2.1 Soil and manure samples collection

Soil and manure samples were collected from the Tullimba farm (University of New England, Armidale, NSW, Australia) located 40 km southwest of Armidale, NSW. The Tullimba farm is spread over 740 ha and has approximately 1000 head of cattle. Carboxylic ionophore (Monensin) is the only antibiotic included in standard cattle diets on the Tullimba farm (Table 1). A total of 500 soil subsamples and 130 manure subsamples were collected from different sites on the Tullimba farm between April and June 2019 (Figure 1).

Soil samples were collected from five paddocks of various sizes (Figure 1). Soil sites 1–3 were chosen due to the regular application of composted cattle manure from the feedlot area every 5 months. Before sampling, site 1 was compost-treated in June and November 2018. Sites 2 and 3 were treated with compost in September 2018 and February 2019. Sites 4 and 5 had not been treated with composted manure and therefore were selected as control. Three hundred subsamples were collected from treatment sites 1, 2, and 3. While 200 subsamples were collected from control sites 4 and 5. Systematic grid-square sampling patterns were used to collect soil samples (de Zorzi et al., 2008). Twenty, 22-mm diameter cores (up to 15 cm deep) were collected from each of the five plots of 50 square meters (100 samples) set up within 4 ha of land in each site.

Manure samples were collected from the different sites. Sixty subsamples of fresh manure (i.e., still soft and warm) were collected from the six sites (paddock and feedlot), each site grazing 50–100 head of cattle. Thirty subsamples of fresh manure from paddock sites were collected and 10 subsamples were combined into a composite sample for testing. Similarly, 30 subsamples of fresh manure from feedlot were also collected and 10 subsamples were combined into a single sample for testing. While 100 manure subsamples were composted for several months onsite by the farm and used for pasture fertilization. Of these composted samples, 33 subsamples were combined into a single sample for analysis. In the lab, these soil and manure samples were kept at 4°C before culturing. Five grams of each sample was placed in a 15-mL screw cap tube and archived at −80°C for DNA extraction.

2.2 Isolation of ARB from field samples

Suspensions of soil and manure samples were made by placing 5 g of soil or manure into a sterile flask with 45 mL of sterile distilled water, shaken at 28°C, 220 rpm for 1 h (Yang et al., 2016) and the slurries were filtered through sterile gauze. The resuspended soil and compost filtrates were serially diluted from 10⁻¹ to 10⁻⁴, in sterile distilled water under aseptic conditions.

Diluted manure and soil filtrates (100 µL) containing bacterial populations were cultured on nutrient agar plates (agar 15 g/L, peptone 5.0 g/L, NaCl 5.0 g/L, yeast extract 2.0 g/L, and beef extract 1.0 g/L) (Snyder & Atlas, 2006), each supplemented with a specific antibiotic, tetracycline (256 µg/mL), kanamycin (256 µg/mL), streptomycin (2000 µg/mL), ciprofloxacin (256 µg/mL), trimethoprim (256 µg/mL), cefotaxime (256 µg/mL), chloramphenicol (256 µg/mL), erythromycin (256 µg/mL), ampicillin (256 µg/mL), colistin (256 µg/mL), and monensin (8 µg/mL). Bacteria were defined as ARB if they were able to grow in the presence of the antibiotics used.

2.3 Molecular identification of bacterial species isolated from soil and manure

The genomic DNA of ARB isolates was extracted using the Isolate II Genomic DNA Kit (Bioline), according to the manufacturer's instructions. The quantity and concentration of DNA were determined using a Nanodrop 8000c spectrophotometer (Thermofisher Scientific). The bacterial 16S ribosomal RNA (rRNA) gene was amplified using primer pairs (1369-F, 5′-CGGTGAATACGTTCYCGG-3′ and 1541-R, 5′-AAGGAGGTGATCCRGCCGCA-3′) as described previously (Huang et al., 2015). The PCR reaction and amplification program were described in Anisimova & Yarullina (2019). The PCR products were run on a 1% agarose gel and stained with GelRed (Biotium). The PCR product was purified using a QIAquick PCR Purification Kit (Qiagen) according to the manufacturer's instructions and was sequenced at Australian Genome Research Facility using Sanger sequencing technology. Basic Local Alignments Tool (BLAST) algorithm was used to analyze sequence data.

2.4 Statistical analysis

Two-way analysis of variance with Bonferroni post hoc tests was performed on CFU data using Prism version 8.3.1 (GraphPad) to determine statistical significance (i.e., p < 0.05).

2.5 Analysis of bacterial diversity

Soil microbial DNA was extracted from −80°C archived samples using DNeasy PowerSoil Pro Kit (Qiagen), according to the manufacturer's instructions. Fecal manure was extracted from −80°C archived samples using QIAamp DNA Stool Mini Kit (Qiagen). Microbial profiling of extracted DNA was performed by Illumina MiSeq. The 2 × 300 bp paired-end sequencing (Ramaciotti Centre for Genomics, Sydney, Australia) of the 16S V3–V4 region of the rRNA gene (primers 341F–805R), from which was obtained a total of 2,948,799 merged sequences. Bacterial 16S rRNA amplicon analysis was performed using QIIME 2 2021.11 (Bolyen et al., 2019). Demultiplexed sequences were denoised with DADA2 (Callahan et al., 2016) (via q2-dada2). Taxonomy was assigned to amplicon sequence variants (ASVs) using the q2-feature-classifier naïve Bayes taxonomy classifier (Bokulich et al., 2018) against the SILVA v138 rRNA gene database (Glöckner et al., 2017; Quast et al., 2013; Yilmaz et al., 2014). Metabarcode data were imported to R Core Team (2021; 4.1.0 2021-05-18) using the packages qiime2R (Bisanz, 2018), phyloseq (McMurdie & Holmes, 2013), and the tidyverse (Wickham et al., 2019). Alpha- and beta-diversity analysis of metabarcoding data was performed on the relative abundance of 16S ASVs after samples were rarefied to a common depth (3155 reads per sample). Shannon–Wiener indices were calculated using the vegan package (Oksanen et al., 2020), and the difference between treatments was assessed by the Kruskal–Wallis rank-sum test, followed by pairwise comparison with Dunn's test. Bray–Curtis distances were calculated with vegan (Oksanen et al., 2020), and visualized with ape (Paradis & Schliep, 2019) and differences between treatments were assessed through permutational multivariate ANOVA (999 permutations).

2.6 Plasmid extraction and bacterial transformation

The plasmid pUCITD-AMP Golden Gate plasmid containing the synthetic erythromycin resistance gene (ermF) from IDT was used as a positive control for the transformation of DH5α cells. Bacterial plasmids were extracted using the QIAprep Spin Miniprep Kit, according to the manufacturer's instructions (Qiagen). Mix and Go chemically competent DH5α Escherichia coli cells were transformed with purified circular plasmids according to the manufacturer's instructions (Zymo Research) and colony PCRs for the ermF gene inserts were performed using forward: 5′-TCGTTTTACGGGTCAGCACTTTAC-3′ and reverse: 3′-TTTCAGGGACAACTTCCAGCA-5′ primers.

2.7 Disk diffusion assay

Disk diffusion assays were adapted from the EUCAST method (Matuschek et al., 2014), and bacteria were grown at 28°C overnight on Mueller–Hinton agar plates (Oxoid). Colonies of the isolate were selected and suspended in sterile saline (0.15 M) to the ideal concentration to create a turbidity comparison that matched 0.5 McFarland reagent (Thermo Scientific) (~1–2 × 10⁸ CFUs/mL). A sterile swab was submerged into the suspension and spread across Mueller–Hinton agar plates. Four disks containing antibiotics that included erythromycin (15 µg), gentamycin (10 µg), tetracycline (30 µg), and streptomycin (25 µg) as controls (Thermo Scientific) were applied to each Mueller–Hinton agar plate. The plates were incubated aerobically for 12–20 h at 28°C. Zones of inhibition were measured as the diameter of complete inhibition, which included the disk edge.

3.1 Manure

Fresh paddock and feedlot manure, as well as composted manure from cattle treated prophylactically with monensin, were evaluated for microbial diversity and the abundance of ARB. As shown in Figure 2a, fresh paddock manure had higher bacterial alpha diversity compared to fresh feedlot and composted manure. Furthermore, Figure 2b showed an overall trend of increased ARB abundance in fresh paddock manure, compared to feedlot and composted manure. While the alpha diversity of feedlot and composted manure were similar, the composted manure had the lowest number of ARB observed, a trend that was consistent across all the tested antibiotics (Figure 2b). Unsurprisingly, the greatest number of ARB CFUs in the paddock, feedlot, and composted manure samples were observed for monensin, but also for trimethoprim, ampicillin, and cefotaxime. In contrast, the lowest ARB CFU was observed for ciprofloxacin (Figure 2b). Hence, fresh manure from the paddock had greater bacterial diversity but only intermediate levels of antibiotic resistance, suggesting that the increase in bacterial diversity is associated with a decrease in ARB abundance.

3.2 Bacterial and antibiotic-resistant diversity in compost-treated soil

To examine the effect of the composted manure application on paddock soils, the bacterial diversity and antibiotic resistance of compost-treated and untreated control paddock soil were determined. There was no significant difference in alpha diversity (i.e., richness and evenness of species present) between untreated paddock soil and compost-treated paddock soil (Figure 3a). However, beta-diversity analysis (Figure 3b) demonstrated distinct bacterial community composition differences between the control, uncomposted paddock soil and the compost-treated paddock soil. In addition, within the compost-treated paddock, two discrete bacterial population clusters were present. Both the compost-treated and control soil paddocks were dominated by Proteobacteria, Actinobacteriota, Chloroflexi, and Acidobacteriota phyla with differences in respective relative abundance means (relative atomic mass) of 0.34, 0.22, 0.18, and 0.14, respectively, for the control paddock soil and 0.36, 0.30, 0.1, and 0.1 in the compost-treated paddock soil (Figure 3c).

A significantly (p < 0.01) greater abundance of ARB isolates was detected in the compost-treated paddock soils compared to the control soils for all of the antibiotics examined (Figure 3d). In compost-treated soils, high numbers of isolates had resistance to ampicillin (1.4 × 10⁵ CFU/g soil), trimethoprim (1.4 × 10⁵ CFU/g soil), monensin (1.3 × 10⁵ CFU/g soil), and ciprofloxacin (1.3 × 10⁵ CFU/g soil). By contrast, tetracycline had the lowest number of observed CFUs with 4.6 × 10⁴ CFU/g soil in both the compost-treated and control soils, but resistance to tetracycline was higher in the control soil compared to the compost-treated soil. (Figure 3d). Interestingly, samples from the control soil sites showed relatively high CFUs with resistance to the antibiotics: cefotaxime (1.0 × 10⁵ CFU/g soil), ampicillin (1.0 × 10⁵ CFU/g soil), and monensin (0.9 × 10⁵ CFU/g soil) (Figure 3d). Thus, the abundance of ARB in compost-treated soil was significantly greater compared to control, untreated soil but the bacterial diversity structure of the compost-treated and control bacterial soil communities was different.

3.3 16S identification of resistant isolates

The 16S sequence analysis was used to identify the ARB isolates from compost-treated and control soils as well as from fresh manure to track diversity. The results found significant differences in the identity of resistant isolates analyzed from each group. At the phylum level, resistant manure isolates mostly belonged to the Proteobacteria phylum (74.2%) mainly of the genus Pseudomonas, followed by Bacillota (previously named Firmicutes, 12.9%), Actinomycetota (previously named Actinobacteria, 9.7%), and Bacteroidota (previously named Bacteroidetes, 3.2%). The resistant isolates from the soil microbial community largely belonged to the phyla Bacillota (57.1%) mainly from the genus Bacillus, and Proteobacteria (35.7%), followed by Actinomycetota (4.8%) and Bacteroidetes (2.4%). A complete list of the identified isolates and their resistance profiles is outlined in Appendix 1. Notably, the manure isolate Pseudomonas (accession number: NR_025228.1) and soil isolate Bacillus (accession number: NR_152692.1) demonstrated resistance against five different antibiotics. Because ARGs may be located on mobile genetic elements such as a plasmid, isolations from Pseudomonas (NR_025228.1) and Bacillus (NR_152692.1) isolates were performed. Figure 4 illustrated that a plasmid was successfully extracted from Bacillus (NR_152692.1). When this plasmid was transformed to antibiotic-susceptible DH5α E. coli, it conferred antibiotic resistance activity against erythromycin but not to gentamycin, tetracycline, or streptomycin (Figure 4 and Table 2). The erythromycin resistance phenotype was further validated by sequence analysis demonstrating the plasmid carried a genetic element with 94% homology to the erythromycin resistance gene ermF (NG_047824.1) and 99.6% homology for the ermF protein sequence (AAF68230.1). Collectively, these data suggested that in Bacillus (NR_152692.1) erythromycin resistance was conferred by a mobile plasmid element, whereas the genes conferring resistance against gentamycin, tetracycline, or streptomycin remain unknown, but may be integrated into the genome of Bacillus wiedmannii FSL W8-0169.

4.1 ARB and microbial diversity in compost-treated and untreated soils

The purpose of this study was to follow the development of antibiotic resistance from cattle to agricultural soils, via the application of different forms of animal manure as a soil treatment. Analysis of animal manure revealed a decreasing abundance of ARB from feedlot to paddock and finally stored manure, which contained low ARB abundance.

In agreement with a previous study, animal manure, particularly cattle manure, contains a substantial number of ARB even if the animal has not received antibiotics previously (Pu et al., 2019; Udikovic-Kolic et al., 2014). Additionally, higher ARB loads have also been attributed to cattle in feedlots due to the accumulation of antibiotic residue in animal products (Chee-Sanford et al., 2009; Heuer et al., 2011) as well as the limited space leading to increased transmission of ARB via direct contact among animals or through the ingestion of fecal contamination in feed and water sources (Netthisinghe et al., 2013). These conditions may act to apply selection pressure, thereby reducing the overall bacterial diversity of feedlot cattle compared to fresh manure from cattle grazed on paddocks, as has been observed here and in previous studies (Chang et al., 2021; Gaballah et al., 2021; Khan et al., 2008). However, composting manure significantly decreased the abundance and diversity of ARB mainly due to altered environmental conditions such as sunlight, temperature, pH values, and moisture, which all play a significant role in the degradation of antibiotics and the survival of the microorganisms (Chee-Sanford et al., 2009; Koike et al., 2007; Martinez, 2009; Netthisinghe et al., 2013; Sengeløv, 2003). For these reasons, this study observed that the antibiotic resistance rates of bacteria in stored manure were very low compared to fresh manure from feedlot sites.

Spreading composted manure on paddocks produced divergent bacterial communities with a significantly higher abundance of ARB. This is in accordance with other observations in the literature (Pu et al., 2019; L. Wang et al., 2019). Considering that the compost treatment contained low amounts of live ARB, this raises the question, what is the resistance transfer mechanism, if it is not through the application of live ARB?

One hypothesis is that antibiotic resistance is transferred via resistant DNA or ARG mobile elements. Previous studies demonstrated that ARG and mobile genetic elements remain in the compost material following aerobic composting, long after the bacterial decay (Pu et al., 2019; Xie et al., 2016; Zhang et al., 2016). Once spread on the soil, ARG may disseminate and reach the soil microbial community, conferring antibiotic resistance through horizontal gene transfer (Pu et al., 2019). Analysis of the common resistant isolates identified in this study revealed a novel plasmid carrying the erythromycin resistance gene, with demonstrated capacity to confer resistance to erythromycin in other microbes from different genera. Further analysis should be aimed at characterizing antibiotic resistance-carrying mobile genetic elements that are present in compost.

Another mechanism could be due to antibiotic residues within the animal manure that are introduced into the soil with compost application, thus contributing to the accumulation of antibiotics in the soil (Chang et al., 2015; Speicher, 2016), and creating selection pressure that leads to the observed increase in ARB. Previous studies have found that significant concentrations of antibiotics are accumulated in commercial compost of animal manure (J. Wang et al., 2015; Xie et al., 2016). Further studies are needed to understand the mechanisms of antibiotic resistance transfer from compost to the soil and to develop strategies to minimize this process. Interestingly, compost application not only acted to increase ARB abundance but also influenced the overall diversity of the soil microbial population. Further research is needed to identify which components in the compost are the main facilitators of this effect.

4.2 Effect of compost treatment on the development of specific antibiotic resistance

The highest number of ARB in manure samples from feedlot and compost samples was for monensin (Figure 2). This is not surprising as monensin was given prophylactically to the cattle. Following monensin, high resistance was observed for the antibiotics, cefotaxime, and ampicillin, which inhibit cell wall synthesis (Figure 2). This is consistent with previous investigations that reported how bacterial isolates from cattle manure such as Salmonella spp., E. coli, and Pseudomonas spp. showed high resistance levels for ampicillin, cefotaxime, and monensin (Camiade et al., 2020; Ibrahim et al., 2016; Xu et al., 2018).

Interestingly, we detected a high abundance of ARBs in the untreated soil. In the case of tetracycline, ARB abundance in the untreated soil was higher than the one observed in the treated soil (Figure 3). This high resistance may be a result of historical antibiotic use, before the documented history of these soils; alternatively, this soil could be naturally high in tetracycline-resistant bacteria whose abundance was decreased in the treated soils as a result of continuing compost application.