|dc.description.abstract||Animal feed can potentially expose animals to pathogens which may lead to infection or colonization of the gastrointestinal tract (GIT). Feeds can become contaminated with bacterial pathogens during harvesting/transportation of the ingredients, processing at the feed mill, transportation to the farm or during storage. Pathogenic bacteria have the potential to colonize the GIT of animals leading to either disease of the animal or contamination of the carcasses during processing. The objectives of these studies were to: (1) establish if commercially manufactured animal feed serves as a source of contamination of Salmonella, E. coli and Clostridium perfringens, and (2) to identify the unknown group of bacteria with Clostridium like characteristics found in animal feeds and determine its pathogenicity when used in a necrotic enteritis model. Four commercial feed mills and the research feed mill at Auburn University were sampled. A total of 292 samples (132 of feed ingredients and 160 of mixed feeds) were collected over two sampling periods (Nov-Mar and Apr-Oct) from 5 different locations within each feed mill: ingredient receiving, post mixing, post pelleting (hot sample), post cooling and at loadout. All samples were assayed for Salmonella, E. coli and C. perfringens using selective media. Resulting colony forming unit counts were log10 transformed and then analyzed using ANOVA, if significant (P<0.05), means were separated using Tukey HSD. Additionally, four typical colonies for Clostridial spp. were isolated from each sample and cultured onto blood agar plates to determine the ability of the isolates to produce typical hemolysis. During this isolation process, an unknown group of bacteria with similar characteristics to those of the Clostridium genus were identified. These bacteria were rod-shaped, gram positive, grew under anaerobic conditions, formed spores, produced lecithinase, and some produced double-zone beta hemolysis on sheep blood agar, but, when tested at a molecular level by PCR, they did not have the alpha-toxin gene characteristic for Clostridium perfringens. Consequently, some of those isolates were selected for 16S rRNA gene sequencing analysis for bacterial identification.
For experiment (1), the data shows that the Clostridial spp. counts were significantly higher on the feed ingredients: peanut meal and corn gluten meal (3.91 log10 and 2.61 log10, respectively), and the E. coli counts were significantly higher on the feed ingredients: peanut meal and corn meal (4.15 log10 and 2.85 log10, respectively) when compared to the rest of the feed ingredients sampled (P<0.05). When the contamination levels of ingredients between feed mills were compared, the corn meal samples collected from feed mills B, C and E had the highest Clostridial spp. counts from both sampling periods. No statistical difference was observed on the E. coli counts for corn meal between feed mills during both sampling periods. For soybean meal, the samples collected between Nov-Mar were higher on the E. coli counts on feed mill A (1.00 log10), and for the samples collected between Apr-Oct feed mill C was the most contaminated on both, Clostridial spp. counts and E. coli counts (2.48 log10 and 2.18 log10, respectively). For the processed samples, it was determined that feed mills B and E had the highest Clostridial spp. counts during both sampling periods in all the stages of processing, and for the E. coli counts, only the post mixing and post pelleting stages showed a significant difference being feed mill C (Nov-Mar) the least contaminated on the post mixing stage (1.65 log10) and feed mill B (Apr-Oct) the most contaminated on the post pelleting stage of processing (1.58 log10). Recontamination with E. coli after the pelleting process was observed for feed mills A and B. Salmonella contamination was not detected in the feed ingredients or processed feed throughout the study. Finally, from the processed feed C. perfringens was isolated and confirmed through PCR from two different samples obtained from feed mills A and D. For experiment (2), the sequencing results identified the different isolates as Clostridium argentinense—which is a bacterium ubiquitous to soil that can produce a neuroparalytic toxin— or Bacillus proteolyticus—which is a bacterium associated to marine environments including fish processing wastes. The necrotic enteritis model trial showed that strain of B. proteolyticus evaluated was not able to induce disease and that the C. perfringens isolate recovered from feed mill D was able to induce slight necrotic enteritis lesions (1 score). For future research, greater attention should be placed on E. coli and Clostridial spp. contaminated animal feed as a possible source of disease transmission to the animals and the pathogenicity of Clostridium argentinense should be evaluated on a live animal trial.||en_US