Research Project: Alternatives to Antibiotics and Genomics of Antimicrobial Resistance to Control Foodborne Pathogens in Poultry

Location: Poultry Microbiological Safety and Processing Research Unit

2024 Annual Report

Objectives
1. Identify and determine the presence and contributing factors for antimicrobial resistant foodborne bacteria in poultry and poultry-associated environments. 1.1. Determine the association of antimicrobial resistance (AR) in foodborne bacteria with resistance to biocides, metals, coccidiostats, and ionophores used in poultry husbandry and processing. 1.2. Evaluate the bacterial metagenome of retail poultry. 1.3. Identify and evaluate markers (resistance genes, genetic elements, virulence genes) to define outbreak and persistent foodborne bacteria in poultry. 1.4. Identify antimicrobial resistance gene cassettes (ARCs) and accumulation on plasmids. 2. Identify and evaluate biological and chemical intervention products and alternatives to antimicrobials to control or reduce foodborne pathogens in poultry. 2.1. Develop, validate, and produce multi-subunit vaccines to control Salmonella and Campylobacter in broiler chickens. 2.2. Develop antimicrobial peptides (AMP) as alternatives to antibiotics to reduce foodborne pathogens associated with poultry. 2.3. Identify and develop broad-spectrum bacteriocins to eliminate foodborne pathogens in poultry. 2.4. Utilize phage isolation, whole-genome sequencing (WGS), and metagenomics to identify lytic phage that target Salmonella and pathogenic Escherichia coli.

Approach
Microbial contamination of food products from poultry continues to be a leading cause of foodborne illness. Antibiotics have been used to treat bacterial infections since the mid-twentieth century. Because of their efficacy in treating and preventing disease, antimicrobials have also been widely used in poultry production contributing to antimicrobial resistance (AR) in foodborne pathogens and commensal bacteria. AR among these bacteria has the potential to compromise therapy and remains a global threat to human health. This research project represents a merger of two teams of scientists to provide solutions to colonization of poultry with human pathogens and AR in foodborne pathogens and commensal bacteria from poultry. Two major approaches will be employed: 1) development of alternatives to antibiotics for use in combating foodborne pathogens, and 2) investigations to accurately understand attributes of antimicrobial resistant foodborne pathogens and commensals. Alternatives to antibiotics include vaccines to control foodborne pathogens in live birds while innovative antimicrobial peptides, bacteriocins, and lytic phage will modulate the poultry microbiome to reduce or eliminate colonization by harmful bacteria from poultry to minimize AR and reduce risk to human health. Data generated on resistance to biocides, metals, coccidiostats, and ionophores used in poultry production and processing is a specific concern to the USDA Food Safety and Inspection Service (FSIS). Research designed to determine ecological niches of foodborne bacteria and identify genetic characteristics facilitating transfer of resistance or a fitness advantage will also benefit FSIS. According to FSIS, increased knowledge of the microbial ecology of antimicrobial resistant pathogens on poultry will result in data that the poultry industry can utilize in development of improved pathogen management strategies. Identification of genetic markers which support survival, persistence, and dissemination of foodborne pathogens, especially those that are resistant to antimicrobials, is critical to this research priority. Data and technology from the proposed research will be used to assist other Federal agencies and the poultry and agricultural biotechnology industry in addressing AR in poultry resulting in safer products for the consumer.

Progress Report
Previous work evaluated the use of biocides in sequential combination under sub-objective 1.1. Peracetic acid (PAA), cetylpiridinium chloride (CPC), sodium hypochlorite, and calcium hypochlorite were tested in a 96 well plate format using 2-fold serial dilutions with the first compound to be tested. After inoculation using a mixture of Salmonella serotypes (Enteritidis, Kentucky, Typhimurium, I 4,[5],12:i-, and Infantis), additional chemicals were added at 15 second intervals, followed by neutralizing buffered peptone water. After treatment, bacterial cells were enumerated by spreading on blood agar plates. For all serotype mixtures, when PAA was applied first, the tolerance was lower than when it was applied after a chloring compound. Tolerance to CPC was high across all serotype mixtures and appeared to reduce the efficacy of PAA and therefore may be excluded in the future. Campylobacter was also tested using the assay, but the results were not consistent due to growth conditions required for Campylobacter. This project was discontinued by mutual agreement with USDA-FSIS. Prediction of co-occurrence of antimicrobial resistance using machine learning in Salmonella and Enterococcus relating to sub-objective 1.1 continued. Initial datasets consisted of approximately 4,471 Salmonella isolates and 5,992 Enterococcus isolates from retrospective antimicrobial susceptibility testing surveillance datasets collected by the Food and Drug Administration. Two data sets of 21,601 isolates obtained from the Centers for Disease Control and Prevention and 19,802 isolates obtained from FSIS were added. Causal directed graphs were generated from the interactions of individual sample sources across time, allowing for changes within antimicrobial resistance co-occurrence developing across time to be captured. Bayesian parameters were used to estimate and evaluate values of the conditional probability distributions of the Salmonella network, using datasets detailing time and sample source. This allowed for detailed predictions of any dataset feature based on presence or absence of information from any other combination of dataset features. Additional analysis methods to identify independent assessment of co-occurrence patterns included Random Forests for prediction of feature importance and dataset clustering using KModes to identify the underlying dataset structure. Consistent identification of antimicrobial resistance co-occurrence interaction patterns across analysis methods supports the validity of the ongoing analysis. Under sub-objectives 1.2 and 1.3, year 2 quarterly collection of conventional (n=80) and “no antibiotic ever” (NAE, n=80) retail poultry products was completed. Bacterial metagenomes of retail poultry were analyzed under sub-objective 1.2. DNA extractions continued from whole product rinsates from conventional and NAE retail poultry products. Shotgun sequencing was done followed by metagenomic analysis. Taxonomic comparison of the microbial composition of non-incubated versus incubated samples showed that overnight incubation increased relative abundance of each organism and population diversity to detectable limits while also decreasing abundance of contaminant reads. Antimicrobial, metal, and biocide resistance genes and mobile genetic elements in the samples were also analyzed. For sub-objective 1.3, conventional samples were positive for Salmonella (13.8%), Campylobacter (7.5%), Escherichia coli (98.8%), Enterococcus (96.3%), and Staphylococcus (71.3%). NAE products had higher levels of Salmonella (15%) and Enterococcus (97.5%) and lower levels of Campylobacter (6.3%), E. coli (93.4%), and Staphylococcus (50%). No Escherichia albertii were recovered from any of the poultry products. Salmonella serogroups included B, C1, C2/C3, D1/D2, E, and unidentified; 31% of Salmonella were resistant to two or more of the drugs tested for antimicrobial susceptibility. Both Campylobacter coli and C. jejuni were identified among the Campylobacter isolates. Enterococcus faecalis was the dominant enterococcal species from both product types. The ARS Salmonella Infantis Working Group continued under sub-objective 1.3. The working group was formed in collaboration with the Office of National Programs (ONP) to address the threat of S. Infantis and pESI to U.S. poultry. A strategic planning meeting was conceived with ONP, the National Chicken Council, and the National Turkey Federation and held July 31, 2023. The meeting was attended by over 60 representatives from the poultry industry. A PCR assay was developed for the FSIS to detect S. Infantis and pESI under sub-objective 1.3. The assay was developed to detect the unique clustered regularly interspaced short palindromic repeats (CRISPR) sequence in the Infantis chromosome and the pESI repA gene. Analysis of nearly 200 Salmonella isolates confirmed that the PCR assay could accurately detect S. Infantis and pESI. In collaboration with scientists from a commercial company, projected accuracy and specificity of the PCR assays were determined using bioinformatic analysis of over 30,000 Salmonella sequences in the public database. This analysis showed that the CRISPR target was more than 95% accurate and that the pESI PCR assays that detect the unique repA gene from pESI was greater than 99% accurate. A real-time detection assay was developed for these two genes to meet the FSIS requirements. In addition, primers and probes were designed to detect Salmonella serotypes Typhimurium, Enteritidis, and 4,5,12, [i]- (monophasic). This assay will detect the top four key performance indicator serotypes of Salmonella as identified by the FSIS as well as the pESI plasmid. Work continued on recombinant Salmonella sub-unit protein vaccines to test in chickens to determine if these proteins could induce humoral immune responses in hosts relating to sub-objective 2.1. Three isotypes of broiler immunoglobulins against the recombinant proteins were detected and quantification was done using automated capillary immunoblot assays. The recombinant flagellar proteins triggered high levels of IgY and IgM responses in most vaccinated birds but not the unvaccinated group indicating the antigenicity of the recombinant proteins. Alternatively, the recombinant fimbrial proteins barely induced antibody responses in vaccinated chickens. Differences of the flagellar and fimbrial proteins in triggering antibody responses may be due to the sizes of their molecular masses. The former are greater than 40 kDa while the latter are about 20 kDa. A single injection of the recombinant proteins was enough to induce chickens to produce high levels of the IgY in sera which lasted for three weeks post vaccination. This finding seems in contrast to the immunological doctrine that a booster vaccination be required to have high and long-lasting IgY responses. A single shot with high, long-lasting responses will simplify the vaccine program for poultry producers. Under sub-objective 2.1, a collaboration with Emory University, Atlanta, Georgia, also continued to develop and produce an mRNA-based vaccine against S. Infantis for use in poultry. Under sub-objective 2.3, work continued on identification of bacteriocins using genome mining of whole genome sequences from multidrug resistant Enterococcus isolates from poultry. Genomes were screened using the program antiSMASH which provides rapid genome-wide identification, annotation, and analysis of secondary metabolite biosynthesis gene clusters in bacterial genomes. Staphylococcus xylosus isolates that inhibited growth of several Gram-negative and Gram-positive bacteria from a previous study were tested against S. Infantis. Broth supernatants from ten isolates were tested, but none had the ability to inhibit or reduce Salmonella growth. For sub-objective 2.4, genome sequences from NAE studies and 100 genome sequences of extended-spectrum beta-lactamase resistant Salmonella were analyzed by PHASTER to identify phage sequences within their genomes. A wide variety of phage and phage genes were cataloged in these Salmonella isolates. Broiler cecal contents were also collected for bacteriophage isolation according to the standard protocol. Salmonella Typhimurium, Enteritidis and Heidelberg were used as targets. No phage was detected from the cecal contents with Salmonella plaque assays. Relating to sub-objective 2.4, work continued on developing Bdellovibrio as a probiotic against S. Infantis, complementary to the development of bacteriophage to reduce levels of Salmonella. Bdellovibrio are predatory bacteria that kill other Gram-negative bacteria, suggesting it could be a used to reduce the levels of Salmonella. Twenty new Bdellovibrio were isolated from soil and water samples collected in the Athens, Georgia, area. Ten of the isolates were tested in a predation assay on the top 10 Salmonella serotypes that infect humans, including S. Infantis with and without the pESI plasmid. Bdellovibrio strains were able to kill all Salmonella tested. Predation efficiency assays were done on two isolates which were able to reduce the levels of Salmonella by 2 log in 24 hours. Ten of the Bdellovibrio were subjected to whole genome sequencing, followed by phylogenetic analysis. These ten isolates were found to be very diverse representing different species of Bdellovibrio which were divergent from the type strain HD100 and from the genome sequences found in the National Center for Biotechnology Information database. These findings indicate that Bdellovibrio may be ubiquitous in the environment, are genetically highly divergent, and can kill Salmonella. Plate assays using S. Infantis with pESI as prey yielded plaques demonstrating the activity of predatory bacteria. Several of these plaques had a morphology indicating the presence of bacteriophage.

Accomplishments
1. Detection and distribution of phage CRISPR sequences in Campylobacter genomes. Application of bacteriophages in feed during poultry production has been used as an alternative to antibiotics to reduce Campylobacter contamination in poultry products. The presence of the bacteriophage sequences in the clustered regularly interspaced short palindromic repeats (CRISPR) spacers in Campylobacter is critical as they may confer bacterial resistance to bacteriophage treatment. ARS researchers in Athens, Georgia, tested 178 Campylobacter jejuni isolates from chicken livers and 87% were found to harbor the CRISPR sequences. Analysis of these sequences revealed that most spacer sequences were homologous to Campylobacter bacteriophage DA10. This study provided important information for the future development of effective bacteriophage use in Campylobacter mitigation during poultry production.

2. Pastured poultry farms are a source of Campylobacter. Campylobacter is a leading cause of foodborne illnesses. Pasture farming is an important source of agricultural production for small farming communities; however, information on the microbiological safety of pasture-raised poultry products is limited. ARS researchers in Athens, Georgia, in collaboration with Mansoura University, Egypt, investigated antibiotic resistance, genetic relatedness, and virulence of thermophilic Campylobacter isolates from pasture-raised poultry farms from the southeastern United States. Campylobacter isolates from pasture-raised farms were resistant to antibiotics and genetically related to Campylobacter isolates commonly associated with humans. This data is informative to the poultry industry concerning different poultry sources of Campylobacter as pasture-raised broiler flocks, like conventionally reared broiler flocks, are indicated as a potential vector for antibiotic-resistant and pathogenic strains of thermophilic Campylobacter to consumers.

3. Establishing an antimicrobial resistance monitoring program for fresh surface water. Antimicrobial resistance (AMR) is a critical problem that limits our ability to fight infectious disease, resulting in longer illnesses, or even death. Although the National Antimicrobial Resistance Monitoring System (NARMS) monitors antimicrobial resistant bacteria in retail meats, humans, and food animals, little is known about antibiotic resistance from the environment. ARS researchers in Athens, Georgia, Beltsville, Maryland, Clay Center, Nebraska, and Albany, California, investigated a pilot effort to add an environmental component to NARMS focused on monitoring surface waters. The effort was divided into five areas: defining objectives and questions, designing study/sampling design, selecting AMR indicators, establishing analytical methods, and developing data management/analytics/metadata plans. For each of these areas, the consensus among the scientific community and literature was reviewed and carefully considered prior to the development of this environmental monitoring program. This pilot study will provide a comprehensive picture of how resistant strains are related spatially and temporally within a watershed and help assess how anthropogenic drivers and intervention strategies impact the transmission of AMR within human, animal, plants, and environmental systems. Data from this research will aide in development of a robust surface water monitoring program in NARMS with the goal of assessing risks associated with AMR pathogens in surface water.

4. Comparison of methods for the detection of low levels of Salmonella in surface waters. Recovery of antibiotic resistant Salmonella from water is important for the environmental component of the National Antimicrobial Resistance Monitoring System (NARMS). To date, there is not a specific or standardized method used to recover Salmonella from surface water or agricultural irrigation water. ARS researchers from Athens, Georgia, Beltsville, Maryland, Clay Center, Nebraska, and Albany, California, compared several microbiological methods to recover low levels of Salmonella from surface water to identify the single method that could be used in national surveillance of rivers and streams. Three methods were compared to recover low levels of fluorescently labeled Salmonella Typhimurium from inoculated surface water – bulk water, vertical Modified Moore Swab, and the modified Standard Method 9260.B2. Sixty water samples were compared across four laboratories from five different sites. S. Typhimurium were recovered more often using the modified Standard Method than the bulk water and vertical Modified Moore Swab. The modified Standard Method was prioritized for Salmonella recovery from surface water in laboratory settings. This research benefits NARMS environmental surveillance as well as farmers who can utilize these methods to improve their surveillance of agricultural water.

5. Development of a detection method for the Salmonella Infantis outbreak strain. Salmonella is a human bacterial pathogen that can cause foodborne illnesses and outbreaks. There are more than 2,500 serotypes of Salmonella; however, these serotypes vary in prevalence among poultry meat samples. Recently, Salmonella Infantis has become the most frequently isolated serotype from chicken and turkey meat samples. Most of the S. Infantis being detected carry a novel DNA plasmid called pESI. Detecting this plasmid and confirming the serotype Infantis is important for epidemiological investigations and for poultry producers monitoring production. ARS researchers in Athens, Georgia, developed a PCR protocol to detect the pESI plasmid and confirm the Infantis serotype of Salmonella isolates. Primers were tested bioinformatically to predict specificity, sensitivity, and precision. Fifty-four isolates of Salmonella serotypes Infantis, Senftenberg, and Alachua were tested, with and without the pESI plasmid. Isolates of 31 additional serotypes were also screened to confirm specificity to Infantis. This PCR improved upon existing detection assays due to the accuracy, greater than 95% specificity, sensitivity, and precision, as well as a reduced number of reactions needed for positive detection. This assay is less expensive and more accessible than previously available detection assays which required more reactions, equipment, or whole genome sequencing.

6. Impact of antibiotic resistance and wastewater treatment plants on surface water. Surface water serves as a source for dissemination and exchange of antibiotic resistance (AR) and there is growing concern about the presence of AR contaminants in the environment. ARS researchers in Athens, Georgia, in collaboration with the University of Georgia, evaluated prevalence, diversity, and sources of AR contaminants in surface water and the impact of wastewater treatment plant effluents on the levels of AR contaminants in receiving water. AR contaminants were prevalent and widely distributed in a mixed-use watershed that is representative of much of the land use in the United States. The removal efficiency of antibiotics and AR genes in wastewater treatment plants showed that the plants were not able to completely remove AR contaminants, indicating that wastewater treatment plants are a source of AR contaminants in receiving water. Much of AR contamination was present in streams not associated with known wastewater discharges, suggesting that there are other sources of AR contamination. Understanding the widespread occurrence and abundance of medically important antibiotics, pathogenic bacteria, the genes associated with resistance to these antibiotics, and the potential risk to the local populations exposed to these water sources is important for environmental agencies and risk assessors.

7. Tulathromycin treatment increases multidrug resistant Mannheimia haemolytica in heifers. Bovine respiratory disease (BRD) is a leading cause of illness in feedlot and stocker calves and is commonly associated with the bacterium, Mannheimia haemolytica. Metaphylaxis, administering antimicrobials to all animals at high risk of developing BRD, is an effective way of controlling the disease. Increasing prevalence of multidrug resistant M. haemolytica may reduce effectiveness of metaphylaxis. ARS researchers in Athens, Georgia, in collaboration with Mississippi State University, determined the effect of tulathromycin metaphylaxis and BRD treatment on antimicrobial resistance in M. haemolytica isolated from stocker calves. Crossbred beef heifers at high risk for BRD were randomly assigned to either receive metaphylaxis or a control group. Animals in the metaphylaxis group were more likely to have multidrug resistant isolates of Mannheimia. Animals in the control group had three times higher odds of being treated for BRD. The study demonstrated that tulathromycin metaphylaxis increased risk of shedding multidrug resistant M. haemolytica and that the multidrug resistance is associated with integrative conjugative elements carrying antimicrobial resistance genes. This may have important implications for the future ability of antimicrobials to control and treat BRD, which will be an important consideration for cattle farmers, regulatory agencies, pharmaceutical industry, and veterinarians.

8. Antimicrobial resistance in Salmonella Enteritidis from poultry and humans in Burkina Faso. Salmonella is a foodborne pathogen found globally; however, little is known about the genetic characteristics of Salmonella serotypes in developing countries. ARS researchers in Athens, Georgia, in collaboration with the University of Ouagadougou, Africa, used whole-genome sequencing to characterize four Salmonella Enteritidis isolated from poultry and humans in Burkina Faso. Antimicrobial resistance genes, chromosomal mutations, and mobile genetic elements were detected. The four strains were resistant to multiple classes of antibiotics including beta-lactams, tetracyclines, and folates; whole-genome sequencing identified nine resistance genes and two plasmids in the strains. These results showed that poultry can be a reservoir for multidrug resistant Salmonella Enteritidis strains in Burkina Faso helping to fill an important data gap on antimicrobial resistance found in this area of Africa.

Review Publications
Awad, A., Yeh, H., Ramadan, H., Rothrock Jr, M.J. 2023. Genotypic characterization, antimicrobial susceptibility and virulence determinants of Campylobacter jejuni and Campylobacter coli isolated from pastured poultry farms. Frontiers in Microbiology. 14:1271551. https://doi.org/10.3389/fmicb.2023.1271551.
Yeh, H., Cox Jr, N.A., Hinton Jr, A., Berrang, M.E. 2024. Detection and distribution of clustered regularly interspaced short palindromic repeats (CRISPRs) in Campylobacter jejuni isolates from chicken livers. Journal of Food Protection. 87(4):100250. https://doi.org/10.1016/j.jfp.2024.100250.
Franklin, A.M., Weller, D.L., Durso, L.M., Bagley, M., Davis, B.C., Frye, J.G., Grim, C., Ibekwe, A.M., Jahne, M., Keely, S.P., Kraft, A.L., McConn, B.R., Mitchell, R., Ottesen, A., Sharma, M., Strain, E., Tadesse, D., Tate, H., Wells, J., Williams, C.F., Cook, K.L., Kabera, C., McDermott, P., Garland, J. 2024. A one health approach for monitoring antimicrobial resistance: Developing a national freshwater pilot effort. Frontiers in Water. 6. Article 1359109. https://doi.org/10.3389/frwa.2024.1359109.
Mcconn, B.R., Kraft, A.L., Durso, L.M., Ibekwe, A.M., Frye, J.G., Wells, J., Tobey, E.M., Ritchie, S.M., Williams, C.F., Cook, K.L., Sharma, M. 2024. An analysis of culture-based methods used for the detection and isolation of Salmonella spp., Escherichia coli, and Enterococcus spp. from surface water: a systematic review. Science of the Total Environment. 927. Article 172190. https://doi.org/10.1016/j.scitotenv.2024.172190.
Kraft, A., Wells, J., Frye, J.G., Ibekwe, A.M., Durso, L.M., Hiott, L.M., East, C.L., McConn, B., Franklin, A.M., Boczek, L.A., Garland, J.L., Kabrera, C., McDermott, P., Ottesen, A.R., Zheng, J., Cook, K.L., Sharma, M. 2023. A comparison of methods to detect low levels of Salmonella enterica in surface waters to support antimicrobial resistance surveillance efforts performed in multiple laboratories. Science of the Total Environment. https://doi.org/10.1016/j.scitotenv.2023.167189.
Han, J., Aljahdali, N., Zhao, S., Tang, H., Harbottle, H., Hoffmann, M., Frye, J.G., Foley, S.L. 2024. Infection biology of Salmonella enterica. EcoSal Plus. https://doi.org/10.1128/ecosalplus.esp-0001-2023.
Mcmillan, E.A., Hiott, L.M., Carrica, J.A., Machado, M.P., Bailey, S.J., Dutta, V., Jackson, C.R., Frye, J.G. 2023. Polymerase Chain Reaction for the in vitro detection of the pESI plasmid associated with the globally circulating Salmonella Infantis outbreak strain. Letters in Applied Microbiology. https://doi.org/10.1093/lambio/ovad088.
Cho, S., Hiott, L.M., Read, Q.D., Damashek, J., Westrich, J., Edwards, M., Seim, R.F., Glinsk, D.A., Mcdonald, J.M., Ottesen, E.A., Lipp, E.K., Henderson, M.W., Jackson, C.R., Frye, J.G. 2023. Distribution of antibiotic resistance in a mixed-use watershed and the impact of wastewater treatment plants on antibiotic resistance in surface water. Antibiotics. 2023(12):1586. https://doi.org/10.3390/antibiotics.
Crosby, W.B., Karisch, B.B., Hiott, L.M., Pinnell, L.J., Pittman, A., Frye, J.G., Jackson, C.R., Loy, D.J., Epperson, W.B., Blanton, J., Morley, P.S., Woolams, A.R. 2023. Tulathromycin metaphylaxis increases nasopharyngeal isolation of multidrug resistant Mannheimia haemolytica in stocker heifers. Frontiers in Veterinary Science. 10. Article 1256997. https://doi.org/10.3389/fvets.2023.1256997.
Kagambega, A., Ramadan, H., Dione, M., Bouda, S.C., Hiott, L.M., Mcmillan, E.A., Sharma, P., Gupta, S.K., Jackson, C.R., Frye, J.G. 2024. Genome analysis of Salmonella enterica serovar enteritis isolated from poultry and humans in Burkina Faso. Microbiology Resource Announcements. 2024:e01024. https://doi.org/10.1128/mra.01024-23.
Ejaz, M., Syed, M., Jackson, C.R., Sharif, M., Faryal, R. 2023. Epidemiology of Staphylococcus aureus non-susceptible to vancomycin in South Asia. Antibiotics. 12(6). Article 12060972. https://doi.org/10.3390/antibiotics12060972.