intestine on acute lactic acidosis in cattle. Dtsch Tierarztl Wochenschr 107, 359363.[Medline]
Received 22 April 2002;
revised 29 August 2002;
accepted 4 September 2002.
HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
INT J SYST EVOL MICROBIOL | MICROBIOLOGY | J GEN VIROL |
J MED MICROBIOL | ALL SGM JOURNALS |