The administration and utility of antibiotics to control and treat bovine respiratory disease (BRD) in beef cattle feedlots is a growing concern. Antimicrobial resistance (AMR) among BRD-associated bacterial pathogens has been the subject of cultivation-dependent and cultivation-independent surveillance. Bacterial genome sequencing and metagenomic approaches facilitate the characterization of AMR in the beef industry; however, the current collection of cattle-associated AMR research programs lack connections to each other. A more integrated view of how antimicrobial use (AMU) is related to resistance at a gene level is needed. We sought to establish a catalog of commonly observed AMR genes (ARGs) in opportunistic bacterial pathogens that contribute to BRD using publicly available data sets that were generated by the scientific community with and without AMU in mind. The presence of these clinically relevant ARGs appeared to differ by geography. Greater sampling in North America facilitated the generation of a list of ARGs often encoded by Mannheimia haemolytica and Pasteurella multocida . Detection of clinically relevant ARGs in shotgun metagenomes of cattle-associated and accessible feedlot samples such as water, soil, and feces was possible but limited by relative sequence read abundance. An exception was the tylosin esterase-encoding gene estT , which is among the most frequently observed ARGs in M. haemolytica and feedlot-related metagenomic data sets. Finally, by re-evaluating studies on the impact of AMU on AMR in beef production systems, we show that conventional practices, including in-feed antibiotic use, increase the relative abundance of ARGs in animal-derived samples. IMPORTANCE: This retrospective analysis delivers a list of ARGs found in opportunistic pathogens that contribute to BRD. The high incidence of BRD in North America is linked to the origin and implementation of metaphylaxis to mitigate detrimental animal losses at feedlots. Notably, ARGs commonly observed in these pathogens isolated in North America were not conserved across the globe, underscoring the relationship between regional AMU and AMR. A positive relationship was also observed between the relative abundance of ARGs in cattle-associated metagenomes with greater exposure to antibiotics. Overall, this analysis should help to guide future surveillance efforts and experimental designs to more directly evaluate the impacts of feedlot practices on AMR.

Bovine respiratory disease (BRD) is the primary disease of cattle and is responsible for most of the antibiotic use in the beef industry, both for metaphylaxis and treatment. Infection prevention and targeted treatments would benefit from detecting and identifying bacterial pathogens and, ideally, assessing antibiotic sensitivity. Here, we report success refining targeted metagenomics by hybridization capture sequencing (CapSeq) to detect and genotype bacterial pathogens and genes for antibiotic resistance in BRD. We developed a novel multi-locus sequence typing (MLST) strategy to both enhance assay sensitivity and provide epidemiological data compatible with traditional typing assays. Phylogenetically informed panel design by ProbeTools coupled with hybridization capture enabled the sequencing of novel genotypes and elucidated coinfection by distinct strains of Mycoplasmopsis bovis . CapSeq was orders of magnitude more sensitive than classical metagenomics at a scale compatible with low-throughput DNA sequencing capacity in diagnostic laboratories. Iterative design and testing of longer typing loci confirmed improvements in locus detection through a strategy applicable to future target selection. CapSeq detected bacterial pathogens that evaded detection by traditional metagenomics, and it demonstrated greater range and sensitivity for antibiotic resistance gene detection. IMPORTANCE: Shotgun metagenomic sequencing for infectious disease diagnostics and surveillance benefits from simultaneous detection of multiple pathogens in a sample. Adding a hybridization probe capture step to the metagenomics workflow enriches targeted loci to improve the sensitivity of pathogen detection without compromising the ability to detect pathogen variants. Our custom probe panel enables multi-locus sequence typing of bovine respiratory disease agents as well as capture of antibiotic resistance genes, which improves the sensitivity of metagenomic testing and provides genotyping data compatible with traditional assays. This study demonstrates the utility of new design principles for probe panels; it also demonstrates how targeted metagenomics provides important new insights into co-infections and is equally useful for surveillance of environmental reservoirs of disease agents.

Mycoplasma bovis is the most common mycoplasma associated with cattle diseases worldwide. However, other seemingly less virulent Mycoplasma spp. such as M. bovigenitalium and M. bovirhinis have also been associated with mycoplasmosis. The study objective was to compare the adhesion and cellular invasion characteristics of these bovine Mycoplasma spp. using Madin–Darby Bovine Kidney (MDBK) epithelial cells. MDBK cells were separately infected with 12 M. bovis strains and one strain each of M. bovigenitalium and M. bovirhinis. Following infection, a gentamicin protection assay was performed and the cells lysed at 6 and 54 h post-infection. The MDBK cell lysates were cultured for Mycoplasma spp. and qPCR was used to estimate the average number of Mycoplasma bacterial cells that infected each MDBK cell (Myc/Cell ratio). Confocal and electron microscopy studies using M. bovis mNeonGreen strain were also performed. All 14 Mycoplasma strains multiplied within the MDBK cells, a finding confirmed by microscopy studies of the M. bovis mNeonGreen strain. Unexpectedly, the M. bovis strains, obtained from diseased and asymptomatic cattle and bison, had lower Myc/Cell ratios than M. bovirhinis and M. bovigenitalium strains. These findings suggest that the ability for mycoplasmas to invade and replicate within host cells does not account for the differences in virulence between species.

Septic arthritis (SA) is a cause of lameness in cattle attributed to bacterial infections. Mycoplasmopsis bovis is the best known and characterized etiological agent of SA; however, cases caused by diverse bacteria have been reported. Accordingly, we surveyed bacteria associated with septic and healthy joints from animals in western Canadian feedlots. Microbial community profiling showed that M. bovis was the most frequently detected pathogen in septic joints, followed by Metamycoplasma alkalescens and Trueperella pyogenes . In most cases, disease was ostensibly caused by a single pathogen, though polymicrobial infections and complex communities were also observed in DNA isolated from septic joints. The application of enhanced metagenomics by target DNA hybridization capture sequencing (CapSeq) provided more robust pathogen detection and characterization. CapSeq revealed resistance determinants that escaped detection using a conventional shotgun metagenomic approach. Notably, a series of nucleotide polymorphisms to M. bovis rrs , rrl , gyrA , and parC gene sequences were observed that confer resistance to macrolides and oxytetracycline-resistant T. pyogenes were also apparent in the CapSeq data. Complementary pathogen isolation, whole-genome sequencing, and phenotyping efforts were focused on the two most prominent pathogens, M. bovis and M. alkalescens , and corroborated metagenomic data sets. IMPORTANCE: Informed antimicrobial use for the treatment of septic arthritis (SA) has been limited by overlooking the potential diversity of causative agents and our knowledge of their antimicrobial resistance (AMR) profiles. This survey begins to provide epidemiological insights, offering renewed appreciation of Metamycoplasma alkalescens as an etiological agent of SA and highlighting the prominence of important AMR determinants. Finally, the survey suggests that our knowledge of even the identities of the causative agents of SA is incomplete.

DNA capture-based sequencing by hybridization probe enrichment (CapSeq) affords high-throughput and targeted gene-to-organism detection. Despite the specificity designed into CapSeq platforms, untargeted DNA represents additional genes and organisms of interest. To this end, de novo metagenome assemblies of CapSeq datasets present opportunities for discovery. We report an example here. CapSeq of samples that originated from cattle suffering from septic arthritis (SA) resulted in detection and assembly of two ungulate tetraparvovirus 1 genomes. These coincidental detections and assemblies were obtained from a platform designed to characterize bacterial, not viral, pathogens. Access to a sample matched shotgun metagenomic dataset corroborated the result. Importantly, SA is a significant source of lameness in beef cattle that lacks effective and approved treatments, and the list of aetiological agents is incomplete. This metagenomic observation calls attention to the possible implication of viral contributors to bovine SA, which is treated as a bacterial disease. The results exemplify the potential application of unbiased CapSeq data mining to support clinical discoveries.

We present closed circular genome sequences of Histophilus somni , Mannheimia haemolytica, and Pasteurella multocida associated with bovine respiratory disease in feedlot cattle. The data will contribute to better understanding of disease and the development of potential mitigation strategies.