A survey of the fecal bacteria of bison (Bison bison) for potential pathogens and antimicrobial susceptibility of bison-origin E. coli

Abstract

An observational study determined the normal fecal bacterial flora of clinically healthy bison, detected the presence of common potential zoonotic pathogens, and determined the antimicrobial susceptibility of isolated E. coli strains. Ninety-six fecal samples from 10 captive herds were cultured for aerobic, anaerobic, facultative, and microaerophillic bacteria. Nineteen major genera of gram-positive and 8 genera of gram-negative bacteria were identified. Salmonella spp. were not detected but some of the isolated bacteria are potential gastrointestinal pathogens. Minimum inhibitory concentrations (MIC) of 24 antimicrobials were determined for the E. coli isolated. Nearly all were susceptible to 23 of the 24 antimicrobials but there was a reduced susceptibility to sulphonamide. There were fewer resistant strains than were reported in recent studies of generic E. coli from cattle living in the same area.

Introduction

Bison (Bison bison) farming is a growing industry in North America. According to Statistics Canada, the Canadian farmed bison population increased from 144 000 in 2001 to an estimated 225 000 in 2007, reflecting an annual increase of almost 8% (1). Despite this growth and their recent emergence as a significant farmed species, relatively little is known about their biology. For instance, fecal flora from traditionally farmed livestock such as cattle has been well studied but fecal bacterial populations from bison have not. Studies have been conducted to characterize bison rumen and conjunctival microflora, and to search for potential respiratory tract pathogens and zoonotic bison carcass contaminants (2–7), but the importance of these bacterial species in pathological conditions of bison is poorly understood. Without information about normal bacterial populations, the interpretation of bacterial culture results from diseases of the intestinal tract or other organ systems in bison is difficult and must be based on assumptions that bison flora is similar to that of cattle (3,5,8–11). In this study our aim was to create some context for the findings of bacterial cultures of the bison intestinal tract.

Antimicrobial resistant strains of bacteria have become a concern in human and veterinary medicine, in the treatment of clinical disease and in food safety (12,13). In particular, antimicrobial use in animal agriculture has been suggested as a contributing factor to bacterial antimicrobial resistance found in human infections (14). The North American bison industry markets their product as relatively “organic” or natural. This suggests that antimicrobial use in bison production is minimal, thereby reducing the risk of encountering resistant bacteria in clinical diseases of bison and in bison food products for human consumption (1,15). The objectives of this study were to identify by culture of live-animal fecal samples, the normal aerobic, anaerobic, and microaerophilic fecal bacteria of bison, to detect the presence of common enteropathogens and potential zoonotic foodborne pathogens, and to determine the susceptibilities of bison-derived generic E. coli to antimicrobial drugs used in human and veterinary medicine.

Materials and methods

Ninety-six fecal samples from apparently healthy adult bison in 10 captive herds were cultured. Opportunistic samples were obtained from cooperative bison producers, mostly from the Saskatoon area of Saskatchewan, Canada, but included 12 samples from 2 zoo populations. Where possible, a minimum of 10 animals per herd were sampled. Samples were from females (n = 61), males (n = 8), or animals of undetermined gender (n = 27). Fresh feces were received from some herds but in others, duplicate bacterial swab samples were collected directly from the rectum and kept refrigerated if they were to be processed within 24 h, or immediately frozen at −80°C if they were to be kept for several days. Swabs, or 0.5 g of feces, were combined with 4.5 mL of peptone buffer to create a semi-quantitative dilution of 10⁻¹, then subsequent dilutions from 10⁻² to 10⁻⁵ were made. Diluted samples were streaked with a 10-μL loop onto each of the general and selective media under the culture conditions shown in Table 1, which shows the target bacterial species for that particular medium. In addition to the media incubated under aerobic conditions, an anaerobic environment was provided for inocula on selective media for Clostridium and Fusobacterium spp. These plates were incubated in anaerobic jars (Anaerocult A, EM Science, EM Industries, Gibbstown, New Jersey, USA) at 37°C for 24 h and re-examined for growth at 48 h. Microaerophillic conditions were created for incubation of Campylobacter selective media by placing the plates into sealed bags filled with a gas mixture of 5% O₂, 10% CO₂, balanced with 85% N (Praxair Canada) and placing the bags into a CO₂ incubator (Thermo Scientific, Waltham, Massachusetts, USA) so that if the bags leaked during incubation there would be a better chance of bacterial survival until the bags could be refilled. All isolates were placed into 1.0 mL of litmus milk and frozen at −80°C for later identification.

Table 1.

Summary of media and culture conditions used to isolate various bacterial species from bison rectal swabs or feces

		Incubation
Target bacteria	Dilution	Time (h)a	Temp (°C)	Media
Aerobic
Total bacteria	10−1—10−5	24	37.0	Sheep blood agarc
Enterobacter spp.	10−1—10−4	24	37.0	McConkey agarc
Staphylococcus aureus	10−1—10−2	24	37.0	Mannitol salt agarc
Listeria spp.	10−1—10−3	24—48	37.0	Listeria selective agard
Enterococcus spp.	10−1—10−2	24—48	30.0	Kanamycin-aesculin azide agard
Yersinia spp.	10−1—10−2	48	37.0	Cefsulodin-Irgasan-Novobiocine
Yeast, filamentous fungi	10−1—10−3	24—48	37.0	Sabouraud agarc
Salmonella spp.	10−1	12—18	37.0	Selenite enrichment brothc
Salmonella spp.	10−1—10−2	24	37.0	Hektoen or XLDf agarc
Anaerobic
Clostridium spp.	10−1—10−4	24	37.0	Sheep blood agar with 0.01% neomycin sulfatec
Clostridium difficile	10−1—10−2	24—48	37.0	C. difficile selective agarc
Fusobacterium spp.	10−1—10−2	24 + 48b	37.0	Egg yolk base-crystal violet-phenethylalcohol agarg
Microaerophillic
Campylobacter spp.	10−1—10−3	48 + 72b	37.0	Campylobacter selective agarc

^aTime at which bacterial growth was assessed.

^bRe-examined at 48 or 72 h.

^cManufactured by Becton Dickson, Oakville, Ontario.

^dManufactured by Oxoid MP 1750, Nepean, Ontario.

^eManufactured by Quelab Laboratories, Montreal, Quebec.

^fXylose lysine desoxycholate^greference 33.

Colonies obtained from cultures were quantified and identified by standard microbiology laboratory techniques. Morphological characteristics such as Gram stain, motility at room temperature, cellular morphology, colony characteristics (shape, size, elevation, and margin type), hemolytic effect, and growth on selective media were recorded. Conventional biochemical tests were used to identify the isolated bacteria, following standard identification schedules (16–18). API 20E and API 20A systems (bioMerieux-Vitek, Hazelwood, Missouri, USA) were used to identify Enterobacteriaceae and Clostridium-like organisms respectively.

Quantification of bacterial populations was done by a standard plate count method (19). Each swab was weighed and placed into a sterile tube, and peptone buffer was added according to the weight obtained. Samples were serially diluted from 10⁻¹ to 10⁻⁵ in peptone buffer. Dilutions were vortexed and diluted samples were streaked with a 10-μL loop onto each agar plate medium. Plates containing between 1 and 100 colony forming units (CFU) were selected and the number of bacteria was estimated.

Minimum inhibitory concentrations (MICs) for selected antimicrobials were determined for the E. coli isolates using the agar dilution method in accordance with the Clinical Laboratory Standards Institute (CLSI) standards (20,21). The same standards were used for the antimicrobial breakpoints and interpretation, except for those belonging to neomycin and streptomycin. Information for these was derived from Dowling (22). Lacking interpretive data for bison, cattle information was used when available; otherwise, information for humans was used (21).

Data were entered into an Excel database (Microsoft Corporation, One Microsoft Way, Redmond, Washington, USA) and summary statistics were generated and reported.

Results

Table 2 shows the frequency of isolation from 96 samples and mean counts [colony forming units (CFU)/g of feces] of aerobic and anaerobic bacteria, filamentous fungi, and yeast. Among these, 19 major genera of gram-positive and 8 genera of gram-negative bacteria were identified. The most common isolates are shown according to their Gram stain characteristics in Table 3.

Table 2.

Frequency of isolation and mean counts [CFU/g of feces (wet weight)] of major aerobic and anaerobic bacteria, filamentous fungi and yeast found in the feces of 96 bison

Bacterium, fungi, or yeast	Frequency (%)	CFU
Aerobic
Bacillus spp.	96 (100)	5.8 × 105
Escherichia coli	73 (76.0)	3.1 × 106
Arthrobacter spp.	51 (53.0)	3.1 × 104
Brevibacterium spp.	35 (36.5)	2.0 × 105
Micrococcus spp.	34 (35.4)	7.0 × 106
Enterococcus spp.	26 (27.1)	2.4 × 107
Streptococcus milleri group	24 (25.0)	7.8 × 106
Other Enterobacteriaceaea	22 (21.9)	3.4 × 105
Microbacterium spp.	22 (21.9)	6.5 × 105
Mucor sp. (fungus)	19 (19.8)	3.0 × 104
Enterococcus faecalis	15 (15.6)	5.3 × 105
Aeromonas spp.	14 (14.6)	4.4 × 105
Lactobacillus spp.	12 (12.5)	5.5 × 105
Other α hemolytic Streptococcus	11 (11.5)	2.6 × 107
Dermabacter spp.	10 (10.4)	1.0 × 105
Staphylococcus spp.	7 (7.3)	4.9 × 104
Enterococcus durans	6 (6.3)	2.8 × 104
Corynebacterium spp.	6 (6.3)	1.6 × 106
Acinetobacter spp.	6 (6.3)	1.0 × 104
Yersinia spp.	4 (4.2)	1.0 × 104
Pasteurella spp.	4 (4.2)	2.8 × 106
Yersinia enterocolitica	2 (2.1)	3.5 × 104
Rhodotorula rubra (yeast)	2 (2.1)	9.5 × 104
Staphylococcus aureus	1 (1.0)	1.0 × 104
Moraxella spp.	1 (1.0)	1.0 × 103
Turicella spp.	1 (1.0)	2.7 × 104
Listeria-like	7 (7.3)	1.0 × 103
Anaerobic
Clostridium perfringens	6 (6.3)	8.1 × 103
Clostridium spp.	6 (6.3)	5.3 × 103

CFU — colony-forming units.

^aEnterobacteriacaeae isolated sporadically and in very low quantities (10⁵ in total) were Escherichia fergusonii, Citrobacter freundii, Morganella morganii, Proteus rettgeri, Klebsiella spp., and Cronobacter (“Enterobacter”) zakazakii.

Table 3.

The most commonly isolated bacteria from 96 samples of bison feces, classified according to Gram stain characteristics

Bacterium	Number of isolates
Gram-positive
Bacillus spp.	239
Enterococcus spp.	64
Arthrobacter spp./Brevibacterium spp.	52
Microbacterium spp.	34
Micrococcus spp.	33
α hemolytic Streptococcus	30
Coagulase –ve Staphylococcus	14
Corynebacterium spp.	12
Lactobacillus spp.	9
Listeria — like	7
Clostridium perfringens	6
Clostridium spp.	4
Gram-negative
E. coli	255
Enterobacter	22
Aeromonas spp.	9
Pasteurella spp.	6
Serratia spp.	5
Yersinia spp.	6

The isolates were screened to identify common bovine and human enteropathogenic species, such as Salmonella spp., Clostridium perfringens and Clostridium difficile, as well as zoonotic bacteria such as Yersinia enterocolitica, Klebisella spp., Citrobacter spp., Campylobacter spp., Bacillus cereus and Aeromonas spp. Among those identified after screening were Aeromonas spp., Citrobacter spp., Y. enterocolitica, Klebsiella spp., C. perfringens, and E. coli.

Minimum inhibitory concentrations (MIC) of 24 antibiotics were determined for 255 E. coli isolates (Table 4). Nearly all isolates (> 92%) were susceptible to 23 of the 24 antibiotics but the isolates showed a high prevalence of reduced susceptibility to sulfonilamide (52%).

Table 4.

Minimal inhibitory concentration (MIC) of antimicrobials against 255 E. coli bacterial isolates from bison feces (μg/mL)

Antimicrobial	Number of strains and MIC (μg/mL)										Resistance breakpointa (μg/mL)	Percent resistant
	< 0.05	1	2	4	8	16	32	64	128	> 512
Amikacin					250	5					≥ 32	0
Ampicillin				181	3	68	3				≥ 16	28
Aztreonam				253	2						≥ 16	0
Carbenicillin					251	1	1	2			≥ 32	1
Cefazolin				248	5	1	1				≥ 16	1
Cefotaxime				255							≥ 16	0
Ceftiofur		254			1						≥ 4	0.3
Ceftriaxone				255							≥ 16	0
Cefuroxime			241	7	1	1	5				≥ 8	3
Cephalothin				236	16	2	1				≥ 16	1
Chloramphenicol				254			1				≥ 16	1
Ciprofloxacin	255										≥ 2	0
Enrofloxacin	255										≥ 1	0
Florfenicol		142	71	35	6	1					≥ 8	3
Gentamicin			255								≥ 8	0
Imipenem			255								≥ 8	0
Kanamycin					255						≥ 32	0
Nalidixic Acid			253	2							≥ 32	0
Neomycin			252	3							≥ 16	0
Streptomycin			243	2	4	6					≥ 8	0
Sulfonilamide									121	134	≥ 512	48
Tetracycline			235		5	15					≥ 8	8
Ticarcillin					245	7	1	2			≥ 32	2
Trimethoprim				250	4	1					≥ 16	1

^aBreakpoints were based on references 20 and 21.

Discussion

Bison feces appear to contain similar aerobic bacterial populations as cattle (9–11). The anaerobic culture was geared to detect specific potential pathogens such as C. difficile and C. perfringens and a comprehensive culture of the anaerobic flora was not done. Only C. perfringens was obtained and a study examining the anaerobic population of bison feces is needed. This is difficult with traditional methods because anaerobic microbes are not easily cultured and are expensive to identify by biochemical methods. However, new techniques such as a bacterial tag-encoded FLX 16s rDNA amplicon pyrosequencing (bTEFAP) could be used (23).

Knowledge of the bacteria normally present in the bison gastrointestinal tract and detection of potential pathogens provide the basis of a rational antimicrobial therapeutic approach to cases of diarrhea in this species. There was only sporadic antimicrobial resistance and resistance of E. coli to only 1 of 24 antimicrobials suggests a low prevalence of resistance in the normal intestinal flora of bison as compared with recently reported resistance found in generic E. coli isolated from cattle in the same area (24,25). The relatively high prevalence of resistance to sulfonamide seen in E. coli from the present study is similar to that described in the studies cited herein, even though sulfonamides are rarely, if ever, used to treat bison. Resistance to sulfonamide has been associated with virulence factors of strains in human and animal infections (25). Among the genes sul1, sul2, and sul3 that encode sulfonamide resistance, sul2 was detected most often in bovine strains (24,25). If this is also the case with bison strains it would be important because such genes might be transmissible among E. coli strains in bison and it is known that these genes are transmissible to human pathogenic strains of E. coli including uropathogenic E. coli (26). The enhancement of bacterial fitness conferred by some sul2-coding plasmids in the absence of selective pressure has been demonstrated (27). If the bison E. coli strains harbor such plasmids, this may partially explain the high prevalence of sulphonamide resistance in these strains.

The low level of resistance to other antimicrobial drugs corresponds with similar studies (7) and might be attributed to the relatively low levels of antimicrobials used in the routine production of bison (1). This has perhaps facilitated the existence of antimicrobial naïve populations of E. coli although, at least in cattle under feedlot conditions, there may not be a significant association between total amount of antimicrobial use and antimicrobial resistance (13). Bison are most often farmed extensively on pasture rather than intensively in feedlots and are not subjected to the crowded conditions that favor the sharing and distribution of resistant organisms. Another speculation for the low level of antimicrobial resistance found in our study may be that there has been very little co-mingling of bison and cattle populations.

Identification of the normal bison flora may allow recognition of any zoonotic risk to humans who might contact fecal material or food products potentially contaminated with bison feces. Strains of E. coli were found in 76% of samples but the isolates were not tested for virulence factors. A 2007 fecal survey of bison found an overall E. coli O157:H7 prevalence of 47.4%, but the overall prevalence of other serotypes or total E. coli isolated in the study were not indicated (28). That study used abattoir samples, which are not comparable to samples collected from animals on pasture.

Yersinia enterocolitica and C. perfringens were other pathogenic species detected, although the numbers were relatively low, presumably making the zoonotic risk low. The infective dose for Y. enterocolitica is 10⁹, and for food poisoning due to enterotoxin-producing C. perfringens it is 10⁷ (29,30).

The appearance of 7 Listeria-like organisms, although isolated in low numbers (10³) on selective media, warrants further investigation (31). On the selective media, the colonies had the morphological appearance of Listeria sp., and phenotypically they fit the general description of the genus, but they did not match any of the known species. The Listeria-like organisms were clustered within a single herd and their existence is therefore likely related to conditions on that particular farm. Due to the potential pathogenicity of these organisms, further genetic characterization studies are planned.

Salmonella were not found, although they have previously been identified as a carcass contaminant in slaughtered bison (6,7). In a 2007 review Gill (32) suggested that the Salmonella found on carcasses in such studies may have originated from in-plant sources of post slaughter contamination rather than from the gastrointestinal tracts of the slaughtered bison. The semi-quantitative aspect of the study was done with the aim of establishing the potential risk of putative pathogens which were targeted by using selective and differential media. Only Y. enterocolitica was detected in numbers suggestive of a potential zoonotic risk.

In summary, culturable fecal bacteria of bison were surveyed and were found to be similar to bacteria from cattle (9–11). It appears that, with the exception of sulfonamides, resistance to antimicrobials is not a significant feature of bison fecal E. coli. Studies on the genes coding for sulfonamide resistance in bison are warranted.