- Email: [email protected]
Feed additives shift gut microbiota and enrich antibiotic resistance in swine gut
STOTEN-24330; No of Pages 9 Science of the Total Environment xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv
Feed additives shift gut microbiota and enrich antibiotic resistance in swine gut Yi Zhao a,b, Jian-Qiang Su a, Xin-Li An a, Fu-Yi Huang a, Christopher Rensing a,c, Kristian Koefoed Brandt b, Yong-Guan Zhu a,d,⁎ a
Key Laboratory of Urban Environment and Health, Institute of Urban Environment, Chinese Academy of Sciences, 1799 Jimei Road, Xiamen 361021, China Department of Plant and Environmental Sciences, Faculty of Science, University of Copenhagen, Thorvaldsensvej 40, 1871 Frederiksberg, Denmark Institute of Environmental Microbiology, College of Resources and Environment, Fujian Agriculture and Forestry University, Fuzhou 350002, Fujian, China d State Key Laboratory of Urban and Regional Ecology, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing 100085, China b c
H I G H L I G H T S
G R A P H I C A L
A B S T R A C T
• Feed additives and the gut microbiome were investigated at 3 Chinese pig farms. • Multiple metals and 79 antibiotics were analyzed as potential feed additives. • 285 antibiotic resistance genes (ARGs) were quantified by high throughput qPCR. • ARGs were highly diverse and abundant in swine feces from Chinese pig farms. • Feed additives and gut microbiota were major determinants shaping ARG profiles.
a r t i c l e
i n f o
Article history: Received 25 August 2017 Received in revised form 11 October 2017 Accepted 12 October 2017 Available online xxxx Editor: Jay Gan Keywords: Antibiotic resistance High-throughput qPCR 16S rRNA sequencing Swine gut microbiota ARG chip Network analysis
a b s t r a c t Antibiotic resistance genes (ARGs) are emerging environmental contaminants posing a threat to public health. Antibiotics and metals are widely used as feed additives and could consequently affect ARGs in swine gut. In this study, high-throughput quantitative polymerase chain reaction (HT-qPCR) based ARG chip and nextgeneration 16S rRNA gene amplicon sequencing data were analyzed using multiple statistical approaches to profile the antibiotic resistome and investigate its linkages to antibiotics and metals used as feed additives and to the microbial community composition in freshly collected swine manure samples from three large-scale Chinese pig farms. A total of 146 ARGs and up to 1.3 × 1010 total ARG copies per gram of swine feces were detected. ARGs conferring resistance to aminoglycoside, macrolide-lincosamide-streptogramin B (MLSB) and tetracycline were dominant in pig gut. Total abundance of ARGs was positively correlated with in-feed antibiotics, microbial biomass and abundance of mobile genetic elements (MGEs) (P b 0.05). A significant correlation between microbial communities and ARG profiles was observed by Procrustes analysis. Network analysis revealed that Bacteroidetes and Firmicutes were the most dominant phyla co-occurring with specific ARGs. Partial redundancy analysis indicated that the variance in ARG profiles could be primarily attributed to antibiotics and metals in feed (31.8%), gut microbial community composition (23.3%) and interaction between feed additives and community composition (16.5%). These results suggest that increased levels of in-feed additives could aggravate the enrichment of ARGs and MGEs in swine gut. © 2017 Elsevier B.V. All rights reserved.
⁎ Corresponding author at: Key Laboratory of Urban Environment and Health, Institute of Urban Environment, Chinese Academy of Sciences, No. 1799 Jimei Road, Xiamen 361021, China. E-mail addresses: [email protected], [email protected] (Y.-G. Zhu).
https://doi.org/10.1016/j.scitotenv.2017.10.106 0048-9697/© 2017 Elsevier B.V. All rights reserved.
Please cite this article as: Zhao, Y., et al., Feed additives shift gut microbiota and enrich antibiotic resistance in swine gut, Sci Total Environ (2017), https://doi.org/10.1016/j.scitotenv.2017.10.106
2
Y. Zhao et al. / Science of the Total Environment xxx (2017) xxx–xxx
1. Introduction The frequent occurrence and spread of multidrug-resistant pathogenic bacteria is a worldwide challenge and can be largely attributed to the use of antibiotics in animals and humans (Goossens et al., 2005). Antibiotic resistance genes (ARGs) are increasingly regarded as emerging contaminants (Pruden et al., 2006) providing a threat to human and environment health (Ashbolt et al., 2013; Zhu et al., 2017b). China is one of the largest consumers of antibiotics in the world (Van Boeckel et al., 2014) with a significant fraction of all antibiotics being used for animal disease treatment and growth promotion (Yun-peng and Yue, 2008). Antibiotics were first used as feed additives in China in the mid-1970s, but consumption has increased dramatically concomitant with the economic growth in recent decades. A previous report revealed that almost half of the annual antibiotic production in China, totaling about 97,000 tons, went into the animal industry (Collignon and Voss, 2015). Furthermore, metals such as copper and zinc are also approved as feed additives under the stipulation in China for growth promotion and disease control purposes (Cang, 2004). It has been observed that these metals in some cases may cause coselection of ARGs in gut microbiota, but further studies are needed to fully elucidate the links between metals in feed and ARGs in gut microbiota (Yazdankhah et al., 2014). By contrast, it is well established that use of antibiotics in animals may contribute to the antibiotic resistance crisis in humans (Ashbolt et al., 2013; Hoelzer et al., 2017). Hence, the swine gut microbiota is now considered an important reservoir of pathogens with the potential to transfer antibiotic resistant pathogens to humans (Gibson et al., 2015; Heuer et al., 2011). Indeed, the overuse of antibiotics in Chinese animals may possibly explain why Chinese people harbor higher numbers of ARGs in their guts than Europeans (Hu et al., 2013). Apart from direct human exposure to ARGs present in swine gut via contaminated meat (Perreten et al., 1997), ARGs in swine gut may also spread to humans via pig-house dust (Hamscher et al., 2003) or via manure application and wastewater discharge. Although antibiotics and ARGs originating from Chinese swine farms have been documented and characterized (Cheng et al., 2013; Zhu et al., 2013), little is known about the relative importance of different abiotic and biotic factors in shaping the swine gut antibiotic resistome. Hence, the aim of this study was to perform high-throughput profiling of the ARGs in fecal samples from Chinese pig farms and to investigate the effect of antibiotic and metals in swine feed on the antibiotic resistome and its association with the microbial community. To this end, the prevalence and diversity of ARGs and selected mobile genetic elements known to carry ARGs were characterized in fresh fecal matter samples obtained from three large-scale Chinese pig farms using highthroughput quantitative PCR chip technology. ARG profiles were subsequently linked to microbial community composition as analyzed by 16S rRNA amplicon sequencing and to concentrations of antibiotics and metals in swine feed using partial redundancy analysis, Procrustes analysis, and network analysis. Linear discriminative analysis (LDA) Effect Size (LEfSe) (Segata et al., 2011) was further used for highdimensional ARG data and indicator ARGs and microbe discovery (Goecks et al., 2010).
biological replicates of fresh fecal samples were taken immediately after excretion. The four replicate fecal samples were derived from four individual adult pig siblings for biological consistency. The four biological replicates of sampled pigs in each group had all received the same swine feed. All collected seven swine feed and 28 fresh fecal matter samples were sealed in individual 300-mL sterile plastic sampling boxes, immediately transported to the laboratory on dry ice, and stored at −20 °C for further analysis. 2.2. Chemical analysis of antibiotics and metals in swine feed A total of 79 different antibiotics were analyzed in swine feed based on a previously published protocol (Huang et al., 2013). Analyzed included sulfonamides, tetracyclines, chloramphenicols, macrolidelincosamide-streptogramin B (MLSB) and other veterinary antibiotics. All targeted antibiotics and their method detection limits (MDL) are listed in Table S1. Metals including copper, zinc, cadmium, lead, mercury, chromium and arsenic in swine feed were analyzed using inductively coupled plasma mass spectrometry (ICP-MS) according to ISO/TS 16965:2013(E). 2.3. DNA extraction High-quality DNA was extracted from fecal samples (0.5 g) using the MoBio PowerFecal® DNA Isolation Kit according to the instruction manual. The quality and concentration of DNA were checked by agarose gel electrophoresis and ultraviolet-absorbance (ND1000, Nanodrop, Thermo Fisher Scientific Inc.). DNA was diluted to 50 ng per μL using sterile PCR-grade water and stored at −20 °C until further analysis. 2.4. High-throughput 16S rRNA gene sequencing The V4-V5 region of 16S rRNA gene (Claesson et al., 2010) was amplified, purified, quantified, pooled and sequenced on Illumina Miseq platform at Novogene, Beijing, China (Xu et al., 2014). Each 50 μl PCR reaction contained 25 μl TaKaRa ExTaq, 0.5 μl bovine serum albumin, 1 μl of each primer, 1 μl DNA template and 21.5 μl PCR-grade water. The thermal cycles consisted of initial enzyme activation at 94 °C for 3 mins, followed by 30 cycles of denaturation at 94 °C for 30s, annealing at 58 °C for 1 min and extenstion at 72 °C for 1 min, and a final extension at 72 °C for 5 mins. Demultiplexing was performed and low-quality or ambiguous read were removed to downstream clean sequence generation (Novogene). Quantitative Insights Into Microbial Ecology (QIIME) workflow was used to analyze 16S rRNA gene sequence data. The de novo open-reference operational taxonomic unit (OTU) picking process was conducted by QIIME following the online tutorial and script documentation with 93% as similarity level by UCLUST clustering (Edgar, 2010). Ribosomal Database Project (RDP) classifier was used to assign taxonomic data to each representative sequence by default with a confidence threshold of 80%. Singletons were removed prior to downstream analysis. The within community diversity (alpha diversity) and rarefaction curves were computed using observed OTUs, phylogenetic diversity metrics (PD whole tree) and Shannon index. 2.5. HT-qPCR and data processing
2. Materials and methods 2.1. Sampling A total of 28 fecal samples representing seven groups (populations) with four replicates of swine receiving seven different feed types were collected in 2014 from three Chinese large-scale swine farms. The three large-scale swine farms were located in three Chinese provinces including Liaoning (LiaoYang/LY farms), Zhejiang (JiaXing/JX farms) and Hunan (HengYang/HY farms). For each group, four swine feed samples were mixed on site into one pooled swine feed sample, while four
A total of 296 primer sets targeting 285 ARGs, 8 transposase genes, 2 integron-integrase genes and the 16S rRNA gene were used (Table S2). All HT-qPCR reactions were performed on the Wafergen SmartChip Real-time PCR system as described previously (Ouyang et al., 2015; Su et al., 2015). For each run, a non-template negative control (NTC) was included. All qPCRs were conducted with technical triplicates. Individual PCR reactions having amplification efficiencies outside the 1.8–2.3 range or r2 below 0.99 were discarded. A threshold cycle (CT) of 31 was used to define the detection limit for individual PCR reactions. Based on
Please cite this article as: Zhao, Y., et al., Feed additives shift gut microbiota and enrich antibiotic resistance in swine gut, Sci Total Environ (2017), https://doi.org/10.1016/j.scitotenv.2017.10.106
Y. Zhao et al. / Science of the Total Environment xxx (2017) xxx–xxx
absolute quantification performed in this study, this means the genes will be detected using HT-qPCR chip when its copy number is estimated to be N6000–50,000 copies per gram of feces. Only genes detected in all 3 technical replicates and in all 4 fecal samples per treatment group were considered positive and used for further data analysis. Relative gene copy numbers were calculated with the formula: relative gene copy numbers = 10(31-CT)/(10/3) as described previously (Looft et al., 2012), where 31 refers to the threshold cycle (CT) chosen to define the detection limit. The normalized ARG copy numbers (ARG copy number per bacterial cell) were calculated as follows: normalized ARG copy numbers = (relative ARG copy number)/(4 × relative 16S rRNA gene copy numbers), where 4 is the average number of 16S rRNA-encoding genes per bacterium (Klappenbach et al., 2001). Absolute ARG copy numbers (ARG copy numbers per gram of feces) were determined by using standard curve method of quantification with the Roche LightCycler Real-time qPCR 480 system, as described previously (Ouyang et al., 2015). Relative abundance of specific ARG and OTU refers to the percentage of individual abundance among total ARG/OTU abundance. 2.6. Statistical analysis Averages and standard deviations were calculated using Excel 2013 (Microsoft Office Professional Plus 2013, Microsoft, USA). Linear discriminative analysis (LDA) Effect Size (LEfSe) (Segata et al., 2011) was performed using Galaxy as explanation tool for high-dimensional data and indicator microbe/gene discovery (Goecks et al., 2010). Principle component analysis (PCA), partial redundancy analysis (pRDA), Procrustes analysis, single environmental factor test, and Mantel test were conducted on RStudio (Version 0.99.486 – © 2009–2015 RStudio, Inc.) using Community Ecology Package “Vegan” (version 2.3–0) for ecological statistical analysis. Bar chart and scatter diagrams were generated by OriginPro 2015. Correlation analysis and significance tests were performed using IBM SPSS Statistics version 19. Venn diagram was created by Creatly software (Cinergix, Pty Ltd). Network analysis was plotted using Cytoscape 3.2.1. Relative abundance data for both ARGs and OTUs were used as input data for LDA, PCA, pRDA, Procrustes analysis, Mantel test and network analysis. Data of total number of ARGs detected and absolute abundance of 16S/ARGs were used for correspondent correlation analysis and significance tests. The flowchart was generated using ProcessOn software. 3. Results 3.1. Antibiotics and metals as additives in swine feed Antibiotics detected in swine feed samples (Table 1). Seven different antibiotics were detected among all the feed samples and the concentrations varied from 1.9 to 1159 μg per kg of dry weight. Olaquindox was found in most of the feed samples at a concentration up to 1159 μg/kg. Florfenicol was detected in two out of 7 feed samples at concentrations between 4.9 and 16.5 μg/kg. The remaining five detected antibiotics (sulfamethazine, sulfaisodimidine, ofloxacin, nicarbazin and enrofloxacin) were each detected in only one group at concentrations between 0.2 and 5.1 μg/kg (Table 1). Concentrations of copper and zinc averaged 90.2 μg Cu/g dry wt and 179.3 μg Zn/g dry wt, respectively, and ranged from 6.1 to 186.3 μg Cu/g dry wt and 65.6 to 285.6 μg Zn/ g dry wt (Table 2). Group HY3 and JX3 were found to be highly medicated with antibiotics at 1159 and 935 μg/kg, while the other groups were lowly medicated with total in-feed antibiotics in the range of 2–52 μg/kg. 3.2. Microbial community composition in swine gut The number of 16S rRNA gene copies ranged from 2.59 × 109 to 8.52 × 109 copies per gram of fresh wet fecal matter (Fig. S1). A total
3
Table 1 Antibiotics detected and their concentrations in swine feeds. Group Detected antibiotic HY1 Florfenicol HY2 Sulfamethazine Sulfaisodimidine Olaquindox
Concentration (μg/kg)
Total concentration (μg/kg)
4.9
4.9
0.2 0.2
1.9
1.5 HY3 Olaquindox
1159
1159
LY Olaquindox
32
32
JX1 Florfenicol
16.5
16.5
JX2 Ofloxacin Olaquindox Nicarbazin
0.5 46.5 5.1
52.1
JX3 Enrofloxacin Olaquindox
1.1 934
935.1
of 1,331,801 high quality filtered sequences were obtained from all 28 samples with a range of 8472 to 85,719 sequences per samples. These sequences were clustered into 18,171 OTUs at the 97% similarity level and then assigned to Bacteria (97.9%), Archaea (1.5%) and other unassigned (0.6%). Bacteroidetes and Firmicutes were the two dominating phyla in all seven groups with N75% of the sequences belonging to these taxa (Fig. S2). The dominant classes were Clostridia/Firmicutes, Bacilli/ Firmicutes and Bacteroidia/Bacteroidetes with relative abundances up to 44.5%, 31.4%, and 43.6%, respectively. OTU level rarefaction curves showed a similar species richness among groups (Fig. S3). Similar results were obtained when evaluating diversity using PD whole tree, Chao1 estimator, and the Shannon index (Fig. S3). Principal component analysis (PCA) displayed distinct clustering of replicate samples from each group demonstrating similar microbial community composition (Fig. S4). However, the microbial communities from different groups (HY3, LY, JX1, JX2 and JX3) also shared a certain degree of similarity, whereas HY1 and HY2 formed another loose cluster. To characterized the difference of gut microbiota between groups and discover indicator taxa, LDA Effect Size (LEfSe) algorithm module was used to couple statistical significance with biological consistency and effect size estimation. Potential indicator taxa are represented in Fig. 1 with their LDA scores in Fig. S5. These taxa displayed statistically significant differential relative abundance between groups. Every
Table 2 Metals detected in swine feed and their concentration.
HY1 HY2 HY3 LY JX1 JX2 JX3
Cu (μg/g)
Zn (μg/g)
Cd (μg/g)
Pb (μg/g)
Hg (μg/g)
Cr (μg/g)
As (μg/g)
6.1 108.2 186.3 24.12 25.1 129.4 151.9
65.6 131.4 207.2 178.9 123.7 262.4 285.6
0.063 0.11 0.083 0.05 0.05 0.04 0.045
0.78 0.32 1 0.43 0.71 0.84 0.76
0.015 0.007 0.007 0.03 0.009 0.04 0.011
0.71 0.33 0.62 1.05 0.39 1.61 1.06
0.242 0.295 0.197 0.254 0.457 0.336 0.294
Please cite this article as: Zhao, Y., et al., Feed additives shift gut microbiota and enrich antibiotic resistance in swine gut, Sci Total Environ (2017), https://doi.org/10.1016/j.scitotenv.2017.10.106
4
Y. Zhao et al. / Science of the Total Environment xxx (2017) xxx–xxx
Fig. 1. Indicator taxa with statistically significant differential abundance between treatment groups as revealed by linear discriminative analysis (LDA) Effect Size (LEfSe) analysis of swine gut microbiota 16S rRNA gene sequence data. Each node refers to a single taxon. Each color refers to each group and its corresponding indicator bacteria. Each circle's diameter is proportional to the taxa levels (from phylum to family). This representation employs the Ribosomal Database Project (RDP) taxonomy. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
group had at least one potential indicator except for HY2 (Fig. 1). All potential indicator taxa of HY1, HY3, and JX3 belonged to the phylum Firmicutes. Other indicators included taxa from Spirochaetes, Bacteroidetes. Partial redundancy analysis was used to characterize the effect of feed additives on the microbial community. Results showed variance in the microbial community could be mainly explained by in-feed antibiotics and metals (53.99%), while MGEs explaining 11.50%. 3.3. ARGs in gut microbiota A total of 156 genes including 146 antibiotic resistance genes (ARGs), 8 transposon-transposase genes and 2 class 1 integron– integrase genes were detected among all samples. The number of ARGs detected in each group varied from 48 to 126 (Fig. 2). The detected resistance genes represented all major resistance mechanisms: antibiotic deactivation (44.5%), efflux pumps (34.3%) and cellular protection (19.2%) (Fig. 2) and conferred resistance to major classes of antibiotics: multidrug (22%), aminoglycoside (20%), beta-lactam (14%), tetracycline (14%), MLSB (14%), vancomycin (8%), sulfonamide (2%), chloramphenicol (2%) and others (4%) (Fig. S6). Swine gut microbiota from HY3 harbored a significantly higher number of ARGs than the other groups (P b 0.05), where multidrug and aminoglycoside resistance genes accounted for 23% and 19%, respectively, followed by MLSB (16%), tetracycline (14%) and beta-lactam resistance genes (14%). ARGs detected in these fresh pig manures were highly abundant, ranging from 2.72 × 109 to 1.34 × 1010 copies per gram (Fig. 3A). The normalized copy numbers (copies per bacterial cell) varied from 4.51
to 13.53, with an average of 6.85 ARGs per bacterial cell (Fig. 3B). The highest absolute abundance of ARG (copy number per g of fecal matter) was found in the highly medicated HY3 group, contributed mostly by tetracycline (4.77 × 109), MLSB (3.41 × 109), and aminoglycoside resistance genes (4.02 × 109). To compare the similarity and dissimilarity of ARGs from different groups, the relative abundance of specific ARGs as a percentage of total ARG abundance across all samples was examined. The overall patterns of ARGs were revealed by PCA. The seven treatment groups could not be distinguished using PCA indicating that they shared similar ARG profiles (Fig. S7). Among 285 targeted ARGs, 32 shared ARGs were found in all samples (Table S3). Unique unshared ARGs were also found with the highest number of unique ARGs for one group being 26 for HY3 (Fig. 4), contributed by multidrug and beta-lactam resistance genes (RGs). LEfSe characterized the indicator ARGs in each group displaying significantly different abundance compared to all other groups (Fig. S8). All indicator ARGs found in HY1 (ermT, ermX, lnuB, vatE) were genes conferring resistance to the MLSB class of antibiotics. A total of 4 indicator ARGs including 3 tetracycline resistance genes (tet(32), tetL, tetO) and 1 aminoglycoside resistance gene (aadD) were observed for HY2. HY3 harbored 8 indicator ARGs conferring resistance to aminoglycosides (aacA/aphD, aphA1(aka kanR)), multiple drugs (mexF, oprD), tetracyclines (tetPB, tetQ), beta-lactams (fox5) and streptomycin ARG (sat4) with efflux pump as resistance mechanism. Other indicator ARGs included aminoglycoside RGs - aadA2 for JX1; aac(6′)-Ib(aka aacA4), aadA2, aadA, aph(2′)-Id, cmxA, and qacEdelta1–01 for JX2; aadA1,tetG,and tetX for JX3.
Please cite this article as: Zhao, Y., et al., Feed additives shift gut microbiota and enrich antibiotic resistance in swine gut, Sci Total Environ (2017), https://doi.org/10.1016/j.scitotenv.2017.10.106
Y. Zhao et al. / Science of the Total Environment xxx (2017) xxx–xxx
5
Fig. 2. Antibiotic resistance genes (ARGs) detected in the seven different treatment groups of swine gut microbiota. ARGs were classified based on (A) antibiotics to which they confer resistance and (B) mechanism of resistance. MLSB as abbreviation for macrolide-lincosamide-streptogramin B.
3.4. Relationship between feed additives, microbial community composition, and ARGs The total concentration of antibiotics in feed was significantly correlated to the absolute abundance of total ARGs in manure (Pearson's coefficient = 0.868, P b 0.05), and the number of different ARGs detected was significantly correlated to total concentration of copper in feed (Pearson's coefficient = 0.871, P b 0.05) (Table S4). Meanwhile, microbial abundance (16S rRNA absolute gene copy numbers) was strongly correlated with total ARG abundance (Pearson's coefficient = 0.65, P b 0.01), with ARGs conferring resistance to each major class of antibiotics except sulfonamides (Table S5). Moreover, the absolute abundance of total MGEs, integrons, and transposons were significantly correlated with total ARG abundance and ARGs conferring resistance to different classes of antibiotics (Table S5). The abilities of antibiotics and metals in swine feed and phylumlevel microbial community composition to explain the relative abundance of ARGs were also evaluated by partial redundancy analysis (pRDA). A total of 31.84% of the variance in the ARG data set could be explained by antibiotic and metal concentrations in swine feed, with
microbial community composition and MGEs explaining 23.3% and 2.4%, respectively (Fig. 5). Interactions between swine feed and microbial community accounted for 16.5% of the variation. Using Monte Carlo permutation test to verify, sorting accepted pRDA results with PseudoF value 6.30 (Significance b0.001, Permutation: free, Number of permutations = 999). Single environmental factor test was conducted to reveal the correlation between each single explanatory variable and ARG variance. Single factors that positively correlated with change of ARGs included Bacteroidetes, Firmicutes, copper, and arsenic (Significance b0.01, Permutation: free, Number of permutations = 999). Procrustes analysis exhibited a good-fit correlation between ARG profiles and microbial community composition on the basis of BrayCurtis dissimilarity metrics (Procrustes sum of squares M2 = 0.6698, r = 0.5746, Significance: 4e-04, Permutation: free, Number of permutations: 9999) (Fig. 6). To further verify, Mantel test was performed and results confirmed a significant correlation between ARG profiles and microbial community composition (r = 0.4436, P b 0.001, Permutation: free, Number of permutations: 999). Hence, the antibiotic resistome did not decouple from the microbiome. Network analysis was performed to reveal the co-occurrence pattern between ARGs and microbial OTUs.
Please cite this article as: Zhao, Y., et al., Feed additives shift gut microbiota and enrich antibiotic resistance in swine gut, Sci Total Environ (2017), https://doi.org/10.1016/j.scitotenv.2017.10.106
6
Y. Zhao et al. / Science of the Total Environment xxx (2017) xxx–xxx
Fig. 3. Abundance of antibiotic resistance genes (ARGs) in the seven different treatment groups of swine gut microbiota. (A) Absolute copy numbers of ARGs conferring resistance to specific classes of antibiotics per gram of fecal sample, with error bars showing calculated standard deviation (s.d.) of four sampling replicates on each group (n = 4). (B) the normalized copy numbers of ARGs presented as total ARG copies per bacterial cell with error bar showing calculated s.d. (n = 4).
Specific ARGs conferring resistance to aminoglycosides, beta-lactams, MLSB, and tetracyclines could be linked to Bacteriodetes, Firmciutes, and Spirochaetes (Fig. 7). 4. Discussion Although time-resolved dynamics of the swine gut microbiome and antibiotic resistome were not obtained, the results of the this study demonstrate that swine gut microbiota from Chinese pig farms harbored diverse and abundant ARGs in both high-medicated (HY3, JX3) and low-medicated groups (other five groups). The presented results also support a potential role of Cu in co-selection of ARGs. Analyses demonstrated that increased levels of in-feed additives affected the composition of the gut microbiota and aggravated the enrichment of ARGs. Feed additives and microbial community were the dominant determinants of shaping the swine gut antibiotic resistome. Highly diverse and abundant ARGs were found in pig fecal samples of both high-medicated and low-medicated group (Figs. 2 & 3). The 32 ARGs shared by all 28 investigated microbiomes belong to the most
abundant type of ARGs conferring resistance to aminoglycosides, MLSB and tetracyclines. Consistent with the findings from the present study, the genes aadA and tet family genes were previously suggested as the most common and abundant ARGs in manure (Binh et al., 2009; Munir and Xagoraraki, 2011). About 100 additional ARGs were detected in the present study compared to the number of ARGs detected in a swine manure control sample from pigs that received no known antibiotics from a previous study employing the same HT-qPCR based ARG chip (Zhu et al., 2013). These findings indicate that even small dosages of antibiotics can induce higher levels of ARGs in swine gut. Recent studies have demonstrated that even subinhibitory levels of antibiotic can induce different expression of genes causing enhancement of toxin and biofilm production in potential human pathogens (Goh et al., 2002; Goneau et al., 2015; Hoffman et al., 2005; Wong et al., 2000). In addition to increasing the level of ARG, a small dosage of antibiotics in feed additives for prophylaxis could thus also pose a risk by promoting the virulence and biofilm formation during bacterial infection. In-feed copper was significantly correlated to the diversity of ARGs in gut microbiota (P b 0.05) and together with arsenic were shown to
Please cite this article as: Zhao, Y., et al., Feed additives shift gut microbiota and enrich antibiotic resistance in swine gut, Sci Total Environ (2017), https://doi.org/10.1016/j.scitotenv.2017.10.106
Y. Zhao et al. / Science of the Total Environment xxx (2017) xxx–xxx
7
Fig. 6. Procrustes analysis showing the correlation between antibiotic resistance gene (ARG) and microbial community structure (ARG/operational taxonomic unit distribution patterns) on the basis of Bray-Curtis dissimilarity metrics (M2 = 0.6698, r = 0.5746, P b 0.0001, 9999 free permutations). Red line connecting the two data sets of groups. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) Fig. 4. Number of unique unshared antibiotic resistance genes (ARGs) in swine gut microbiota originating from seven different treatment groups. MLSB as abbreviation for macrolide-lincosamide-streptogramin B.
be the two dominant single factors shifting the ARG profiles (P b 0.01), suggesting roles of Cu and As as co-selecting agents. A number of studies have implicated that metals play an important role in co-selecting ARGs (Baker-Austin et al., 2006; Hu et al., 2016; Zhu et al., 2013). Toxic metals, such as copper and zinc, may even exert a stronger selection pressure
Fig. 5. Partial redundancy analysis differentiating the effect of antibiotic and metal concentrations in swine feed, swine gut microbial community composition and mobile gene elements on the overall ARG profiles.
for antibiotic resistance than the specific antibiotic itself in certain cases (Song et al., 2017). It should be noted that much higher concentrations of metals than observed in this study have been observed in pig fecal matter obtained from other pig farms (Jensen et al., 2016). A key finding of this study is that the ARGs present in swine gut depended on the composition of the microbial community, indicating that swine gut microbiome could be an important conduit for transferring ARGs into the environment. A similar conclusion was drawn by a recent study of differentially medicated pigs in USA (Looft et al., 2012). The strong association between the resistome and microbiome has also been found in urban sewage (Su et al., 2017), human gut (Penders et al., 2013; Su et al., 2017) and soil (Forsberg et al., 2014). And due to the anthropogenic activities in last decades, the microbes are moving globally (Zhu et al., 2017a) and ARGs spread into our environments with unprecedented diversity and abundance (Zhu et al., 2017b). Therefore, importance of monitoring and control of microbes and ARGs in reservoirs like animal gut should not be underestimated. To the best of our knowledge, our study is the first to use LDA as a tool to discover potential indicator ARGs and indicator microbial taxa in fecal samples. LDA and network analysis gave consistent results. pRDA results showed that in-feed antibiotics and metals, and microbial community in total explained 71.6% of the variance in ARG pattern. Moreover, 54.0% of changes in gut microbiota were contributed by antibiotics and metals in feed. Collectively, these results imply that swine gut antibiotic resistome was largely shaped by the enrichment of certain microbial taxa rather than by horizontal gene transfer of ARGs. However, this does not rule out the importance of HGT which considered a vital pathway for dissemination of ARGs and causing emergence of multidrug resistant bacteria in different environments (Dzidic and Bedeković, 2003; Martínez, 2008). This is consistent with the observed positive correlation between MGEs and multidrug resistance genes found in this study (Pearson's r = 0.94, P b 0.001). Multidrugresistant bacteria (MDB) as the most threatening type of antibioticresistant bacteria (ARB) to public health includes superbug methicillin-resistant Staphylococcus aureus (MRSA), which is responsible for several difficult-to-treat infections in humans (Sharma et al., 2016). Moreover, diverse mobile ARGs induced in swine gut, enriched and transferred into the environment by manure fertilization and farm wastewater, could pose a risk to human health by potential HGT of ARGs to human-associated pathogens (Ashbolt et al., 2013; Baquero et al., 2008). Hence, management options for mitigating the dissemination of such manure-borne ARGs are critical (Pruden et al., 2013).
Please cite this article as: Zhao, Y., et al., Feed additives shift gut microbiota and enrich antibiotic resistance in swine gut, Sci Total Environ (2017), https://doi.org/10.1016/j.scitotenv.2017.10.106
8
Y. Zhao et al. / Science of the Total Environment xxx (2017) xxx–xxx
Fig. 7. Network analysis revealing the co-occurrence pattern between antibiotic resistance genes (ARGs) and microbial taxa at phylogenetic resolution from phylum to genus (16S rRNA gene sequence data). The nodes were colored according to the ARG classes and different phyla, and the size of the node is proportional to the degree of the connection (number of correlation). The edge (connection) shows significant correlation (P b 0.01), and edge line width represents corresponding Spearman's correlation coefficient.
Composting and lagoons have been shown quite effective treatments of ARGs in manure (Storteboom et al., 2007). However, the persistence of some ARGs has been reported after treatment (Sharma et al., 2009). Therefore, better manure treatment technologies are needed to prevent dissemination of ARGs to agricultural soils and harvested crops. However, the ultimate solution may be to replace antibiotics by other feed additives such as phytogenic products, probiotics, and prebiotics (Gaggia et al., 2010; Windisch et al., 2008).
21 specific ARGs and their potential hosts. Hence, the swine gut microbial community may serve as an important reservoir of ARGs of relevance to public health. Conflicts of interests The authors declare no conflicts of interests. Acknowledgements
5. Conclusions The presented results clearly indicate that antibiotics and metals used as feed additives in large-scale Chinese farming operation shifted the swine gut microbiome and significantly enhanced the diversity and abundance of ARGs in swine gut. The taxonomic composition of the swine gut community also played an important role in shaping the ARG profile, and network analysis revealed co-occurrence patterns of
The authors thank Jie Xu for help with sample collection, and Xiuli Hao for advice on figure plotting. This work was supported financially by National Natural Science Foundation of China (41571130063, 21210008); and REMEDIATE (Improved decision-making in contaminated land site investigation and risk assessment) Marie-Curie Innovation Training Network from the European Union's Horizon 2020 Programme (grant agreement n. 643087).
Please cite this article as: Zhao, Y., et al., Feed additives shift gut microbiota and enrich antibiotic resistance in swine gut, Sci Total Environ (2017), https://doi.org/10.1016/j.scitotenv.2017.10.106
Y. Zhao et al. / Science of the Total Environment xxx (2017) xxx–xxx
Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.scitotenv.2017.10.106. References Ashbolt, N.J., Amezquita, A., Backhaus, T., Borriello, P., Brandt, K.K., Collignon, P., et al., 2013. Human Health Risk Assessment (HHRA) for environmental development and transfer of antibiotic resistance. Environ. Health Perspect. 121, 993–1001. Baker-Austin, C., Wright, M.S., Stepanauskas, R., McArthur, J.V., 2006. Co-selection of antibiotic and metal resistance. Trends Microbiol. 14, 176–182. Baquero, F., Martínez, J.-L., Cantón, R., 2008. Antibiotics and antibiotic resistance in water environments. Curr. Opin. Biotechnol. 19, 260–265. 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, 427–433. Cang, L., 2004. Heavy metals pollution in poultry and livestock feeds and manures under intensive farming in Jiangsu Province, China. J. Environ. Sci. 16, 371–374. Cheng, W., Chen, H., Su, C., Yan, S., 2013. Abundance and persistence of antibiotic resistance genes in livestock farms: a comprehensive investigation in eastern China. Environ. Int. 61, 1–7. Claesson, M.J., Wang, Q., O'Sullivan, O., Greene-Diniz, R., Cole, J.R., Ross, R.P., et al., 2010. Comparison of two next-generation sequencing technologies for resolving highly complex microbiota composition using tandem variable 16S rRNA gene regions. Nucleic Acids Res. e200, 38. Collignon, P., Voss, A., 2015. China, what antibiotics and what volumes are used in food production animals? Antimicrob. Resist. Infect. Control 4, 16. Dzidic, S., Bedeković, V., 2003. Horizontal gene transfer-emerging multidrug resistance in hospital bacteria. Acta Pharmacol. Sin. 24, 519–526. Edgar, R.C., 2010. Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26, 2460–2461. Forsberg, K.J., Patel, S., Gibson, M.K., Lauber, C.L., Knight, R., Fierer, N., et al., 2014. Bacterial phylogeny structures soil resistomes across habitats. Nature 509, 612–616. Gaggia, F., Mattarelli, P., Biavati, B., 2010. Probiotics and prebiotics in animal feeding for safe food production. Int. J. Food Microbiol. 141 (Suppl. 1), S15–28. Gibson, M.K., Crofts, T.S., Dantas, G., 2015. Antibiotics and the developing infant gut microbiota and resistome. Curr. Opin. Microbiol. 27, 51–56. Goecks, J., Nekrutenko, A., Taylor, J., 2010. Galaxy: a comprehensive approach for supporting accessible, reproducible, and transparent computational research in the life sciences. Genome Biol. 11, 1. Goh, E.-B., Yim, G., Tsui, W., McClure, J., Surette, M.G., Davies, J., 2002. Transcriptional modulation of bacterial gene expression by subinhibitory concentrations of antibiotics. Proc. Natl. Acad. Sci. 99, 17025–17030. Goneau, L.W., Hannan, T.J., MacPhee, R.A., Schwartz, D.J., Macklaim, J.M., Gloor, G.B., et al., 2015. Subinhibitory antibiotic therapy alters recurrent urinary tract infection pathogenesis through modulation of bacterial virulence and host immunity. MBio 6, e00356–15. Goossens, H., Ferech, M., Vander Stichele, R., Elseviers, M., Group EP, 2005. Outpatient antibiotic use in Europe and association with resistance: a cross-national database study. Lancet 365, 579–587. Hamscher, G., Pawelzick, H.T., Sczesny, S., Nau, H., Hartung, J., 2003. Antibiotics in dust originating from a pig-fattening farm: a new source of health hazard for farmers? Environ. Health Perspect. 111, 1590–1594. Heuer, H., Schmitt, H., Smalla, K., 2011. Antibiotic resistance gene spread due to manure application on agricultural fields. Curr. Opin. Microbiol. 14, 236–243. Hoelzer, K., Wong, N., Thomas, J., Talkington, K., Jungman, E., Coukell, A., 2017. Antimicrobial drug use in food-producing animals and associated human health risks: what, and how strong, is the evidence? BMC Vet. Res. 13, 211. Hoffman, L.R., D'Argenio, D.A., MacCoss, M.J., Zhang, Z., Jones, R.A., Miller, S.I., 2005. Aminoglycoside antibiotics induce bacterial biofilm formation. Nature 436, 1171–1175. Hu, Y., Yang, X., Qin, J., Lu, N., Cheng, G., Wu, N., et al., 2013. Metagenome-wide analysis of antibiotic resistance genes in a large cohort of human gut microbiota. Nat. Commun. 4, 2151. Hu, H.W., Wang, J.T., Li, J., Li, J.J., Ma, Y.B., Chen, D., et al., 2016. Field-based evidence for copper contamination induced changes of antibiotic resistance in agricultural soils. Environ. Microbiol. 18, 3896–3909. Huang, Y., Cheng, M., Li, W., Wu, L., Chen, Y., Luo, Y., et al., 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.
9
Jensen, J., Larsen, M.M., Bak, J., 2016. National monitoring study in Denmark finds increased and critical levels of copper and zinc in arable soils fertilized with pig slurry. Environ. Pollut. 214, 334–340. 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, 181–184. Looft, T., Johnson, T.A., Allen, H.K., Bayles, D.O., Alt, D.P., Stedtfeld, R.D., et al., 2012. In-feed antibiotic effects on the swine intestinal microbiome. Proc. Natl. Acad. Sci. U. S. A. 109, 1691–1696. Martínez, J.L., 2008. Antibiotics and antibiotic resistance genes in natural environments. Science 321, 365–367. Munir, M., Xagoraraki, I., 2011. Levels of antibiotic resistance genes in manure, biosolids, and fertilized soil. J. Environ. Qual. 40, 248. 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, 5697–5707. Penders, J., Stobberingh, E.E., Savelkoul, P.H., Wolffs, P.F., 2013. The human microbiome as a reservoir of antimicrobial resistance. Front. Microbiol. 4, 87. Perreten, V., Schwarz, F., Cresta, L., Boeglin, M., Dasen, G., Teuber, M., 1997. Antibiotic resistance spread in food. Nature 389, 801–802. Pruden, A., Pei, R., Storteboom, H., Carlson, K.H., 2006. Antibiotic resistance genes as emerging contaminants: studies in northern Colorado. Environ. Sci. Technol. 40, 7445–7450. Pruden, A., Larsson, D.G., Amezquita, A., Collignon, P., Brandt, K.K., Graham, D.W., et al., 2013. Management options for reducing the release of antibiotics and antibiotic resistance genes to the environment. Environ. Health Perspect. 121, 878–885. Segata, N., Izard, J., Waldron, L., Gevers, D., Miropolsky, L., Garrett, W.S., et al., 2011. Metagenomic biomarker discovery and explanation. Genome Biol. 12, R60. Sharma, R., Larney, F.J., Chen, J., Yanke, L.J., Morrison, M., Topp, E., et al., 2009. Selected antimicrobial resistance during composting of manure from cattle administered subtherapeutic antimicrobials. J. Environ. Qual. 38, 567–575. Sharma, M., Nunez-Garcia, J., Kearns, A.M., Doumith, M., Butaye, P.R., Argudin, M.A., et al., 2016. Livestock-associated methicillin resistant Staphylococcus aureus (LA-MRSA) clonal complex (CC) 398 Isolated from UK animals belong to European lineages. Front. Microbiol. 7, 1741. Song, J., Rensing, C., Holm, P.E., Virta, M., Brandt, K.K., 2017. Comparison of metals and tetracycline as selective agents for development of tetracycline resistant bacterial communities in agricultural soil. Environ. Sci. Technol. 51, 3040–3047. Storteboom, H.N., Kim, S.C., Doesken, K.C., Carlson, K.H., Davis, J.G., Pruden, A., 2007. Response of antibiotics and resistance genes to high-intensity and low-intensity manure management. J. Environ. Qual. 36, 1695–1703. Su, J.Q., Wei, B., Ou-Yang, W.Y., Huang, F.Y., Zhao, Y., HJ, Xu, et al., 2015. Antibiotic resistome and its association with bacterial communities during sewage sludge composting. Environ. Sci. Technol. 49, 7356–7363. Su, J.Q., An, X.L., Li, B., Chen, Q.L., Gillings, M.R., Chen, H., et al., 2017. Metagenomics of urban sewage identifies an extensively shared antibiotic resistome in China. Microbiome 5, 84. Van Boeckel, T.P., Gandra, S., Ashok, A., Caudron, Q., Grenfell, B.T., Levin, S.A., et al., 2014. Global antibiotic consumption 2000 to 2010: an analysis of national pharmaceutical sales data. Lancet Infect. Dis. 14, 742–750. Windisch, W., Schedle, K., Plitzner, C., Kroismayr, A., 2008. Use of phytogenic products as feed additives for swine and poultry. J. Anim. Sci. 86, E140–8. Wong, C.S., Jelacic, S., Habeeb, R.L., Watkins, S.L., Tarr, P.I., 2000. The risk of the hemolytic– uremic syndrome after antibiotic treatment of Escherichia coli O157: H7 infections. N. Engl. J. Med. 342, 1930–1936. Xu, H.-J., Wang, X.-H., Li, H., Yao, H.-Y., J-Q, Su, Zhu, Y.-G., 2014. Biochar impacts soil microbial community composition and nitrogen cycling in an acidic soil planted with rape. Environ. Sci. Technol. 48, 9391–9399. Yazdankhah, S., Rudi, K., Bernhoft, A., 2014. Zinc and copper in animal feed - development of resistance and co-resistance to antimicrobial agents in bacteria of animal origin. Microb. Ecol. Health Dis. 25. Yun-peng, W., Yue, M., 2008. Potential public hazard of using antibiotics in livestock industry. Chin. J. Antibiot. 9, 001. Zhu, Y.G., Johnson, T.A., Su, J.Q., Qiao, M., Guo, G.X., Stedtfeld, R.D., et al., 2013. Diverse and abundant antibiotic resistance genes in Chinese swine farms. Proc. Natl. Acad. Sci. U. S. A. 110, 3435–3440. Zhu, Y.-G., Gillings, M., Simonet, P., Stekel, D., Banwart, S., Penuelas, J., 2017a. Microbial mass movements. Science 357, 1099–1100. Zhu, Y.-G., Zhao, Y., Li, B., Huang, C.-L., Zhang, S.-Y., Yu, S., et al., 2017b. Continental-scale pollution of estuaries with antibiotic resistance genes. Nat. Microbiol. 2, 16270.
Please cite this article as: Zhao, Y., et al., Feed additives shift gut microbiota and enrich antibiotic resistance in swine gut, Sci Total Environ (2017), https://doi.org/10.1016/j.scitotenv.2017.10.106
Contents lists available at ScienceDirect
Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv
Feed additives shift gut microbiota and enrich antibiotic resistance in swine gut Yi Zhao a,b, Jian-Qiang Su a, Xin-Li An a, Fu-Yi Huang a, Christopher Rensing a,c, Kristian Koefoed Brandt b, Yong-Guan Zhu a,d,⁎ a
Key Laboratory of Urban Environment and Health, Institute of Urban Environment, Chinese Academy of Sciences, 1799 Jimei Road, Xiamen 361021, China Department of Plant and Environmental Sciences, Faculty of Science, University of Copenhagen, Thorvaldsensvej 40, 1871 Frederiksberg, Denmark Institute of Environmental Microbiology, College of Resources and Environment, Fujian Agriculture and Forestry University, Fuzhou 350002, Fujian, China d State Key Laboratory of Urban and Regional Ecology, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing 100085, China b c
H I G H L I G H T S
G R A P H I C A L
A B S T R A C T
• Feed additives and the gut microbiome were investigated at 3 Chinese pig farms. • Multiple metals and 79 antibiotics were analyzed as potential feed additives. • 285 antibiotic resistance genes (ARGs) were quantified by high throughput qPCR. • ARGs were highly diverse and abundant in swine feces from Chinese pig farms. • Feed additives and gut microbiota were major determinants shaping ARG profiles.
a r t i c l e
i n f o
Article history: Received 25 August 2017 Received in revised form 11 October 2017 Accepted 12 October 2017 Available online xxxx Editor: Jay Gan Keywords: Antibiotic resistance High-throughput qPCR 16S rRNA sequencing Swine gut microbiota ARG chip Network analysis
a b s t r a c t Antibiotic resistance genes (ARGs) are emerging environmental contaminants posing a threat to public health. Antibiotics and metals are widely used as feed additives and could consequently affect ARGs in swine gut. In this study, high-throughput quantitative polymerase chain reaction (HT-qPCR) based ARG chip and nextgeneration 16S rRNA gene amplicon sequencing data were analyzed using multiple statistical approaches to profile the antibiotic resistome and investigate its linkages to antibiotics and metals used as feed additives and to the microbial community composition in freshly collected swine manure samples from three large-scale Chinese pig farms. A total of 146 ARGs and up to 1.3 × 1010 total ARG copies per gram of swine feces were detected. ARGs conferring resistance to aminoglycoside, macrolide-lincosamide-streptogramin B (MLSB) and tetracycline were dominant in pig gut. Total abundance of ARGs was positively correlated with in-feed antibiotics, microbial biomass and abundance of mobile genetic elements (MGEs) (P b 0.05). A significant correlation between microbial communities and ARG profiles was observed by Procrustes analysis. Network analysis revealed that Bacteroidetes and Firmicutes were the most dominant phyla co-occurring with specific ARGs. Partial redundancy analysis indicated that the variance in ARG profiles could be primarily attributed to antibiotics and metals in feed (31.8%), gut microbial community composition (23.3%) and interaction between feed additives and community composition (16.5%). These results suggest that increased levels of in-feed additives could aggravate the enrichment of ARGs and MGEs in swine gut. © 2017 Elsevier B.V. All rights reserved.
⁎ Corresponding author at: Key Laboratory of Urban Environment and Health, Institute of Urban Environment, Chinese Academy of Sciences, No. 1799 Jimei Road, Xiamen 361021, China. E-mail addresses: [email protected], [email protected] (Y.-G. Zhu).
https://doi.org/10.1016/j.scitotenv.2017.10.106 0048-9697/© 2017 Elsevier B.V. All rights reserved.
Please cite this article as: Zhao, Y., et al., Feed additives shift gut microbiota and enrich antibiotic resistance in swine gut, Sci Total Environ (2017), https://doi.org/10.1016/j.scitotenv.2017.10.106
2
Y. Zhao et al. / Science of the Total Environment xxx (2017) xxx–xxx
1. Introduction The frequent occurrence and spread of multidrug-resistant pathogenic bacteria is a worldwide challenge and can be largely attributed to the use of antibiotics in animals and humans (Goossens et al., 2005). Antibiotic resistance genes (ARGs) are increasingly regarded as emerging contaminants (Pruden et al., 2006) providing a threat to human and environment health (Ashbolt et al., 2013; Zhu et al., 2017b). China is one of the largest consumers of antibiotics in the world (Van Boeckel et al., 2014) with a significant fraction of all antibiotics being used for animal disease treatment and growth promotion (Yun-peng and Yue, 2008). Antibiotics were first used as feed additives in China in the mid-1970s, but consumption has increased dramatically concomitant with the economic growth in recent decades. A previous report revealed that almost half of the annual antibiotic production in China, totaling about 97,000 tons, went into the animal industry (Collignon and Voss, 2015). Furthermore, metals such as copper and zinc are also approved as feed additives under the stipulation in China for growth promotion and disease control purposes (Cang, 2004). It has been observed that these metals in some cases may cause coselection of ARGs in gut microbiota, but further studies are needed to fully elucidate the links between metals in feed and ARGs in gut microbiota (Yazdankhah et al., 2014). By contrast, it is well established that use of antibiotics in animals may contribute to the antibiotic resistance crisis in humans (Ashbolt et al., 2013; Hoelzer et al., 2017). Hence, the swine gut microbiota is now considered an important reservoir of pathogens with the potential to transfer antibiotic resistant pathogens to humans (Gibson et al., 2015; Heuer et al., 2011). Indeed, the overuse of antibiotics in Chinese animals may possibly explain why Chinese people harbor higher numbers of ARGs in their guts than Europeans (Hu et al., 2013). Apart from direct human exposure to ARGs present in swine gut via contaminated meat (Perreten et al., 1997), ARGs in swine gut may also spread to humans via pig-house dust (Hamscher et al., 2003) or via manure application and wastewater discharge. Although antibiotics and ARGs originating from Chinese swine farms have been documented and characterized (Cheng et al., 2013; Zhu et al., 2013), little is known about the relative importance of different abiotic and biotic factors in shaping the swine gut antibiotic resistome. Hence, the aim of this study was to perform high-throughput profiling of the ARGs in fecal samples from Chinese pig farms and to investigate the effect of antibiotic and metals in swine feed on the antibiotic resistome and its association with the microbial community. To this end, the prevalence and diversity of ARGs and selected mobile genetic elements known to carry ARGs were characterized in fresh fecal matter samples obtained from three large-scale Chinese pig farms using highthroughput quantitative PCR chip technology. ARG profiles were subsequently linked to microbial community composition as analyzed by 16S rRNA amplicon sequencing and to concentrations of antibiotics and metals in swine feed using partial redundancy analysis, Procrustes analysis, and network analysis. Linear discriminative analysis (LDA) Effect Size (LEfSe) (Segata et al., 2011) was further used for highdimensional ARG data and indicator ARGs and microbe discovery (Goecks et al., 2010).
biological replicates of fresh fecal samples were taken immediately after excretion. The four replicate fecal samples were derived from four individual adult pig siblings for biological consistency. The four biological replicates of sampled pigs in each group had all received the same swine feed. All collected seven swine feed and 28 fresh fecal matter samples were sealed in individual 300-mL sterile plastic sampling boxes, immediately transported to the laboratory on dry ice, and stored at −20 °C for further analysis. 2.2. Chemical analysis of antibiotics and metals in swine feed A total of 79 different antibiotics were analyzed in swine feed based on a previously published protocol (Huang et al., 2013). Analyzed included sulfonamides, tetracyclines, chloramphenicols, macrolidelincosamide-streptogramin B (MLSB) and other veterinary antibiotics. All targeted antibiotics and their method detection limits (MDL) are listed in Table S1. Metals including copper, zinc, cadmium, lead, mercury, chromium and arsenic in swine feed were analyzed using inductively coupled plasma mass spectrometry (ICP-MS) according to ISO/TS 16965:2013(E). 2.3. DNA extraction High-quality DNA was extracted from fecal samples (0.5 g) using the MoBio PowerFecal® DNA Isolation Kit according to the instruction manual. The quality and concentration of DNA were checked by agarose gel electrophoresis and ultraviolet-absorbance (ND1000, Nanodrop, Thermo Fisher Scientific Inc.). DNA was diluted to 50 ng per μL using sterile PCR-grade water and stored at −20 °C until further analysis. 2.4. High-throughput 16S rRNA gene sequencing The V4-V5 region of 16S rRNA gene (Claesson et al., 2010) was amplified, purified, quantified, pooled and sequenced on Illumina Miseq platform at Novogene, Beijing, China (Xu et al., 2014). Each 50 μl PCR reaction contained 25 μl TaKaRa ExTaq, 0.5 μl bovine serum albumin, 1 μl of each primer, 1 μl DNA template and 21.5 μl PCR-grade water. The thermal cycles consisted of initial enzyme activation at 94 °C for 3 mins, followed by 30 cycles of denaturation at 94 °C for 30s, annealing at 58 °C for 1 min and extenstion at 72 °C for 1 min, and a final extension at 72 °C for 5 mins. Demultiplexing was performed and low-quality or ambiguous read were removed to downstream clean sequence generation (Novogene). Quantitative Insights Into Microbial Ecology (QIIME) workflow was used to analyze 16S rRNA gene sequence data. The de novo open-reference operational taxonomic unit (OTU) picking process was conducted by QIIME following the online tutorial and script documentation with 93% as similarity level by UCLUST clustering (Edgar, 2010). Ribosomal Database Project (RDP) classifier was used to assign taxonomic data to each representative sequence by default with a confidence threshold of 80%. Singletons were removed prior to downstream analysis. The within community diversity (alpha diversity) and rarefaction curves were computed using observed OTUs, phylogenetic diversity metrics (PD whole tree) and Shannon index. 2.5. HT-qPCR and data processing
2. Materials and methods 2.1. Sampling A total of 28 fecal samples representing seven groups (populations) with four replicates of swine receiving seven different feed types were collected in 2014 from three Chinese large-scale swine farms. The three large-scale swine farms were located in three Chinese provinces including Liaoning (LiaoYang/LY farms), Zhejiang (JiaXing/JX farms) and Hunan (HengYang/HY farms). For each group, four swine feed samples were mixed on site into one pooled swine feed sample, while four
A total of 296 primer sets targeting 285 ARGs, 8 transposase genes, 2 integron-integrase genes and the 16S rRNA gene were used (Table S2). All HT-qPCR reactions were performed on the Wafergen SmartChip Real-time PCR system as described previously (Ouyang et al., 2015; Su et al., 2015). For each run, a non-template negative control (NTC) was included. All qPCRs were conducted with technical triplicates. Individual PCR reactions having amplification efficiencies outside the 1.8–2.3 range or r2 below 0.99 were discarded. A threshold cycle (CT) of 31 was used to define the detection limit for individual PCR reactions. Based on
Please cite this article as: Zhao, Y., et al., Feed additives shift gut microbiota and enrich antibiotic resistance in swine gut, Sci Total Environ (2017), https://doi.org/10.1016/j.scitotenv.2017.10.106
Y. Zhao et al. / Science of the Total Environment xxx (2017) xxx–xxx
absolute quantification performed in this study, this means the genes will be detected using HT-qPCR chip when its copy number is estimated to be N6000–50,000 copies per gram of feces. Only genes detected in all 3 technical replicates and in all 4 fecal samples per treatment group were considered positive and used for further data analysis. Relative gene copy numbers were calculated with the formula: relative gene copy numbers = 10(31-CT)/(10/3) as described previously (Looft et al., 2012), where 31 refers to the threshold cycle (CT) chosen to define the detection limit. The normalized ARG copy numbers (ARG copy number per bacterial cell) were calculated as follows: normalized ARG copy numbers = (relative ARG copy number)/(4 × relative 16S rRNA gene copy numbers), where 4 is the average number of 16S rRNA-encoding genes per bacterium (Klappenbach et al., 2001). Absolute ARG copy numbers (ARG copy numbers per gram of feces) were determined by using standard curve method of quantification with the Roche LightCycler Real-time qPCR 480 system, as described previously (Ouyang et al., 2015). Relative abundance of specific ARG and OTU refers to the percentage of individual abundance among total ARG/OTU abundance. 2.6. Statistical analysis Averages and standard deviations were calculated using Excel 2013 (Microsoft Office Professional Plus 2013, Microsoft, USA). Linear discriminative analysis (LDA) Effect Size (LEfSe) (Segata et al., 2011) was performed using Galaxy as explanation tool for high-dimensional data and indicator microbe/gene discovery (Goecks et al., 2010). Principle component analysis (PCA), partial redundancy analysis (pRDA), Procrustes analysis, single environmental factor test, and Mantel test were conducted on RStudio (Version 0.99.486 – © 2009–2015 RStudio, Inc.) using Community Ecology Package “Vegan” (version 2.3–0) for ecological statistical analysis. Bar chart and scatter diagrams were generated by OriginPro 2015. Correlation analysis and significance tests were performed using IBM SPSS Statistics version 19. Venn diagram was created by Creatly software (Cinergix, Pty Ltd). Network analysis was plotted using Cytoscape 3.2.1. Relative abundance data for both ARGs and OTUs were used as input data for LDA, PCA, pRDA, Procrustes analysis, Mantel test and network analysis. Data of total number of ARGs detected and absolute abundance of 16S/ARGs were used for correspondent correlation analysis and significance tests. The flowchart was generated using ProcessOn software. 3. Results 3.1. Antibiotics and metals as additives in swine feed Antibiotics detected in swine feed samples (Table 1). Seven different antibiotics were detected among all the feed samples and the concentrations varied from 1.9 to 1159 μg per kg of dry weight. Olaquindox was found in most of the feed samples at a concentration up to 1159 μg/kg. Florfenicol was detected in two out of 7 feed samples at concentrations between 4.9 and 16.5 μg/kg. The remaining five detected antibiotics (sulfamethazine, sulfaisodimidine, ofloxacin, nicarbazin and enrofloxacin) were each detected in only one group at concentrations between 0.2 and 5.1 μg/kg (Table 1). Concentrations of copper and zinc averaged 90.2 μg Cu/g dry wt and 179.3 μg Zn/g dry wt, respectively, and ranged from 6.1 to 186.3 μg Cu/g dry wt and 65.6 to 285.6 μg Zn/ g dry wt (Table 2). Group HY3 and JX3 were found to be highly medicated with antibiotics at 1159 and 935 μg/kg, while the other groups were lowly medicated with total in-feed antibiotics in the range of 2–52 μg/kg. 3.2. Microbial community composition in swine gut The number of 16S rRNA gene copies ranged from 2.59 × 109 to 8.52 × 109 copies per gram of fresh wet fecal matter (Fig. S1). A total
3
Table 1 Antibiotics detected and their concentrations in swine feeds. Group Detected antibiotic HY1 Florfenicol HY2 Sulfamethazine Sulfaisodimidine Olaquindox
Concentration (μg/kg)
Total concentration (μg/kg)
4.9
4.9
0.2 0.2
1.9
1.5 HY3 Olaquindox
1159
1159
LY Olaquindox
32
32
JX1 Florfenicol
16.5
16.5
JX2 Ofloxacin Olaquindox Nicarbazin
0.5 46.5 5.1
52.1
JX3 Enrofloxacin Olaquindox
1.1 934
935.1
of 1,331,801 high quality filtered sequences were obtained from all 28 samples with a range of 8472 to 85,719 sequences per samples. These sequences were clustered into 18,171 OTUs at the 97% similarity level and then assigned to Bacteria (97.9%), Archaea (1.5%) and other unassigned (0.6%). Bacteroidetes and Firmicutes were the two dominating phyla in all seven groups with N75% of the sequences belonging to these taxa (Fig. S2). The dominant classes were Clostridia/Firmicutes, Bacilli/ Firmicutes and Bacteroidia/Bacteroidetes with relative abundances up to 44.5%, 31.4%, and 43.6%, respectively. OTU level rarefaction curves showed a similar species richness among groups (Fig. S3). Similar results were obtained when evaluating diversity using PD whole tree, Chao1 estimator, and the Shannon index (Fig. S3). Principal component analysis (PCA) displayed distinct clustering of replicate samples from each group demonstrating similar microbial community composition (Fig. S4). However, the microbial communities from different groups (HY3, LY, JX1, JX2 and JX3) also shared a certain degree of similarity, whereas HY1 and HY2 formed another loose cluster. To characterized the difference of gut microbiota between groups and discover indicator taxa, LDA Effect Size (LEfSe) algorithm module was used to couple statistical significance with biological consistency and effect size estimation. Potential indicator taxa are represented in Fig. 1 with their LDA scores in Fig. S5. These taxa displayed statistically significant differential relative abundance between groups. Every
Table 2 Metals detected in swine feed and their concentration.
HY1 HY2 HY3 LY JX1 JX2 JX3
Cu (μg/g)
Zn (μg/g)
Cd (μg/g)
Pb (μg/g)
Hg (μg/g)
Cr (μg/g)
As (μg/g)
6.1 108.2 186.3 24.12 25.1 129.4 151.9
65.6 131.4 207.2 178.9 123.7 262.4 285.6
0.063 0.11 0.083 0.05 0.05 0.04 0.045
0.78 0.32 1 0.43 0.71 0.84 0.76
0.015 0.007 0.007 0.03 0.009 0.04 0.011
0.71 0.33 0.62 1.05 0.39 1.61 1.06
0.242 0.295 0.197 0.254 0.457 0.336 0.294
Please cite this article as: Zhao, Y., et al., Feed additives shift gut microbiota and enrich antibiotic resistance in swine gut, Sci Total Environ (2017), https://doi.org/10.1016/j.scitotenv.2017.10.106
4
Y. Zhao et al. / Science of the Total Environment xxx (2017) xxx–xxx
Fig. 1. Indicator taxa with statistically significant differential abundance between treatment groups as revealed by linear discriminative analysis (LDA) Effect Size (LEfSe) analysis of swine gut microbiota 16S rRNA gene sequence data. Each node refers to a single taxon. Each color refers to each group and its corresponding indicator bacteria. Each circle's diameter is proportional to the taxa levels (from phylum to family). This representation employs the Ribosomal Database Project (RDP) taxonomy. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
group had at least one potential indicator except for HY2 (Fig. 1). All potential indicator taxa of HY1, HY3, and JX3 belonged to the phylum Firmicutes. Other indicators included taxa from Spirochaetes, Bacteroidetes. Partial redundancy analysis was used to characterize the effect of feed additives on the microbial community. Results showed variance in the microbial community could be mainly explained by in-feed antibiotics and metals (53.99%), while MGEs explaining 11.50%. 3.3. ARGs in gut microbiota A total of 156 genes including 146 antibiotic resistance genes (ARGs), 8 transposon-transposase genes and 2 class 1 integron– integrase genes were detected among all samples. The number of ARGs detected in each group varied from 48 to 126 (Fig. 2). The detected resistance genes represented all major resistance mechanisms: antibiotic deactivation (44.5%), efflux pumps (34.3%) and cellular protection (19.2%) (Fig. 2) and conferred resistance to major classes of antibiotics: multidrug (22%), aminoglycoside (20%), beta-lactam (14%), tetracycline (14%), MLSB (14%), vancomycin (8%), sulfonamide (2%), chloramphenicol (2%) and others (4%) (Fig. S6). Swine gut microbiota from HY3 harbored a significantly higher number of ARGs than the other groups (P b 0.05), where multidrug and aminoglycoside resistance genes accounted for 23% and 19%, respectively, followed by MLSB (16%), tetracycline (14%) and beta-lactam resistance genes (14%). ARGs detected in these fresh pig manures were highly abundant, ranging from 2.72 × 109 to 1.34 × 1010 copies per gram (Fig. 3A). The normalized copy numbers (copies per bacterial cell) varied from 4.51
to 13.53, with an average of 6.85 ARGs per bacterial cell (Fig. 3B). The highest absolute abundance of ARG (copy number per g of fecal matter) was found in the highly medicated HY3 group, contributed mostly by tetracycline (4.77 × 109), MLSB (3.41 × 109), and aminoglycoside resistance genes (4.02 × 109). To compare the similarity and dissimilarity of ARGs from different groups, the relative abundance of specific ARGs as a percentage of total ARG abundance across all samples was examined. The overall patterns of ARGs were revealed by PCA. The seven treatment groups could not be distinguished using PCA indicating that they shared similar ARG profiles (Fig. S7). Among 285 targeted ARGs, 32 shared ARGs were found in all samples (Table S3). Unique unshared ARGs were also found with the highest number of unique ARGs for one group being 26 for HY3 (Fig. 4), contributed by multidrug and beta-lactam resistance genes (RGs). LEfSe characterized the indicator ARGs in each group displaying significantly different abundance compared to all other groups (Fig. S8). All indicator ARGs found in HY1 (ermT, ermX, lnuB, vatE) were genes conferring resistance to the MLSB class of antibiotics. A total of 4 indicator ARGs including 3 tetracycline resistance genes (tet(32), tetL, tetO) and 1 aminoglycoside resistance gene (aadD) were observed for HY2. HY3 harbored 8 indicator ARGs conferring resistance to aminoglycosides (aacA/aphD, aphA1(aka kanR)), multiple drugs (mexF, oprD), tetracyclines (tetPB, tetQ), beta-lactams (fox5) and streptomycin ARG (sat4) with efflux pump as resistance mechanism. Other indicator ARGs included aminoglycoside RGs - aadA2 for JX1; aac(6′)-Ib(aka aacA4), aadA2, aadA, aph(2′)-Id, cmxA, and qacEdelta1–01 for JX2; aadA1,tetG,and tetX for JX3.
Please cite this article as: Zhao, Y., et al., Feed additives shift gut microbiota and enrich antibiotic resistance in swine gut, Sci Total Environ (2017), https://doi.org/10.1016/j.scitotenv.2017.10.106
Y. Zhao et al. / Science of the Total Environment xxx (2017) xxx–xxx
5
Fig. 2. Antibiotic resistance genes (ARGs) detected in the seven different treatment groups of swine gut microbiota. ARGs were classified based on (A) antibiotics to which they confer resistance and (B) mechanism of resistance. MLSB as abbreviation for macrolide-lincosamide-streptogramin B.
3.4. Relationship between feed additives, microbial community composition, and ARGs The total concentration of antibiotics in feed was significantly correlated to the absolute abundance of total ARGs in manure (Pearson's coefficient = 0.868, P b 0.05), and the number of different ARGs detected was significantly correlated to total concentration of copper in feed (Pearson's coefficient = 0.871, P b 0.05) (Table S4). Meanwhile, microbial abundance (16S rRNA absolute gene copy numbers) was strongly correlated with total ARG abundance (Pearson's coefficient = 0.65, P b 0.01), with ARGs conferring resistance to each major class of antibiotics except sulfonamides (Table S5). Moreover, the absolute abundance of total MGEs, integrons, and transposons were significantly correlated with total ARG abundance and ARGs conferring resistance to different classes of antibiotics (Table S5). The abilities of antibiotics and metals in swine feed and phylumlevel microbial community composition to explain the relative abundance of ARGs were also evaluated by partial redundancy analysis (pRDA). A total of 31.84% of the variance in the ARG data set could be explained by antibiotic and metal concentrations in swine feed, with
microbial community composition and MGEs explaining 23.3% and 2.4%, respectively (Fig. 5). Interactions between swine feed and microbial community accounted for 16.5% of the variation. Using Monte Carlo permutation test to verify, sorting accepted pRDA results with PseudoF value 6.30 (Significance b0.001, Permutation: free, Number of permutations = 999). Single environmental factor test was conducted to reveal the correlation between each single explanatory variable and ARG variance. Single factors that positively correlated with change of ARGs included Bacteroidetes, Firmicutes, copper, and arsenic (Significance b0.01, Permutation: free, Number of permutations = 999). Procrustes analysis exhibited a good-fit correlation between ARG profiles and microbial community composition on the basis of BrayCurtis dissimilarity metrics (Procrustes sum of squares M2 = 0.6698, r = 0.5746, Significance: 4e-04, Permutation: free, Number of permutations: 9999) (Fig. 6). To further verify, Mantel test was performed and results confirmed a significant correlation between ARG profiles and microbial community composition (r = 0.4436, P b 0.001, Permutation: free, Number of permutations: 999). Hence, the antibiotic resistome did not decouple from the microbiome. Network analysis was performed to reveal the co-occurrence pattern between ARGs and microbial OTUs.
Please cite this article as: Zhao, Y., et al., Feed additives shift gut microbiota and enrich antibiotic resistance in swine gut, Sci Total Environ (2017), https://doi.org/10.1016/j.scitotenv.2017.10.106
6
Y. Zhao et al. / Science of the Total Environment xxx (2017) xxx–xxx
Fig. 3. Abundance of antibiotic resistance genes (ARGs) in the seven different treatment groups of swine gut microbiota. (A) Absolute copy numbers of ARGs conferring resistance to specific classes of antibiotics per gram of fecal sample, with error bars showing calculated standard deviation (s.d.) of four sampling replicates on each group (n = 4). (B) the normalized copy numbers of ARGs presented as total ARG copies per bacterial cell with error bar showing calculated s.d. (n = 4).
Specific ARGs conferring resistance to aminoglycosides, beta-lactams, MLSB, and tetracyclines could be linked to Bacteriodetes, Firmciutes, and Spirochaetes (Fig. 7). 4. Discussion Although time-resolved dynamics of the swine gut microbiome and antibiotic resistome were not obtained, the results of the this study demonstrate that swine gut microbiota from Chinese pig farms harbored diverse and abundant ARGs in both high-medicated (HY3, JX3) and low-medicated groups (other five groups). The presented results also support a potential role of Cu in co-selection of ARGs. Analyses demonstrated that increased levels of in-feed additives affected the composition of the gut microbiota and aggravated the enrichment of ARGs. Feed additives and microbial community were the dominant determinants of shaping the swine gut antibiotic resistome. Highly diverse and abundant ARGs were found in pig fecal samples of both high-medicated and low-medicated group (Figs. 2 & 3). The 32 ARGs shared by all 28 investigated microbiomes belong to the most
abundant type of ARGs conferring resistance to aminoglycosides, MLSB and tetracyclines. Consistent with the findings from the present study, the genes aadA and tet family genes were previously suggested as the most common and abundant ARGs in manure (Binh et al., 2009; Munir and Xagoraraki, 2011). About 100 additional ARGs were detected in the present study compared to the number of ARGs detected in a swine manure control sample from pigs that received no known antibiotics from a previous study employing the same HT-qPCR based ARG chip (Zhu et al., 2013). These findings indicate that even small dosages of antibiotics can induce higher levels of ARGs in swine gut. Recent studies have demonstrated that even subinhibitory levels of antibiotic can induce different expression of genes causing enhancement of toxin and biofilm production in potential human pathogens (Goh et al., 2002; Goneau et al., 2015; Hoffman et al., 2005; Wong et al., 2000). In addition to increasing the level of ARG, a small dosage of antibiotics in feed additives for prophylaxis could thus also pose a risk by promoting the virulence and biofilm formation during bacterial infection. In-feed copper was significantly correlated to the diversity of ARGs in gut microbiota (P b 0.05) and together with arsenic were shown to
Please cite this article as: Zhao, Y., et al., Feed additives shift gut microbiota and enrich antibiotic resistance in swine gut, Sci Total Environ (2017), https://doi.org/10.1016/j.scitotenv.2017.10.106
Y. Zhao et al. / Science of the Total Environment xxx (2017) xxx–xxx
7
Fig. 6. Procrustes analysis showing the correlation between antibiotic resistance gene (ARG) and microbial community structure (ARG/operational taxonomic unit distribution patterns) on the basis of Bray-Curtis dissimilarity metrics (M2 = 0.6698, r = 0.5746, P b 0.0001, 9999 free permutations). Red line connecting the two data sets of groups. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) Fig. 4. Number of unique unshared antibiotic resistance genes (ARGs) in swine gut microbiota originating from seven different treatment groups. MLSB as abbreviation for macrolide-lincosamide-streptogramin B.
be the two dominant single factors shifting the ARG profiles (P b 0.01), suggesting roles of Cu and As as co-selecting agents. A number of studies have implicated that metals play an important role in co-selecting ARGs (Baker-Austin et al., 2006; Hu et al., 2016; Zhu et al., 2013). Toxic metals, such as copper and zinc, may even exert a stronger selection pressure
Fig. 5. Partial redundancy analysis differentiating the effect of antibiotic and metal concentrations in swine feed, swine gut microbial community composition and mobile gene elements on the overall ARG profiles.
for antibiotic resistance than the specific antibiotic itself in certain cases (Song et al., 2017). It should be noted that much higher concentrations of metals than observed in this study have been observed in pig fecal matter obtained from other pig farms (Jensen et al., 2016). A key finding of this study is that the ARGs present in swine gut depended on the composition of the microbial community, indicating that swine gut microbiome could be an important conduit for transferring ARGs into the environment. A similar conclusion was drawn by a recent study of differentially medicated pigs in USA (Looft et al., 2012). The strong association between the resistome and microbiome has also been found in urban sewage (Su et al., 2017), human gut (Penders et al., 2013; Su et al., 2017) and soil (Forsberg et al., 2014). And due to the anthropogenic activities in last decades, the microbes are moving globally (Zhu et al., 2017a) and ARGs spread into our environments with unprecedented diversity and abundance (Zhu et al., 2017b). Therefore, importance of monitoring and control of microbes and ARGs in reservoirs like animal gut should not be underestimated. To the best of our knowledge, our study is the first to use LDA as a tool to discover potential indicator ARGs and indicator microbial taxa in fecal samples. LDA and network analysis gave consistent results. pRDA results showed that in-feed antibiotics and metals, and microbial community in total explained 71.6% of the variance in ARG pattern. Moreover, 54.0% of changes in gut microbiota were contributed by antibiotics and metals in feed. Collectively, these results imply that swine gut antibiotic resistome was largely shaped by the enrichment of certain microbial taxa rather than by horizontal gene transfer of ARGs. However, this does not rule out the importance of HGT which considered a vital pathway for dissemination of ARGs and causing emergence of multidrug resistant bacteria in different environments (Dzidic and Bedeković, 2003; Martínez, 2008). This is consistent with the observed positive correlation between MGEs and multidrug resistance genes found in this study (Pearson's r = 0.94, P b 0.001). Multidrugresistant bacteria (MDB) as the most threatening type of antibioticresistant bacteria (ARB) to public health includes superbug methicillin-resistant Staphylococcus aureus (MRSA), which is responsible for several difficult-to-treat infections in humans (Sharma et al., 2016). Moreover, diverse mobile ARGs induced in swine gut, enriched and transferred into the environment by manure fertilization and farm wastewater, could pose a risk to human health by potential HGT of ARGs to human-associated pathogens (Ashbolt et al., 2013; Baquero et al., 2008). Hence, management options for mitigating the dissemination of such manure-borne ARGs are critical (Pruden et al., 2013).
Please cite this article as: Zhao, Y., et al., Feed additives shift gut microbiota and enrich antibiotic resistance in swine gut, Sci Total Environ (2017), https://doi.org/10.1016/j.scitotenv.2017.10.106
8
Y. Zhao et al. / Science of the Total Environment xxx (2017) xxx–xxx
Fig. 7. Network analysis revealing the co-occurrence pattern between antibiotic resistance genes (ARGs) and microbial taxa at phylogenetic resolution from phylum to genus (16S rRNA gene sequence data). The nodes were colored according to the ARG classes and different phyla, and the size of the node is proportional to the degree of the connection (number of correlation). The edge (connection) shows significant correlation (P b 0.01), and edge line width represents corresponding Spearman's correlation coefficient.
Composting and lagoons have been shown quite effective treatments of ARGs in manure (Storteboom et al., 2007). However, the persistence of some ARGs has been reported after treatment (Sharma et al., 2009). Therefore, better manure treatment technologies are needed to prevent dissemination of ARGs to agricultural soils and harvested crops. However, the ultimate solution may be to replace antibiotics by other feed additives such as phytogenic products, probiotics, and prebiotics (Gaggia et al., 2010; Windisch et al., 2008).
21 specific ARGs and their potential hosts. Hence, the swine gut microbial community may serve as an important reservoir of ARGs of relevance to public health. Conflicts of interests The authors declare no conflicts of interests. Acknowledgements
5. Conclusions The presented results clearly indicate that antibiotics and metals used as feed additives in large-scale Chinese farming operation shifted the swine gut microbiome and significantly enhanced the diversity and abundance of ARGs in swine gut. The taxonomic composition of the swine gut community also played an important role in shaping the ARG profile, and network analysis revealed co-occurrence patterns of
The authors thank Jie Xu for help with sample collection, and Xiuli Hao for advice on figure plotting. This work was supported financially by National Natural Science Foundation of China (41571130063, 21210008); and REMEDIATE (Improved decision-making in contaminated land site investigation and risk assessment) Marie-Curie Innovation Training Network from the European Union's Horizon 2020 Programme (grant agreement n. 643087).
Please cite this article as: Zhao, Y., et al., Feed additives shift gut microbiota and enrich antibiotic resistance in swine gut, Sci Total Environ (2017), https://doi.org/10.1016/j.scitotenv.2017.10.106
Y. Zhao et al. / Science of the Total Environment xxx (2017) xxx–xxx
Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.scitotenv.2017.10.106. References Ashbolt, N.J., Amezquita, A., Backhaus, T., Borriello, P., Brandt, K.K., Collignon, P., et al., 2013. Human Health Risk Assessment (HHRA) for environmental development and transfer of antibiotic resistance. Environ. Health Perspect. 121, 993–1001. Baker-Austin, C., Wright, M.S., Stepanauskas, R., McArthur, J.V., 2006. Co-selection of antibiotic and metal resistance. Trends Microbiol. 14, 176–182. Baquero, F., Martínez, J.-L., Cantón, R., 2008. Antibiotics and antibiotic resistance in water environments. Curr. Opin. Biotechnol. 19, 260–265. 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, 427–433. Cang, L., 2004. Heavy metals pollution in poultry and livestock feeds and manures under intensive farming in Jiangsu Province, China. J. Environ. Sci. 16, 371–374. Cheng, W., Chen, H., Su, C., Yan, S., 2013. Abundance and persistence of antibiotic resistance genes in livestock farms: a comprehensive investigation in eastern China. Environ. Int. 61, 1–7. Claesson, M.J., Wang, Q., O'Sullivan, O., Greene-Diniz, R., Cole, J.R., Ross, R.P., et al., 2010. Comparison of two next-generation sequencing technologies for resolving highly complex microbiota composition using tandem variable 16S rRNA gene regions. Nucleic Acids Res. e200, 38. Collignon, P., Voss, A., 2015. China, what antibiotics and what volumes are used in food production animals? Antimicrob. Resist. Infect. Control 4, 16. Dzidic, S., Bedeković, V., 2003. Horizontal gene transfer-emerging multidrug resistance in hospital bacteria. Acta Pharmacol. Sin. 24, 519–526. Edgar, R.C., 2010. Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26, 2460–2461. Forsberg, K.J., Patel, S., Gibson, M.K., Lauber, C.L., Knight, R., Fierer, N., et al., 2014. Bacterial phylogeny structures soil resistomes across habitats. Nature 509, 612–616. Gaggia, F., Mattarelli, P., Biavati, B., 2010. Probiotics and prebiotics in animal feeding for safe food production. Int. J. Food Microbiol. 141 (Suppl. 1), S15–28. Gibson, M.K., Crofts, T.S., Dantas, G., 2015. Antibiotics and the developing infant gut microbiota and resistome. Curr. Opin. Microbiol. 27, 51–56. Goecks, J., Nekrutenko, A., Taylor, J., 2010. Galaxy: a comprehensive approach for supporting accessible, reproducible, and transparent computational research in the life sciences. Genome Biol. 11, 1. Goh, E.-B., Yim, G., Tsui, W., McClure, J., Surette, M.G., Davies, J., 2002. Transcriptional modulation of bacterial gene expression by subinhibitory concentrations of antibiotics. Proc. Natl. Acad. Sci. 99, 17025–17030. Goneau, L.W., Hannan, T.J., MacPhee, R.A., Schwartz, D.J., Macklaim, J.M., Gloor, G.B., et al., 2015. Subinhibitory antibiotic therapy alters recurrent urinary tract infection pathogenesis through modulation of bacterial virulence and host immunity. MBio 6, e00356–15. Goossens, H., Ferech, M., Vander Stichele, R., Elseviers, M., Group EP, 2005. Outpatient antibiotic use in Europe and association with resistance: a cross-national database study. Lancet 365, 579–587. Hamscher, G., Pawelzick, H.T., Sczesny, S., Nau, H., Hartung, J., 2003. Antibiotics in dust originating from a pig-fattening farm: a new source of health hazard for farmers? Environ. Health Perspect. 111, 1590–1594. Heuer, H., Schmitt, H., Smalla, K., 2011. Antibiotic resistance gene spread due to manure application on agricultural fields. Curr. Opin. Microbiol. 14, 236–243. Hoelzer, K., Wong, N., Thomas, J., Talkington, K., Jungman, E., Coukell, A., 2017. Antimicrobial drug use in food-producing animals and associated human health risks: what, and how strong, is the evidence? BMC Vet. Res. 13, 211. Hoffman, L.R., D'Argenio, D.A., MacCoss, M.J., Zhang, Z., Jones, R.A., Miller, S.I., 2005. Aminoglycoside antibiotics induce bacterial biofilm formation. Nature 436, 1171–1175. Hu, Y., Yang, X., Qin, J., Lu, N., Cheng, G., Wu, N., et al., 2013. Metagenome-wide analysis of antibiotic resistance genes in a large cohort of human gut microbiota. Nat. Commun. 4, 2151. Hu, H.W., Wang, J.T., Li, J., Li, J.J., Ma, Y.B., Chen, D., et al., 2016. Field-based evidence for copper contamination induced changes of antibiotic resistance in agricultural soils. Environ. Microbiol. 18, 3896–3909. Huang, Y., Cheng, M., Li, W., Wu, L., Chen, Y., Luo, Y., et al., 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.
9
Jensen, J., Larsen, M.M., Bak, J., 2016. National monitoring study in Denmark finds increased and critical levels of copper and zinc in arable soils fertilized with pig slurry. Environ. Pollut. 214, 334–340. 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, 181–184. Looft, T., Johnson, T.A., Allen, H.K., Bayles, D.O., Alt, D.P., Stedtfeld, R.D., et al., 2012. In-feed antibiotic effects on the swine intestinal microbiome. Proc. Natl. Acad. Sci. U. S. A. 109, 1691–1696. Martínez, J.L., 2008. Antibiotics and antibiotic resistance genes in natural environments. Science 321, 365–367. Munir, M., Xagoraraki, I., 2011. Levels of antibiotic resistance genes in manure, biosolids, and fertilized soil. J. Environ. Qual. 40, 248. 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, 5697–5707. Penders, J., Stobberingh, E.E., Savelkoul, P.H., Wolffs, P.F., 2013. The human microbiome as a reservoir of antimicrobial resistance. Front. Microbiol. 4, 87. Perreten, V., Schwarz, F., Cresta, L., Boeglin, M., Dasen, G., Teuber, M., 1997. Antibiotic resistance spread in food. Nature 389, 801–802. Pruden, A., Pei, R., Storteboom, H., Carlson, K.H., 2006. Antibiotic resistance genes as emerging contaminants: studies in northern Colorado. Environ. Sci. Technol. 40, 7445–7450. Pruden, A., Larsson, D.G., Amezquita, A., Collignon, P., Brandt, K.K., Graham, D.W., et al., 2013. Management options for reducing the release of antibiotics and antibiotic resistance genes to the environment. Environ. Health Perspect. 121, 878–885. Segata, N., Izard, J., Waldron, L., Gevers, D., Miropolsky, L., Garrett, W.S., et al., 2011. Metagenomic biomarker discovery and explanation. Genome Biol. 12, R60. Sharma, R., Larney, F.J., Chen, J., Yanke, L.J., Morrison, M., Topp, E., et al., 2009. Selected antimicrobial resistance during composting of manure from cattle administered subtherapeutic antimicrobials. J. Environ. Qual. 38, 567–575. Sharma, M., Nunez-Garcia, J., Kearns, A.M., Doumith, M., Butaye, P.R., Argudin, M.A., et al., 2016. Livestock-associated methicillin resistant Staphylococcus aureus (LA-MRSA) clonal complex (CC) 398 Isolated from UK animals belong to European lineages. Front. Microbiol. 7, 1741. Song, J., Rensing, C., Holm, P.E., Virta, M., Brandt, K.K., 2017. Comparison of metals and tetracycline as selective agents for development of tetracycline resistant bacterial communities in agricultural soil. Environ. Sci. Technol. 51, 3040–3047. Storteboom, H.N., Kim, S.C., Doesken, K.C., Carlson, K.H., Davis, J.G., Pruden, A., 2007. Response of antibiotics and resistance genes to high-intensity and low-intensity manure management. J. Environ. Qual. 36, 1695–1703. Su, J.Q., Wei, B., Ou-Yang, W.Y., Huang, F.Y., Zhao, Y., HJ, Xu, et al., 2015. Antibiotic resistome and its association with bacterial communities during sewage sludge composting. Environ. Sci. Technol. 49, 7356–7363. Su, J.Q., An, X.L., Li, B., Chen, Q.L., Gillings, M.R., Chen, H., et al., 2017. Metagenomics of urban sewage identifies an extensively shared antibiotic resistome in China. Microbiome 5, 84. Van Boeckel, T.P., Gandra, S., Ashok, A., Caudron, Q., Grenfell, B.T., Levin, S.A., et al., 2014. Global antibiotic consumption 2000 to 2010: an analysis of national pharmaceutical sales data. Lancet Infect. Dis. 14, 742–750. Windisch, W., Schedle, K., Plitzner, C., Kroismayr, A., 2008. Use of phytogenic products as feed additives for swine and poultry. J. Anim. Sci. 86, E140–8. Wong, C.S., Jelacic, S., Habeeb, R.L., Watkins, S.L., Tarr, P.I., 2000. The risk of the hemolytic– uremic syndrome after antibiotic treatment of Escherichia coli O157: H7 infections. N. Engl. J. Med. 342, 1930–1936. Xu, H.-J., Wang, X.-H., Li, H., Yao, H.-Y., J-Q, Su, Zhu, Y.-G., 2014. Biochar impacts soil microbial community composition and nitrogen cycling in an acidic soil planted with rape. Environ. Sci. Technol. 48, 9391–9399. Yazdankhah, S., Rudi, K., Bernhoft, A., 2014. Zinc and copper in animal feed - development of resistance and co-resistance to antimicrobial agents in bacteria of animal origin. Microb. Ecol. Health Dis. 25. Yun-peng, W., Yue, M., 2008. Potential public hazard of using antibiotics in livestock industry. Chin. J. Antibiot. 9, 001. Zhu, Y.G., Johnson, T.A., Su, J.Q., Qiao, M., Guo, G.X., Stedtfeld, R.D., et al., 2013. Diverse and abundant antibiotic resistance genes in Chinese swine farms. Proc. Natl. Acad. Sci. U. S. A. 110, 3435–3440. Zhu, Y.-G., Gillings, M., Simonet, P., Stekel, D., Banwart, S., Penuelas, J., 2017a. Microbial mass movements. Science 357, 1099–1100. Zhu, Y.-G., Zhao, Y., Li, B., Huang, C.-L., Zhang, S.-Y., Yu, S., et al., 2017b. Continental-scale pollution of estuaries with antibiotic resistance genes. Nat. Microbiol. 2, 16270.
Please cite this article as: Zhao, Y., et al., Feed additives shift gut microbiota and enrich antibiotic resistance in swine gut, Sci Total Environ (2017), https://doi.org/10.1016/j.scitotenv.2017.10.106