Changes in antibiotic concentrations and antibiotic resistome during commercial composting of animal manures

Environmental Pollution 219 (2016) 182e190

Contents lists available at ScienceDirect

Environmental Pollution journal homepage: www.elsevier.com/locate/envpol

Changes in antibiotic concentrations and antibiotic resistome during commercial composting of animal manures* Wan-Ying Xie a, Xin-Ping Yang a, Qian Li a, Long-Hua Wu b, Qi-Rong Shen a, Fang-Jie Zhao a, c, * a

Jiangsu Key Laboratory for Organic Waste Utilization, Jiangsu Collaborative Innovation Center for Solid Organic Waste Resource Utilization, College of Resources and Environmental Sciences, Nanjing Agricultural University, Nanjing 210095, China Key Laboratory of Soil Environment and Pollution Remediation, Institute of Soil Science, Chinese Academy of Sciences, Nanjing 210008, China c Rothamsted Research, Harpenden, Hertfordshire AL5 2JQ, UK b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 15 August 2016 Received in revised form 11 October 2016 Accepted 15 October 2016

The over-use of antibiotics in animal husbandry in China and the concomitant enhanced selection of antibiotic resistance genes (ARGs) in animal manures are of serious concern. Thermophilic composting is an effective way of reducing hazards in organic wastes. However, its effectiveness in antibiotic degradation and ARG reduction in commercial operations remains unclear. In the present study, we determined the concentrations of 15 common veterinary antibiotics and the abundances of 213 ARGs and 10 marker genes for mobile genetic elements (MGEs) in commercial composts made from cattle, poultry and swine manures in Eastern China. High concentrations of fluoroquinolones were found in the poultry and swine composts, suggesting insufficient removal of these antibiotics by commercial thermophilic composting. Total ARGs in the cattle and poultry manures were as high as 1.9 and 5.5 copies per bacterial cell, respectively. After thermophilic composting, the ARG abundance in the mature compost decreased to 9.6% and 31.7% of that in the cattle and poultry manure, respectively. However, some ARGs (e.g. aadA, aadA2, qacED1, tetL) and MGE marker genes (e.g. cintI-1, intI-1 and tnpA-04) were persistent with high abundance in the composts. The antibiotics that were detected at high levels in the composts (e.g. norfloxacin and ofloxacin) might have posed a selection pressure on ARGs. MGE marker genes were found to correlate closely with ARGs at the levels of individual gene, resistance class and total abundance, suggesting that MGEs and ARGs are closely associated in their persistence in the composts under antibiotic selection. Our research shows potential disseminations of antibiotics and ARGs via compost utilization. © 2016 Elsevier Ltd. All rights reserved.

Keywords: Antibiotics Antibiotic resistance genes Resistome Manure Compost

1. Introduction Antibiotics are commonly used in animal husbandry to treat and prevent diseases, as growth promoters and to improve feed efficiency (Jechalke et al., 2014; Sarmah et al., 2006). Generally, the antibiotics fed to the animals are poorly absorbed and large proportions can be excreted into feces as the parent compounds or in bioactive forms (Jechalke et al., 2014; Sarmah et al., 2006). In China,

*

This paper has been recommended for acceptance by Klaus Kummerer. * Corresponding author. Jiangsu Key Laboratory for Organic Waste Utilization, Jiangsu Collaborative Innovation Center for Solid Organic Waste Resource Utilization, College of Resources and Environmental Sciences, Nanjing Agricultural University, Nanjing 210095, China. E-mail address: [email protected] (F.-J. Zhao). http://dx.doi.org/10.1016/j.envpol.2016.10.044 0269-7491/© 2016 Elsevier Ltd. All rights reserved.

nearly half of the antibiotics consumed (162,000 tons in 2013) were used in animal husbandry, of which substantial amounts could end up in manure (Zhang et al., 2015). Concentrations of some veterinary antibiotics in manures were found to reach hundreds or even thousands of mg kg
1 in some parts of China (Zhao et al., 2010). Therefore, direct land applications of these antibioticcontaminated manures may introduce substantial amounts of antibiotic residues into arable lands. Although antibiotics are used mainly to control bacterial infections, some bacterial groups play important roles in antibiotic production or in their degradation (Dantas et al., 2008; Martinez, 2008; Sarmah et al., 2006). During these processes, the antibiotic resistant bacteria (ARB) rely on antibiotic resistance genes (ARGs) for protection from the compounds. The widespread use of antibiotics accelerates the development of antibiotic resistance and

W.-Y. Xie et al. / Environmental Pollution 219 (2016) 182e190

facilitates the maintenance of the resistance at high levels. ARGs could disseminate into a broader range of microbial communities by horizontal gene transfer (HGT) via mobile genetic elements (MGEs), such as plasmids, integrons and transposons (Blair et al., 2014; WHO, 2014). Due to the poorly regulated administration of veterinary antibiotics, China's animal production industries contribute greatly to the enhanced antibiotic resistance in the environmental matrices, such as watercourses and soils (Larson, 2015; Zhu et al., 2013). Assessment of environmental antibiotic resistance relies greatly on the detection of ARGs and gene markers associated with MGEs, such as genes for transposases and class 1 integron-integrase (Gillings et al., 2014; Luby et al., 2016; Zhu et al., 2013). Manure composting is a common and effective pathway for hazard reduction prior to land application (Bernal et al., 2009). The process has been found to be effective for the degradation of spiked antibiotics in laboratory-scale and pilot-scale studies (Ho et al., 2013; Mitchell et al., 2015; Selvam et al., 2012; Wang et al., 2015). However, little information is available with respect to the levels of antibiotic residues in commercial composts of animal manures in China. Furthermore, the dynamics of ARGs during composting appear to be complex, with a number of studies showing contradictory observations (Selvam et al., 2012; Wang et al., 2015; Zhu et al., 2013). Therefore, investigations on the commercial composts with respect to the antibiotic contents and ARG abundances are necessary, especially in regions with high antibiotic usage, such as Eastern China (Zhang et al., 2015). In the present study, we determined the concentrations of 15 common veterinary antibiotics using liquid chromatographyelectrospray ionization tandem mass spectrometry (LC-ESI-MS/ MS) and the abundances of 213 ARGs and 10 MGE marker genes by high-throughput quantitative PCR (HT-qPCR) in the commercial composts made from cattle, poultry and swine manures in Eastern China, during different phases of composting. Correlations between individual ARGs and MGE marker genes and their relationships with each antibiotic were analyzed to reveal possible mechanisms of ARG dissemination. 2. Materials and methods 2.1. Sampling Manures, thermophilic composts and mature composts were sampled from four large-scale commercial composting companies in Jiangsu province, China, in 2014. Three of the four companies produce aerobic and thermophilic commercial composts exclusively from manures of cattle, poultry and swine, respectively, with the additions of rice straw as the bulking agent. The fourth company uses a mixture of mainly poultry and some other animal manures. All composting operation lasted for about 1 month with a thermophilic phase (temperature > 60  C) of approximately 2 weeks. Different types of samples were taken on the same day in each company. Manures to be composted were taken from the raw material pools. Composts were sampled at the thermophilic and mature phases during composting according to the method described by Stevens (2010). At each sampling time point, 12 samples were taken from the top, middle, and bottom of the composting piles, mixed thoroughly and subsampled into 3 replicates. All samples from the aforementioned 3 companies contained 3 replicates. These samples were named in the format of animal type (C, P and S for cattle, poultry and swine, respectively), followed by the composting phase (M, T and C for the raw manure, thermophilic compost and mature compost, respectively). Samples from the fourth company that uses mixed manures were taken with just one replicate and named as PM1, PT1 and PC1, and the data

183

were included only for correlation analysis. All samples were kept in ice boxes and transported to the laboratory as soon as possible. Samples were then freeze-dried, homogenized by sieving through a 0.15-mm mesh and stored at
80  C for further analysis. 2.2. Determination of antibiotics The concentrations of fifteen antibiotics, including 4 tetracyclines [tetracycline (TC), oxytetracycline (OTC), chlortetracycline (CTC), doxycycline (DOC)], 4 fluoroquinolones [norfloxacin (NFC), ofloxacin (OFC), ciprofloxacin (CFC), enrofloxacin (EFC)], and 7 sulfonamides [sulfadiazine (SDZ), sulfamethoxazole (SMX), sulfadimidine (SM2), sulfamonomethoxine (SMM), sulfaquinoxaline (SQX), sulfamethoxydiazine (SM), sulfaclozine (SCZ)], were determined using liquid chromatography-electrospray ionization tandem mass spectrometry (LC-ESI-MS/MS) as described previously (Huang et al., 2013). Briefly, 0.2 g freeze dried samples were weighed into 50 mL centrifuge tubes, immersed in 20 mL EDTA-SPB (Sodium Phosphate Buffer), and left to stand overnight. The tubes were shaken in the dark at 200 rpm for 30 min, ultra-sonicated for 15 min, and centrifuged at 3500 g for 10 min. The supernatants were kept in brown glass bottles and the residues were extracted with 10 mL extraction buffer twice more. Supernatants from the 3 extractions were combined, diluted to 500 mL with Milli-Q water (18.2 MU cm, Millipore, Bedford, MA), and passed through solid phase extraction cartridges (Oasis HLB, 6 cc/500 mg, Waters, Watford, UK) which had been preconditioned with 10 mL of methanol and 10 mL of super pure water. The cartridges were rinsed with 10 mL super pure water, dried for 20 min with a gentle flow of nitrogen gas, and eluted with 10 mL methanol (supplemented with 0.1% formic acid). The eluents were concentrated with a flow of nitrogen gas, dissolved in 1.0 mL methanol and water (v:v ¼ 3:2, supplemented with 0.1% formic acid), and passed through 0.22 mm membrane filters (Millex, Millipore Corp., Billerica, MA). The solutions were kept in 1.5 mL amber vials before analysis. Antibiotics in the samples were quantified by LC-ESI-MS/ MS with calibration standards and internal standards as described by Huang et al. (2013). Internal standards (tetracycline-D5 for tetracyclines, ciprofloxacin-D8 for fluoroquinolones and sufadimethoxine-D6 for sulfonamides) were spiked into the sample matrices (100 mg kg
1) prior to the extraction and into calibration standards (100 mg L
1) (Huang et al., 2013). Average recoveries for tetracyclines, fluoroquinolones and sulfonamides in manures and composts were 96.6 ± 9.3%, 121 ± 18.9% and 95.4 ± 5.0%, respectively. 2.3. DNA extraction and purification DNA was extracted from 0.3 g freeze-dried and ground samples each time using a PowerSoil DNA Isolation Kit (MoBio Laboratories, Carlsbad, CA) according to the manufacturer's instruction. Three to ten aliquots of each sample were extracted and pooled to obtain enough DNA for analysis. The DNA extracts were purified with a PowerClean DNA Clear-up Kit (MoBio Laboratories, Carlsbad, CA). Purification of DNA from swine manure failed to yield enough purified DNA probably due to a high level of impurities in the DNA extracts and relatively low capacity of the PowerClean DNA Clearup Kit in the cleaning process under this circumstance. For this reason, ARGs could not be determined in the swine manure (SM) samples. The absorbance at 230, 260 and 280 nm of the purified DNA samples was determined by a NanoDrop 2000C spectrophotometer (Thermo Scientific, Wilmington, USA). The quality of the purified DNA was deemed acceptable if A260/A280 was greater than 1.8 and A260/A230 greater than 1.7. DNA concentrations were measured with a Quant-iT PicoGreen double-stranded DNA

184

W.-Y. Xie et al. / Environmental Pollution 219 (2016) 182e190

(dsDNA) assay kit (Invitrogen, USA). The DNA was stored at
80  C until further analysis. 2.4. Real-time quantitative PCR of 16S rRNA genes The absolute copy numbers of 16S rRNA genes were quantified using Roche LightCycle 480 (Roche) according to a previous study (Ouyang et al., 2015). Briefly, the reaction system contained 10 mL of SYBR Premix Ex TaqII (2  ) (Takara Bio, Dalian, China), 1 mL of each primer (10 mmol L
1, the primers were the same as that in the HTqPCR described below), 0.5 mL of bovine serum albumin (20 mg mL
1), and approximately 10 ng of the DNA template. The amplifying conditions were as follows: 95  C for 5 min; followed by 40 cycles of 95  C for 15 s, 60  C for 60 s, 72  C for 15 s for elongation and fluorescence detection. A plasmid with the same fragment of 16S rRNA gene prepared with verified plasmid-constructing method was used as an external quantitative standard (Xie et al., 2014). PCR was conducted in triplicate with standard curves and negative controls in every 96-well plate. The average of the triplicate Ct values was used for copy number calculation. Copy numbers of 16S rRNA genes were used to calculate the absolute abundances of ARGs or MGE marker genes as described below. 2.5. High-throughput quantitative PCR of antibiotic resistance genes Quantification of ARGs was performed using a HT-qPCR platform (Wafergen SmartChip Real-time PCR system) as described previously (Chen et al., 2016; Looft et al., 2012; Zhu et al., 2013). This platform can simultaneously quantify 296 verified primers (Table S1), including 285 for ARGs, 10 for MGE marker genes (2 integrase genes and 8 transposase genes) and 1 for 16S rRNA genes on a chip with 5184 nanowells (Wang et al., 2014). All the primers were previously verified through BLAST homology search, amplification efficiency check and type strain test (Looft et al., 2012; Zhu et al., 2013). For some genes, multiple primers were designed for amplification. A total of 285 primers were included for 212 ARGs in the detection profile. The PCR conditions were the same as those described previously (Chen et al., 2016; Wang et al., 2014). Technical triplicates for each sample and one reagent control were included on one chip. A threshold of Ct < 31 and amplifying efficiency range of 80%e120% in technical triplicates were used to identify positive amplifications, and the average Ct values from the technical triplicates were used for further calculations. The absolute abundance (copies/g DW) of each ARG in each sample was calculated by multiplying the absolute copy number of 16S rRNA genes by the relative copy number of this ARG, which was calculated through the following equation (Ouyang et al., 2015): Relative copy number ¼

ð31
Ct

Þ=ð10=3Þ

ðARGÞ 10 ð31
Ctð16SÞ Þ=ð10=3Þ 10

, where Ct(ARG) and Ct(16S) were both

retrieved from the same sample on the chip. Multiple primers were combined for the corresponding gene. For samples with biological replicates, genes detected in all three replicates were taken as positive and included in further analysis. Because of the possible overlap amplification by the primers intI-1 and cintI-1 (Gillings et al., 2014), only intI-1 was included in the calculation of total abundances of MGE marker genes along with the 8 transposase genes. 2.6. Data analysis Calculations of abundances, means and standard deviations (SD) were conducted in Microsoft Office Excel 2010. The platform of R (version 3.1.3) (R Core Team, 2015) was employed to analyze the adiversity (Shannon index) and the structure of ARGs (principal

coordinate analysis (PCoA) based on Bray-Curtis dissimilarity) with the “Vegan” (2.3e0) package. Heatmaps of ARG profiles were constructed using the package “pheatmap” (version 1.0.8). Statistical analyses of multiple comparison and correlations were conducted in SPSS 16.0. In order to minimize the variance of DNA extraction, the absolute abundances of total ARGs and total MGE marker genes were normalized to those of 16S rRNA genes (4.1 copies per cell) to obtain copy numbers per cell (copies/cell) for comparisons between different samples (Klappenbach et al., 2001; Su et al., 2015). Correlation and regression analyses were based on log10-transformed abundances to stabilize the variance. 3. Results 3.1. Occurrence and concentrations of antibiotics All the 15 antibiotics were detected in at least one of the samples (Table 1). TC, DOC, NFC, CFC and SDZ were found in all the samples. Manure samples, especially cattle manures, contained higher concentrations of tetracyclines than fluoroquinolones and sulfonamides. Concentrations of total tetracyclines in all compost samples were significantly (P < 0.01) lower than those in the corresponding manures. Similar results were obtained for total fluoroquinolones and sulfonamides in cattle samples. However, in poultry and swine compost samples, fluoroquinolones (OFC, CFC and EFC in poultry composite samples; NFC and OFC in swine compost samples) were detected as the dominant antibiotics with concentrations of several mg kg
1, which were significantly higher than those in the corresponding manures. A significant (P < 0.01) decline of total antibiotic concentration after composting was observed only in cattle samples, whereas the total antibiotic concentrations in composts made from poultry and swine manures were significantly higher than those in the manures (Table 1). 3.2. Diversity of ARGs In total, 161 ARGs were detected in all samples, ranging from 65 to 137 in each sample. The number of ARGs detected in the cattle and poultry manures (CM and PM) was larger than those in the corresponding compost samples collected at the thermophilic phase (CT and PT) and the mature phase (CC and PC) (Fig. 1). Similarly, a-diversity in terms of Shannon index decreased significantly in the compost samples compared to the corresponding manure samples (Fig. 1). For cattle and poultry manure composts, a-diversity in the mature compost was significantly lower than that in the thermophilic samples. Because of the impurity of the DNA samples from the swine manure, ARGs could not be determined. However, composts made from swine manure at both the thermophilic and mature phase showed broadly similar numbers and a-diversity of ARGs to those made from cattle or poultry manures (Fig. 1). Both the number and the a-diversity of ARGs increased significantly from the thermophilic phase to the mature phase in the swine manure compost (Fig. 1). PCoA analysis revealed significantly different structures of ARG compositions in the compost samples from those in the manures (ANOSIM, P < 0.05). All compost samples distributed closely along the first coordinate, which explained 53.8% of the variance, whereas the cattle and poultry manures were distant from the compost samples on this coordinate (Fig. S1). The second coordinate (explaining 15.8% of the variance) significantly separated the cattle manure composts from the poultry and swine manure composts; the latter two were close together (Fig. S1). ARGs conferring resistance to multidrug, aminoglycoside, betalactam, tetracycline, vancomycin, MLSB (macrolide, lincosamide and streptogramin B), FCA (fluoroquinolone and chloramphenicol

W.-Y. Xie et al. / Environmental Pollution 219 (2016) 182e190

185

Table 1 Concentrations of antibiotics (mg kg
1 DW) in the manure and compost samples. Data in the parentheses are standard deviation (n ¼ 3).

Tetracyclines

Fluoroquinolones

Sulfonamides

Total antibiotics

TC OTC CTC DOC Total NFC OFC CFC EFC Total SDZ SMX SM2 SMM SCX SM SCZ Total

CM

CT

CC

PM

PT

PC

SM

ST

SC

80.1(9.9) 2817.5(140.8) 48.3(1.2) 71.8(15.0) 3017.7(132.3) 22.5(9.2) 9.0(6.1) 107.1(23.9)
29.9(1.1) 52.5(6.3)
28.2(2.5)
33.3(1.1) 13.3(2.9) 185.9(49.2) 33.1(3.5) 265.6(45.3) 38.5(22.1) 135.3(20.8) 34.9(31.6) 18.2(5.6) 226.9(18.2) 30.0(6.2) 3.1(2.8) 9.7(7.9) 10.3(6.6) 20.1(22.3) 17.6(2.5) 29.0(4.2) 119.7(38.0) 612.5(10.9)

31.3(2.6) 7.1(1.8) 56.6(22.3) 26.5(3.4) 121.5(23.8) 245.8(27.7) 2641.0(402.3) 292.9(21.9) 1445.5(60.8) 4625.2(452.0) 4.0(1.3)
32.7(4.8) 7.2(3.1) 57.6(17.8) 30.1(1.0) 127.6(15.3) 110.2(8.1) 870.8(30.4) 729.3(17.9) 3633.0(118.3) 5343.4(135.6) 3.6(0.8)
95.0(24.3) 19.4(12.3) 80.8(13.9) 94.8(20.8) 289.9(42.7) 8.0(13.9) 20.4(17.3) 202.2(32.7)
39.9(8.0) 38.2(13.1) 69.1(47.1) 36.5(3.7) 183.7(51.0) 1191.8(359.8) 1801.6(494.3) 103.1(33.9) 18.2(7.2) 3114.7(888.4) 9.5(7.5)
25.6(0.6) 10.5(0.7) 31.6(8.4) 24.3(1.6) 92.0(9.5) 491.9(61.0) 936.3(102.7) 41.0(4.1)
MDL: method detection limit. CM, PM and SM represent cattle manure, poultry manure and swine manure, respectively. CT, PT and ST represent thermophilic compost made from cattle manure, poultry manure and swine manure respectively; CC, PC and SC represent mature compost made from cattle manure, poultry manure and swine manure, respectively.

absolute abundances of ARGs and MGE marker genes were normalized to those of 16S rRNA genes (Fig. 2). Abundances of ARG and MGE marker genes in the compost samples were significantly lower than those in the corresponding manures (1.9 and 5.5 copies/ cell for the cattle manure and poultry manure, respectively). ARG

Fig. 1. The numbers of ARGs detected in manures or manure composts and the Shannon index. Data for swine manure (SM) were not available because of the lack of enough purified DNA. ARGs were grouped according to the antibiotic classes they confer resistance to.

in the present study) and sulfonamide were detected in all samples (Fig. S2). Among these ARG classes, genes conferring resistance to aminoglycoside (42%), MLSB (19%), tetracycline (15%) and multidrug (14%) were the most frequently detected. As to the resistance mechanisms, antibiotic deactivation, efflux pump and cellular protection accounted for 50%, 25% and 22% of the total abundances, respectively (Fig. S2).

3.3. Abundances and compositions of ARGs and MGE marker genes The absolute abundances of ARGs and MGE marker genes in all manure and compost samples ranged from 109 to 1010 and from 108 to 1010 copies g
1 DW, respectively. In order to minimize the effect of the variation in DNA extraction among different samples, the

Fig. 2. Normalized abundances of ARGs (A) and MGE marker genes (B). The absolute abundances of ARGs and MGE marker genes were divided by the corresponding absolute abundance of 16S rRNA genes and then multiplied with the average copy number of 16S rRNA genes per cell (4.1). Data for swine manure SM were not available because of the lack of enough purified DNA.

186

W.-Y. Xie et al. / Environmental Pollution 219 (2016) 182e190

abundances in the cattle and poultry mature composts (0.2 and 1.8 copies/cell, respectively) were 9.6% and 31.7% of those in the corresponding manures, respectively. For MGE marker genes, the abundances in the cattle and poultry mature composts (0.04 and 0.75 copies/cell, respectively) were 18.1% and 40.6% of those in the corresponding manures (0.22 and 1.9 copies/cell for the cattle manure and poultry manure, respectively), respectively. Differences between the thermophilic composts and mature composts were either not significant or not constant (Fig. 2). Cattle manure composts contained significantly lower (P < 0.01, 2 tailed t-test) ARGs and MGE marker genes than the composts of poultry and swine manures. Significant changes of gene abundance were observed at the level of ARG class. For cattle samples, all ARG classes in the composts (thermophilic and mature phases) were less abundant than in the cattle manure, especially genes conferring resistance to tetracyclines, MLSB, beta-lactams and vancomycin. A similar trend was observed in poultry samples, especially for genes conferring resistance to tetracyclines, beta-lactams, and vancomycin (Fig. 3). At the level of individual genes, different ARGs showed distinct responses in terms of absolute abundance changes during the composting process (Tables S2 and S3). Some prominent ARGs in the manures decreased significantly in abundance in the mature composts, such as aacA/aphD, tetO and Tp614 in both cattle and poultry samples (Table S2). There were more ARGs and MGE marker genes in the poultry samples showing a significant decrease in abundance during composting than in the cattle samples (Table S2). However, some genes showed little abundance change during composting, especially for poultry (20 genes) and swine samples (23 genes) (Table S3). Genes such as aadA, aadA2, qacED1,

tetL, cintI-1, intI-1 and tnpA-04 were detected at high concentrations in all the manure and compost samples (Table S3). 3.4. Correlation analysis Significantly positive correlation was observed between total ARGs and total MGE marker genes (Table 2), which showed a strong linear relationship (R2 ¼ 0.87, P < 0.01; Fig. S3). Total MGE marker genes and the integrase genes cintI-1 and intI-1 showed high correlation (r > 0.80, P < 0.01) with gene classes conferring resistance to multidrug, aminoglycoside, FCA (mostly Chloramphenicol in the current study) and sulfonamide. In addition, significant correlations were observed between transposase gene tnpA and some ARG groups (Table 2). Meanwhile, no significant correlation was observed between the abundance of 16S rRNA genes and that of ARGs or MGE marker genes (Table 2). Correlation analysis was further conducted on the abundances of individual genes to identify highly-correlated genes (Pearson correlation coefficient r > 0.8) shown in a heatmap (Fig. 4). Several correlation “islands” harboring highly-correlated genes were observed, among which islands 1, 2, 3 and 4 contained genes detected in all the samples. Islands 1, 2 and 4 each contained ARGs in different resistance classes, whereas island 3 contained both ARGs (qacED1, aadA2, ermX, cmlA1) and MGE marker genes (cintI-1, intI-1 and tnpA-04) (Fig. 4). The genes inside each island all showed significantly linear relationship (P < 0.01) with each other. Outside these islands, drfA1 and strB also had high correlation coefficients with intI-1, cintI-1, aadA2, and qacED1 (Fig. 4). These highlycorrelated genes may result from co-occurrence of bacterial strains or genetic elements harboring these genes. Correlations between ARGs and the concentrations of antibiotics were analyzed at the levels of classes and individuals (Table 2; Table S4). The coefficients between total antibiotics and total ARGs was 0.58 (P < 0.01, Pearson). At the class level, no significant correlation was observed between the three categories of antibiotics (tetracyclines, sulfonamides and fluoroquinolones) and their corresponding resistance genes (Table 2). However, it is noteworthy that fluoroquinolones showed significantly positive correlations with ARG classes conferring no direct resistance to fluoroquinolones (Table 2). At the individual level, significantly negative or positive correlations were observed between individual antibiotics and genes (Table S4). For example, SDZ and SMX correlated negatively (r <
0.6, P < 0.01) with intI-1, tnpA-04, cmlA1 and ermX. On the other hand, 28, 18 and 16 genes showed significant positive correlations (r > 0.6, P < 0.01) with OFC, CTC and NFC, respectively (Table S4). Notably, NFC and OFC exhibited significant linear relationships with cintI-1, intI-1, tnpA-04 and some abundant ARGs with no direct relationship to fluoroquinolones, such as aadA2, cmlA1, ermX, qacED1 and tetG (Figs. S4 and S5). Tetracycline resistance gene tetM correlated significantly with TC, CTC and DOC, while tetPA, tetPB, tetR and tetT correlated significantly with CTC. 4. Discussion 4.1. High levels of antibiotics in the animal manures and manure composts

Fig. 3. Heatmap of the MGE marker genes and ARGs in different resistance classes. The absolute abundances were divided by the corresponding copy numbers of 16S rRNA genes. In order to obtain positive values for better displaying, the relative abundances were multiplied by 106 and ln transformed. The heatmap was generated in R with the package “pheatmap”. The columns were clustered with the method “ward.D2” based on the Bray-Curtis distance. The rows were arranged manually to show the changes according to the compost phases.

Over use of antibiotics in animal husbandry is a world-wide problem (Van Boeckel et al., 2015). China is one of the largest consumer of antibiotics around the world (Zhang et al., 2015).In the present study, high concentrations (up to mg kg
1) of tetracyclines were found in cattle manure, whereas fluoroquinolones were present at high levels in composts made from swine or poultry manures (Table 1). Tetracyclines are commonly used in animal feeds in China, and have been reported with high occurrence in

W.-Y. Xie et al. / Environmental Pollution 219 (2016) 182e190

187

Table 2 Pearson's correlation coefficients between ARG groups and 16S rRNA genes, MGE marker genes and antibiotic categories. ARG groups

Multidrug Aminoglycoside Beta-lactam Tetracycline Vancomycin MLSB FCA Fluoroquinolone resistance Chloramphenicol resistance Sulfonamide Other Total ARGs

16S rRNA genes MGE marker genes

0.22 0.33
0.28
0.27
0.17 0.02 0.22
0.06 0.22 0.33 0.08 0.12

Antibiotics

cintI-1 intI-1

tnpA-02 tnpA-04 tnpA-05 IS613

Total fluoroquinolones tetracyclines sulfonamides Total MGEs antibiotics

0.95 a 0.91 a 0.33 0.65 a 0.28 0.65 a 0.92 a 0.30 0.92 a 0.88 a 0.41 b 0.87 a

0.61 a 0.66 a 0.45 b 0.43 b
0.01 0.24 0.41 b 0.25 0.41 b 0.58 a 0.49 b 0.53 a

0.97 a 0.92 a 0.51 a 0.74 a 0.43 b 0.78 a 0.88 a 0.40 0.89 a 0.93 a 0.60 a 0.93 a

0.92 a 0.86 a 0.24 0.50 a 0.28 0.61 a 0.86 a 0.32 0.86 a 0.87 a 0.44 b 0.80 a

0.78 a 0.72 a
0.11 0.36 0.10 0.49 a 0.86 a 0.07 0.86 a 0.72 a 0.18 0.66 a

0.75 a 0.73 a 0.76 a 0.59 a 0.67 a 0.83 a 0.62 a 0.59 a 0.62 a 0.81 a 0.68 a 0.81 a

0.45 b 0.43 b 0.61 a 0.71 a 0.46 b 0.47 b 0.36 0.19 0.36 0.38 0.58 a 0.56 a

0.67 a 0.64 a
0.01 0.52 a 0.20 0.60 a 0.84 a
0.07 0.84 a 0.61 a 0.18 0.67 a


0.20
0.07 0.49 a 0.37 0.09
0.01
0.19
0.11
0.19
0.24 0.00 0.04

0.11 0.16 0.24 0.56 a 0.27 0.25 0.23
0.26 0.23 0.05 0.06 0.30

0.45 b 0.49 b 0.19 0.67 a 0.19 0.49 a 0.63 a
0.19 0.63 a 0.36 0.08 0.58 a

The absolute abundances or concentrations were log10 transformed prior to the analysis. MLSB, Macrolide-Lincosamide-Streptogramin B resistance genes. FCA, fluoroquinolone, quinolone, florfenicol, chloramphenicol, and amphenicol resistance genes. FCA resistance genes contain fluoroquinolone resistance genes (qnrA) and chloramphenicol resistance genes (catA1, catB3, catB8, cfr, cmlA, cmx(A), and floR). Data in bold are coefficients >0.8. a Significance level at P < 0.01 (2-tailed). b Significance level P < 0.05 (2-tailed).

manures from previous studies (Qiao et al., 2012; Zhao et al., 2010). Fluoroquinolones are the most widely used veterinary antibiotics in China according to a market survey (Zhang et al., 2015). The results from the present study therefore reflect the usage of veterinary antibiotics in China. Previous studies showed considerable efficacy of composting in the degradation of the spiked tetracyclines, fluoroquinolones and sulfonamides (Ho et al., 2013; Mitchell et al., 2015; Wang et al., 2015). Similarly, significant declines in the concentrations of antibiotics from the manure to composts were detected for all the 3 categories of antibiotics in cattle samples and for tetracyclines in poultry and swine composts in the present study. However, large increases were observed for fluoroquinolones after composting in poultry and swine samples (Table 1). Transformation of antibiotic metabolites N-acetylsulfadiazine back to sulfadiazine has been reported during manure storage in previous studies based on isotope tracking (Heuer et al., 2008; Lamshoft et al., 2010). However, no similar transformation was reported for fluoroquinolone metabolites. Fluoroquinolones can be easily adsorbed by organic matter (Sukul and Spiteller, 2007). It is possible that during composting, mineralization of organic matter released some of the adsorbed fluoroquinolones, resulting in an apparent increase in their concentrations. Another possibility for the apparent increase in the concentrations of fluoroquinolones is the heterogeneity of the raw materials, which is likely to be substantial for commercial composting operations dealing with large quantities of raw materials. In the present study, samples of manures and composts from different composting phases were collected in a factory on the same day. This sampling strategy was simple to conduct, but could have introduced more heterogeneity to the samples used in this study. Nevertheless, the results suggested that antibiotics, especially fluoroquinolones, were not sufficiently degraded during industrial-scale thermophilic composting. Although the antibiotic concentrations in the manures and composts in the present study were lower than those reported in previous studies (Hu et al., 2010; Huang et al., 2013; Qiao et al., 2012; Zhao et al., 2010), repeated applications of these manures or composts could still introduce considerable amounts of antibiotics into soils. 4.2. Abundant and diverse ARGs and MGE marker genes in the manures and composts In contrast to the concentrations of antibiotics, the diversity and abundances of ARGs and MGE marker genes significantly declined

from manure to compost for both cattle and poultry manures (Figs. 1 and 2). Although ARGs in swine manure could not be determined, the ARG diversity and abundance in the composts made from swine manure were at the same level as those in the cattle and poultry composts, suggesting similar reductions during composting of swine manure. The reduction could result from the dilution by the addition of straws and/or decrease in the ARG abundance during the composting process. In order to achieve an initial C/N ratio of around 25:1 for optimal composting performance, the addition of straw as the bulking material is usually less than 50% by weight, given its high C/N ratio (Liu et al., 2011; Zhu et al., 2004). However, the reduction of ARGs and MGE marker genes in term of normalized abundance were more than 68% and 59%, respectively. Therefore, these observations suggest that the observed decrease in ARG abundance was not just a simple dilution effect, but was at least partly due to the effect of thermophilic composting, a process that is known to kill the majority of bacteria in the manures (Bernal et al., 2009). Despite the general decreasing trend in the total ARG abundance, individual genes behaved differently during composting, with some showing large decreases in the abundance while others being persistent (Table S2 and Table S3). Noticeable examples for significant attenuation included beta-lactam resistance genes such as blaCMY, blaCTX-M and blaTEM, and genes associated with MGEs such as IS613, tnpA-01, tnpA-03, tnpA-05 and Tp614 in poultry samples. In contrast, ARGs such as aadA, aadA2, qacED1 and tetL appeared to be highly persistent in all 3 types of mature composts. These ARGs deserve more attention as they may pose a greater risk of environmental dissemination. In many cases, antibiotics used in animal husbandry are in the same classes as those used in humans, and ARGs found in human pathogens have been found to share a close phylogenic relationship with those in livestock production (Finley et al., 2013). Additionally, the persistence of integrase genes (intI-1) and transposase genes (tnpA-04) could add further risks of ARG transfers from the compost utilization into human pathogens (Djordjevic et al., 2013; Domingues et al., 2012; Gillings et al., 2008). Future studies should investigate the optimal composting conditions to achieve more comprehensive eliminations of resistance elements. 4.3. Correlations between ARGs and MGE marker genes in the manures and composts The ARGs were found to have close relationships with MGE

188

W.-Y. Xie et al. / Environmental Pollution 219 (2016) 182e190

Fig. 4. Identification of the highly-correlated ARGs and MGE marker genes. The Heatmap was clustered with correlation coefficients (Pearson) above 0.8, with those under 0.8 all assigned to 0.0. Highly-correlated genes clustered together to form correlation “islands” marked in black squares, among which the “islands” containing genes found in all the samples were marked with 1, 2, 3 and 4. Island 1 contains oprJ, oprD, fox5, pikR2 and pncA. Island 2 contains aadA9, tetG, erm(35) and cmx(A). Island 3 contains qacED1, aadA2, ermX, cmlA1, cintI-1, intI-1 and tnpA-04. Island 4 contains drfA1, strB, tetR, floR, aac(60 )-Ib(akaaacA4) and catB3.

maker genes on the basis of total abundances, gene class abundances and individual gene abundances. Mobile genetic elements are critical factors in the exchanges of multiple genes between fecal and environmental bacteria (Djordjevic et al., 2013; Domingues et al., 2012; Goldstein et al., 2001) and the genes for integrase (intI-1 or cintI-1) and transposase (tnpA) are often used as marker genes for MGEs (Johnson et al., 2016; Zhu et al., 2013). The high correlations between ARGs and MGE marker genes indicate the important roles of MGEs in the persistence and/or proliferation of ARGs in the animal wastes, which was further emphasized by the identification of the “correlation island” consisting of genes associated with MGEs (cintI-1, intI-1 and tnpA-04) and ARGs in different resistance classes (qacED1, aadA2, ermX and cmlA1) (Fig. 4). These MGE-associated genes and ARGs were persistent from manures to

the composts (Table S3). In previous studies, resistance genes such as qacED1, aadA2 and cmlA1 have been detected on class 1 integrons in animal manures and manured soils (Binh et al., 2009; Chuanchuen et al., 2007; Johnson et al., 2016; Srinivasan et al., 2008). Class 1 integrons are able to capture different ARGs in the form of circular cassette to construct a tandem array of ARGs using the integrase encoded by intI-1 or cintI-1 (Gillings et al., 2014; Partridge et al., 2009), while the closely related transposons are critical for the further dissemination of these integron cassettes (Heuer et al., 2012; Partridge et al., 2009). Our observations suggest possible prevalence of multiple ARG-harboring integron cassettes in close relationship to transposons in the animal wastes, and therefore could be an important pathway of environmental dissemination of multiple antibiotic resistance.

W.-Y. Xie et al. / Environmental Pollution 219 (2016) 182e190

4.4. The role of antibiotics in the ARG selection

Academy of Sciences for their support in ARG determination.

Both positive and negative correlations were observed between individual antibiotics and ARGs (Table S4). Therefore, correlations between the total ARG abundance and the overall antibiotic concentration may be less meaningful than those between specific genes and antibiotics. At the level of individual genes and antibiotics, the significantly negative correlations observed suggest that some of the antibiotics, SDZ for instance, might still be active to inhibit specific bacterial groups (Lamshoft et al., 2007). On the other hand, the positive correlations indicate possible selective pressures conveyed by the antibiotics, which have been frequently observed in previous studies (Zhu et al., 2013). Co-selection of diverse ARGs by the antibiotics was evident from the high correlation coefficients between OFC or NFC and genes in the resistance classes other than fluoroquinolone resistance genes. Similar observations from a test of in-feed antibiotics on swine intestinal microbiome and a survey on ARG profiles in manure-amended soils were reported in previous studies (Looft et al., 2012; Zhu et al., 2013). The phenomenon might be caused by the prevalence of multiple ARG-flanking mobile genetic elements in the manures and composts (Binh et al., 2009; Chuanchuen et al., 2007; Zhu et al., 2013), because MGE marker genes (cintI-1, intI-1 and tnpA-04) also showed high correlations with OFC or NFC concentration. When any one of these closely located genes on MGEs was selected by OFC or NFC, co-selection of the adjacent genes became inevitable. In this case, it appears that antibiotics in the animal manure or composts and MGEs act synergistically in the proliferation and persistence of high levels of resistance genes.

Appendix A. Supplementary data

5. Conclusions Relatively high levels of fluoroquinolones were detected in poultry or swine manure-based composts. Repeated applications of these composts could lead to buildup of antibiotic residues in the soil. Commercial thermophilic composting was effective in reducing the abundances of some ARGs and MGE marker genes. However, other ARGs and MGE marker genes were found to be persistent from manures to composts. Optimal operation conditions and durations of composting should be investigated in the future for better performance in antibiotic and ARG eliminations. Antibiotics, norfloxacin and ofloxacin in particular, and MGEs appeared to play important roles in the proliferation and persistence of high levels of some ARGs in the manures and composts. Multiple ARG-harboring class 1 integron cassettes in close relationship with transposons appear to be prevalent in the animal manures and manure-based composts, which may pose a high risk of ARG dissemination into the environment. Results in the present study indicate an urgent need for a comprehensive stewardship of antibiotic usage in animal husbandry and to improve the efficiency of commercial thermophilic composting in the removals of antibiotics and ARGs, in China and other regions with high usage of antibiotics. Acknowledgement This research was supported by China Postdoctoral Science Foundation (2015M580440), “the Fundamental Research Funds for the Central Universities” (KYZ201518), Jiangsu Planned Projects for Postdoctoral Research Funds (1501062B), and Jiangsu Collaborative Innovation Center for Solid Organic Waste Resource Utilization (R0201300356). The authors sincerely thank Yu-Juan Huang and Lu Yang from Institute of Soil Science, Chinese Academy of Sciences for her help in antibiotic quantification, and Drs. Yong-Guan Zhu and Jian-Qiang Su from Institute of Urban Environment, Chinese

189

Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.envpol.2016.10.044. References Bernal, M.P., Alburquerque, J.A., Moral, R., 2009. Composting of animal manures and chemical criteria for compost maturity assessment: a review. Bioresour. Technol. 100, 5444e5453. Binh, C.T., Heuer, H., Kaupenjohann, M., Smalla, K., 2009. Diverse aadA gene cassettes on class 1 integrons introduced into soil via spread manure. Res. Microbiol 160, 427e433. Blair, J.M.A., Webber, M.A., Baylay, A.J., Ogbolu, D.O., Piddock, L.J.V., 2014. Molecular mechanisms of antibiotic resistance. Nat. Rev. Microbiol. 13, 42e51. Chen, Q., An, X., Li, H., Su, J., Ma, Y., Zhu, Y.G., 2016. Long-term field application of sewage sludge increases the abundance of antibiotic resistance genes in soil. Environ. Int. 92e93, 1e10. Chuanchuen, R., Khemtong, S., Padungtod, P., 2007. Occurrence of qacE/qacEDelta1 genes and their correlation with class 1 integrons in salmonella enterica isolates from poultry and swine. Se. Asian. J. Trop. Med. 38, 855e862. Dantas, G., Sommer, M.O.A., Oluwasegun, R.D., Church, G.M., 2008. Bacteria subsisting on antibiotics. Science 320, 100e103. Djordjevic, S.P., Stokes, H.W., Roy Chowdhury, P., 2013. Mobile elements, zoonotic pathogens and commensal bacteria: conduits for the delivery of resistance genes into humans, production animals and soil microbiota. Front. Microbiol. 4, 86. Domingues, S., Harms, K., Fricke, W.F., Johnsen, P.J., da Silva, G.J., Nielsen, K.M., 2012. Natural transformation facilitates transfer of transposons, integrons and gene cassettes between bacterial species. PLoS Pathog. 8, e1002837. Finley, R.L., Collignon, P., Larsson, D.G., McEwen, S.A., Li, X.Z., Gaze, W.H., ReidSmith, R., Timinouni, M., Graham, D.W., Topp, E., 2013. The scourge of antibiotic resistance: the important role of the environment. Clin. Infect. Dis. 57, 704e710. Gillings, M.R., Gaze, W.H., Pruden, A., Smalla, K., Tiedje, J.M., Zhu, Y.-G., 2014. Using the class 1 integron-integrase gene as a proxy for anthropogenic pollution. ISME J. 9, 1269e1279. Gillings, M.R., Krishnan, S., Worden, P.J., Hardwick, S.A., 2008. Recovery of diverse genes for class 1 integron-integrases from environmental DNA samples. FEMS Microbiol. Lett. 287, 56e62. Goldstein, C., Lee, M.D., Sanchez, S., Hudson, C., Phillips, B., Register, B., Grady, M., Liebert, C., Summers, A.O., White, D.G., Maurer, J.J., 2001. Incidence of class 1 and 2 integrases in clinical and commensal bacteria from livestock, companion animals, and exotics. Antimicrob. Agents Chemother. 45, 723e726. Heuer, H., Binh, C.T., Jechalke, S., Kopmann, C., Zimmerling, U., Krogerrecklenfort, E., Ledger, T., Gonzalez, B., Top, E., Smalla, K., 2012. IncP-1epsilon plasmids are important vectors of antibiotic resistance genes in agricultural systems: diversification driven by class 1 integron gene cassettes. Front. Microbiol. 3, 2. € ft, M., Smalla, K., Matthies, M., Spiteller, M., 2008. Fate of Heuer, H., Focks, A., Lamsho sulfadiazine administered to pigs and its quantitative effect on the dynamics of bacterial resistance genes in manure and manured soil. Soil Biol. Biochem. 40, 1892e1900. Ho, Y.B., Zakaria, M.P., Latif, P.A., Saari, N., 2013. Degradation of veterinary antibiotics and hormone during broiler manure composting. Bioresour. Technol. 131, 476e484. Hu, X., Zhou, Q., Luo, Y., 2010. Occurrence and source analysis of typical veterinary antibiotics in manure, soil, vegetables and groundwater from organic vegetable bases, northern China. Environ. Pollut. 158, 2992e2998. Huang, Y., Cheng, M., Li, W., Wu, L., Chen, Y., Luo, Y., Christie, P., Zhang, H., 2013. Simultaneous extraction of four classes of antibiotics in soil, manure and sewage sludge and analysis by liquid chromatography-tandem mass spectrometry with the isotope-labelled internal standard method. Anal. Methods 5, 3721. Jechalke, S., Heuer, H., Siemens, J., Amelung, W., Smalla, K., 2014. Fate and effects of veterinary antibiotics in soil. Trends Microbiol. 22, 536e545. Johnson, T.A., Stedtfeld, R.D., Wang, Q., Cole, J.R., Hashsham, S.A., Looft, T., Zhu, Y.G., Tiedje, J.M., 2016. Clusters of antibiotic resistance genes enriched together stay together in swine agriculture. mBio 7, e02214ee02215. Klappenbach, J.A., Saxman, P.R., Cole, J.R., Schmidt, T.M., 2001. rrndb: the ribosomal RNA operon copy number database. Nucleic Acids Res. 29, 181e184. Lamshoft, M., Sukul, P., Zuhlke, S., Spiteller, M., 2007. Metabolism of 14C-labelled and non-labelled sulfadiazine after administration to pigs. Anal. Bioanal. Chem. 388, 1733e1745. Lamshoft, M., Sukul, P., Zuhlke, S., Spiteller, M., 2010. Behaviour of (14)C-sulfadiazine and (14)C-difloxacin during manure storage. Sci. Total Environ. 408, 1563e1568. Larson, C., 2015. China's lakes of pig manure spawn antibiotic resistance. Science 347, 704. Liu, D., Zhang, R., Wu, H., Xu, D., Tang, Z., Yu, G., Xu, Z., Shen, Q., 2011. Changes in biochemical and microbiological parameters during the period of rapid composting of dairy manure with rice chaff. Bioresour. Technol. 102, 9040e9049.

190

W.-Y. Xie et al. / Environmental Pollution 219 (2016) 182e190

Looft, T., Johnson, T.A., Allen, H.K., Bayles, D.O., Alt, D.P., Stedtfeld, R.D., Sul, W.J., Stedtfeld, T.M., Chai, B., Cole, J.R., Hashsham, S.A., Tiedje, J.M., Stanton, T.B., 2012. In-feed antibiotic effects on the swine intestinal microbiome. Proc. Natl. Acad. Sci. U.S.A. 109, 1691e1696. Luby, E., Ibekwe, A.M., Zilles, J., Pruden, A., 2016. Molecular methods for assessment of antibiotic resistance in agricultural ecosystems: prospects and challenges. J. Environ. Qual. 45, 441e453. Martinez, J.L., 2008. Antibiotics and antibiotic resistance genes in natural environments. Science 321, 365e367. Mitchell, S.M., Ullman, J.L., Bary, A., Cogger, C.G., Teel, A.L., Watts, R.J., 2015. Antibiotic degradation during thermophilic composting. Water, Air, Soil Pollut. 226. Ouyang, W.Y., Huang, F.Y., Zhao, Y., Li, H., Su, J.Q., 2015. Increased levels of antibiotic resistance in urban stream of Jiulongjiang River, China. Appl. Microbiol. Biotechnol. 99, 5697e5707. Partridge, S.R., Tsafnat, G., Coiera, E., Iredell, J.R., 2009. Gene cassettes and cassette arrays inmobile resistance integrons. FEMS Microbiol. Rev. 33, 757e784. Qiao, M., Chen, W., Su, J.Q., Zhang, B., Zhang, C., 2012. Fate of tetracyclines in swine manure of three selected swine farms in China. J. Environ. Sci. 24, 1047e1052. R Core Team, 2015. R: a Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. Sarmah, A.K., Meyer, M.T., Boxall, A.B., 2006. A global perspective on the use, sales, exposure pathways, occurrence, fate and effects of veterinary antibiotics (VAs) in the environment. Chemosphere 65, 725e759. Selvam, A., Xu, D., Zhao, Z., Wong, J.W., 2012. Fate of tetracycline, sulfonamide and fluoroquinolone resistance genes and the changes in bacterial diversity during composting of swine manure. Bioresour. Technol. 126, 383e390. Srinivasan, V., Nam, H.M., Sawant, A.A., Headrick, S.I., Nguyen, L.T., Oliver, S.P., 2008. Distribution of tetracycline and streptomycin resistance genes and class 1 integrons in Enterobacteriaceae isolated from dairy and nondairy farm soils. Microb. Ecol. 55, 184e193. Stevens, R., 2010. U.S. EPA's 2006-2007 Targeted National Sewage Sludge Survey. In: Halden, R.U. (Ed.), Contaminants of Emerging Concern in the Environment: Ecological and Human Health Considerations. American Chemical Society, Washington DC, USA, pp. 173e198.

Su, J.Q., Wei, B., Ou-Yang, W.-Y., Huang, F.Y., Zhao, Y., Xu, H.-J., Zhu, Y.G., 2015. Antibiotic resistome and its association with bacterial communities during sewage sludge composting. Environ. Sci. Technol. 49, 7356e7363. Sukul, P., Spiteller, M., 2007. Fluoroquinolone Antibiotics in the Environment. In: Ware, G.W. (Ed.), Reviews of Environmental Contamination and Toxicology. Springer, New York, pp. 131e162. Van Boeckel, T.P., Brower, C., Gilbert, M., Grenfell, B.T., Levin, S.A., Robinson, T.P., Teillant, A., Laxminarayan, R., 2015. Global trends in antimicrobial use in food animals. Proc. Natl. Acad. Sci. U.S.A. 112, 5649e5654. Wang, F.H., Qiao, M., Su, J.Q., Chen, Z., Zhou, X., Zhu, Y.G., 2014. High throughput profiling of antibiotic resistance genes in urban park soils with reclaimed water irrigation. Environ. Sci. Technol. 48, 9079e9085. Wang, J., Ben, W., Zhang, Y., Yang, M., Qiang, Z., 2015. Effects of thermophilic composting on oxytetracycline, sulfamethazine, and their corresponding resistance genes in swine manure. Environ. Sci. Process Impacts 17, 1654e1660. WHO, 2014. Antimicrobial Resistance: Global Report on Surveillance 2014. World Health Organization. Xie, W.Y., Su, J.Q., Zhu, Y.G., 2014. Arsenite oxidation by the phyllosphere bacterial community associated with Wolffia australiana. Environ. Sci. Technol. 48, 9668e9674. Zhang, Q.Q., Ying, G.G., Pan, C.G., Liu, Y.S., Zhao, J.L., 2015. Comprehensive evaluation of antibiotics emission and fate in the river basins of China: source analysis, multimedia modeling, and linkage to bacterial resistance. Environ. Sci. Technol. 49, 6772e6782. Zhao, L., Dong, Y.H., Wang, H., 2010. Residues of veterinary antibiotics in manures from feedlot livestock in eight provinces of. China. Sci. Total Environ. 408, 1069e1075. Zhu, N., Deng, C., Xiong, Y., Qian, H., 2004. Performance characteristics of three aeration systems in the swine manure composting. Bioresour. Technol. 95, 319e326. Zhu, Y.G., Johnson, T.A., Su, J.Q., Qiao, M., Guo, G.X., Stedtfeld, R.D., Hashsham, S.A., Tiedje, J.M., 2013. Diverse and abundant antibiotic resistance genes in Chinese swine farms. Proc. Natl. Acad. Sci. U. S. A. 110, 3435e3440.