UC Davis

Antimicrobial resistance (AMR) is one of the most serious public health threats of the century. Resistant bacteria and AMR genes (ARGs) can spread through human and animal populations through pathways such that selective pressures in one population inextricably impacts others on the One Health continuum. This dissertation takes a farm-to-fork approach on AMR by evaluating the distribution and risk factors for AMR in retail products, AMR co-selection in food-producing animals, and genomic profiles of these bacteria.In Chapter 1, a cross-sectional study was conducted to assess the distribution and AMR profiles of Salmonella from retail meat products in California. From multivariable logistic regression, season of purchase and meat type were significantly associated with the isolation of Salmonella. Whole genome sequencing (WGS) characterized Salmonella isolates into 14 distinct serotypes corresponding to 17 MLST patterns. Diverse ARGs including those of high public health significance and putative plasmids were identified. The IncFIB(pN55391) replicon previously reported in connection to the worldwide dissemination of pESI-like mega plasmid carriage in an emerged S. Infantis clone was detected in four of the six multidrug-resistant (MDR) isolates. In Chapter 2, Escherichia coli from samples in Chapter 1 were assessed to gain further insight on the clinical and epidemiologic risks associated with AMR in retail meat products from California. Phenotypic resistance to ampicillin, gentamicin, streptomycin, and tetracycline were significantly associated with meat type, with poultry counterparts (chicken or ground turkey) exhibiting higher odds for resistance compared to non-poultry meats (beef and pork). Clustering analysis and co-occurrence networks revealed that genomic AMR determinants of E. coli from retail meat were highly heterogeneous with sparsity of shared gene networks and minimally driven by retail-level factors of meat type, season of purchase, packaging, and antibiotic label claims. In Chapter 3, the impact of dietary zinc supplementation in pre-weaned dairy calves on phenotypic AMR of fecal Enterococcus spp. and E. coli was investigated. Accelerated failure time (AFT) models were constructed to determine the association between zinc treatment and AMR, with exponentiated coefficients adapted for minimum inhibitory concentration (MIC) values instead of time representing the degree of change in AMR (MIC Ratio, MR). Zinc supplementation did not significantly alter the MIC in Enterococcus spp. for 13 tested antimicrobials and in E. coli for azithromycin and ceftriaxone. However, a significant reduction in E. coli MIC values was observed for ciprofloxacin (MR= 0.17, 95% CI 0.03–0.97) and nalidixic acid (MR= 0.28, 95% CI 0.15–0.53) for zinc-treated compared to placebo-treated calves. In Chapter 4, whole-genome comparative analysis was conducted to investigate the host-microbe interface of MDR E. coli from dairy calves. The pangenome of E. coli was open, with all-by-all genome similarity comparisons clustering primarily by sequence type (ST) rather than host factors of diarrheal disease, zinc supplementation, and antimicrobial exposure. E. coli lacked the typical virulence factors of diarrheagenic strains, however virulence factors overlapping with those in major pathotypes were identified, with the most prevalent genes corresponding to iron acquisition. Dietary zinc exposure was not associated with the selection of individual ARGs, however significant associations between the occurrence of certain ARGs and metal resistance genes were identified. Collectively, this dissertation provides greater insight into the epidemiology of AMR in enteric bacteria of public health significance. This improved understanding of the distribution and drivers of AMR in food products and food-producing animals will inform future AMR monitoring and control strategies by supporting more targeted approaches to mitigate AMR from farm-to-fork.