738 Journal o f Food Protection, Vol. 77, No. 5, 2014, Pages 738-744 doi:10.4315/0362-028X.JFP-13-433 Copyright © , International Association for Food Protection
A Mixture of Lactobacillus casei, Lactobacillus lactis, and Paenibacillus polymyxa Reduces Escherichia coli 0157:H7 in Finishing Feedlot Cattle KIM STANFORD,' SUSAN BACH,2 JOHN BAAH,3 a n d TIM McALLISTER3* 'Alberta Agriculture and Rural Development, Lethbridge, Alberta, Canada T1J 4V6; Agriculture and Agri-Food Canada, Summerland, British Columbia, Canada VOH IZO; and 3Agriculture and Agri-Food Canada, Lethbridge, Alberta, Canada T1J 4V6 MS 13-433: Received 9 October 2013/Accepted 9 January 2014
ABSTRACT A direct-fed microbial (DFM) containing Paenibacillus polymyxa, Lactobacillus casei, and Lactobacillus lactis was fed to cattle (n = 120) to determine impacts on shedding and survival of Escherichia coli 0157:H7 in feces. Cattle were individually penned and fed diets containing 0 (control), 4 x 107 CFU (DFM-4), 8 x 107 CFU (DFM-8), or 1.2 x 108 CFU (DFM-12) lactobacilli per kg of dietary dry matter over 84-day fall-winter growing and 140-day spring-summer finishing periods. Fecal grab samples were collected from cattle at 28-day intervals, E. coli 0157:H7 was detected by immunomagnetic separation, and isolates were compared by pulsed-field gel electrophoresis. During the growing period, feces negative for E. coli 0157 from each dietary treatment were inoculated with 105 CFU/g nalidixic acid-resistant E. coli 0157:H7 and were incubated at 4 and 22°C for 11 weeks. Fecal pH and fecal dry matter were measured on days 0, 1,3, and 7 and weekly thereafter, with E. coli 0157:H7 enumerated through dilution plating. Treatment with DFMs did not affect survival of E. coli 0157:H7 in feces or fecal pH (P > 0.05). Only one steer was positive for E. coli 0157:H7 during the growing period, but during the finishing period, DFM-8 and DFM-12 reduced the prevalence of E. coli 0157:H7 in feces (P < 0.05). Feeding DFMs also reduced the frequency of individual steers shedding E. coli 0157:H7 during finishing (P < 0.05), with control steers shedding E. coli 0157:H7 up to four times, whereas DFM-12 steers shed E. coli 0157:H7 a maximum of twice. Treatment with DFMs influenced pulsed-field gel electrophoresis profiles; steers that were fed DFM-8 and DFM-12 shed more diverse subtypes of E. coli 0157:H7 than did control or DFM-4 steers. Because a companion study found linear improvement in performance with increasing dosage of DFMs in the first 28 days of the growing period, targeted use of DFM-12 during this time and for the final 1 or 2 weeks prior to slaughter may optimize performance and reduce E. coli 0157:H7 while minimizing feed costs.
Direct-fed microbials (DFMs) are live microorganisms that have been used to improve growth performance of cattle (3, 10) or to reduce colonization of the bovine gastrointestinal tract with zoonotic or other pathogens (7, 24). DFMs are thought to function by a number of mechanisms, including production of inhibitory compounds such as acids or bacteriocins (33), competitive exclusion (7), improvement of rumen fermentation parameters (3), the blocking of quorum sensing (25), improvement of host mucosal immunity (16, 34), or by other as yet undefined mechanisms (36). Because the microbial populations of the bovine gastrointestinal tract are exceedingly complex and many interrelationships among organisms are poorly characterized, the efficacy of DFMs has varied (2, 39, 40). Age of cattle impacts the efficacy of DFMs, with more positive outcomes in younger calves for growth perfor mance (13) and pathogen control (7). Young calves stressed after weaning and/or transportation may show an optimal response to DFMs (15). Dosage of live organisms in the * Author for correspondence. Tel: 403-317-2240; Fax 403-382-3156; E-mail: [email protected].
DFMs is critical, and efficacy for improving growth performance has varied within a range of 1 or 2 log CFU per head per day (43). As well, efficacy for growth improvement may vary according to the strain of both the organism in the DFM and the pathogen (1). The properties of the microbial species in the DFM, such as acid tolerance, adhesion to gut tissues, and survival in the final feed formulation, can also influence efficacy for both pathogen control and growth performance (34). Increased efficacy has been noted in DFMs containing multiple compared with single organisms, possibly due to synergistic mechanisms of action (33), and a DFM including Lactobacillus acidophilus and Propionibacterium freudenreichii has been widely adopted for improvement of growth performance in larger feedlots in the United States (7). Because few reports have identified DFMs that offer concurrent control of pathogens and improved growth performance in feedlot cattle, the primary objective of the current study was to determine the appropriate dosage of a novel DFM containing a mixture of Lactobacillus casei, Lactobacillus lactis, and Paenibacillus polymyxa for reduc tion of Escherichia coli 0157:H7 in feces of feedlot cattle
J. Food Prot., Vol. 77, No. 5
DFMS FOR CONTROL OF E. COLI 0157:H7 IN FEEDLOT CATTLE
over a complete feeding cycle. A companion study (3) demonstrated that this DFM improved both average daily gain and feed conversion efficiency of calves during the first 84 days of feeding. A secondary objective was to determine whether the DFM had a residual impact on the survival of E. coli 0157:H7 in experimentally inoculated feces. MATERIALS AND METHODS Animals, facilities, and diets. Cattle and diets were those used in the study of Baah et al. (3), and cattle were handled according to the guidelines of the Canadian Council on Animal Care (30). Briefly, 120 Hereford x Angus steers (initial body weight 280 + 15.5 kg) were purchased at auction and were transported to the individual feeding barn of the Agriculture and Agri-Food Canada Research Centre in Lethbridge, Alberta. Steers were individually weighed and randomly assigned to one of four treatment groups for the duration of the study: 0 (control), 4 x 107 CFU (DFM-4), 8 x 107 CFU (DFM-8), or 1.2 x 108 CFU(DFM12) of a mixture of L. casei and L. lactis per kg of dietary dry matter (DM). Paenibacillus polymyxa was also present in the DFMs (28), although numbers in the DFMs were not determined. Each treatment was housed in a separate wing of the facility to avoid cross-contamination among treatments. Steers were individ ually penned and watered without contact among animals and were given ad libitum access to feed and water. For the first 84 days (12 November through 2 February), steers received a growing diet consisting of 60% barley silage, 35% barley grain, and 5% vitamin-mineral supplement on a DM basis. Over the next 28 days, steers were transitioned to a finishing diet by gradually increasing the proportion of dietary barley to 86% and reducing barley silage to 9%. Steers received finishing diets for 140 days (3 February to 23 June) before shipment to slaughter. Other than the DFMs, no microbial or antimicrobial agents were included in diets. A Data Ranger feed mixer (American Calan, Northwood, NH) was used to prepare diets, starting with control and ending with the highest level of DFM. The feed mixer was thoroughly cleaned after each day’s feeding to avoid cross-contamination among treatments. DFM. The DFM (RE3, Basic Environmental Systems and Technology Inc., Edmonton, Alberta, Canada) was shipped in liquid form and stored at 5°C until it was added to diets. Dosage of DFM in diets was verified by enumeration of Lactobacillus spp. after serial dilution of 1 ml of DFM in 9 ml of buffered peptone water, plating in duplicate onto de Man Rogosa Sharpe agar, and incubation at 37°C for 48 h. Lactobacillus spp. colonies were small and translucent and were enumerated from plates containing 30 to 300 colonies. The DFM contained a mixed culture of L. casei, L. lactis, and P. polymyxa and also included fermentation products of these organisms. Collection of fecal samples and detection of E. coli 0157:H7 in cattle. Fecal grab samples (10 g) were rectally collected from each steer at the start of the growing phase of the study (day 0) and every 28 days thereafter until steers were shipped to slaughter. A sterile glove was used to collect each sample, with samples transported to the laboratory within 1 h of collection. Feces (1 g) were then enriched for 6 h at 37°C in modified tryptic soy broth containing 20.0 mg/liter novobiocin (Sigma, Oakville, Ontario, Canada) and 1.5 g/liter dipotassium phosphate. Immunomagnetic separation was performed using Dynabeads anti-E. coli 0157 (Dynal, Lake Success, NY) according to manufacturer’s instructions. A 50-pl aliquot of bead-bacteria mixture was plated onto sorbitol MacConkey agar supplemented with 2.5 mg/liter
739
potassium tellurite and 0.05 mg/liter cefixime (CT-SMAC), and plates were incubated for 18 to 24 h at 37°C. A maximum of five non-sorbitol-fermenting (gray) colonies from each plate were tested for the presence of the 0157 antigen using an 0157 latex kit (Oxoid, Nepean, Ontario, Canada). Inoculation of feces with E. coli 0157:H7. Four strains of nalidixic acid-resistant E. coli 0157:H7 (E318N and R508N, R. Johnson, Public Health Agency of Canada, Guelph, Ontario; E32511N and H4420N, V. Gannon, Public Health Agency of Canada, Lethbridge, Alberta) were each grown in 100 ml of TSB containing 50 pg/liter nalidixic acid for 18 h at 37°C with agitation (150 rpm). Bacteria were sedimented by centrifugation (4,000 x g, 12 min), washed three times in phosphate-buffered saline (PBS; pH 7.2), and resuspended in PBS. Cells were adjusted with PBS to an optical density of 0.5 at 640 nm using a spectrophotometer (Genesys 20, Cole-Parmer, Montreal, Quebec, Canada) to reach an approximate concentration of 108 CFU/ml. The four strains of E. coli 0157:H7 were combined in equal numbers verified by enumeration on CT-SMAC containing 50 gg/ml nalidixic acid (CT-SMACnal). Feces (500 g) were collected during the growing phase from multiple cattle on each of the four treatments (0,4, 8, and 12 x 107 CFU Lactobacillus per kg of dietary DM), were mixed by hand, and were tested for presence of E. coli 0157 by immunomagnetic separation, as previously described. Feces negative for E. coli 0157 from each diet were then mixed in separate sterile stomacher bags for 8 min in a Stomacher 400 laboratory blender (Seward Medical, London, UK). Inoculum (3 ml) of the four-strain mixture of E. coli 0157:H7 was added to 297 g of feces to reach a concentration of 105 CFU E. coli 0157:H7 per g of feces and was mixed in stomacher bags at high speed for 4 min. Incubation of feces and enumeration and detection of E. coli 0157:H7. Triplicate stomacher bags of inoculated feces from cattle receiving each of the four treatments were incubated at 4 and 22°C. Stomacher bags were sampled on days 0, 1, 3, and 7 postinoculation and weekly thereafter for an 11-week experimental period. At each sampling time, subsamples of feces (10 g) were removed; 3 g was diluted in deionized water for determination of fecal pH using a PerpHect pH meter (model 301, Orion Research Inc., Boston, MA), and the remainder was oven dried for 24 h at 105°C for determination of fecal DM. At each sampling time, a 1-g fecal subsample was serially diluted in 9 ml of PBS and was enumerated by plating 100 pi of the appropriate dilution onto duplicate plates of CT-SMACnal, followed by incubation for 18 h at 37°C. For all microbial counts, plates yielding 30 to 300 colonies were used to determine E. coli 0157:H7 populations. Three non-sorbitol-fermenting colonies on each CT-SMACnal plate were tested using a latex kit (Oxoid), and PCR methodology described by Gannon et al. (17) was used to detect the presence of eae, stx, and fliC genes in 10% of latex positive isolates from both the feeding and inoculation studies. Isolates with all three genes were confirmed as E. coli 0157:H7. PFGE of isolates from DFM feeding study. All isolates from feces of DFM-fed cattle confirmed as E. coli 0157:H7 (n = 87) were subtyped by pulsed-field gel electrophoresis (PFGE) using Xbal restriction, according to the standard 1-day protocol as previously described by Bach et al. (4). Banding patterns were viewed with UV illumination and were photographed using the Speedlight Platinum Gel Documentation System (Bio-Rad, Mis sissauga, Ontario, Canada). Banding patterns in the digital images were classed as unique or were grouped into restriction endonuclease digestion pattern clusters (REPC; 90% or greater
740
J. Food Prot., Vol. 77, No. 5
STANFORD ET AL.
homology) using Dice similarity coefficients, unweighted pair group methods arithmetic average algorithm, 1% position tolerance, and 0.5% optimization (Bionumerics 3.5, Applied Maths BVBA, Sin-Martens-Latem, Belgium). Statistical analyses. For all statistical tests, significance was set at P < 0.05, using commercially available software (SAS 9.1, SAS Institute, Cary, NC). Steer was the experimental unit for binomially distributed data generated from immunomagnetic separa tion, and data were analyzed using a repeated measures analysis within the GLIMMIX procedure of SAS, with treatment and day of sampling in the model. For GLIMMIX analyses, model-adjusted means (back-transformed to original scale) were reported. Bacterial counts were log transformed and, along with fecal pH and DM, were compared using repeated measures analysis within the mixed model procedure of SAS. For mixed model analyses, treatment and incubation temperature were fixed effects, with day of sampling the repeated measure. Frequencies of steers shedding E. coli 0157:H7 single or multiple times within treatment groups were analyzed using Fisher’s exact test within the FREQ procedure of SAS.
4°C
Day
22 °C
RESULTS AND DISCUSSION Survival of E . coli 0157 :H7 in feces from DFM-fed cattle. A lthough changes in pH or volatile fatty acid concentrations in the bovine hindgut have been suggested to influence prevalence and survival o f E. coli 0 1 5 7 :H 7 (19), treatm ent w ith D FM s did not influence survival o f E. coli 0 1 5 7 :H 7 in feces (Fig. 1) or fecal pH (Fig. 2). Production of organic acids is thought to be a primary mechanism for lactobacilli to prevent colonization of the bovine gastrointestinal tract by pathogenic bacteria (33), but no difference in fecal pH was evident in feces from control steers and those offered DFMs. Similarly, antibacterial activities described for P. polymyxa (28) did not impact E. coli 0157:H 7 in feces o f DFM-fed cattle. Although not measured in the present study, fecal survival of lactobacilli and P. polymyxa may have been transient, and mechanisms whereby organisms in the DFM control E. coli 0157:H 7 may perhaps only be expressed within the gastroin testinal tract. Further investigation of DFM organism survival is warranted because enhanced viability o f organisms in feces could provide an additional mechanism for limiting contamina tion of the environment or carcasses with fecal-bome pathogens. In contrast to DFM treatm ent, incubation tem perature o f feces influenced survival o f E. coli 0 1 5 7 :H 7 and fecal pH. At 4°C, num bers o f E. coli 0 1 5 7 :H 7 declined linearly over tim e, reaching 2 log CFU /g by day 77. In contrast, num bers o f E. coli 0 1 5 7 :H 7 in feces incubated at 22°C rem ained equivalent to or higher than those on day 0 for the com plete experim ental period. R apid grow th o f E. coli 0 1 5 7 :H 7 resulted in equally rapid changes in pH o f feces. After an initial decline, feces incubated at 4°C linearly increased in pH over the experim ental period, from 6.0 to a pH o f 8.7 on day 77. In contrast, feces held at 22°C initially dropped to pH 5.4, rapidly rebounded to pH 8.1 by day 14, and then rem ained steady at a pH betw een 8.3 and 8.7 for the rem ainder o f the study. A lthough a pH < 6 .5 associated with feeding high-grain diets has reduced survival o f E. coli 0 1 5 7 :H 7 in feces (5), survival o f E. coli 0 1 5 7 :H 7 in the present study was not related to fecal pH, in accord w ith a previous study o f steers receiving barley-based diets (18).
Day FIGURE 1. Effect o f incubation temperature on survival o f naladixic acid-resistant E. coli 0157:H7 inoculated into feces (10s CFUIg) on day 0. Treatment groups: CON, diets without DFMs; DFM-4, 4 x 107 CFUIkg DM mixed culture o /L . casei and L. lactis; D FM S, 8 x 107 CFUIkg DM mixed culture o f L. casei and L. lactis; DFM -12,1.2 x 108 CFUIkg DM mixed culture o f L. casei and L. lactis. All DFMs also included an unknown concentration o f P. polymyxa. A lthough increased survival o f E. coli 0 1 5 7 :H 7 in feces incubated at 22°C com pared w ith 4°C w ould coincide with the enhanced prevalence observed in cattle feces during sum m er as com pared to w inter m onths (41), seasonal declines in shedding have also been observed in w arm er clim ates at am bient tem peratures that w ould support growth o f E. coli 0 1 5 7 :H 7 (12). A dditionally, “ reverse seasonal ity ” has been observed, in w hich fecal prevalence o f E. coli 0 1 5 7 :H 7 was higher in cattle in w inter as com pared with sum m er m onths (29, 39). A s show n in the present study, am bient tem perature does influence survival o f E. coli 0 1 5 7 :H 7 in feces; but, because m ultiple interacting factors likely im pact shedding o f E. coli 0 1 5 7 :H 7 by cattle and survival o f the organism in feces, the relationship am ong these factors and their relative im pact is as yet unclear. Results o f the present study w ould support those o f K udva et al. (21) and Echeverry et al. (11), who also reported a gradual decline num bers o f E. coli 0 1 5 7 :H 7 in
J. Food Prot, Vol. 77, No. 5
DFMS FOR CONTROL OF E. COLI 0157:H7 IN FEEDLOT CATTLE
4 °C
741
rv
x A
in rH
20 -
a
Day T re a tm e n t g roup2
FIGURE 3. Effect o f DFM treatment on mean prevalence o f E. coli 0157:H7 detected at 28-day intervals in feces o f steers (n = 120) during a 140-day finishing period. Treatment groups with different letters (a, b) differ (P < 0.05). Treatment groups: CON, diets without DFMs; D FM S, 4 x 107 CFUIkg DM mixed culture o f L. casei and L. lactis; D FM S, 8 x 107 CFUIkg DM mixed culture o f L. casei and L. lactis; DFM-12, 1.2 x lO 8 CFUIkg DM mixed culture o f L. casei and L. lactis. All DFMs also included an unknown concentration o /P . polymyxa.
22 °C
Day FIGURE 2. Effect o f incubation temperature on pH o f feces inoculated with 10s CFUIg naladixic acid-resistant E. coli 0157.H7 on day 0. Treatment groups: CON, diets without DFMs; DFM-4, 4 x 107 CFUIkg DM mixed culture o fL . casei and L. lactis; D FM S, 8 x 107 CFUIkg DM mixed culture o fh . casei and L. lactis; DFM-12, 1.2 x 10s CFUIkg DM mixed culture o fh . casei and L. lactis. All DFMs also included an unknown con centration ofP . polymyxa.
feces inoculated with 5 log CFU/g and incubated at 4°C; however, those researchers also reported declines in numbers of E. coli 0157:H7 in feces incubated at 23°C for 28 and 7 days, respectively. Long-term survival of E. coli 0157:H7 in feces (>150 days) has previously been reported. After incubation for 37 weeks at 20°C, E. coli 0157:H7 was detected in feces of steers receiving barleybased diets (18), whereas Bach et al. (5) reported stable populations of E. coli 0157:H7 in feces of steers incubated at 22°C for 12 weeks with no impact of grain source (barley versus com) on survival. Differences among studies in survival of E. coli 0157:H7 in feces are likely related to variation among strains used for inoculation and desiccation and other factors in feces (21). Desiccation of feces was minimized in the present study by incubation in sealed stomacher bags with minimal free air space, resulting in a relatively constant fecal DM over the experiment (data not shown). Results of the present study demonstrate long-term survival of E. coli 0157:H7 in feces over a range of
temperatures and fecal pH in the absence of desiccation. Because 4°C is used for storage of raw beef products (22), results of the present study also demonstrate that even after 77 days of refrigerated storage, E. coli 0157:H7 may be present in numbers sufficient to cause significant human disease. Impact of DFM on shedding E. coli 0157:H7 by feedlot cattle. Possibly due to seasonality of shedding, only one steer was positive for E. coli 0157:H7 during the winter months of the growing period, and results instead are presented for the spring and summer finishing period, when shedding of E. coli 0157:H7 is more prevalent in Alberta feedlot cattle (41). During the finishing period, both DFM-8 and DFM-12 treatments reduced (P < 0.05) prevalence of E. coli 0157:H7 in feces compared with control (Fig. 3). As well, frequency of shedding E. coli 0157:H7 by individual steers was also influenced by DFM treatment (P < 0.05). For control, numbers of steers with one, two, or three positive fecal samples during the finishing period did not differ, and one steer shed E. coli 0157:H7 on four occasions (Fig. 4). In contrast, for the DFM-12 group, 12 steers shed E. coli 0157 once, with only 1 steer shedding twice. The majority of cattle are thought to shed E. coli 0157:H7 intermittently and at low numbers (< 100 CFU/g feces (14)). Persistent or high-level shedders of E. coli 0157:H7 are thought to arise from enhanced colonization of the gastrointestinal tract compared with more typical, inter mittent shedders (8, 23). Colonization of the gastrointestinal tract with E. coli 0157:H7 may have been reduced by DFM8 and DFM-12 treatments, although pathogen transmission dynamics are influenced by the number of cattle in the pen (20) and would differ between individual and group-penned cattle. Because numerous factors may influence efficacy of DFMs, including stage of production and age of animals (7,
742
J. Food Prot., Vol. 77, No. 5
STANFORD ET AL.
s
a
g
e
s
*
*
*
StMT 169
REPC U
190 198 284 193 202 237 176 208 212 192 228 188 188 174
B B B C C U D D D E E F F U
Tnatnwnt* DFM-4
■ Steers positive once = Steers positive tw ice ■ Steers positive 3X ■ Steers positive 4X
T re a tm e n t G ro u p 2
FIGURE 4. Effect o f DFM treatment on cumulative frequency o f individual steers being fecal positive fo r E. coli 0157:H7 during the 140-day finishing period from fecal grab samples collected every 28 days. Columns within each treatment group with different letters (a, b, c) differ (P < 0.05). Treatment groups: CON, diets without DFMs; DFM-4, 4 x I0 7 CFUIkg DM mixed culture o fL . casei and L. lactis; D FM S, 8 x 107 CFUIkg DM mixed culture o f L. casei and L. lactis; DFM-12, 1.2 x 108 CFUIkg DM mixed culture o f L. casei and L. lactis. All DFMs also included an unknown concentration o f P. polymyxa. 13), further evaluation is warranted for DFM-8 and DFM-12 treatments in group-penned cattle. The relatively limited impact of DFM-4 compared with the other DFM treatments for control of E. coli 0157:H7 is not surprising because other Lactobacillus- based DFMs have demonstrated marked changes in efficacy over a 1- to 2-log range in dosage (40, 43, 44). Assuring delivery of precise levels of live microorganisms to all cattle within pens containing hundreds of animals is not without challenges. Compared to DFMs that are shipped and stored frozen to maintain microbial viability and are reconstituted before feeding (13), the DFM of the present study would be relatively easy to use under commercial feedlot conditions because it could be directly added to feed trucks without additional mixing or handling.
Impact of DFMs on PFGE patterns of E. coli 0157:H7 isolates. It was not possible to determine whether steers showed consistent transitions in subtypes of E. coli 0157:FT7 shed over time. Many steers shed E. coli 0157:H7 only once, and other steers shed multiple isolates of E. coli 0157:H7 within the same REPC (Fig. 5). The predominant REPC, REPC A, contained isolates from 79% of cattle shedding E. coli 0157:H7, and PFGE profiles of the isolates from cattle varied according to DFM treatment. For control and DFM-4 treatments, PFGE profiles were similar; 94 to 100% of these steers shed isolates of E. coli 0157:H7 belonging to REPC A. In contrast, isolates of E. coli 0157:H7 from DFM-8 and DFM-12 steers showed increasing divergence from the most predominant REPC. For DFM-8, 75% of steers shed E. coli 0157:H7 isolates only in REPC A, with this proportion falling to 50% for isolates from DFM-12.
DFM-12 DFM-12 DFM-8 CONTROL CONTROL CONTROL DFM-12 CONTROL DFM-4 DFM-12 DFM-12 DFM-8 DFM-8 DFM-4
FIGURE 5. Pulsed-field gel electrophoresis dendrogram showing impacts o f dietary treatments (CONTROL, diets without DFMs; D FM S, 4 x 107 CFUIkg DM mixed culture o f L. casei and L. lactis; D F M S, 8 x 107 CFUIkg DM mixed culture o f L. casei and L. lactis; DFM-12, 1.2 x 10s CFUIkg DM mixed culture o f L. casei and L. lactisj on restriction endonuclease pattern clusters (REPC) with 90% or greater similarity and unique isolates (
A Mixture of Lactobacillus casei, Lactobacillus lactis, and Paenibacillus polymyxa Reduces Escherichia coli 0157:H7 in Finishing Feedlot Cattle KIM STANFORD,' SUSAN BACH,2 JOHN BAAH,3 a n d TIM McALLISTER3* 'Alberta Agriculture and Rural Development, Lethbridge, Alberta, Canada T1J 4V6; Agriculture and Agri-Food Canada, Summerland, British Columbia, Canada VOH IZO; and 3Agriculture and Agri-Food Canada, Lethbridge, Alberta, Canada T1J 4V6 MS 13-433: Received 9 October 2013/Accepted 9 January 2014
ABSTRACT A direct-fed microbial (DFM) containing Paenibacillus polymyxa, Lactobacillus casei, and Lactobacillus lactis was fed to cattle (n = 120) to determine impacts on shedding and survival of Escherichia coli 0157:H7 in feces. Cattle were individually penned and fed diets containing 0 (control), 4 x 107 CFU (DFM-4), 8 x 107 CFU (DFM-8), or 1.2 x 108 CFU (DFM-12) lactobacilli per kg of dietary dry matter over 84-day fall-winter growing and 140-day spring-summer finishing periods. Fecal grab samples were collected from cattle at 28-day intervals, E. coli 0157:H7 was detected by immunomagnetic separation, and isolates were compared by pulsed-field gel electrophoresis. During the growing period, feces negative for E. coli 0157 from each dietary treatment were inoculated with 105 CFU/g nalidixic acid-resistant E. coli 0157:H7 and were incubated at 4 and 22°C for 11 weeks. Fecal pH and fecal dry matter were measured on days 0, 1,3, and 7 and weekly thereafter, with E. coli 0157:H7 enumerated through dilution plating. Treatment with DFMs did not affect survival of E. coli 0157:H7 in feces or fecal pH (P > 0.05). Only one steer was positive for E. coli 0157:H7 during the growing period, but during the finishing period, DFM-8 and DFM-12 reduced the prevalence of E. coli 0157:H7 in feces (P < 0.05). Feeding DFMs also reduced the frequency of individual steers shedding E. coli 0157:H7 during finishing (P < 0.05), with control steers shedding E. coli 0157:H7 up to four times, whereas DFM-12 steers shed E. coli 0157:H7 a maximum of twice. Treatment with DFMs influenced pulsed-field gel electrophoresis profiles; steers that were fed DFM-8 and DFM-12 shed more diverse subtypes of E. coli 0157:H7 than did control or DFM-4 steers. Because a companion study found linear improvement in performance with increasing dosage of DFMs in the first 28 days of the growing period, targeted use of DFM-12 during this time and for the final 1 or 2 weeks prior to slaughter may optimize performance and reduce E. coli 0157:H7 while minimizing feed costs.
Direct-fed microbials (DFMs) are live microorganisms that have been used to improve growth performance of cattle (3, 10) or to reduce colonization of the bovine gastrointestinal tract with zoonotic or other pathogens (7, 24). DFMs are thought to function by a number of mechanisms, including production of inhibitory compounds such as acids or bacteriocins (33), competitive exclusion (7), improvement of rumen fermentation parameters (3), the blocking of quorum sensing (25), improvement of host mucosal immunity (16, 34), or by other as yet undefined mechanisms (36). Because the microbial populations of the bovine gastrointestinal tract are exceedingly complex and many interrelationships among organisms are poorly characterized, the efficacy of DFMs has varied (2, 39, 40). Age of cattle impacts the efficacy of DFMs, with more positive outcomes in younger calves for growth perfor mance (13) and pathogen control (7). Young calves stressed after weaning and/or transportation may show an optimal response to DFMs (15). Dosage of live organisms in the * Author for correspondence. Tel: 403-317-2240; Fax 403-382-3156; E-mail: [email protected].
DFMs is critical, and efficacy for improving growth performance has varied within a range of 1 or 2 log CFU per head per day (43). As well, efficacy for growth improvement may vary according to the strain of both the organism in the DFM and the pathogen (1). The properties of the microbial species in the DFM, such as acid tolerance, adhesion to gut tissues, and survival in the final feed formulation, can also influence efficacy for both pathogen control and growth performance (34). Increased efficacy has been noted in DFMs containing multiple compared with single organisms, possibly due to synergistic mechanisms of action (33), and a DFM including Lactobacillus acidophilus and Propionibacterium freudenreichii has been widely adopted for improvement of growth performance in larger feedlots in the United States (7). Because few reports have identified DFMs that offer concurrent control of pathogens and improved growth performance in feedlot cattle, the primary objective of the current study was to determine the appropriate dosage of a novel DFM containing a mixture of Lactobacillus casei, Lactobacillus lactis, and Paenibacillus polymyxa for reduc tion of Escherichia coli 0157:H7 in feces of feedlot cattle
J. Food Prot., Vol. 77, No. 5
DFMS FOR CONTROL OF E. COLI 0157:H7 IN FEEDLOT CATTLE
over a complete feeding cycle. A companion study (3) demonstrated that this DFM improved both average daily gain and feed conversion efficiency of calves during the first 84 days of feeding. A secondary objective was to determine whether the DFM had a residual impact on the survival of E. coli 0157:H7 in experimentally inoculated feces. MATERIALS AND METHODS Animals, facilities, and diets. Cattle and diets were those used in the study of Baah et al. (3), and cattle were handled according to the guidelines of the Canadian Council on Animal Care (30). Briefly, 120 Hereford x Angus steers (initial body weight 280 + 15.5 kg) were purchased at auction and were transported to the individual feeding barn of the Agriculture and Agri-Food Canada Research Centre in Lethbridge, Alberta. Steers were individually weighed and randomly assigned to one of four treatment groups for the duration of the study: 0 (control), 4 x 107 CFU (DFM-4), 8 x 107 CFU (DFM-8), or 1.2 x 108 CFU(DFM12) of a mixture of L. casei and L. lactis per kg of dietary dry matter (DM). Paenibacillus polymyxa was also present in the DFMs (28), although numbers in the DFMs were not determined. Each treatment was housed in a separate wing of the facility to avoid cross-contamination among treatments. Steers were individ ually penned and watered without contact among animals and were given ad libitum access to feed and water. For the first 84 days (12 November through 2 February), steers received a growing diet consisting of 60% barley silage, 35% barley grain, and 5% vitamin-mineral supplement on a DM basis. Over the next 28 days, steers were transitioned to a finishing diet by gradually increasing the proportion of dietary barley to 86% and reducing barley silage to 9%. Steers received finishing diets for 140 days (3 February to 23 June) before shipment to slaughter. Other than the DFMs, no microbial or antimicrobial agents were included in diets. A Data Ranger feed mixer (American Calan, Northwood, NH) was used to prepare diets, starting with control and ending with the highest level of DFM. The feed mixer was thoroughly cleaned after each day’s feeding to avoid cross-contamination among treatments. DFM. The DFM (RE3, Basic Environmental Systems and Technology Inc., Edmonton, Alberta, Canada) was shipped in liquid form and stored at 5°C until it was added to diets. Dosage of DFM in diets was verified by enumeration of Lactobacillus spp. after serial dilution of 1 ml of DFM in 9 ml of buffered peptone water, plating in duplicate onto de Man Rogosa Sharpe agar, and incubation at 37°C for 48 h. Lactobacillus spp. colonies were small and translucent and were enumerated from plates containing 30 to 300 colonies. The DFM contained a mixed culture of L. casei, L. lactis, and P. polymyxa and also included fermentation products of these organisms. Collection of fecal samples and detection of E. coli 0157:H7 in cattle. Fecal grab samples (10 g) were rectally collected from each steer at the start of the growing phase of the study (day 0) and every 28 days thereafter until steers were shipped to slaughter. A sterile glove was used to collect each sample, with samples transported to the laboratory within 1 h of collection. Feces (1 g) were then enriched for 6 h at 37°C in modified tryptic soy broth containing 20.0 mg/liter novobiocin (Sigma, Oakville, Ontario, Canada) and 1.5 g/liter dipotassium phosphate. Immunomagnetic separation was performed using Dynabeads anti-E. coli 0157 (Dynal, Lake Success, NY) according to manufacturer’s instructions. A 50-pl aliquot of bead-bacteria mixture was plated onto sorbitol MacConkey agar supplemented with 2.5 mg/liter
739
potassium tellurite and 0.05 mg/liter cefixime (CT-SMAC), and plates were incubated for 18 to 24 h at 37°C. A maximum of five non-sorbitol-fermenting (gray) colonies from each plate were tested for the presence of the 0157 antigen using an 0157 latex kit (Oxoid, Nepean, Ontario, Canada). Inoculation of feces with E. coli 0157:H7. Four strains of nalidixic acid-resistant E. coli 0157:H7 (E318N and R508N, R. Johnson, Public Health Agency of Canada, Guelph, Ontario; E32511N and H4420N, V. Gannon, Public Health Agency of Canada, Lethbridge, Alberta) were each grown in 100 ml of TSB containing 50 pg/liter nalidixic acid for 18 h at 37°C with agitation (150 rpm). Bacteria were sedimented by centrifugation (4,000 x g, 12 min), washed three times in phosphate-buffered saline (PBS; pH 7.2), and resuspended in PBS. Cells were adjusted with PBS to an optical density of 0.5 at 640 nm using a spectrophotometer (Genesys 20, Cole-Parmer, Montreal, Quebec, Canada) to reach an approximate concentration of 108 CFU/ml. The four strains of E. coli 0157:H7 were combined in equal numbers verified by enumeration on CT-SMAC containing 50 gg/ml nalidixic acid (CT-SMACnal). Feces (500 g) were collected during the growing phase from multiple cattle on each of the four treatments (0,4, 8, and 12 x 107 CFU Lactobacillus per kg of dietary DM), were mixed by hand, and were tested for presence of E. coli 0157 by immunomagnetic separation, as previously described. Feces negative for E. coli 0157 from each diet were then mixed in separate sterile stomacher bags for 8 min in a Stomacher 400 laboratory blender (Seward Medical, London, UK). Inoculum (3 ml) of the four-strain mixture of E. coli 0157:H7 was added to 297 g of feces to reach a concentration of 105 CFU E. coli 0157:H7 per g of feces and was mixed in stomacher bags at high speed for 4 min. Incubation of feces and enumeration and detection of E. coli 0157:H7. Triplicate stomacher bags of inoculated feces from cattle receiving each of the four treatments were incubated at 4 and 22°C. Stomacher bags were sampled on days 0, 1, 3, and 7 postinoculation and weekly thereafter for an 11-week experimental period. At each sampling time, subsamples of feces (10 g) were removed; 3 g was diluted in deionized water for determination of fecal pH using a PerpHect pH meter (model 301, Orion Research Inc., Boston, MA), and the remainder was oven dried for 24 h at 105°C for determination of fecal DM. At each sampling time, a 1-g fecal subsample was serially diluted in 9 ml of PBS and was enumerated by plating 100 pi of the appropriate dilution onto duplicate plates of CT-SMACnal, followed by incubation for 18 h at 37°C. For all microbial counts, plates yielding 30 to 300 colonies were used to determine E. coli 0157:H7 populations. Three non-sorbitol-fermenting colonies on each CT-SMACnal plate were tested using a latex kit (Oxoid), and PCR methodology described by Gannon et al. (17) was used to detect the presence of eae, stx, and fliC genes in 10% of latex positive isolates from both the feeding and inoculation studies. Isolates with all three genes were confirmed as E. coli 0157:H7. PFGE of isolates from DFM feeding study. All isolates from feces of DFM-fed cattle confirmed as E. coli 0157:H7 (n = 87) were subtyped by pulsed-field gel electrophoresis (PFGE) using Xbal restriction, according to the standard 1-day protocol as previously described by Bach et al. (4). Banding patterns were viewed with UV illumination and were photographed using the Speedlight Platinum Gel Documentation System (Bio-Rad, Mis sissauga, Ontario, Canada). Banding patterns in the digital images were classed as unique or were grouped into restriction endonuclease digestion pattern clusters (REPC; 90% or greater
740
J. Food Prot., Vol. 77, No. 5
STANFORD ET AL.
homology) using Dice similarity coefficients, unweighted pair group methods arithmetic average algorithm, 1% position tolerance, and 0.5% optimization (Bionumerics 3.5, Applied Maths BVBA, Sin-Martens-Latem, Belgium). Statistical analyses. For all statistical tests, significance was set at P < 0.05, using commercially available software (SAS 9.1, SAS Institute, Cary, NC). Steer was the experimental unit for binomially distributed data generated from immunomagnetic separa tion, and data were analyzed using a repeated measures analysis within the GLIMMIX procedure of SAS, with treatment and day of sampling in the model. For GLIMMIX analyses, model-adjusted means (back-transformed to original scale) were reported. Bacterial counts were log transformed and, along with fecal pH and DM, were compared using repeated measures analysis within the mixed model procedure of SAS. For mixed model analyses, treatment and incubation temperature were fixed effects, with day of sampling the repeated measure. Frequencies of steers shedding E. coli 0157:H7 single or multiple times within treatment groups were analyzed using Fisher’s exact test within the FREQ procedure of SAS.
4°C
Day
22 °C
RESULTS AND DISCUSSION Survival of E . coli 0157 :H7 in feces from DFM-fed cattle. A lthough changes in pH or volatile fatty acid concentrations in the bovine hindgut have been suggested to influence prevalence and survival o f E. coli 0 1 5 7 :H 7 (19), treatm ent w ith D FM s did not influence survival o f E. coli 0 1 5 7 :H 7 in feces (Fig. 1) or fecal pH (Fig. 2). Production of organic acids is thought to be a primary mechanism for lactobacilli to prevent colonization of the bovine gastrointestinal tract by pathogenic bacteria (33), but no difference in fecal pH was evident in feces from control steers and those offered DFMs. Similarly, antibacterial activities described for P. polymyxa (28) did not impact E. coli 0157:H 7 in feces o f DFM-fed cattle. Although not measured in the present study, fecal survival of lactobacilli and P. polymyxa may have been transient, and mechanisms whereby organisms in the DFM control E. coli 0157:H 7 may perhaps only be expressed within the gastroin testinal tract. Further investigation of DFM organism survival is warranted because enhanced viability o f organisms in feces could provide an additional mechanism for limiting contamina tion of the environment or carcasses with fecal-bome pathogens. In contrast to DFM treatm ent, incubation tem perature o f feces influenced survival o f E. coli 0 1 5 7 :H 7 and fecal pH. At 4°C, num bers o f E. coli 0 1 5 7 :H 7 declined linearly over tim e, reaching 2 log CFU /g by day 77. In contrast, num bers o f E. coli 0 1 5 7 :H 7 in feces incubated at 22°C rem ained equivalent to or higher than those on day 0 for the com plete experim ental period. R apid grow th o f E. coli 0 1 5 7 :H 7 resulted in equally rapid changes in pH o f feces. After an initial decline, feces incubated at 4°C linearly increased in pH over the experim ental period, from 6.0 to a pH o f 8.7 on day 77. In contrast, feces held at 22°C initially dropped to pH 5.4, rapidly rebounded to pH 8.1 by day 14, and then rem ained steady at a pH betw een 8.3 and 8.7 for the rem ainder o f the study. A lthough a pH < 6 .5 associated with feeding high-grain diets has reduced survival o f E. coli 0 1 5 7 :H 7 in feces (5), survival o f E. coli 0 1 5 7 :H 7 in the present study was not related to fecal pH, in accord w ith a previous study o f steers receiving barley-based diets (18).
Day FIGURE 1. Effect o f incubation temperature on survival o f naladixic acid-resistant E. coli 0157:H7 inoculated into feces (10s CFUIg) on day 0. Treatment groups: CON, diets without DFMs; DFM-4, 4 x 107 CFUIkg DM mixed culture o /L . casei and L. lactis; D FM S, 8 x 107 CFUIkg DM mixed culture o f L. casei and L. lactis; DFM -12,1.2 x 108 CFUIkg DM mixed culture o f L. casei and L. lactis. All DFMs also included an unknown concentration o f P. polymyxa. A lthough increased survival o f E. coli 0 1 5 7 :H 7 in feces incubated at 22°C com pared w ith 4°C w ould coincide with the enhanced prevalence observed in cattle feces during sum m er as com pared to w inter m onths (41), seasonal declines in shedding have also been observed in w arm er clim ates at am bient tem peratures that w ould support growth o f E. coli 0 1 5 7 :H 7 (12). A dditionally, “ reverse seasonal ity ” has been observed, in w hich fecal prevalence o f E. coli 0 1 5 7 :H 7 was higher in cattle in w inter as com pared with sum m er m onths (29, 39). A s show n in the present study, am bient tem perature does influence survival o f E. coli 0 1 5 7 :H 7 in feces; but, because m ultiple interacting factors likely im pact shedding o f E. coli 0 1 5 7 :H 7 by cattle and survival o f the organism in feces, the relationship am ong these factors and their relative im pact is as yet unclear. Results o f the present study w ould support those o f K udva et al. (21) and Echeverry et al. (11), who also reported a gradual decline num bers o f E. coli 0 1 5 7 :H 7 in
J. Food Prot, Vol. 77, No. 5
DFMS FOR CONTROL OF E. COLI 0157:H7 IN FEEDLOT CATTLE
4 °C
741
rv
x A
in rH
20 -
a
Day T re a tm e n t g roup2
FIGURE 3. Effect o f DFM treatment on mean prevalence o f E. coli 0157:H7 detected at 28-day intervals in feces o f steers (n = 120) during a 140-day finishing period. Treatment groups with different letters (a, b) differ (P < 0.05). Treatment groups: CON, diets without DFMs; D FM S, 4 x 107 CFUIkg DM mixed culture o f L. casei and L. lactis; D FM S, 8 x 107 CFUIkg DM mixed culture o f L. casei and L. lactis; DFM-12, 1.2 x lO 8 CFUIkg DM mixed culture o f L. casei and L. lactis. All DFMs also included an unknown concentration o /P . polymyxa.
22 °C
Day FIGURE 2. Effect o f incubation temperature on pH o f feces inoculated with 10s CFUIg naladixic acid-resistant E. coli 0157.H7 on day 0. Treatment groups: CON, diets without DFMs; DFM-4, 4 x 107 CFUIkg DM mixed culture o fL . casei and L. lactis; D FM S, 8 x 107 CFUIkg DM mixed culture o fh . casei and L. lactis; DFM-12, 1.2 x 10s CFUIkg DM mixed culture o fh . casei and L. lactis. All DFMs also included an unknown con centration ofP . polymyxa.
feces inoculated with 5 log CFU/g and incubated at 4°C; however, those researchers also reported declines in numbers of E. coli 0157:H7 in feces incubated at 23°C for 28 and 7 days, respectively. Long-term survival of E. coli 0157:H7 in feces (>150 days) has previously been reported. After incubation for 37 weeks at 20°C, E. coli 0157:H7 was detected in feces of steers receiving barleybased diets (18), whereas Bach et al. (5) reported stable populations of E. coli 0157:H7 in feces of steers incubated at 22°C for 12 weeks with no impact of grain source (barley versus com) on survival. Differences among studies in survival of E. coli 0157:H7 in feces are likely related to variation among strains used for inoculation and desiccation and other factors in feces (21). Desiccation of feces was minimized in the present study by incubation in sealed stomacher bags with minimal free air space, resulting in a relatively constant fecal DM over the experiment (data not shown). Results of the present study demonstrate long-term survival of E. coli 0157:H7 in feces over a range of
temperatures and fecal pH in the absence of desiccation. Because 4°C is used for storage of raw beef products (22), results of the present study also demonstrate that even after 77 days of refrigerated storage, E. coli 0157:H7 may be present in numbers sufficient to cause significant human disease. Impact of DFM on shedding E. coli 0157:H7 by feedlot cattle. Possibly due to seasonality of shedding, only one steer was positive for E. coli 0157:H7 during the winter months of the growing period, and results instead are presented for the spring and summer finishing period, when shedding of E. coli 0157:H7 is more prevalent in Alberta feedlot cattle (41). During the finishing period, both DFM-8 and DFM-12 treatments reduced (P < 0.05) prevalence of E. coli 0157:H7 in feces compared with control (Fig. 3). As well, frequency of shedding E. coli 0157:H7 by individual steers was also influenced by DFM treatment (P < 0.05). For control, numbers of steers with one, two, or three positive fecal samples during the finishing period did not differ, and one steer shed E. coli 0157:H7 on four occasions (Fig. 4). In contrast, for the DFM-12 group, 12 steers shed E. coli 0157 once, with only 1 steer shedding twice. The majority of cattle are thought to shed E. coli 0157:H7 intermittently and at low numbers (< 100 CFU/g feces (14)). Persistent or high-level shedders of E. coli 0157:H7 are thought to arise from enhanced colonization of the gastrointestinal tract compared with more typical, inter mittent shedders (8, 23). Colonization of the gastrointestinal tract with E. coli 0157:H7 may have been reduced by DFM8 and DFM-12 treatments, although pathogen transmission dynamics are influenced by the number of cattle in the pen (20) and would differ between individual and group-penned cattle. Because numerous factors may influence efficacy of DFMs, including stage of production and age of animals (7,
742
J. Food Prot., Vol. 77, No. 5
STANFORD ET AL.
s
a
g
e
s
*
*
*
StMT 169
REPC U
190 198 284 193 202 237 176 208 212 192 228 188 188 174
B B B C C U D D D E E F F U
Tnatnwnt* DFM-4
■ Steers positive once = Steers positive tw ice ■ Steers positive 3X ■ Steers positive 4X
T re a tm e n t G ro u p 2
FIGURE 4. Effect o f DFM treatment on cumulative frequency o f individual steers being fecal positive fo r E. coli 0157:H7 during the 140-day finishing period from fecal grab samples collected every 28 days. Columns within each treatment group with different letters (a, b, c) differ (P < 0.05). Treatment groups: CON, diets without DFMs; DFM-4, 4 x I0 7 CFUIkg DM mixed culture o fL . casei and L. lactis; D FM S, 8 x 107 CFUIkg DM mixed culture o f L. casei and L. lactis; DFM-12, 1.2 x 108 CFUIkg DM mixed culture o f L. casei and L. lactis. All DFMs also included an unknown concentration o f P. polymyxa. 13), further evaluation is warranted for DFM-8 and DFM-12 treatments in group-penned cattle. The relatively limited impact of DFM-4 compared with the other DFM treatments for control of E. coli 0157:H7 is not surprising because other Lactobacillus- based DFMs have demonstrated marked changes in efficacy over a 1- to 2-log range in dosage (40, 43, 44). Assuring delivery of precise levels of live microorganisms to all cattle within pens containing hundreds of animals is not without challenges. Compared to DFMs that are shipped and stored frozen to maintain microbial viability and are reconstituted before feeding (13), the DFM of the present study would be relatively easy to use under commercial feedlot conditions because it could be directly added to feed trucks without additional mixing or handling.
Impact of DFMs on PFGE patterns of E. coli 0157:H7 isolates. It was not possible to determine whether steers showed consistent transitions in subtypes of E. coli 0157:FT7 shed over time. Many steers shed E. coli 0157:H7 only once, and other steers shed multiple isolates of E. coli 0157:H7 within the same REPC (Fig. 5). The predominant REPC, REPC A, contained isolates from 79% of cattle shedding E. coli 0157:H7, and PFGE profiles of the isolates from cattle varied according to DFM treatment. For control and DFM-4 treatments, PFGE profiles were similar; 94 to 100% of these steers shed isolates of E. coli 0157:H7 belonging to REPC A. In contrast, isolates of E. coli 0157:H7 from DFM-8 and DFM-12 steers showed increasing divergence from the most predominant REPC. For DFM-8, 75% of steers shed E. coli 0157:H7 isolates only in REPC A, with this proportion falling to 50% for isolates from DFM-12.
DFM-12 DFM-12 DFM-8 CONTROL CONTROL CONTROL DFM-12 CONTROL DFM-4 DFM-12 DFM-12 DFM-8 DFM-8 DFM-4
FIGURE 5. Pulsed-field gel electrophoresis dendrogram showing impacts o f dietary treatments (CONTROL, diets without DFMs; D FM S, 4 x 107 CFUIkg DM mixed culture o f L. casei and L. lactis; D F M S, 8 x 107 CFUIkg DM mixed culture o f L. casei and L. lactis; DFM-12, 1.2 x 10s CFUIkg DM mixed culture o f L. casei and L. lactisj on restriction endonuclease pattern clusters (REPC) with 90% or greater similarity and unique isolates (
