Genomic Characterization of Resident Bacteria from Selected Swine Feed Mills in the Midwest Region of the United States

Abstract

Knowledge of resident bacteria present within animal food manufacturing facilities, as well as the understanding of antimicrobial resistance genes (ARGs), can help inform cleaning and sanitation practices to reduce microbial risks for the animal and human food chains. The goal of this study was to characterize resident bacteria previously isolated in selected swine feed mills within the Midwest region of the United States at multiple seasonal timepoints. About 121 resident bacterial isolates were identified and classified into six genera, consisting of 68 Enterobacter, 34 Citrobacter, four Cronobacter, eight Klebsiella, three Proteus, and four Pseudomonas, in addition to 30 Escherichia coli and 33 Salmonella characterized previously. Among the 121 isolates discussed within the current work, the majority (48; 39.7%) came from environmental samples taken from either non-feed contact surfaces such as floors and feed contact surfaces (38; 31.4%) or transient surfaces such as brooms and workers’ shoes (20; 16.5%), and only 12 samples (9.9%) came directly from finished feed. A range of ARGs and metal tolerance genes (MTGs) were identified. This research provides an initial framework to understand the diversity of resident flora in feed mill facilities in the Midwest region and identify areas to focus on for housekeeping and sanitization within mills. This information is critical to developing feed safety strategies and preventing antimicrobial resistance spread in the farm-to-table continuum.

Keywords

antimicrobial resistance resident bacteria WGS swine feed mills Midwest

Introduction

Understanding the diversity of resident bacteria in the animal food processing environment during operation and after cleaning and sanitation is important to assess the food safety risks associated with animal food commodities. Several studies indicated that the survival of foodborne pathogens is heavily influenced by non-pathogenic resident bacteria (Giaouris et al, 2015). Resident bacteria are usually isolated from non-selective media and defined as the bacterial community that persists over time in production facilities (Møretrø and Langsrud, 2017). Enterobacterales are usually used as indicators for hygiene and sanitation (Marriott et al., 2018). A recent survey, which sampled 647 pet food and 378 animal feed samples, reported that this indicator group was widely present (Ge et al, 2020). These bacteria may also have a role in antimicrobial resistance (AMR) spread in animal food-related commodities, which is a growing and pressing concern for both human and animal health (World Health Organization (WHO), 2015). While characterization of antimicrobial resistance genes (ARGs) and metal tolerance genes (MTGs) has been done on specific bacterial pathogens present within feed mills and ingredients (Davies and Wales, 2010; Torres et al, 2011; Burns et al, 2015; Munoz et al, 2021; Magossi et al, 2020), little is known about the overarching bacterial presence and associated ARGs and MTGs within facilities. Therefore, in this study, we characterized resident bacteria (mostly Enterobacterales) isolated previously from feed mills located in the Midwest region of the United States with the goal of assessing the food safety risks associated with animal food commodities.

Material and Methods

Sample collection and isolation

This study is an extension of our previous survey conducted in six commercial swine feed mills located in the Midwest region (Iowa, Missouri, Kansas, and Oklahoma) of the United States during three different seasons: fall (October of 2018), late winter/early spring (March of 2019), and summer (June of 2019) (Magossi et al, 2020). The selection of the mills was based on location and the feed produced. Sampling sites were selected from areas of high pedestrian traffic, ingredient processing, and dust accumulation, following production flow (Magossi et al., 2019). A total of 405 samples were collected across a range of sampling sites, including finished feed, floors, brooms, and feed manufacturing equipment. Each site was analyzed for the presence of pathogens and commensals. From all the presumptive positive samples, at least two colonies were selected for further investigation (Schwan et al, 2023). Isolates were stored in Trypticase Soy Broth (TSB; Hardy Diagnostics, Santa Maria, CA, USA) containing glycerol at −80°C until analysis. On day 1 of experiments, isolates were retrieved from −80°C and grown on Trypticase Soy Agar (TSA, Hardy Diagnostics, Santa Maria, CA, USA) or Blood Agar (Hardy Diagnostics, Santa Maria, CA, USA) at 37°C overnight.

Bacterial speciation

The VITEK 2 Compact system (bioMérieux, Marcy l’Etoile, France) or matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry on the MALDI Biotyper (Bruker, Billerica, MA) was used for the identification of bacterial genera and species following the manufacturer’s instructions.

Antimicrobial susceptibility testing (AST)

AST against 14 antimicrobial agents was performed by broth microdilution (Schwan et al, 2023). The MICs were interpreted using the CLSI standards and NARMS breakpoints valid for Enterobacterales (Clinical and Laboratory Standards Institute, 2019; Food and Drug Administration, 2015). Multidrug resistance (MDR) was defined as resistance to ≥3 antimicrobial classes.

WGS characterization

DNA extraction, library preparation, sequencing, and in silico analysis were performed as previously described (Domesle et al, 2021) with some modifications. Genomic DNAs were extracted using the Qiagen DNeasy Blood & Tissue kit (Germantown, MD, USA). Nextera XT Library Preparation kit (Illumina, San Diego, CA, USA) was used for preparing paired-end libraries, and WGS was carried out on the MiSeq sequencer, using a 2 × 300-cycle MiSeq reagent V3 kit (Illumina). Trimming and de novo assemblies were obtained with Shovill version 9 (https://github.com/tseemann/shovill), available in the GalaxyTrakr pipeline (http://galaxytrakr.org/) with default parameters (Zhang et al., 2019).

Detection of ARGs and MTGs

ARGs and MTGs were identified using AMRFinderPlus v3.12.8 (https://www.ncbi.nlm.nih.gov/pathogens/antimicrobial-resistance/AMRFinder/) with default parameters (Feldgarden et al., 2021).

Data availability

The accession numbers and additional genomic information for the isolates can be found under the BioProject number PRJNA561052 at the NCBI, National Library of Medicine.

This study was approved by the KSU Institutional Review Board.

Results and Discussion

Bacterial species characterization

Out of 405 samples, a total of 212 isolates were recovered, which were classified into 16 genera by WGS. Of these 212 isolates, 28 isolates were not assembled as they belonged to uncommon Enterobacterales, including Kosakonia cowanii, Leclercia adecarboxylata, Lelliottia amnigena, Pantoea spp., and Rahnella aquatilis. Isolates belonging to E. coli (n = 30) and Salmonella (n = 33) were characterized in an earlier report (Schwan et al., 2023). In the present study, we further investigated the remaining 121 isolates, which consisted of 34 Citrobacter, 4 Cronobacter, 68 Enterobacter, 8 Klebsiella, 3 Proteus, and 4 Pseudomonas.

Isolate location profile within animal feed manufacturing facilities

The isolates obtained in this study (n = 121) were primarily found through environmental swabbing, with only 12 derived from finished feed (Tables 1–6). Most of the environmental samples were taken on hard, stationary surfaces, with slightly more isolates coming from non-feed contact surfaces like floors than those on feed contact surfaces (38/121), such as receiving grates, fat inlet valves, pellet mills, and load-out equipment within the different facilities. Twenty (20/121) isolates were obtained from transient surfaces like worker shoes and brooms (Tables 1–6).

Table 1.