Location: Meat Safety and Quality
2025 Annual Report
Objectives
Objective 1: Evaluate longitudinal ecology of foodborne pathogens in food-animal production continuum.
Sub-objective 1.A: Determine the population diversity of Shiga toxigenic Escherichia coli in a closed cattle feedlot.
Sub-objective 1.B: Determine the population dynamics of Salmonella at cattle feeding operations.
Objective 2: Application of bioinformatic tools to identify factors that contribute to virulence and persistence in foodborne pathogens.
Sub-objective 2.A: Development of machine learning approaches for predicting Shiga toxigenic E. coli and Salmonella pathogenicity in humans.
Sub-objective 2.B: In vitro pathogenicity assays and transcriptomic analyses to examine putative virulence factor contribution to Salmonella enterica pathogenicity, in order to increase our understanding of the strains encountered in production agriculture that have the greatest potential impact on human health.
Sub-objective 2.C: Characterization of environmental impacts on pathogen resistance to antimicrobials and sanitizers.
Objective 3: Development and validation of tools that enable regulators, food-animal producers and processors to monitor high-risk foodborne pathogens.
Approach
Foodborne illness and the resulting loss of productivity in the United States are reportedly greater than $14 billion a year. While research efforts have resulted in significant strides in tracking contamination entry points and identifying mitigation strategies, outlier events continue to occur, and complete prevention of foodborne pathogens entering the food chain remains an elusive goal. Moreover, concerns persist among regulators and health care advocates that antimicrobial use during animal production may impact antimicrobial resistance levels and potential for transfer to foodborne pathogens. Accordingly, the research described here aims to provide new information about these issues by 1) increasing our understanding of both the genomic diversity and persistence of pathogens over space and time in agricultural settings, which will improve foodborne illness traceback investigations 2) improve understanding of the movement of antimicrobial resistance genes among natural reservoirs and foodborne pathogens; 3) using machine learning to identify predictive markers that can be used to rapidly screen samples or isolates for important phenotypic characteristics and further phenotypically characterizing strains predicted to be more pathogenic or persistent in production settings; and 4) developing tools to monitor high-risk foodborne pathogens, including methods for rapidly estimating levels of foodborne pathogens in meat products. The information generated by this research will facilitate the development of solutions to decrease the incidence of pathogen exposure from the meat food chain. The results of this research will be of interest to food regulatory agencies, the pathogen testing industry, livestock and meat processing industries, agricultural and biomedical scientists, and public health professionals. Major beneficiaries of the successful realization and manifestation of the research goals would ultimately be consumers of a safer food supply.
Progress Report
We completed the fourth sample collection at the U.S. Meat Animal Research Center (USMARC) feedlot in Clay Center, Nebraska, and cultured Shiga toxin-containing Escherichia coli (E. coli)O157:H7 (STEC O157:H7) from the samples. The DNA from these strains will be extracted and sequenced during fiscal year (FY) 2026. The pens with culturable STEC O157:H7 in 2024, had a prevalence of 18.9% which is almost half the prevalence of years 2021, 2022, and 2023. The same trend held true with the overall feedlot prevalence of STEC O157:H7 in 2024 being half that of previous years. Weather conditions play an important part in STEC O157:H7 prevalence. The weather for the previous three years has been consistent and unseasonably dry. Consequently, calves have weaned into the feedlot at an earlier age than normal due to drought conditions. However, for 2024, there were timely rains that allowed the cattle to be on pasture until the normal weaning time in September. The additional moisture should have increased the overall prevalence of STEC O157:H7 in the feedlot, but this wasn’t the case. The overall prevalence was lower during each of the sampling months compared to the previous years. We assumed that weather was one of the main drivers in STEC O157:H7 prevalence in our feedlot and we still think that assumption is correct. However, there appears to be additional drivers that we are not measuring that is having an impact on prevalence. Another driver of STEC O157:H7 prevalence that we see in the feedlot is the percentage of pens with cattle will increase the number of pens with STEC O157:H7. It remains to be seen what our weather pattern will be for next season and if we can identify any additional drivers of STEC O157:H7 prevalence in our feedlot. The prevalence of STEC O157:H7 in calves weaned into the feed yard in 2024 was similar to 2021 but less than the high of that seen in 2023. The prevalence from the 8 weaning groups was quite variable with two groups accounting for 84% of the positive samples and 4 groups having no positive samples. When looking at cattle movement of the two high prevalence herds before weaning, they were grazing in pastures and not supplemented with silage like previous years. Calves from one herd have consistently been shedding STEC O157:H7 at weaning. The prevalence of 35% STEC O157:H7 positive calves is significantly higher than the published average prevalence of between 5 and 10%. Interestingly, the weaned calves from this herd were shedding an STEC O157:H7 strain that we haven’t seen during previous STEC O157:H7 samplings. This strain has a different virulence gene genotype not seen previously at USMARC. This isolate was cultured from feedlot pens where the calves were weaned into during the last pen surface sampling. It will be interesting to see how this strain competes with the resident clade 2 strains in the feedlot to see if it is able to persist.
The sequencing and analysis were completed for year four pen surface samples and weaned calf fecal samples. A total of 384 strains from the 2024 sampling were sequenced that represent the 185 cultured strains and additional strains picked from the same plate as the original culture strains. The sequencing results from 2024 were intermediate to the previous years. Clade 2 represented 84% of the sequenced isolates compared to 90% for 2023 and 72% for 2022. The percentage of clade 3 strains from 2024 were like 2022 with 5% and 6.8%, respectively. This is different from 2023 when 19% of the strains belonged to clade 3. Unlike 2022 and 2023, no clade 4 strains were found in the 2024 samples. An example of the dominance of clade 2 strains during the 2023 and 2024 sampling was demonstrated by the lack of strains from other clades in the second, third, fourth and fifth sampling times. Strains from other clades were only cultured from pen surface material after calves were weaned into the feedlot and the first feedlot sampling of the next year.
To identify targets responsible for STEC O26:H11 virulence and identification, complete closed genome sequences from 84 STEC O26:H11 strains were either sequenced or downloaded from the National Center for Biotechnology Information (NCBI) database. These isolates were placed into two groups according to their designation as human or environmental, sequence type (ST) 21 or ST29, and whether they contained a Shiga toxin gene or not. Each group represents a different method of identifying if a strain can cause disease in humans. The groupings were analyzed using a microbial pan-genome wide association (GWAS) study to identify genes that associate with characteristics from each group. Effector proteins are important in the pathogenicity of bacteria as they promote colonization of and persistence in the host. There were 2,811 core genes (found in all strains) and 9,349 accessory genes (found in one or more strains but not all) identified from the 84 STEC O26:H11. The GWAS study comparing the source being either human or environmental or whether it contained a Shiga toxin gene or not did not identify any genes that were significantly associated with either (Bonferroni p=<0.05). When this same analysis was run with the source strains for STEC O157:H7, there were seven genes that associated with isolates from human strains. This indicates that those genes may play a role in STEC O157:H7 ability to cause disease in humans. For STEC O26:H11 it appears that there is no one set of genes responsible for causing disease in humans but there are different sets of genes depending upon the strains. Humans infected with ST29 have a higher propensity to have more severe disease than those with ST21. This is thought to be due to the presence of Shiga toxin 2 in ST29 strains. When ST21 was compared to ST29, there was a strong association of Shiga toxin 1 genes, bacteriophage associated genes, type VI secretion genes and hypothetical genes but none of them were unique to one sequence type or the other. Interestingly, Shiga toxin 2 wasn’t associated with ST29. Overall, the results of these computational analyses suggest that there isn’t one virulence factor that is responsible for causing disease, rather it is likely a combination of factors within each STEC O26:H11 strain that contributes to the success of that pathogen and its ability to cause illness in humans.
Our previous results suggest that the multispecies bacterial communities at processing plants have higher influence on pathogen tolerance/adaptation to sanitizers than the pathogen strains’ intrinsic properties. Sanitizer effectiveness could be affected by the interactions between the pathogens and the natural multispecies microorganisms often persisting as mixed biofilms whose compositions may vary depending upon the plant processing activities, animal types/sources, and the selective pressure caused by the daily cleaning practice and sanitization reagents. Therefore, we shifted our aim to use metagenomic analysis of the microbial composition of the mixed biofilm communities isolated from beef and pork plants to understand how the different microbial communities might influence pathogen survival and adaption to sanitization. To that end, we characterized natural microorganisms collected from floor drains at various areas at multiple beef and pork plants and analyzed their impact on pathogen sanitizer tolerance. We observed the pathogens were able to integrate efficiently into the multispecies biofilms attached on contact surfaces even under low temperatures commonly seen in processing facilities. S. enterica colonized in mixed biofilms more effectively than E. coli O157:H7. Scanning electron microscope analysis showed that the contact surface topography may impact the mixed biofilm morphology and bacterial tolerance. Interestingly, after treatment with the multicomponent sanitizer, overall higher survival and post-sanitization recovery of the pathogen cells were observed in the treated pork plant samples than those in the beef plant samples. Metagenomic analysis of the multispecies bacterial communities showed that Pseudomonadaceae, Halomonadaceae and Enterobacteriaceae were the three most abundant families across all samples. No significant difference in genus compositions between the beef and pork plants or among the drain areas was observed. However, substantial variations in the percentages of species’ relative abundance were observed among the samples. Our study indicates that sanitization processes and the resulting pathogen inactivation and prevalence prevention that are described for the different types of the processing facilities should be analyzed on a case-by-case basis to provide practical information for meat processors to monitor and prevent contamination at plants.
Accomplishments
1. Development and validation of an assay to detect more human pathogenic versions of Salmonella. Current detection of Salmonella is based on serotype which is a characteristic of the surface of the bacterium. While this catches most of the virulent Salmonella, it doesn’t catch all of them and requires costly and laborious testing. ARS scientists in Clay Center, Nebraska, and a commercial partner developed an assay that uses multiple targets that are found in virulent Salmonella regardless of serotype. The commercial prototype assay shows high accuracy and repeatability. The highly pathogenic Salmonella assay is scheduled for commercial release in the 4th quarter of 2025 and is anticipated to improve human health by facilitating removal and thereby reducing exposure to versions of Salmonella more likely to cause human illness. This assay will provide revolutionary improvement in detecting and mitigating pathogenic Salmonella in the food chain and lead to reductions in the 1.35 million foodborne illnesses caused by Salmonella annually in the United States.
Review Publications
Harhay, D.M., Brader, K.D., Katz, T.S., Harhay, G.P., Bono, J.L., Bosilevac, J.M., Wheeler, T.L. 2025. A novel approach for detecting Salmonella enterica strains frequently attributed to human illness - development and validation of the highly pathogenic Salmonella (HPS) multiplex PCR assay. Frontiers in Microbiology. 15. Article 1504621. https://doi.org/10.3389/fmicb.2024.1504621.
Schmidt, J.W., Carlson, A., Bosilevac, J.M., Harhay, D., Arthur, T.M., Brown, T., Wheeler, T.L., Vipham, J.L. 2024. Evaluation of methods for identifying poultry wing rinses with Salmonella concentrations greater than or equal to 10 CFU/mL. Journal of Food Protection. 87(11). Article 100362. https://doi.org/10.1016/j.jfp.2024.100362.
Schmidt, J.W., Wu, W., Harhay, D.M., Wheeler, T.L. 2025. Identification of chicken component samples containing Salmonella concentrations greater than or equal to 1 CFU/g. Meat and Muscle Biology. 9(1). Article 18993. https://doi.org/10.22175/mmb.18993.