Introduction

The increased demand for food of animal origin has resulted in growing numbers of food-producing animals producing large quantities of manure; indeed, a single dairy cow can produce 54 kg of wet manure per day (Girotto and Cossu 2017), which is equivalent to 7–8% of the body weight of the cow (Font-Palma 2019). Such production has raised environmental concerns which are driving food animal producers to identify economically and environmentally-safe solutions for manure disposal.

Cattle manure management is especially important for Europe, because approximately 21% of global cattle milk production takes place in the European Union (EU), with almost half of this EU production being concentrated in Germany, France, and Poland (FAO 2020). Although, in Europe, regulations limiting antimicrobial use in animal production are restrictive, and antibiotics are applied only to treat bacterial infections, in other regions, antibiotics are still used as growth promoters (Dibner and Richards 2005; Centner 2016; Hu and Cowling 2020). Even though antibiotics are used for medicinal purposes, many cases of antibiotic overuse or misuse have been reported in the animal production sector, and this has been associated with the emergence of antibiotic resistance in bacteria. After ingestion, only a small amount of the consumed antibiotics is absorbed or metabolized from the animal intestine, with about 75% of the applied antibiotic being excreted into the feces or urine: the exact percentage depends on the antimicrobial class (Sarmah et al. 2006; Amarakoon et al. 2016; Spielmeyer 2018; Filippitzi et al. 2019). Limitation of antibiotic use seems to be a proper way to reduce the severity of antibiotic resistance occurrence (Kaur Sodhi and Singh 2022).

Due to the breeding programs applied in the dairy industry, modern dairy cows are now highly productive; however, they are also prone to a range of infections, particularly those related to the mammary gland (Zalewska et al. 2020). In such cases, bovine mastitis, dry cow therapy, and udder health disorders in dairy cattle are often treated using various antibiotics, including cephalosporins, penicillins, and macrolide-lincosamides (Barlow 2011). In fact, these conditions are the most common cause of antibiotic use, followed by respiratory, reproductive, and gastrointestinal diseases (Ferroni et al. 2020). Antimicrobial prescription practices vary by country, and cattle are also often administered aminoglycosides, amphenicol, quinolones, and sulfonamides, often combined with trimethoprim (Merle et al. 2012; Olmos Antillón et al. 2020; Ferroni et al. 2020; Diana et al. 2020). In addition, dairy cattle demonstrate significantly higher mean annual antibiotic consumption than beef cattle (Merle et al. 2012; Ferroni et al. 2020).

The gastrointestinal tract of dairy cattle is inhabited by a diverse bacterial community, whose composition and function play key roles in animal wellness and production efficiency. The composition of the community varies along the gastrointestinal tract; however, the most common bacteria overall belong to the phyla Bacillota (65%), Bacteroidota (15%), and Pseudomonadota (13%) (Mao et al. 2015). Although antibiotics intended to target specifically pathogens, other non-target bacteria are also affected (Mann et al. 2021). Antibiotic treatment of cattle induces shifts in the taxonomic structure and biodiversity of the gastrointestinal microbiota (Beyi et al. 2021); however, this effect varies between the compartments of the digestive tract, with the rumen being the most heavily impacted (Thomas et al. 2011). Unfortunately, this rich microbiota may also harbor antibiotic resistance genes (ARGs). Animal feces can therefore become a source of manure-borne antimicrobial resistance (AMR) spread, especially in more intensive livestock farming scenarios, characterized by both high amounts of bacteria and significant antimicrobial selective pressure, promoting the occurrence and enhancement of antibiotic-resistant bacteria (ARB) and ARGs (Ferroni et al. 2020; Huygens et al. 2021). AMR determinants such as ARB, ARGs, and mobile genetic elements (MGEs) can contaminate surface water, groundwater, rivers, ponds, lakes, soil, or air through several pathways: (a) directly through runoff from animal production/breeding systems, (b) via the direct application of animal manure on arable fields for crop production, (c) through irrigation water from manure storage ponds and manure treatment facilities, and (d) via air particles (McEachran et al. 2015; Nguyen et al. 2020). During therapy, the antibiotic reaches the highest possible concentration in the treated body for efficient therapy without any toxic effect on the host, but when an antibiotic is excreted into the natural environment, it reaches doses sublethal for many bacteria. In this way, the antibiotic only causes a selection pressure that stimulates the bacteria to develop resistance (Kaur Sodhi and Singh 2022). Antibiotics have been found in almost all environments, making them the pollutant of enormous concern. It has been estimated that until 2050, ARB will constitute the main cause of mortality; thus, new ways to efficiently remove them from the environment as well as new tools to combat already developed AMR are needed (Kaur Sodhi and Singh 2022; Singh et al. 2023).

In 2020, manure production in dairy cattle in Poland totaled approximately 149 million kg of nitrogen content, accounting for about 7.5% of production in the EU (FAO 2020). The easiest disposal strategy for unprocessed manure is direct land application; however, such raw manure might contain antibiotic residues, ARB and ARGs, thereby increasing the risk of AMR being spread into the environment. This poses a significant threat to both human and animal health. The persistence and leaching of antibiotics into the ground and surface waters depend on the physicochemical properties of the antimicrobial compound and the fertilized soil. These compounds can persist for extended periods, and eventually accumulate in crops (Zalewska et al. 2021). The presence of ARGs in the environment has even greater implications compared to the presence of antibiotics itself (Ahmed et al. 2021). Consequently, extensive research on the effects of manure management practices is necessary to understand their effect on microbial communities and the spread of AMR.

The choice of manure management method can influence the microbial community of the manure and reduce the survival of pathogens (Heinonen-Tanski et al. 2006). The most common strategy involves storing the feces for a prolonged time, usually between 4 and 7.5 months; however, this requires sufficient storage capacity (Loyon 2018). An alternative strategy is the composting of manure, which is both economically and environmentally beneficial, as it reduces solid organic waste volume while eliminating pathogens, antibiotic residues, and ARGs (Dolliver et al. 2007; Tasho and Cho 2016; Gou et al. 2018). Composting transforms organic waste into a stable, humus-like product that is easily available for soil amendment. Although composting offers numerous advantages, estimates from 2011 show that only 0.8% (10.4 million tons) of total EU livestock manure was composted. Moreover, only about 8% of the manure used in fields underwent any form of treatment (Foged et al. 2012). In the Netherlands, 75% of manure from calves was applied to agricultural land without any treatment (Berendsen et al. 2018). It is believed that the primary reason for omitting such processing is the lack of adequate storage capacity (Köninger et al. 2021).

The structure of the bacterial community in cattle manure and its resistome (pool of ARGs in the bacterial community) remain relatively unknown, with a lack of comprehensive data on the subject. Until now, little studies have compared the effects of storage and composting of manure originating from the same source; only a few have examined the effects of multiple manure treatment strategies on the same raw manure batch. The aim of the study was to determine how composting and storage of dairy cattle manure impact the ARGs and microbiome of animal feces, with the goal of limiting the potential risk of spreading ARGs in the fields and compare these both manure treatment techniques in term of better preparation manure for land application. The material was collected simultaneously from the same animals, allocating one portion of the sample for storage and another for composting. The analysis considers the type of process as the independent variable, with ARGs relative abundance and bacterial phylum relative abundance as the dependent variables. The changes in microbiome and resistome in composted and stored manure compared to raw material were determined by high-throughput qPCR for ARGs tracking and sequencing of the V3–V4 variable region of the 16S rRNA gene to indicate bacterial community composition. The V3–V4 variable region of 16S rRNA sequencing is a well-known technique for microbial community determination. High-throughput qPCR uses the SmartChip nanowell platform for large-scale detecting and quantifying specific gene sequences (particular primers are already located within wells); it can process 5184 nanowell reactions per run.

Materials and methods

Cattle manure collection

The study was conducted on 50 Polish Holstein–Friesian dairy cows, Black and White variety. The herd was located in the central part of Poland (Masovian voivodeship). The animals were kept in a loose barn with free access to water, and fed the same total mixed ration (TMR) diet ad libitum, consisting of corn silage (75%), concentrates (20%), and hay (5%), supplemented with a mineral and vitamin mixture, according to the INRA system (ruminant feeding system developed at the Instytut National de la Recherche Agronomigue, France) (Brzóska et al. 2014). The cattle manure samples were collected in October 2019. The manure samples were deposited into two plastic drums measuring approximately 20L each. One aliquot of manure was subjected to composting, and the other to storage on a laboratory scale. The farm owner agreed to the manure being sampled and shared the history of the use of antibiotics on the farm.

Manure management strategies

Defined amounts of cattle manure (approx. 5 kg) were mixed with Sitka spruce (4:1 ratio; w/w) and placed in a bucket. The Sitka spruce was sourced from a local hardware store. Containers with manure were placed in the laboratory under a ventilated hood at room temperature. All samples originate from the same batch of animal feces, collected during one sampling separated into two portions of manure—one designated for composting and second for storage. Each of them was in turn divided into three parts and each of them was mixed well with Sitka spruce (three biological replicates per treatment). Then each biological replicate was separated into three technical replicates. During composting, the manure was mixed well each week for 10 weeks, while the stored manure was left undisturbed for 4 months. Although the duration of both processes differs, the composting and storage are both, highly recommended for manure treatment before its land application and both were applied according to commonly used practices—the composting process was completed within 8 weeks (56 days), as noted by Mc Carthy et al. (2011) and the 4-month storage period was selected based on EU legislation. Moreover, the same time intervals were applied by Do et al. (2022, 2023) during their study on manure treatment. Samples for the microbiological and molecular analysis were collected at the beginning (raw/untreated manure), in the middle (a fifth week for composting, 5W; second month for storage, 2 M), and at the end of the process (a tenth week for composting, 10W; fourth month for storage, 4 M). For statistical analysis purposes, samples were divided into five groups representing processes and time points; corresponding technical replicates were collected and pooled together as replicates. A total of 200 samples were collected during the entire experiment. From each group, we gathered 40 samples, encompassing three technical repetitions (represented as pooled samples from technical replicates). The groups included control samples, composted samples at 5 and 10 weeks, and stored samples at 2 and 4 months. DNA isolated from these samples was grouped accordingly and then dispatched for analysis.

The pH, moisture, and temperature were measured in both mixtures at approximately half of their height (~ 15 cm) each week for 10 weeks (enough time to observe the dynamics of changes during composting and to stabilize parameters during storage). The moisture content was adjusted to 50–60% for compost by adding sterile MilliQ water. The moisture, pH, and temperature were measured directly in the containers with treated samples along with the time. The moisture and pH were measured by a dual-purpose pH meter dedicated to environmental samples (pH-meter Stelzner 3000/Soiltester). Temperature was measured by laboratory thermometer (Testo 905-T1).

DNA extraction

The total DNA was extracted using the commercially available FastDNA™ SPIN Kit for Feces (MP Biomedicals, California, USA) according to the manufacturer’s recommendations. The quality and quantity of the extracted DNA were analyzed with a Qubit 4.0 Fluorometer using the dsDNA high-sensitivity assay kit (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) and a Colibri spectrophotometer (Titertek Berthold, Pforzheim, Germany). The DNA samples were isolated in triplicate and then pooled to obtain a single DNA sample for each time point.

Sequencing the variable V3–V4 regions of bacterial 16S rRNA

The bacterial community structure was determined using Illumina technology on the Miseq platform by sequencing the variable V3–V4 regions of bacterial 16S rRNA. The libraries were prepared with Nextera® XT index Kit v2 Set A (Illumina, San Diego, CA, USA). Once prepared, amplicons underwent sequencing, 2 × 300 bp paired-end sequences. PCR and sequencing were performed at the DNA Sequencing and Oligonucleotide Synthesis Facility (Institute of Biochemistry and Biophysics Polish Academy of Sciences). Raw sequences were processed and analyzed using QIIME2 software suite (Bolyen et al. 2019) with the DADA2 option for sequence quality control and the newest release of the SILVA (SILVA SSU database 138, accession date October 2021) ribosomal RNA sequence database for taxonomy assignment (Quast et al. 2013; Yilmaz et al. 2014).

High-throughput qPCR

During the study, we decided to use high-throughput qPCR techniques as a convenient way to track up to 384 defined genes at one run (in parallel) using the SmartChip Real-Time PCR system. In each chip analysis, up to 384 genes, including the 16S rRNA gene as a positive control, ARGs targeting aminoglycoside, amphenicol, beta-lactam, florfenicol, multi-drug efflux pump, macrolide-lincosamides-streptogramin B (MLSB), trimethoprim, tetracycline, vancomycin; genes targeting other antibacterial agents such as nisin, bacitracin; and genes that are associated with MGEs and integrons, can be measured. Primers are located in the well, and each primer set was designed to target a specific region in an ARG.

High-throughput qPCR was performed by an outside firm (Resistomap, Helsinki, Finland) using the qPCR SmartChip Real-Time PCR cycler (Takara; Kusatsu, Japan). The qPCR cycling conditions and initial data processing were carried out as previously described by Wang et al. (2014). The analysis used a 384-well template of primers. All samples used for the analysis met the following criteria: 260/280 ratio of 1.8–2.0 (+ / − 0.1) and a concentration of 10 ng/μl.

The threshold cycle (Ct) values were calculated using default parameters provided by the SmartChip analysis software. All samples for each primer set underwent qPCR efficiency and melting curve analysis. Amplicons with unspecific melting curves and multiple peak profiles were considered false positives and excluded from further analysis. Samples meeting the following criteria were selected for analysis: (1) Ct ≤ 27, (2) at least two replicates, (3) amplification efficiency 1.8–2.2. The relative copy number was determined using an equation originally published by Chen et al. (2016) and adapted to our study conditions. The gene copy numbers were calculated by normalizing the relative copy numbers per 16S rRNA gene copy numbers.

$$\mathrm{Relative\ gene\ copy\ number}\hspace{0.17em}=\hspace{0.17em}{10}^{(27-{\text{Ct}})/(10/3)}$$

The genes and their corresponding primer sequences can be found in Supplementary Materials File 1. The gene names and groupings, as detailed and organized in the table from Supplementary Materials File 1, are consistently referred to throughout the manuscript.

Statistical analysis

The minimal sample number was determined by the G*Power software (version 3.1.9.7) (Faul et al. 2007, 2009). The amplicon data, including ARGs distribution, was analyzed and visualized using the MicrobiomeAnalyst web server (https://www.microbiomeanalyst.ca/)Chong et al. 2020. Data were not rarefied, not scaled and transformed. The differences in microbial community structure were evaluated using the QIIME2 pipeline, based on the Kruskal–Wallis H-test (Shannon and Chao1 indexes). The differences between ARGs groups at each stage of the process were calculated using analysis of variance (ANOVA) with post hoc Tukey’s test (GraphPad Prism 9). The correlations between ARGs and MGEs and phylum and ARGs groups were calculated using Spearman’s correlation coefficient with GraphPad Prism 9. The networks representing correlations were built with Cytoscape (3.9.0 version) (Shannon et al. 2003). The correlations between ARGs and MGEs were considered strong and significant when the absolute value of Spearman’s rank |r|> 0.9 and p < 0.05 (Zhu et al. 2017), and for phylum and ARGs groups, |r|> 0.9 and p < 0.05.

Results

Microbiome composition

During the analysis, we identified 28 amplicon sequence variants (ASV) at phylum level (Table S2 Supplementary Material File 2—mean abundance with standard deviation of each phylum across the treatment). In all samples, the microbial community composition was determined for the untreated (raw) cattle manure and the composted and stored manure; these determinations were performed at the beginning, middle, and end of the process. The humidity, pH, and temperature were checked during the process, to ensure their proper conduct (Table S1 of Supplementary File 2). The composition of bacterial communities at the phylum level is presented in Fig. 1, using the taxon bar plot (additional table with total reads count S1 in Supplementary material file 2), and the core microbiome is presented in Fig. S1 (Supplementary material file 2). Core microbiome consists of members of the community that are common to all communities across analyzed samples (Sharon et al. 2022). In raw manure, Bacillota has the highest abundance, while at the end of both processes, the shift to Pseudomonadota as the most abundant phylum was observed. Interestingly, when physicochemical parameters of raw material changed along with both processes, new phyla were detected. During composting, Myxococcota, Bdellovibrionota, Chloroflexota, Gemmatimonadota, Acidobacteriota, Hydrogenedentes, Armatimonadota, WPS-2, Dependentiae, Abditibacteriota, and Deferribacterota increase its abundance, and become detected by applied method, while Euryarchaeota and Elusimicrobiota abundance decrease, reaching levels below detection limit. During storage, Myxococcota, Bdellovibrionota, Chloroflexota, Gemmatimonadota, and Acidobacteriota become detectable, while Spirochaetota, Desulfobacterota, and Elusimicrobiota prevalence become below the detection level. The richness and diversity of the microbial communities were determined using the Shannon (comparable results, p-value 0.08) and Chao1 (higher value for composted samples, p-value 0.025) indexes (Figs. S2, S3 Supplementary Material File 2). The raw data obtained from sequencing was uploaded to the Sequence Read Archive (SRA) repository (BioProject ID: PRJNA1007935) (NCBI). The PCoA of taxon abundance in all sample groups, calculated based on the Bray–Curtis distance, revealed significant visual separation between the initial (untreated) and treated samples (p-value < 0.01) (Fig. S4 Supplementary Material File 2; the results of pairwise PERMANOVA analysis are presented in Table S2 Supplementary Material File 2).

Fig. 1
figure 1

Relative abundance of phyla across samples; the samples were divided into five groups–control (cow manure raw; sample before treatment), cow manure composting 5 W, cow manure composting 10 W (composted samples after 5 weeks and 10 weeks, respectively), cow manure storage 2 M, and cow manure storage 4 M (stored samples after 2 months and 4 months, respectively); stacked bar plots represent mean value for replicates

Full size image

Diversity and abundance of ARG

During the study, we found a number of 202.67 (mean value from replicates) out of 384 analyzed gene types (particular genes belonging to aminoglycoside, beta-lactam, cross-resistance to macrolides, lincosamides and streptogramin B (MLSB), phenicol, sulfonamide, tetracycline, trimethoprim, vancomycin resistance genes together with integrons, MDR pumps (multi-drug resistance), MGEs, and “other” genes) in raw manure, 76 in composted samples after 5 weeks, 54 after 10 weeks, and 51 in stored samples after 2 months and 39 after 4 months. The relative abundances of the analyzed ARGs classes and MGEs varied as storage and composting processed (Fig. S4 Supplementary Material File 2). Specifically, the relative prevalence of beta-lactam, MLSB, and vancomycin resistance genes fell in the treated manure samples compared to the untreated samples. In contrast, the relative abundance of genes encoding MGE, integrons, and sulfonamide resistance genes increases, but only after composting. However, the total gene count (i.e., total resistance gene abundance normalized per 16S rRNA gene) rapidly decreased following the application of both treatment strategies. Notably, for composting, the gene relative abundance was even lower in the middle of the process than at the end (Fig. 2).

Fig. 2
figure 2

The differences in the amounts of ARGs and MGEs detected during composting and storage compared with raw manure. The following cut-off points for significance within the same process were applied: the values differ significantly at p < 0.01 (indicated as *); the values differ significantly at p < 0.05 (indicated as **); the values do not differ significantly at p > 0.5. Relative abundances of genes were additionally compared at the end of both processes between them; the value with the same letters differs significantly: A, A at p < 0.01.; a, a at p < 0.05 (indicated as A, B, C, D, E, F); the values differ significantly at p < 0.05 (indicated as d)

Full size image

Composting efficiently reduced the quantity of aminoglycoside and tetracycline resistance genes, as well as MGEs, within the initial 5 weeks (twofold, fivefold, 2.5-fold reduction respectively, p < 0.01). However, a subsequent rise in their abundance was observed after 10 weeks of treatment compared to their level after 5 weeks of treatment. Overall, compared to raw material, it still remains lower (1,twofold, threefold, and twofold reduction respectively, p < 0.01). Beta-lactam, vancomycin resistance genes, MLSB-coding genes, and genes responsible for MDR demonstrated a decline at both the 5- (66-fold for beta-lactam resistance genes, 74 for MLSB-coding genes, and sixfold for MDR genes, with vancomycin resistance genes being below detection limit; p < 0.01) and 10-week (156-fold for beta-lactam resistance genes, 80-fold for MLSB-coding genes, and fivefold for MDR genes, with vancomycin resistance genes being below detection limit; p < 0.01) marks, with the rate of reduction being consistent between the two points. The abundance of integron-coding genes decreased after 5 weeks of composting (1.5-fold, p < 0.01), but remained unchanged after 10 weeks compared to raw sample and sample after 5 weeks of composting. Neither phenicol nor trimethoprim resistance gene abundance varied between composting stages. However, the sulfonamide resistance gene count increased after 5 weeks of composting (0.6 fold, p < 0.05), and exhibited an even greater increase after 10 weeks (0.33-fold, p < 0.01).

Our results show that storage effectively diminishes the quantities of aminoglycoside resistance genes (fivefold, p < 0.01), MGEs (ninefold, p < 0.01), and genes responsible for MDR (12-fold, p < 0.01) and MLSB-coding genes (8.5-fold, p < 0.01) after 2 months. However, their numbers subsequently increased after 4 months of treatment, but still remains lower than in raw manure (p < 0.01). The counts of beta-lactam, vancomycin resistance genes, and those labeled as “other” demonstrated a decline both at the 2- (289-fold, 8.5-fold, respectively, with vancomycin resistance genes being below detection limit; p < 0.01) and 4-month marks (579-fold, sevenfold, respectively, with vancomycin resistance genes being below detection limit; p < 0.01), yet the rate of reduction remained consistent between these points. The abundance of genes coding for integrons remained consistent after 2 months of storage but declined after 4 months (threefold, p < 0.01). No noticeable change in the amounts of phenicol, sulfonamide, and trimethoprim resistance genes was observed across the storage periods. The tetracycline resistance genes were lower after both the 2 (eightfold, p < 0.01) and 4-month (fourfold, p < 0.01) periods, but their concentration was notably higher at the end of the fourth month compared to the second.

Both composting and storage processes reduce the abundance of aminoglycoside resistance genes, with storage achieving a greater reduction. A similar trend can be observed for the abundance of MLSB-coding genes, MGEs, and genes classified as “other.” For beta-lactam, tetracycline, vancomycin resistance genes, and genes responsible for MDR, a reduction in gene abundance was noted after both processes. However, composting and storage yielded comparable outcomes by the end of treatment. While storage resulted in a decreased abundance of integron-coding genes relative to untreated samples and 10-week composted manure, composting itself did not appear to affect these genes. For sulfonamide resistance genes, composting increased their amount, while storage did not include any noticeable changes. Nevertheless, after 2 months of composting, the abundance of sulfonamide resistance genes remained elevated compared to post-storage levels. The quantities of the phenicol and trimethoprim resistance genes did not demonstrate any significant variation after the two treatments compared to untreated manure (Fig. 2).

The relative abundances of ARGs and MGEs across all analyzed samples were subjected to principal coordinate analysis (PCoA), using the Bray–Curtis distance method. The results indicate a significant visual separation between the untreated manure (control) and treated samples with regard to the relative abundances of resistance gene classes (p < 0.01).

The core resistome across all analyzed samples consisted of 29 genes. They belonged to the following groups: aminoglycoside resistance genes (aadA5_1, aadA_2, aac3-VI, aacC2, aadA1, strB, aadA2_3, aadA2_1), beta-lactam resistance genes (fox5), phenicol resistance genes (floR_1), sulfonamide resistance genes (sul1_1, sul2_1), tetracycline resistance genes (tetA_2, tetG_1, tetG_2), trimethoprim resistance genes (dfrA1_1, dfrA1_2), integron-coding genes (intI1_3, intI1_1, intI1_4), genes responsible for MDR (mdtH_1, oprJ, acrR_3), MGE (tnpA_2, ISSm2, IncP_oriT, repA), and resistance genes classified as “other” (qacE∆1_2, qacE∆1_3). For both applied treatment strategies, a decreased number of ARGs was observed post-treatment compared to raw manure (53 for compost and 38 for storage) (Fig. S5 Supplementary material file 2).

In all composted and stored samples, almost all ARGs, excluding trimethoprim resistance genes, were found to be reduced at the end of both processes, which means lower diversity of genes after treatment. In addition, the stored samples demonstrated lower numbers of aminoglycoside, beta-lactam, phenicol, sulfonamide, tetracycline resistance genes, and MGEs compared with composting. However, the composted samples demonstrated lower numbers of genes conferring MLSB and MDR resistance compared to the stored samples (Fig. 3).

Fig. 3
figure 3

Comparison between ARGs (assigned to ARGs classes) at the end of both treatment processes

Full size image

In the composted samples, the correlation network analysis revealed 124 nodes connected by 512 edges, representing strong and significant correlations (|r|> 0,99; p < 0.05) (Fig. 4). All correlations were positive. The MGEs (integrons included) exhibited more links than ARGs, indicating their significant role in network formulation. The MGEs (n = 15) included in this network are intI3_1, orf39_IS26, IncN_rep, Tn5, Tp614, tnpA_5, ISAba3, trfA, IS613, intI3_2, tnpA_6, intI1_3, ISSm2, IS1111, and tnpA_2.

Fig. 4
figure 4

ARGs and MGEs interaction network in composted manure. The network is presented as an “organic layout.” Based on Spearman’s rank correlation, a strong and significant correlation is shown, where |r|> 0.99 and p < 0.05. The size of the nodes represents the degree of interaction. The shape of the nodes represents: hexagon, MGE group; square circle, ARG. The gray edges show positive correlations between ARGs and MGEs. The color of the nodes indicates ARG (white) and MGE (gray)

Full size image

In the stored samples, strong and significant correlations were found between MGEs (integrons included) and ARGs levels, manifested as a network containing 144 nodes connected by 643 edges (|r|> 0.99; p < 0.05) (Fig. 5). Every correlation in the network was positive. The MGEs (integrons included) had more connections than ARGs, indicating they play a pivotal role in the formation of the network. The MGEs (n = 16) presented in the network are intI3_2, IS613, IS1111, ISAba3, orf37-IS26, tnpA_1, trfA, tnpA_5, orf39_IS26, Tn5, Tp614, tnpA_6, intI2_2, IncN_rep, intI1_1, and intI3_1.

Fig. 5
figure 5

ARGs and MGEs interaction network in stored manure. The network is presented as an “organic layout.” Based on Spearman’s rank correlation, a strong and significant correlation is shown, where |r|> 0.99 and p < 0.05. The size of the nodes represents the degree of interaction. The shape of the nodes represents: hexagon, MGE group; square circle, ARG. The gray edges show positive correlations between ARGs and MGEs. The color of the nodes indicates ARG (white) and MGE (gray)

Full size image

In composted samples, negative correlations were found between beta-lactam resistance genes and MDR and Spirochartota and Cyanobacteria, and positive ones between MLSB and Acidobacteriota and WPS-2 group (Fig. 6).

Fig. 6
figure 6

Phylum and ARGs group interaction network in composted manure. The network is presented as a “circular layout.” Based on Spearman’s rank correlation, a strong and significant correlation is shown, where |r|> 0.9 and p < 0.05. The size of the nodes represents the degree of interaction. The shape of the nodes represents: hexagon, ARGs group; square circle, bacterial phylum. The gray edges show positive correlations, while red one—negative. The shape of the nodes indicates phylum (square with rounded corners) and ARGs group (hexagon)

Full size image

In stored samples, only negative correlations were identified as follows: between sulfonamide and Bdellovibrionota, Chloroflexota, and Gemmatimonadota, between tetracycline and Bdellovibrionota, Gemmatimonadota, Chloroflexota, and Myxcoccota, between vancomycin and Desulfobacterota and between aminoglycosides and Campilobacterota; however, vancomycin and aminoglycosides create separate clusters (Fig. 7).

Fig. 7
figure 7

Phylum and ARGs group interaction network in stored manure. The network is presented as a “circular layout.” Based on Spearman’s rank correlation, a strong and significant correlation is shown, where |r|> 0.9 and p < 0.05. The size of the nodes represents the degree of interaction. The shape of the nodes represents: hexagon, ARGs group; square circle, bacterial phylum. The gray edges show positive correlations, while red one—negative. The shape of the nodes indicates phylum (square with rounded corners) and ARGs group (hexagon)

Full size image

Discussion

Farms often dispose of animal manure by direct application to the land. Many also employ a range of manure management technologies to reduce environmental pollution before application. These technologies not only reduced the total manure volume but also can be used to produce organic fertilizer, or possibly facilitate biogas production. Two such technologies are storage and composting, after which the feces are often used as organic soil amendments with potential benefits for soil quality and crop yield (Epelde et al. 2018). Such amendments typically contain all nutrients essential for plant growth and can replace or complement inorganic fertilizers in agricultural production systems (Youngquist et al. 2016). However, cattle manure can also serve as a reservoir of environmental ARG pollution (Udikovic-Kolic et al. 2014; Zalewska et al. 2021), and the direct use of livestock manure without proper treatment could promote the dispersal of ARG in arable soil and cultivated plants (Ghosh and LaPara 2007; Heuer et al. 2011; Udikovic-Kolic et al. 2014; Zalewska et al. 2023). The presence of ARGs in soil samples collected from fields amended with manure, such as tetW, tetO, tetT, tetM, tetA, tetL, tetQ, sul1, sul2, and sul3, was confirmed (Zhao et al. 2017). Similar studies from Finland were focused on changes in ARGs and MGEs presence in the soil after fertilization with swine and cow manures. The results have shown ARGs from animal manures in fertilized soil; however, ARGs associated with manure decreased in samples collected after 2 and 6 weeks after fertilization, but compared to unfertilized soil, the ARGs level against disinfectants, aminoglycosides, or vancomycin resistance genes was still elevated (Muurinen et al. 2021). Moreover, the impact of cow and swine manure applications on the soil resistome was examined by Marti et al. (Marti et al. 2014) in Canada. According to the study, compared to non-amended soil, the abundance of targeted genes such as sul1, ermB, strB, int1, and repA was higher in fertilized soils. Plants may naturally uptake bacteria, antibiotics or antibiotic residues from soil, but this ability may additionally increase due to manure application (Christou et al. 2019). This occurrence may exert long-term pressure to facilitate drug resistance and its spread across plant resistome (Chen et al. 2019). ARB associated with soil and manure may enter the plant microbiome by colonizing the vegetable roots, which are in direct contact with soil, or aboveground parts of vegetables, potentially through air particulates or the motility of root endophytes (Guron et al. 2019).

Manure treatment strategies have not been specifically designed to mitigate AMR (Oliver et al. 2020); however, our present findings indicate that composting and storage are effective methods of reducing ARGs from manure before use. It is hard to define which process accomplishes this task better, as despite a generally observed significant reduction in the number of ARGs, particular groups of resistance genes responded differently to the applied strategies. For example, the levels of aminoglycoside resistance genes, integrons and MGEs, genes classified as “other” and sulfonamide resistance genes were lower after storage than composting. In contrast the MLSB levels in the stored samples were higher than those in the composted samples, but they were still lower than those found in the raw material. No differences were observed between the composted and stored samples for beta-lactam, tetracycline, vancomycin resistance gene, and MDR genes; those genes are often prevalent in animal manure (Kühn et al. 2005; Qiu et al. 2022). Similar results for composting were obtained by Gou et al. (2018), who found higher amounts and numbers of ARGs in unprocessed manure compared to composted samples.

Bacterial community structure

Wang and Zeng (2018) reported that bacteria play a greater role in composting than fungi. Pitta et al. (2016) identified several prevalent bacterial phyla in dairy cattle feces and raw manure: Bacteroidota, Bacillota, Pseudomonadota, Actinomycetota, Acidobacteriota, and Spirochaetota. This aligns with our present findings, which also included Spirochaetota and Acidobacteriota, though they were not the most represented phyla. Moreover, Zhou and Yao (2020) report the presence of Pseudomonadota and Bacillota as the main phyla in animal feces.

The decomposition of organic matter starts before composting commences, as microorganisms are introduced unintentionally to the organic matter during grinding, mixing, cutting, and transporting. The composting pile initially rests at an ambient temperature. It then gradually heats up during the mesophilic phase to 25–45 °C, indicating increased microbial activity (Ezugworie et al. 2021). The diversity of microbial community correlates with the metabolic functions of the microbes, e.g., lactic acid-producing bacteria and acetate-oxidizing bacteria are usually detected at the mesophilic phase of the composting process, where they utilize the easily degradable and soluble organic matter. Zhang et al. (2022) have found that Bacillota, Pseudomonadota, Bacteroidota, Chloroflexota, Actinomycetota, and Planctomycetota were the dominant phyla during co-composting of cow manure and maize straw. In addition, Zhang et al. (2020) report the presence of Bacillota, Bacteroidota, Deinococcota, Pseudomonadota, and Actinomycetota in raw manure and composted samples; however, the abundance of Deinococcota and Actinomycetota was much higher in composted samples. Similarly, Zhou and Yao (2020) noted that composting resulted in a shift in the bacterial community, with an increased prevalence of Actinomycetota compared to raw manure. Our present findings also identified Pseudomonadota, Bacillota, Planctomycetota, and Actinomycetota as being the most prevalent during composting; however, their ratio changed with time. Zhang et al. (2016) found the Bacillota, Pseudomonadota, Bacteroidota, and Actinomycetota to predominate in the thermophilic phase of composting. Of these, Bacillota were the dominant phylum due to their adaptability to aerobic condition and heat resistance. Temperatures exceeding 70 °C terminate most microbial activities, forcing certain bacteria to sporulate (Ezugworie et al. 2021).

While our sequencing results differed from those found in some previous studies, the identified phyla remained consistent. Composting can be influenced by a range of physicochemical factors, such as temperature, moisture, pH, aeration, and carbon-to-nitrogen ratio. As such, as there are no strict guidelines on composting, composted samples can differ widely. A crucial role is played by temperature: it governs the pace of biological processes conducted by microorganisms and determines the presence of particular microorganisms; for example, fungi, which are not heat-resistant, are eliminated at temperatures above 50 °C. In addition, a pH of 7–8 is needed for optimal composting, which slows down at low pH. The C/N ratio should also be kept at optimum, as nitrogen is crucial for microbial growth, and carbon is an energy source. Humidity should be kept at 40–60% to ensure the transfer of dissolved nutrients necessary for microbial growth (Ezugworie et al. 2021). As such, the differences in bacterial community structures obtained by different research teams may hence be due to the differences in physicochemical parameters during the process. In our case, the temperature was within the proper range, but the humidity and pH were slightly suboptimal (humidity above 80% and pH below 7).

Changes in antibiotic resistance genes after treatment

Tetracycline resistance genes

The tetracyclines are the most commonly used antibiotics in the livestock production sector; as such, tetracycline resistance genes are the subject of intensive study. These genes appear to be unaffected by manure treatment strategies such as composting or storage, and the detection rate of tetW and tetO may account for up to 80% of fecal samples obtained from cattle raised in grassland-production systems, even without antibiotic usage (Santamaría et al. 2011). The following tetracycline resistance genes were identified in the present study: efflux (tetA, tetA/B, tetG, tetH, tetC, tetL, tetA(P)), protection (tet(36), tet(32), tetO, tetQ, tetW, tetM, tetPB), deactivation (tetX), and regulation (tetR) in raw manure samples. Our findings are partially in line with those of other studies (Jauregi et al. 2021). Zhang et al. (2020) found tetW, tet40, tetO, tetQ, and tetC to be the most prevalent types across all analyzed raw dairy cattle manure samples, accounting for over 40% of detected ARGs. Marti et al. (2013) report the presence of tetM, tetQ, tetS, tetT, tetA, tetB/P, and tetW in cattle manure, and Mu et al. (2015) found tetM, tetO, and tetW in cattle manure, and tetM, tetO, tetQ, and tetW in feces collected per rectum; the latter partially agrees with our present findings indicating these genes to be present in manure.

However, Zhou and Yao (2020) reported a general increase in the abundances of total tetracycline resistance genes, which contradicts our results. Despite this, as in our present study, Zhang et al. (2020) failed to detect tetA, tetW, and tetR, and that total ARGs abundance decreased after composting. Interestingly, Zhang et al. (2020) also indicate a reduction in tetW, tetA, and tetR abundance with composting time, together with a decrease in the concentration of the total amount of ARGs. Staley et al. (2021) did not identify any differences in the concentrations of tetO and tetQ genes between composted and stored samples; however, their level was found to change with sampling season and depth. In addition, Qian et al. (2016) report increased abundance of tetC and tetX, and decreased abundance of tetQ, tetM, and tetW after composting.

Interestingly, while Wang et al. (2019) did not identify any decrease in tetW, tetO, or tetM, Storteboom et al. (2007) indicate that the tetO and tetW genes appeared to gradually decrease during composting; this agrees with our present findings indicating that tetO was present in the middle of composting, but below the detection limit at the end of the process. However, our data indicate that tetW, together with tetA, tetG, tetPB, tetQ, tetM, tetR, tetX, and tetA/B, remained persistent during composting.

Another study examining the tetracycline resistance genes tetO and tetA found no noticeable change in their abundance, even after 1 year of storage (Hurst et al. 2019). These findings are in line with ours, where tetA remained persistent throughout storage, contrary to tetO. Interestingly, Jauregi et al. (2021) found tetW, tetO, tetM, tetB, tet32, tetN, and tetG to be present in stored cattle manure even after 6 months; in contrast, tet32, tetO, tetW, and tetM were below the detection limit after 2 months of storage but reappeared after 4 months.

Tetracycline resistance genes can be transmitted to gram-negative and gram-positive bacteria species, e.g., tetM, tetW, and tetQ originate from gram-positive bacteria and can be transmitted across Gram-positives and Gram-negatives (Chopra and Roberts 2001). Moreover, tet genes are often placed on plasmids or inserted in transposons, facilitating their dissemination (Lima et al. 2020). Jauregi et al. (2021) regard tetM, tetO, tetW, tet32, and tetS as being of particular concern due to their high dissemination potential associated with their co-occurrence with MGEs; they also highlight tet44 and tetW/N/W, which were not detected during our study. Our results partially agree, as our data also indicates some tetracycline resistance genes to be significantly associated with MGEs: tetM (tnpA, IS613, trfA, orf37_IS26, ISAba3, IS1111, IntI3), tetO (tnpA, trfA, orf39_IS26, orf37_IS26, ISAba3, IS1111, Tn5, IntI3, tp614, IS 613), tet32 (intI3); however, neither tetW or tetS were confirmed in samples obtained from either process. We also found tet36, tetPB, tetC, tetA(P), tetC, tetH, and tetL to have a high probability of being exchanged by horizontal gene transfer (HGT) due to co-occurrence with MGEs.

Beta-lactam resistance genes

The following beta-lactam resistance genes were identified in raw manure, classified as deactivation (ampC, blaMOX/blaCMY, blaOCH, blaSHV, blaROB, blaOXY, cphA, cfxA, blaCMY2, ampC/blaDHA, blaFOX, blaCTX-M, blaSFO, blaVIM, blaKPC, blaOXA, blaNDM, blaACC, bla1, blaCMY, blaACT) and protection (penA, pbp2b). Our results are similar to those of Marti et al. (2013), who identified blaOXA, blaCTX-M, blaPSE, and blaVIM, as well as blaTEM, which was not found in the present study. Pitta et al. (2016) report the presence of pbp1A, pbp2, and pbp2B in dairy cattle manure; however, these were not present in our samples. While Zhou and Yao (2020) confirm our observed general decrease in beta-lactam resistance genes, Keenum et al. (2021) found the levels of beta-lactam resistance genes to be elevated after composting.

Hurst et al. (2019) did not identify blaOXA, blaCTX-M, and blaVEB-1 after long-term storage of dairy cattle manure. This is in line with our present findings, which did not identify any beta-lactam resistance genes present in raw manure in the samples subjected to long-term storage and composting, except for the fox5 gene; however, blaVEB was not observed in our raw manure.

The blaTEM and blaCTX-M genes are often associated with IncN plasmids, which are believed to play an essential role in beta-lactamase gene dissemination (Lima et al. 2020). Our study found bla genes to be significantly associated with IncN_rep in composted manure samples but not in stored ones. Moreover, beta-lactam resistance genes are often located in insertion sequences such as ISEcp1 or ISCR1 (Lima et al. 2020). Although our findings did not identify the insertion sequences mentioned above, they did indicate the presence of other MGEs associated with bla genes viz. IS613, trfA, ISAba3, IS1111, intI3, tnpA, Tp614, orf39-IS26, Tn5, IS613; this suggests they have high mobility in both treatment strategies.

Aminoglycoside resistance genes

The aadA, aadA5, aadD, aadA9, aphA1/7, aac(6′)I1, aac3-VI, aacC2, aadA1, aadE, strA, strB, acc, aacC4, aph6, aac(6′)-Ib, aadA2, aph(2′)-Ib, aphA3, aadA9, spcN, aac(6′)-II, aphA3 genes were observed in the present study, all of which are responsible for deactivation. Zhou and Yao (2020) report a general decrease in aminoglycoside resistance genes during treatment, which agreed with our results.

Hurst et al. (2019) also found aac-(6)-lb and aadA1 to be present in manure after long-term storage; however, our present findings indicate that aadA1 was also persistent in manure after storage as well as composting, but aac(6)-lb was not detected after either process. Although aadA1 persisted after composting, its concentration was reduced, as also noted by Zhang et al. (2020).

Moreover, Jauregi et al. (2021) found aad6 and ant6-lb to be of particular concern due to their co-occurrence with MGE and hence their high dissemination potential. However, although these genes were not identified in raw manure in the present study, other genes of potential concern (i.e. co-occurring with MGE) were noted: aac(6″)-Ib (intI3, IS613, tnpA, trfA, orf37-IS26, ISAba3, IS1111, intI3), aac(6″)-Ib (intI3, IS613, tnpA, trfA, orf37-IS26, ISAba3, IS1111), and aphA3 (Tp614, IS613, tnpA, trfA, orf37-IS26, orf39-IS26, ISAba3, IS1111, Tn5, intI3).

MLSB resistance genes

Despite their wide application in livestock, often with lincosamides and streptogramin, the data on resistance genes against macrolides is limited. Mu et al. (2015) identified ermB and ermC in cattle feces (collected per rectum) and manure; however, the study only addressed macrolide resistance genes. Our present data did not confirm the presence of ermB and ermC, but did identify other genes coding MLSB resistance phenotypes in cattle manure, such as ermK, ermF, ereB, lnuB, vatE, erm36, matA/mel, mphA, mphB, vgb, mefA, msrC, pncA, mphC, msrA, lnuC, oleC, carB, and pikR2. In contrast, Wang et al. (2019) found ermB and ermF to be present in raw cattle manure, while Pitta et al. (2016) found carA, mefA, tlrC, and vgaA genes to also be present in feces and manure.

Zhou and Yao (2020) report a general decrease in MLSB resistance genes in total, which agreed with our results. However, Hurst et al. (2019) found ermB, ereC, and mefA to remain persistent in manure after long-term storage; our present findings did not indicate the presence of ermB in dairy cattle manure, and ereB and mefA genes to be removed by long-term storage and composting. Wang et al. (2019) did not identify any changes in ermB and ermF gene abundances during composting, since their levels were too low at the beginning of the process.

Lima et al. (2020) report ermB to be frequently incorporated in the Tn916-Tn1545 family of conjugative transposons. Although ermB was not found in the present study, other MLSB phenotypes were noted including genes co-occurring with MGE: IS613, trfA, ISAba3, IS1111, intI3, and tnpA in compost, and IS613, tnpA, trfA, orf37-IS26, ISAba3, IS1111, intI3, and IncN_rep in storage. The presence of IncN_rep in stored samples should be especially concerning because it suggests the location of erm genes on a broad-range host plasmid from the IncN group.

Resistance genes against other antimicrobials

Although the tetracycline and beta-lactam resistance genes are considered to be most prevalent in dairy and beef cattle feces, Marti et al. (2013) found chloramphenicol resistance genes to be predominant, with tetracycline resistance genes in fourth place. The cmlA and floR genes had high dissemination potential, due to co-occurrence (Jauregi et al. 2021), but our data indicates that cmlA was efficiently eliminated from animal manure after storage or composting. However, we agree that the floR gene should be considered highly mobile as our findings indicate that it could be associated with MGEs such as IS613, tnpA_1, trfA, orf37-IS26, ISAba3, IS1111, and intI3, in both analyzed treatment strategies.

The effect of composting on sulfonamide resistance genes has also been analyzed. Zhou and Yao (2020) report a general decrease in sulfonamide resistance genes, which was in line with our present results; in addition, Mu et al. (2015) also identified sul1, sul2, and sul3 in cattle manure, as noted in the present study. Additionally, our study also identified sul4, folP, and folA. The presence of sul1 and sul2 was also confirmed (Wang et al. 2019); however, their study did not show a reduction in the abundance of those genes after composting.

Furthermore, being often located on transposable elements of self-transposable or mobilizable broad host-range plasmids, sul genes are highly prevalent in a wide range of bacterial species (Lima et al. 2020). The sul2 gene should be especially concerning, because it is often embedded on small mobile plasmid of the IncQ family with a broad host spectrum (Wang et al. 2019). The sul3 gene has also been found in many sources together with class 1 integrons (Lima et al. 2020). Our present findings indicate the sul gene family to be associated with intI3, tnpA, Tp614, orf39-IS26, Tn5, IncN_rep, IS613, trfA, ISAba3, and IS1111 in both processes; however, intI1 was not noted during our study.

In addition, qnrA and qnrB resistance genes were not identified in the present cattle manure samples. Similarly, Mu et al. (2015) were also not able to detect the sought quinolone resistance genes (qnrD and qnrS). However, Wang et al. (2019) detected the presence of qnrB and qnrS genes in raw cattle manure, but did not observe any reduction after composting. Pitta et al. (2016) confirm the presence of vanRA, vanRE, and vanRG resistance genes in raw manure. While vanRA was found together with vanB, vanC, vanC2/vanC3, vanHB, vanRB, vanTC, vanTE, vanTG, vanYB, and vanYD in the present study, vanRE and vanRG were absent. Finally, Zhou and Yao (2020) did not find any change in total vancomycin resistance gene levels after composting, which contradicts our present results indicating a decrease in abundance.

Although significant reductions in ARGs abundances grouped according to antibiotic type (Fig. 2) were noted after composting and storage, the effect of those treatment strategies on particular genes is more complex, and many gaps still need to be addressed. The changes in ARGs abundance during storage or composting may vary depending on the location of the long-term storage pit or composting pile, storage/composting time (age), manure removal techniques (scrap versus flush removal) or even season of storage (Hurst et al. 2019). In addition, during composting, ARGs and MGEs removal can also be influenced by temperature, pH, moisture, or type of waste and bulking material (Pezzolla et al. 2021). Additionally, ARGs levels can also depend on the co-occurrence of ARGs or MGEs, and on various factors affecting bacterial fitness, gene regulation and transfer, such as nutrient accessibility or heavy metal concentration (Oliver et al. 2020). While ARGs concentration also seems to be dependent on on-site pressure, such as the presence of antibiotics or heavy metals, it also appears to be influenced by the specific gene. Zhang et al. (2018) found lower levels of macrolide resistance genes (ermX, ermQ, ermF, ermT, and ermB) to be associated with lower tylosin concentrations, due to the effect of tylosin degradation during composting. Guo et al. (2018) reported that adding cyromazine (a non-antibiotic drug) to samples increases the prevalence of blaCTX-M, blaVIM, and Tn916/1545 genes.

It should not be neglected that microorganisms themselves may generate selection pressure regarding antibiotic resistance (Allen et al. 2010). Bacteria from the Actinomycetota produce diverse secondary metabolites, including antibiotics, to promote their survival (Heul et al. 2018). Zhang et al. (2016) observed a strong relationship between the availability of zinc and the presence of tetQ, tetG, ermB, dfrA, and intI1. Similarly, the presence of bioavailable copper corresponded to the presence of sul1, ermX ermB, dfrA7 and aac(6′)-ib-cr; however, it should be noted that the study was performed in a wastewater treatment plant.

In addition, the abundance of ARGs appears to be contingent on microbial community structure and its shift during processing. While composting is often considered more reliable than storage for reducing pathogen levels in manure, it is also significantly more cost-, time-, and labor-consuming (Staley et al. 2021). Zhang et al. (2018) attribute the changes in ARGs levels observed during treatment to the dynamics of the potential host bacteria, with the mobility and type of gene influencing the host range.

The increase in ARGs noted during composting may occur due to the genes being harbored by ARB during proliferation, or HGT between various ARB (Gou et al. 2018). Indeed, genes conferring resistance to chloramphenicol and lincomycin have been found to undergo HGT from gram-positive Actinomycetota to gram-negative Pseudomonadota, despite the clear phylogenetic and ecological boundaries between them (Jiang et al. 2017). HGT is promoted during composting due to it favoring unfavorable conditions for bacterial growth which induce stress-response mechanisms (Lima et al. 2020).

The differences between the ARGs levels detected during composting or storage may be related to the nature of bacterial succession and the survival of ARG-carrying bacteria during the thermophilic phase. During storage, the bacterial strains harboring ARGs were initially suppressed, with their number increasing later on; however, no differences in sample biodiversity were found between treatment strategies. The scale of the study also matters. While lab-scale reactors are simple to prepare and they have clear value in preliminary testing, they nevertheless cannot adequately simulate full-scale processes, and their findings may not directly apply to field-scale studies (Staley et al. 2021).

The detected ARGs could confer resistance to all major antimicrobial classes, such as aminoglycosides, beta-lactams, MLSB, tetracyclines, and vancomycin; all are critical antibiotic groups, with some being even categorized as “last resort,” life-saving antimicrobials. It is important to remember that while the proposed manure management strategies reduce the risk of ARGs spread, they do not eliminate it entirely, and even if not apparent in treated samples, ARGs may persist in the environment below the detection limit. Even trace amounts of genes may further spread with host species succession or be disseminated via HGT.

Conclusions

Our findings prove that dairy cattle farms can serve as reservoirs of ARGs and MGEs. However, both manure storage and composting were able to reduce aminoglycoside, beta-lactam, tetracycline, MLSB, and vancomycin resistance genes as well as MDR genes, MGEs, and various “other” genes; however, they had less or no effect on phenicol, sulfonamide, or trimethoprim resistance genes.

Composting resulted in greater richness and diversity of the bacterial community compared to storage. In addition, while both proposed manure management strategies influence ARGs and MGEs abundance, particular ARGs respond differently to treatment in highly gene-specific ways, and also appear to depend on the dynamics of the bacterial community harboring ARGs. It is difficult to compare the effect of manure treatment strategies due to the range of possible influences. While both strategies reduce ARGs and MGEs levels, composting provides fertilizer with higher biodiversity, and is devoid of pathogens and parasite eggs, but is more labor- and time-consuming than storage. Our findings are important for improving existing or development of appropriate strategies to minimize ARGs dissemination.