Introduction
Antibiotic resistance is a cause for growing concern because of its increasing prevalence, costs to the economy and disastrous impacts on human health [1]. The transmittance of antibiotic resistance genes has enabled livestock to be a source of resistant infections in humans.[2] In the United States alone, 15.6 million kilograms of antibiotic were sold for food producing livestock in 2015 [3]. This accounts for 80% of all antibiotics sold in the United States [4], [5].
Livestock is a primary concern for antibiotic introduction to the environment because, depending on the antibiotic of interest, 30 – 100% can be excreted as waste [6]–[9] This excreted residual can persist in soil and water and select for antibiotic resistance [10]. The presence of excreted residual of antibiotic resistance genes is of particular concern because of horizontal gene transfer. [ CITE] * Discuss the mechanisms of HGT
Composting is the process of biodegrading organic matter into a humus-like material rich for soil application. It has been identified by the Food Safety Modernization Act (FSMA) as a method of reducing the [11] Three stages of compost communities have been identified [12] tetW and sul1were also discovered.
This study examines three different composting set ups from antibiotic fed cattle. Both dairy and beef manure was collected and composted following the static and turned FSMA guidelines –an anaerobic static – in bench scale composting tumblers [11]. The second trial was a heat controlled, bench scale composting tumbler at the FSMA required thermophilic temperature of >55ºC. This trial focused only on static, dairy manure compost but still compared medicated and untreated manure. The final trial was a full-scale and focused only on static composting of dairy manure with and without antibiotics.
Macrolide, sulfonamide, lincosamide, tetracycline, cephalosporin class antibiotics are regularly used in the cattle industry have been classified as such as tylosin, chlorotetracycline, sulfamethazine, pirlimycin, cephapirin have been classified as which are all regularly used in the cattle industry have been shown to remain in soils for up to XX time. (CITE)
Talk about different types of antibiotics and how ours fit in the WHO standards
Antibiotic resistance is then environmentally selected for and can propagate through the system via horizontal gene transfer.
Why intl1, sul1, tetW – qpcr
Compost was analyzed for manure-derived isolates using MacConkey agar because it selectively factors gram negative enteric bacteria [CITE]. Ten isolate coliforms were then chosen from each of the finished composts and subjected to Kirby-Bauer disk diffusion to determine their resistance profiles.
The purpose of this study was to determine the affect manure composting had on the antibiotic resistance gene abundance and community of medicated beef and dairy cattle.
An important aspect of this work was to test the impact of the Food Safety Modernization Act (FSMA) guidelines for biological soil amendments.
Beef steers are typically treated fed antibiotics as a mix in their diet to prevent disease, treat disorders and ensure they grow quickly.
Methods
The methods for cattle selection, sample selection and first trial composting techniques have been previously published by Ray, Knowlton, Shang, & Xia, 2017. [14]
Manure Production
In summary, manure for composting was procured from eighteen steers and cows selected for their respectively similar body weights, no history of antibiotic treatments and cows stage of lactation. Cattle were housed individually and fed a basal diet of corn silage. All were given free choice of water and ad libitum grain and hay.
Nine Hereford steers were fed a basal diet of corn silage and medicated or non-medicated grain mix. Three steers were each fed 350 mg of chlortetracycline and sulfamethazine per day, three were fed 11 mg Tylosin per kg feed, and three were fed a non-medicated diet. Each steer was fed their respective diet for seven days. Feces and urine were collected from Day 3 to 7 post- treatment. Nine dairy cows were selected for this study; six at peak lactation and three at the end of their current lactation cycle. Three cows were treated with two intermammary doses of 50 mg of pirlimycin at peak lactation; three cows received a single intermammary dose of 300 mg cephapirin at the end of lactation and three received no antibiotic intervention. Urine and feces were collected and composited to obtain a homogenous mixture for each cattle treatment and manure type.
Composting
The first trial composting setup is identical to the method described in Ray, Knowlton, Shang, & Xia, 2017 [14]. In summary, 4 different manures (beef steer manure with and without chlortetracycline, sulfamethazine, and tylosin and dairy cow manure with or without pirlimycin and cephapirin antibiotic treatment) were mixed with alfalfa hay, pine bark mulch, and sawdust and composted in triplicate using both static and turned composting methods yielding 24 independent composters (wet mass = 20-22 kg). Static composters were aerated using an air pump and perforated polyvinyl chloride (PVC) pipes. Turned composters were mixed well by hand daily. Samples were collected on days 0, 4, 7, 14, 21, 28, 35 and 42. On day 0 samples were also taken of each non-composted, raw manure for comparison with the finished composts. Samples were used immediately for analysis by culture techniques and additional samples were stored at -20°C for subsequent molecular analysis.
The heat-controlled trial used dairy manure (with and without pirlimycin and cephapirin) and used the static composting method. Composters were set up in triplicate for a total of 6 composters. In order to achieve a C:N ratio of 25 – 30 and a moisture ratio of 55-60%, manure was combined with grass hay and sawdust. Each composter was aerated using an air pump (Model: DOA-P704-AA, GAST, MI) at a flow rate of 0.05 L s−1 (0.1 ft3 min−1). The pump
was on for 5 min every hour during the thermophilic phase and then for 1 min every hour during the mesophilic phase. Temperature sensors were placed in two locations in each composter to record and monitor moisture and temperature over time. The compost was allowed to self-heat for the first 72 hours. Externally applied heat tape was applied to control compost temperature and was turned on after the initial 72 hours to maintain the thermophilic stage (>55ºC) for 15 days. The mesophilic temperature range (35-45°C) was then maintained for three weeks before allowing compost to cool to room temperature. Samples were collected on days 0, 1, 3, 7, 14, 21, 28, 35, and 42 and raw manure samples were collected on day 0 for comparison with finished composts. Samples were used immediately for analysis by culture techniques and additional sample was stored at -20°C for subsequent molecular analysis.
The full-scale composting occurred outdoors at Kentland Farm in southwest Virginia. The compost piles were formed in early January. Dairy manure (with and without pirlimycin and cephapirin antibiotic treatment) was mixed with alfalfa hay and sawdust at a ratio of 4:1:4.3 to achieve a moisture content of 55-65% and a C:N ratio of 25-30. Compost piles were formed in large metal containers (20 × 8 × 4 ft.) on top of a system of perforated PVC pipes which supplied forced aeration using an atta
ched air pump. A stockpile of antibiotic manure without any amendments or aerat
ion was also formed in a metal container to serve as a non-composted positive control. Compost and stockpile samples were taken on days 0, 4, 7, 14, 28, 42, 56, and 63. Samples were used to enumerate E. coli using the Colilert defined substrate method on days 0 and 63 (www.idexx.com, Wetsbrook, MN). The remaining samples were stored at -20°C for subsequent molecular analysis.
DNA Extraction
DNA from compost and manure samples was extracted using two extraction kits: the FastDNA Spin Kit for Soil, Catalog #116560200 (MP Biomedicals, Solon, OH), which served as the primary extraction kit and was used for all samples, and the ZR Fecal DNA MiniPrep Kit, Catalog #D6010 (Zymo Research Corporation, Irvine, CA) which served as a secondary kit for comparison purposes and was only used to extract 10% of samples. Following blending, 500 mg of compost or manure was aseptically transferred to an appropriate test tube for extraction. Extraction followed manufacturer’s instructions, except that a 2-hour incubation period was added to the protocol immediately following the bead-beating step to optimize lysis of both gram- positive and gram-negative microbial cells, and the final centrifugation step was extended to 3 minutes to ensure complete capture of DNA. The OneStep PCR Inhibitor Removal Kit, Catalog #D6030 (Zymo Research Corporation, Irvine, CA) was used as an additional cleanup step.
Metagenomics Sequencing and Analysis
Metagenomic sequencing permits the analysis of all genes in an environmental sample without requiring specific targeted primers, providing much more information than qPCR. A set of 60 samples were selected from all of the bench-scale and heat-controlled compost and manure samples to provide a cross-section of the overall composting process with time. Samples were selected to be equally representative of all manure types and composting methods at each time point. Samples were submitted to the Virginia Bioinformatics Institute, Blacksburg, VA, for Illumina HiSeq High Output 2×100 paired end cycle clustering and sequencing per lane and run as 5 lanes of 12 samples each. Sequencing results were stitched together and analyzed by uploading to MetaStorm [15]. ARG sequences were identified through comparison with the comprehensive antibiotics resistance database (CARD).
16S rRNA Gene Amplicon Sequencing
All extracted DNA samples from each composting trial were submitted to the Virginia Bioinformatics Institute for 16S rRNA gen amplicon sequencing. Samples were submitted in 2 lanes of 150 samples each sequenced using Illumina MiSeq V3 2×300 paired end cycle clustering and sequencing per lane. Sequencing results were stitched together and taxonomic analysis was performed using QIIME [16]. All singleton reads and chimeric sequences were removed and OTU tables were generated allowing for taxonomic composition analysis. Jackknifed beta diversity analysis was performed to calculate unweighted and weighted UniFrac distance matrices for the comparison of sample taxonomic similarity.
Quantitative Polymerase Chain Reaction (QPCR)
Samples were diluted 1:100 based upon experimentation with the samples. QPCR was performed in triplicate on all diluted samples using the CFX96 Touch Real-Time PCR Detection System (BioRad Laboratories, Hercules, CA) to quantify initial concentrations of 16S rRNA genes, tet(W), intl1 and sul1. Tet(W) and sul1 were chosen because the treated beef cattle were administered chlortetracycline and sulfamethazine, so an effect on tetracycline and sulfonamide resistance was expected; these two genes are also well represented in the current research. Tet(W) codes for ribosomal protection proteins, while sul1 codes for sulfonamide-resistant dihydropteroate synthase. Intl1, encoded the class 1integrons, was chosen because it has shown to be associated with sul1 gene frequency [17]. Primer sequences used were those outlined by Ma et al. [18]. The SsoFast Evagreen Supermix, Catalog #1725204 (BioRad Laboratories, Hercules, CA) was used according to manufacturer instruction to create the qPCR mastermix.
Statistical Analysis
Statistical analyses were performed using R statistics software [19]. Graphics were generated using Microsoft Exel, and R packages ggplot2,cowplot and RColorBrewer. Summary statistics were calculated using the ddply() function in the plyr package. A significance value of α = 0.05 (i.e. p < 0.05) was considered significant. Statistical differences between qPCR ARG concentrations were calculated using the Kruskal-Wallis nonparametric rank test as well as Two sided-Wilcoxon Rank Sum tests. The test was applied to the log starting quantity as well as to the starting quantity normalized to the number of 16S rRNA gene copies. Statistical comparisons of the similarity of ARG profiles of metagenomics data and of similarity of taxonomic class data and UniFrac distances were determined using an analysis of similarities (ANOSIM) test performed using the Primer 6 software [20].
Results and Discussion
Composting has been explored as a potential method of reducing abundance of antibiotic resistance, however the results of this study show that the potential outcomes are complex. Despite following FSMA guidelines for composting, E. coli was not eliminated during initial small-scale composting. The FSMA guidelines for temperature were maintained for small-scale static composting, but not for small-scale turned composting. Despite this discrepancy, neither treatment resulted in reducing fecal coliform counts below the FSMA target of 1000 MPN/g as indicated by the high IDEXX E. coli counts of 9 log MPN/g. (SI 1) QPCR data suggests that the effect of composting on specific ARGs depends on the gene of interest. The two extraction kits were compared for resulting metagenomics compositions and the Fast DNA kit was experimentally determined to be a more efficient (SI 1).
Small Scale Trial 1
The small-scale composting trial achieved the FSMA required 3 days at thermophilic temperatures for the statically composted treatments, but the turned treatments failed to maintain thermophilic temperatures for the required 15 days. This is consistent with other research done trying to maintain these standards. [CITE] IDEXX with Colilert was used to specifically quantify E. coli (Figure 1E,F), which confirmed that E. coli were present in the 3-4 log MPN/g range in the finished day 42 compost. (S1)
QPCR of small scale composting was performed to quantify intl1, sul1 and tet(W) gene copies (Figures 2-4). Relative abundance of sul1, normalized to 16S rRNA genes, varied with time, with day 42 having a significantly higher relative abundance than on day 0 (p<0.007; Kruskal-Wallis). Initial relative abundances varied significantly at day 0 but by day 42 there was no significant variation in compost or cow type (p<0.03, p<0.50; Kruskal-Wallis). The relative abundance of intl1 genes shows the same trend with significant increasing with time and a significant difference in static beef with antibiotics versus without. There was no significant difference between the absolute frequency of sul1 genes and intl1genes (p < 0.092, Wilcox). This is consistent with the findings of Heuer et al. and further associates sul1 with class one integrons [17]. As composting progresses, the sul1 and intl1 genes seemingly propagate throughout the community regardless of antibiotic dosed cows. Relative abundance of tet(W), normalized to 16S rRNA gene abundances, decreased rapidly with time (p<0.0001; Kruskal-Wallis) and static composting was associated with a slightly reduced relative tet(W) abundance as compared to turned composting (p=0.0243; Kruskal-Wallis).
The decrease in tet(W) during the 42 days of composting is consistent with Selvam et al. and Storteboom et al. although the tet(W) did not decrease below detection ([21], [22]). The sul1 relative gene frequency
contradicted the results presented Sel
vam et al. but this could be due to a variety of factors such as differing animal of analysis, and different method on antibiotic administration (spiking versus feeding animals) [21].
Figure 3A-D Sul1 Normalized to 16s- Relative abundance of sul1 genes to 16S genes over time, by composting method and cattle type. Relative abundance only varied significantly for static composted dairy with antibiotics with respect to time (p<0.01; Kruskal-Wallis). Composting was associated with a significant increase in relative abundance on day 42 with respect to raw manure on day 0 (p<0.006; Kruskal-Wallis). There is no significant difference between cow or composting methods on day 42 (p=0.50; Kruskal-Wallis). Error bars represent standard deviation.
Figure 7 A-D Relative abundance of Intl1 genes for the initial small-scale trial by compost type and cow type. Initial values were not different across all compost types and methods. (p< 0.26, Kruskal-Wallis) and increased significantly with time for all compost types and methods. (p< 0.0001; Kruskal-Wallis). There is a significant difference between antibiotic treated beef compost and compost without antibiotics (0.015, Wilcox).
Figure 5A-D Relative abundance of tet(W) genes, normalized to 16S rRNA genes, over time, by composting method. Relative abundance varies significantly with respect to time for all compost types (p<0.0004; Kruskal-Wallis). Composting was associated with a significant decrease in relative abundance on day 42 with respect to raw manure on day 0 (p<0.0002; Kruskal-Wallis). There was no significant difference between static and turned composting or beef and dairy methods on day 42 (p=0.67; Wilcox). On day 0, manure type significantly impacted relative tet(W) abundance, with beef manures having a higher relative abundance than dairy manures (p<0.038; Kruskal-Wallis). There was also a significant impact of manure type on relative tet(W) abundance on day 42 (p<0.018; Kruskal-Wallis).
Metagenomic sequencing for ARGs was performed on small-scale compost samples are shown for each compost and cow type in Figure 7. Taxonomic profiles were also obtained though 16S rRNA Amplicon Sequencing are shown in Figure X. ARG profiles are representeded as stacked bar charts with the relative proportion of each category of ARG shown. On day 0, there is no significant difference between raw manure and the compost mixtures (R =0.028, p<0.28; ANOSIM) (Figure 3.34), but there was a difference in ARG profiles of beef and dairy manures (R = 0.62, p<0.001; ANOSIM). By day 42, the trend reversed, with composting method significantly impacting ARG profile similarity (R = 0.47, p<0.001; ANOSIM) (Figure 3.36) while manure type no longer had an effect (R= -0.01, p<0.51; ANOSIM) (Figure 3.37).
Figure 7 Day 0 versus Day 42 ARG Copies per 16S rRNA genes. Within manure (R=0.147, p=0.005) and compost (R=0.234, p=0.002) types, there was a significant resistance profile change from day 0 to 42.
The metagenomic analysis shows the clear community shift that takes place throughout the 42 day compost. It is consistent with the qPCR findings and most obviously the clear decrease in tetracycline resistance suggests that tetracycline is heavily impacted by composting and is heavily present in cattle regardless of antibiotic dosages. Sulfonimide doses were also found to decrease significantly. It is also interesting that the relative same amount of tetracycline and sulfonamide were present in the same initial concentrations across all cows because only some were given antibiotics. This suggests that the ARGs of interest were already present in cows prior to antibiotic dosing.
The MDS plots show clearly that time plays the most important role in the microbial community shift and for all trials time is the most significant factor in microbial shift.
Figures 3.38 – 3.40 display MDS plots based on the unifrac similarity matrix of ARG abundance profiles for each sample. Samples cluster significantly with respect to time (p<0.001; ANOSIM) (Figure3.38), with day 42 samples clustering with respect to composting method (p=0.012; ANOSIM) (Figure 3.40). There was no apparent effect of manure type by day 42 (p=0.714; ANOSIM) (Figure 3.39).
Figure 3.41 displays the relative abundance of ARGs coding resistance to the four antibiotics considered to be the “highest priority critically important antimicrobials” by the WHO, organized by composting method (58). This data was taken from the metagenomic sequencing results and are based on comparison with the CARD database. Results show that composting increases relative abundance of genes coding resistance to fluoroquinolones (p<0.0001; Kruskal- Wallis) and glycopeptides (p<0.0001; Kruskal-Wallis) by day 42 of composting in both static and turned composts, and a reduction in relative abundance of macrolide resistance genes (p=0.0164; Kruskal-Wallis). Day 42 turned compost had higher abundance of β-lactam resistance genes than static compost (p=0.0102; Kruskal-Wallis).
The results of 16S rRNA gene amplicon sequencing and resulting taxonomic analyses for small scale composting are displayed in Figures 3.47 – 3.54. Figures 3.47 – 3.50 show stacked column charts comparing the relative abundance of each of the 9 most common bacterial classes present in the compost and manure samples, with the remaining, less abundant classes combined into the “other” category. There were significant differences in taxonomic composition with respect to time (p<0.001; ANOSIM) (Figure 3.47); there were also differences between day 42 composted manures as compared to day 0 non-composted raw manure (p<0.001; ANOSIM) as well as differences between taxonomic composition of static and turned composts at day 42 (p<0.001; ANOSIM) (Figure 3.48). Beef and dairy manure composts on day 42 differ significantly in taxonomic composition similarity (p=0.03; ANOSIM) (Figure 3.49). There was also a significant difference in taxonomic composition between the two DNA extraction kits used (p<0.001; ANOSIM) (Figure 3.50).
Heat Controlled Trial 2
During heat controlled composting, resistant fecal coliforms were enumerated using antibiotic-supplemented MacConkey agar containing one of five different antibiotics as well as a control plate without antibiotics, and the ratio of the number of resistant colonies to control colonies for each of the antibiotics was determined (Figures 3.11 – 3.15). The temperature was maintained above thermophilic temperatures for the 3 days required by FSMA for static composting. For the 3rd generation cephalosporins, ceftazidime and cefotaxime (Figure 3.11 and 3.12, respectively), composting significantly increased the proportion of resistant colonies by day 3 (p<0.0001; Zero-Inflated Poisson), before counts fell below detection limits on day 7, while there was no effect of antibiotic treatment. Tetracycline resistance behaved similarly with a sharp increase on day 3 of composting (p<0.0001; Zero-Inflated Poisson) before falling below detection on day 7 (Figure 3.13), but this time antibiotic administration did result in a significant increase in the proportion of tetracycline resistant isolates observed (p=0.0003; Zero-Inflated Poisson). The proportion of resistant isolates to erythromycin, a macrolide, and clindamycin, a lincosamide, behaved differently from the others (Figures 3.14 and 3.15, respectively). In each, there was no effect of composting, while antibiotic administration resulted in higher proportions of resistant isolates (p=0.0287 and p=0.0297 for erythromycin and clindamycin, respectively; Zero-Inflated Poisson), before falling below detection levels.
QPCR data for sul1 and tet(W) ARGs in heat-controlled
composting samples are shown in Figures 3.26 – 3.29
. Relative sul1 abundance (Figure 3.27) varied significantly with time and increased on day 42 compared to day 0 (p<0.0001; Kruskal-Wallis). Relative tet(W) abundance decreased rapidly with time (p<0.0001; Kruskal- Wallis) and antibiotic treatment resulted in higher relative tet(W) abundance than manure from untreated cattle (p=0.0413; Kruskal-Wallis) (Figure 3.29).
Figure 9 – Sum1 and TetW normalized to 16S A) Sul1 Relative abundance of sul1, normalized to 16S rRNA genes, over time by manure type subject to composting. Abundance is not significantly impacted by time (p<0.79; Kruskal- Wallis). There is no significant impact of antibiotic treatment (p=0.3957; Wilcox). B) Relative abundance of tet(W), normalized to 16S rRNA genes, by manure type subject to composting. Time had a significant impact on relative tet(W) abundance (p<0.0007; Wilcox). Relative tet(W) abundance declined at rates of 0.7725 day-1 and 0.7171 day-1 for antibiotic treated and untreated composting methods, respectively, over the first seven days of composting.
Large Scale Trial 3
IDEXX was used to enumerate E. coli on day 0 and day 63 of large scale composting (Figure 2). Composted treatments were maintained at or above thermophilic temperatures for the 3 days required by FSMA. Both composting (static composting vs. stockpiling manure) and prior antibiotic administration had a significant impact on day 63 E. coli counts (p<0.014; Kruskal-Wallis), with counts falling below detection limits for the static compost condition with antibiotics by day 63. The reduction of E. coli below detection indicates a likely decrease in fecal coliforms but it cannot be definitively said whether fecal coliforms fell below the FSMA target of 1000 MPN/g for either the antibiotic treated or untreated composts because fecal coliforms were not measured directly.
QPCR data for sul1 and tet(W) gene abundances during large-scale composting are shown in Figures 3.30 – 3.33. Absolute sul1 abundance significantly increased when composting, as opposed to stockpiling, manure (Figure 3.30; p<0.0041; Kruskal-Wallis), but there was no impact of antibiotic treatment in composted manure sul1 gene abundance (p< 0.70 , Wilcox). Relative abundance of sul1, normalized to 16s rRNA gene abundance, was not higher in composted samples than in stockpiled samples (Figure 3.31, p=0.1285; Wilcox).
Relative abundance of tet(W) did not decrease with time in composted manure samples (p<0.805; Kruskal-Wallis). Relative tet(W) abundance decreased to a greater extent when cattle had previously been administered antibiotics (p=0.016; Wilcox).
Figure 12 A) Relative abundance of sul1, normalized to 16S rRNA genes, over time by treatment type. There was no significant difference between composted and stockpiled manures (p=0.1285; Wilcox). Time was also a significant factor in relative sul1 abundance (p<0.03; Kruskal-Wallis). B) Relative abundance of tet(W) to 16s genes over time by treatment type. Composting did not result in significantly reduced relative abundance relative to stockpiled manure (p<0.805; Kruskal-Wallis). Antibiotic treatment significantly impacted relative tet(W) abundance in composted manures (p=0.016; Wilcox).
Conclusions
Dairy and beef cattle were fed antibiotics and their manure was composted statically and turned at a small scale. The second trial was conducted in response to the unexpected survival of E. Coli through the end of the first small scale trial. This was performed because substantial difficulties were encountered when trying to maintain thermophilic temperatures in the first trial. This seems to be a problem across research laboratories and the feasibility of this standard should be reevaluated [23]–[25]. Field scale compost was also conducted and the trends in qPCR frequency were similar across small scale and field scale. This suggests that small scale studies are relevant in determining what happens in the field. Antibiotics fed to cattle behaved differently than studies where antibiotics were spiked in post cattle.
Composting is shown to radically change the microbial community in
Supplementary Information
16S rRNA sequencing was performed to gain an understanding of the variety present in the samples. Figure S1 shows the log gene copies found in each sample.
Figure S1: 16S Sequenced Data
All extraction kits use varied protocols that are optimized for different purposes. This study compared the taxonomic composition of DNA extracted using two different kits, the FastDNA Spin Kit for Soil by MP Biomedicals, and the Zymo Fecal kit. In all three trials, there was a significant difference in the similarity of taxonomic composition between DNA extracted from the two kits (Figure S2). On the whole, the Zymo fecal kit resulted in a lower relative abundance of the classes actinobacteria, alphaproteobacteria, and clostridia, while having increased relative abundances of the many less abundant classes which are combined into the “other” classes category. This may indicate that the Zymo kit is less able to fully extract DNA from these more abundant classes and that the less abundant species therefore make up a larger proportion of the classes that remain.
Absolute QPCR Figures
Absolute sul1 gene abundance (Figure 2A-D) increased significantly with time (p<0.0001; Kruskal-Wallis) but there was no impact of manure type or significant difference between static and turned composting methods.
Figure 2A-D Abundance of Sul1 genes for the initial small-scale trial by compost type and cow type. Initial values were not different across all compost types and methods. (p< 0.22, Kruskal-Wallis) and increased significantly with time for all compost types and methods. (p< 0.002; Kruskal-Wallis). Time 4 and Time 7 were the only statistically different points across all samples (p<0.01, Kruskal-Wallis). Error bars represent standard deviation.
Absolute tet(W) gene abundance decreased with time and there was no significant difference among composting methods. Manure type was a significant factor in day 0 tet(W) abundance, with beef manures containing more tet(W) gene copies than dairy manures (p<0.039; Kruskal-Wallis). Antibiotic treatment had no effect.
Figure 4A-D Abundance of TetW1 genes for the initial small-scale trial by compost type and cow type. Static Beef, Turned Beef, Static Dairy and Turned Dairy were all significantly different between antibiotic and no antibiotics presence (p<0.03, Kruskal-Wallis). There was a significant difference in static and turned composting for beef but not for dairy.
Absolute intl1 gene abundance increased significantly with time. The antibiotic treated static beef compost differed significantly from static beef compost without antibiotics (0.022, Wilcox).
Figure 6 A-D Abundance of Intl1 genes for the initial small-scale trial by compost type and cow type. Initial values were not different across all compost types and methods. (p< 0.20, Kruskal-Wallis) and increased significantly with time for all compost types and methods. (p< 0.001; Kruskal-Wallis). There is a significant difference between antibiotic treated beef compost and compost without antibiotics (0.022, Wilcox).
Figure 7 Abundance of E. coli measured using IDEXX with Colilert. Time was a significant factor impacting change in abundance (p=0.0222; Kruskal-Wallis). No significant impact of prior antibiotic administration was observed (p=0.
8906; Kruskal-Wallis).
IDEXX was used to enumerate E. coli for the remainder o
f the composting trial (Figure 7). This time, E. coli abundance was reduced to non-detect levels by day 42 of composting, as expected, and there was again no significant impact of prior antibiotic administration. The reduction of E. coli below detection indicates a likely decrease in fecal coliforms but it cannot be definitively said whether fecal coliforms fell below the FSMA target of 1000 MPN/g since they were not measured directly.
Figure 8 A) Abundance of sul1 genes over time by manure type. Abundance is significantly different from time 0 to time 42 (p<0.008; Kruskal-Wallis). Antibiotic treatment does not have a significant impact. (p<0.11; Kruskal-Wallis). B) Abundance of tet(W) over time by manure type subject to composting. Time had a significant effect on tet(W) abundance (p<0.0007; Wilcox), as did antibiotic treatment (p<0.001Wilcox).
Absolute sul1 abundance (Figure 3.26) varied significantly with time (p<0.0001; Kruskal-Wallis) and increased by 0.5 log on day 52 relative to day 0. Antibiotic treatment was not a significant factor. Absolute tet(W) abundance decreased with time (p<0.0001; Kruskal-Wallis) and antibiotic treatment resulted in a significant increase in detected tet(W) gene copies (p=0.0129; Kruskal- Wallis) (Figure 3.28). By contrast, absolute tet(W) gene abundance was reduced with composting, as compared to stockpiled manure (Figure 3.32, p<0.0003; Wilcox) and unexpectedly, antibiotic treatment was associated with a decrease in absolute tet(W) abundance (p=0.016; Wilcox).
Figure 11 A) Effect of treatment type on abundance of sul1 genes with time. There was a significant difference between composted and stockpiled treatments (p<0.0041; Kruskal-Wallis. There was no significant impact of antibiotic treatment on sul1 abundance in composted manures (p< 0.70, Wilcox). Time had a significant impact (p<0.014; Kruskal-Wallis). B) Abundance of tet(W) with respect to time by treatment type. Tet(W) abundance was significantly impacted by time (p=0.0014; Kruskal-Wallis). Composted manures had significantly reduced tet(W) abundance compared to stockpiled manure (p<0.0003; Wilcox). Antibiotic treatment of composted manures significantly impacted tet(W) abundance in those manures. (p<0.016; Wilcox).
Figure S2 : Zymo 16S Copies of all Scales
Generally there was a lot more variability in gene copies detected.
Relative tet(W) abundance declined at rates of 0.4723 day-1 and 0.5321 day-1 for static and turned composting methods, respectively, over the first seven days of composting. -. Comparison of relative abundance of tet(W) to 16S rRNA genes on day 0 with that on day 42, by manure type subject to composting.
References
[1] World Health Organization, “Global action plan on antimicrobial resistance.,” WHO Press, pp. 1–28, 2015.
[2] P. Takhistov and C. M. Bryant, “Protecting the food supply,” Food Technol., vol. 60, no. 7, pp. 34–43, 2006.
[3] US Food and Drug Administration, “2015 SUMMARY REPORT Antimicrobials Sold or Distributed for Use in Food-Producing Animals,” FDA Rep., p. 58, 2016.
[4] US Food and Drug Administration, “2011 SUMMARY REPORT AntimicrobialsAntimicrobials Sold or Distributed for Use in Food-Producing Animals,” FDA Rep., p. 58, 2012.
[5] T. Pham, “Drug Use Review – Systemic Antibacterial Drug Products,” Rep. No. 2012-544, pp. 1–9, 2012.
[6] N. Kemper, “Veterinary antibiotics in the aquatic and terrestrial environment,” Ecol. Indic., vol. 8, no. 1, pp. 1–13, 2008.
[7] K. R. Kim, G. Owens, S. I. Kwon, K. H. So, D. B. Lee, and Y. S. Ok, “Occurrence and environmental fate of veterinary antibiotics in the terrestrial environment,” Water. Air. Soil Pollut., vol. 214, no. 1–4, pp. 163–174, 2011.
[8] G. K. Elmund, S. M. Morrison, D. W. Grant, and M. P. Nevins, “Role of excreted chlortetracycline in modifying the decomposition process in feedlot waste,” Bull. Environ. Contam. Toxicol., vol. 6, no. 2, pp. 129–132, 1971.
[9] A. K. Sarmah, M. T. Meyer, and A. B. A. Boxall, “A global perspective on the use, sales, exposure pathways, occurrence, fate and effects of veterinary antibiotics (VAs) in the environment,” Chemosphere, vol. 65, no. 5, pp. 725–759, 2006.
[10] J. C. Chee-Sanford et al., “Fate and Transport of Antibiotic Residues and Antibiotic Resistance Genes following Land Application of Manure Waste,” J. Environ. Qual., vol. 38, no. 3, p. 1086, 2009.
[11] U.S. Department of Health & Human Services, “Part III Department of Health and Human Services,” Fed. Regist., vol. 80, no. 200, pp. 1–195, 2013.
[12] L. P. Antunes et al., “Microbial community structure and dynamics in thermophilic composting viewed through metagenomics and metatranscriptomics,” Sci. Rep., vol. 6, no. August, p. 38915, 2016.
[13] P. Ray, K. F. Knowlton, C. Shang, and K. Xia, “Method development and validation: Solid phase extraction-ultra performance liquid chromatography-tandem mass spectrometry quantification of pirlimycin in bovine feces and urine,” J. AOAC Int., vol. 97, no. 6, pp. 1730–1736, 2014.
[14] P. Ray, C. Chen, K. F. Knowlton, A. Pruden, and K. Xia, “Fate and Effect of Antibiotics in Beef and Dairy Manure during Static and Turned Composting,” J. Environ. Qual., vol. 46, no. 1, p. 45, 2017.
[15] G. Arango-Argoty, G. Singh, L. S. Heath, A. Pruden, W. Xiao, and L. Zhang, “MetaStorm: A public resource for customizable metagenomics annotation,” PLoS One, vol. 11, no. 9, pp. 1–13, 2016.
[16] F. D. B. and E. K. C. J. Gregory Caporaso, Justin Kuczynski, Jesse Stombaugh, Kyle Bittinger, “QIIME allows analysis of high-throughput community sequencing data,” Nat. Methods, 2010.
[17] H. Heuer, H. Schmitt, and K. Smalla, “Antibiotic resistance gene spread due to manure application on agricultural fields,” Curr. Opin. Microbiol., vol. 14, no. 3, pp. 236–243, 2011.
[18] Y. Ma et al., “Effect of various sludge digestion conditions on sulfonamide, macrolide, and tetracycline resistance genes and class i integrons,” Environ. Sci. Technol., vol. 45, no. 18, pp. 7855–7861, 2011.
[19] R. C. T. (2017). R. A. language and environment for statistical computing. R. F. For and A. U. https://www. R. org/. S. C. Vienna, “R: A language and environment for statistical computing. R Foundation Statistical Computing, Vienna, Austria. URL https://www.R-project.org/.”
[20] C. Lozupone, M. E. Lladser, D. Knights, J. Stombaugh, and R. Knight, “UniFrac: An effective distance metric for microbial community comparison.,” ISME J., vol. 5, no. 2, pp. 169–172, 2011.
[21] A. Selvam, D. Xu, Z. Zhao, and J. W. C. Wong, “Fate of tetracycline, sulfonamide and fluoroquinolone resistance genes and the changes in bacterial diversity during composting of swine manure,” Bioresour. Technol., vol. 126, pp. 383–390, 2012.
[22] H. N. Storteboom, S.-C. Kim, K. C. Doesken, K. H. Carlson, J. G. Davis, and A. Pruden, “Response of Antibiotics and Resistance Genes to High-Intensity and Low-Intensity Manure Management,” J. Environ. Qual., vol. 36, no. 6, p. 1695, 2007.
[23] P. Chandna, L. Nain, S. Singh, and R. C. Kuhad, “Assessment of bacterial diversity during composting of agricultural byproducts,” MBC Microbiol., vol. 13, p. 99, 2013.
[24] N. E. Derby, H. Hakk, F. X. M. Casey, and T. M. DeSutter, “Effects of Composting Swine Manure on Nutrients and Estrogens,” Soil Sci., vol. 176, no. 2, pp. 91–98, 2011.
[25] R. Cáceres, X. Flotats, and O. Marfà, “Changes in the chemical and phys
icochemical properties of the solid fraction of cattle slurry during composting using
different aeration strategies,” Waste Manag., vol. 26, no. 10, pp. 1081–1091, 2006.