Article pubs.acs.org/est
Relationships between Antibiotics and Antibiotic Resistance Gene Levels in Municipal Solid Waste Leachates in Shanghai, China Dong Wu,†,§ Zhiting Huang,†,§ Kai Yang,† David Graham,‡ and Bing Xie*,† †
Shanghai Key Laboratory for Urban Ecological Processes and Eco-Restoration, School of Ecological and Environmental Science, East China Normal University, Shanghai 200241, China ‡ School of Civil Engineering and Geosciences, Newcastle University, Newcastle upon Tyne NE1 7RU, United Kingdom S Supporting Information *
ABSTRACT: Many studies have quantified antibiotics and antibiotic resistance gene (ARG) levels in soils, surface waters, and waste treatment plants (WTPs). However, similar work on municipal solid waste (MSW) landfill leachates is limited, which is concerning because antibiotics disposal is often in the MSW stream. Here we quantified 20 sulfonamide (SA), quinolone (FQ), tetracycline (TC), macrolide (ML), and chloramphenicol (CP) antibiotics, and six ARGs (sul1, sul2, tetQ, tetM, ermB, and mefA) in MSW leachates from two Shanghai transfer stations (TS; sites Hulin (HL) and Xupu (XP)) and one landfill reservoir (LR) in April and July 2014. Antibiotic levels were higher in TS than LR leachates (985 ± 1965 ng/L vs 345 ± 932 ng/L, n = 40), which was because of very high levels in the HL leachates (averaging at 1676 ± 5175 ng/L, n = 40). The mean MLs (3561 ± 8377 ng/L, n = 12), FQs (975 ± 1608 ng/L, n = 24), and SAs (402 ± 704 ng/L, n = 42) classes of antibiotics were highest across all samples. ARGs were detected in all leachate samples with normalized sul2 and ermB levels being especially elevated (−1.37 ± 1.2 and −1.76 ± 1.6 log (copies/16S-rDNA), respectively). However, ARG abundances did not correlate with detected antibiotic levels, except for tetW and tetQ with TC levels (r = 0.88 and 0.81, respectively). In contrast, most measured ARGs did significantly correlate with heavy metal levels (p < 0.05), especially with Cd and Cr. This study shows high levels of ARGs and antibiotics can prevail in MSW leachates and landfills may be an underappreciated as a source of antibiotics and ARGs to the environment.
■
INTRODUCTION
China is a powerful example of this problem because it is a massive producer and user of antibiotics (180 000 t/yr raw antibiotic ingredients used)12 with up to 70% being released directly or indirectly to environment.13 Although antibiotic use in clinical settings is regulated,14 the disposal of antibiotics related to household use and personal care products is less discriminate, usually being disposed with other solid wastes.15 Thus, landfills are often receptacles of antibiotics and, in turn, possible incubators for the selection of resistant bacteria and genes (ARGs) via HGT and other mechanisms due to longterm exposures within landfill ecosystems.11 As such, waste antibiotics and resultant ARGs may pass from MSW into mobile leachates and then into surrounding environments,16 although relative leachate levels and relationships over space and time across MSW networks have not been assessed in great detail. Antibiotics have different physical properties (e.g., solubility and photolytic stability) and effects than ARGs,17,18 but how
Antibiotics have long been used to treat infections in medicine and applied as livestock food additives in animal husbandry.1 However, they are often over used and-or poorly controlled in associated wastes and huge amounts of residues are discharged and released to the environment.2,3 Such use imposes stress on exposed microbial communities to acquire antibiotic resistant genes (ARGs), both in the gut upon initial use and in the environment after release.4 Significant correlations between antibiotic use and the emergence of ARGs have been confirmed, both in environmental samples and in benign and pathogenic culturable bacteria.5 Further, ARGs and resistant strains are found across nature because ARGs readily transfer across bacterial species via horizontal gene transfer (HGT), even without antibiotics being present.6,7 As a consequence, many studies have been performed on the occurrence and fate of antibiotics/ARGs in aquatic, soil and wastewater treatment systems.8−10 However, much less work has been done on residues and genes in municipal solid waste (MSW) landfills and leachates, which is a major omission because unused/ expired drugs are often disposed in solid wastes rather than returned to the vender, as usually recommended.11 © XXXX American Chemical Society
Received: December 16, 2014 Revised: March 11, 2015 Accepted: March 11, 2015
A
DOI: 10.1021/es506081z Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
amplification reactions. The q-PCR (Bio-Rad CFX96) was performed in triplicate to quantify six specific ARGs including sul1, sul2, tetQ, tetM, ermB, mefA, and 16S-rDNA eubacteria. Each reaction was conducted in 25 μL on 96-well plates containing 12.5 μL 2X SYBR Green qPCR Mix (Zoman, Beijing), 1 μL DNA templates, Primier1 and 2 (forward and reverse primers, 0.5 μL each (10 μM)) and 10.5 μL ddH2O. SPE and HPLC-MS/MS Analysis of Antibiotics. The 200 mL leachate samples were initially filtrated through glass fiber filters (0.45 μm, PALL, Port Washington, NY). The filtrate was subsequently extracted by solid-phase extraction (SPE) using HLB cartridges (Waters Oasis HLB, Milford, MA). Prior to extraction, cartridges were preconditioned with 10 mL of methanol and 15 mL of ultrapure water at the rate of 1 mL/ min. Liquid samples were initially acidified to pH 4.0 by 0.1 M citrate buffer solution, and then added with 0.2 g/L Na2EDTA. Finally, this mixture was spiked with the internal standards (100 ng each). Internal standard stock solutions (DrEhrenstorfer, Germany) for each antibiotic (1000 ng/L) were diluted with methanol (CNW, Germany) to prepare working solutions (10 ng/L). The injected internal standards for SAs, FQs, TCs, MLs, and CPs were selected according to previous research,20 and details were provided in SI (section SI-2). The operational flow rate of SPE was kept at 5−10 mL/min. After that, SPE cartridges (column) were eluted by ultrapure (interfering substances removing) and methanol (5 mL/min) respectively and resulting extracts were evaporated using nitrogen gas to 0.5 mL. Ultrapure water (acidified by 0.2% formic acid) was added to each sample to 1.0 mL before high performance liquid chromatograph-tandem mass spectrometry (HPLC−MS) analysis. The HPLC-MS/MS (Waters Acquity UPLC System) analysis was conducted at the flow rate of 0.4 mL/min through testing column (Waters Acquity UPLC BEH). The mobile phase consisted of eluent A (pure water with 0.1% formic acid) and eluent B (acetonitrile with 0.1% formic acid). A triple quadrupole tandem mass spectrometer was used in subsequent MS analysis with a Z-spray electrospray interface (Waters). Both positive (SAs, FQs, TCs, MLs) and negative ion (CPs) modes were adopted in detection of antibiotics, and the instrument operational parameters were presented in the SI. Heavy Metal Analysis. Seven heavy metals, including Pb, Zn, Cu, As, Cd, Cr, and Ni, were chosen for analysis because they have been previously linked with environmental ARGs and-or are frequently detected in leachates.21 Samples were quantified using inductively coupled plasma-atomic emission spectroscopy (ICP-AES, Thermo Scientific-iCAP6300, Waltham, MA). Quality control was provided by parallel analysis of certified reference SOIL GBW-5 (China). The limit of quantification (LOQ) of these heavy metals is 1 μg/L and the details of sample preparation were summarized in the SI. Data Analysis. Samples collected for antibiotics and ARG analyses were perform in duplicate and triplicate, respectively. Broad comparisons of total and antibiotic class levels in the leachates were performed by averaging measured values in different clusters (without consideration of antibiotic), which is valid because identical sample numbers and types were measured for each cluster (e.g., each leachate, each season, etc.). Data analysis and statistical assessments was performed using SPSS 19.0 software (SPSS, Inc., Chicago, IL). Nonparametric tests (i.e., 2-related sample nonparametric Wilcoxon tests (WT) and paired t-tests were employed to detect the statistical significance of ARGs and antibiotics levels among
they relate in MSW leachates is not known. To address this gap, we quantified 20 antibiotics and six ARGs in landfill leachates from two transfer stations (TS) and Laogang landfill reservoirs (LR) in Shanghai. Specifically, we compared in antibiotic levels in two seasons, including selected sulfonamides (SAs), quinolones (FQs), tetracyclines (TCs), macrolides (MLs), and chloramphenicol (CPs). These data were then contrasted with other leachate characteristics, including heavy metal levels, and ARGs for similar antibiotics, which allows, for the first time, a comparative assessment of antibiotics and ARGs from multiple locations within a MSW processing network.
■
MATERIALS AND METHODS Sampling Sites and Samples Pretreatment. Three sites were selected in this study for comparison, including the solid waste transfer stations, Hulin (HL) and Xupu (XP), and one landfill reservoir (LR), which is open-air and exposed to the ambient environment; operated by Laogang Waste Disposal Co., Ltd., Shanghai. The TS are at intermediate sites in the Shanghai municipal network where solid wastes are compressed into containers, which are then shipped to Laogang landfill for final disposal. The HL (cap.1700 t/d) and XP (cap. 1500 t/d) TS are located in north and south of inner city, respectively, whereas the Laogang landfill (located in outer southeast suburbs) typically receives around 10 000 t/d from across greater Shanghai. Leachate samples were collected from access points at the three sites in April and July (2014), and stored at 4 °C upon return the lab. Prior to chemical analysis, samples were prefiltered using sterile 110 mm diameter (11 μm) glass fiber filters (Grade 1, Whatman, UK) to remove larger particles and then filtered again through sterile 0.45 μm membrane (NCM, Whatman, UK), which were aseptically cut into pieces and frozen at −20 °C prior to DNA extraction. DNA Extraction and ARGs Analysis. All frozen filters were slowly thawed, digested and extracted according to the cetyltrimethy-lammonium bromide (CTAB) method using FastDNA kit (K713, Biocolor, China). Extracted DNA were stored at −40 °C prior to polymerase chain reaction (PCR) testing for selected ARGs (Bio-Rad C1000 Touch, USA). Probes and primers employed for ARGs PCRs were chosen or designed for this work; details listed in the Supporting Information (SI) (Table S1). Initial PCR was performed in triplicate with a final reaction volume of 25 μL, consisting of 2.5 μL 10 × PCR buffer (Mg2+ 25 mM), 0.5 μL primer 1, 0.5 μL primer 2 (10 μM), 0.5 μL Taq DNA polymerase (2.5 unit), 0.5 μL dNTPs (10 mM), 1 μL template DNA and 19.5 μL ddH2O. Negative controls (1 μL sterile ddH2O substituting for template DNA) were included in each PCR run. Amplification products and their sizes were collated using agarose gel electrophoresis (Bio-Rad DCode System, Hercules, CA) in 1× TAE buffer at 150 V for 15 min. Plasmids integrated with target ARGs were cloned into Escherichia coil DH5α (Maplechem, Shanghai, China), which were used as calibration standards and copy numbers were determined from six-point standard curves (R2 > 0.980, see SI, Table S2).19 The standard construction and amplification procedure are provided in the SI (section SI-1). 16S-rDNA abundance also were quantified for each sample to normalize detected ARGs (i.e., ARG copies/16S-rDNA copies) to allow for different microbial abundances among samples, matrix effects and DNA extraction efficiencies. Prior to quantitative PCR (qPCR) tests, extracted DNA samples were diluted at three series ratios (1/10, 1/50, 1/100), and the 10 times dilution rate showed lowest related interferences to the B
DOI: 10.1021/es506081z Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
4482.5 338.6 77.3 5319.5 220.5 nd 53.0 46.0 29.3 112.2 196.3 46.6 3.3 3.0 4.9 11.6 8.9 119.1 8488.0 1044.3 6.4 25.3 70.1 94.4 45.8 92.0 18.9 HL XP LR July
CFC
121.4 84.6 60.2 164.9 42.9 25.9
NFC SQ
nd 33.7 10.9 128.3 49.2 659.2
SMT SM
10.3 4.9 659.2 nd 2.8 2186.0
ST SMX
1539.6 46.5 72.3 616.6 72.3 2210.6
SP SD
45.6 27.6 83.6
nd Means “not detected” and this value was estimated as half of detection limit; samples were duplicated. Sulfamethoxazole-D4, norfloxacin-D5, demeclocycline, roxithromycin-D7 and chloramphenicol-D5 were injected as the internal standards for SAs including sulfadiazine (SD), sulfapyridine (SP), sulfamethoxazole (SMX), sulfathiazole (ST), sulfamerazine (SM), sulfamethazine (SMT), and sulfaquinoxaline (SQ); FQs including norfloxacin (NFC), ciprofloxacin (CFC), enrofloxacin, (EFC) and ofloxacin (OFC); TCs including tetracycline (TC), oxytetracycline (OTC), doxycyclinehyclate (DXC), and chlorotetracycline (CTC); MLs including erythromycin (ETM) and roxithromycin (RTM), and CPs including chloramphenicol (CAP), thiamphenicol (TAP), and florfenicol (FF) groups respecitvely. a
951.1 105.9 0.4 40.0 22.5 n.d
FF TAP CAP
879.2 185.8 9.8 1769.5 1658.3 1831.5 440.0 44.9 15.1 496.5 176.8 nd 541.9 nd nd 106.8 25.8 nd nd nd nd 12311.4 793.4 211.0 450.2 578.9 11.3
39800.5 1276.2 4963.2
ETM RTM
608.2 232.6 101.2 nd nd nd
CTC DXC
nd 98.4 nd 425.1 77.8 0.7
OTC TC
19.0 14.5 0.2 477.2 259.2 205.1
OFC EFC
18.7 19.7 29.9
MLs TCs FQs SAs
Table 1. Target Antibiotics Concentration (ng/L) in Leachate Samples of TS and LRa C
HL XP LR
RESULTS AND DISCUSSION Distribution of Antibiotics and ARGs among Leachates. All five classes of antibiotics were detected in leachates from three sampling sites (Table 1). Overall, total antibiotic levels were significantly higher in TS than LR leachates across all sampling (985 ± 1965 ng/L vs 345 ± 932 ng/L, n = 20; WT; p < 0.05) and were particularly high in the HL leachates with mean values of 1676 ± 5175 ng/L (April and July; n = 40). Mean antibiotic levels in the HL leachates were similar between seasonal samplings (1708 ± 4454 vs 1826 ± 3248, n = 20). However, antibiotics levels tended to be higher between April and July in the XP leachates (123 ± 273 vs 280 ± 491 ng/ L, respectively; p = 0.2, n = 20), whereas antibiotic levels were significantly lower over the same period in sampled LR leachates (565 ± 1201 vs 153 ± 395 ng/L, respectively; WT; n = 20, p < 0.05). The mean MLs (3561 ± 8377 ng/L, n = 12), FQs (975 ± 1608 ng/L, n = 24), and SAs (402 ± 704 ng/L, n = 42) classes of antibiotics were highest across all samples. For example, the dominant SA in transfer station leachates was SMX, whereas ST and SP were more prevalent in LR leachates. The dominant FQ and ML antibiotics were OFC and ETM, respectively, which both accounted for over 60% of the total measured levels. Of the less detected antibiotics, CPs and TCs (141 ± 286, n = 18 and 83 ± 163 ng/L, n = 24 respectively,) were particularly present in TS leachates, but were often below detection limits in LR samples. As summarized in Figure 1a, the PCA indicates a broad change in the main antibiotics in leachates from HL (transfer station) and LR (landfill) between April and July. Conversely, less variation in detected antibiotics between samplings was seen in XP leachates. However, no relationship was apparent between patterns of detected antibiotics and ARG types in the PCA, particularly at two transfer stations, XP and HL (Figure 1b). The lack of apparent or consistent patterns between detected ARGs and antibiotics at each leachate sampling site suggests other factors more likely influence the types of ARGs and antibiotics released to the environment. If compare relative ARG abundances at the two transfer stations (Figure 2), we see similar ARG levels in the XP and HL leachates (paired t test, p = 0.14), but antibiotics levels in the leachates of XP were far lower than HL (WT, p < 0.01). A few correlations were observed between selected ARGs and antibiotic (Table 2). Such as significant correlations between tetQ and tetM, and TCs levels (p < 0.05), however, this was the exception rather than the rule. Compared with other target ARGs, sul1 and sul2 (−2.58 ± 1.3 and −1.37 ± 1.1, n = 18, log ARGs/16SrDNA) were at relative high levels in all leachates and total detected SAs (1603.9 ± 1987.3 ng/L, n = 6) were 1−2 orders of magnitude higher than Huangpu River in Shanghai,22 but no significant correlation was observed between either sul gene. Further, although high levels of total target MLs were present the leachates (9940 ± 10420, n = 6), neither ermB nor emfA correlated with detected MLs (p = 0.55 and p = 0.68 respectively).
April
■
61.4 70.6 33.1
CPs
13.4 17.3 0.6
sampling sites. Bivariate correlation and partial bivariate correlation (control for metals) analysis was performed on antibiotics and target ARGs to identify general relationships among sample sites. Principal components analysis (PCA) was used to identify clusters of antibiotics and ARGs (as components), and further assess how they varied among samples. The initial PCA analysis was followed by the varimax rotation to amplify the differences among variables.
117.1 36.9 0.8
Environmental Science & Technology
DOI: 10.1021/es506081z Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
Figure 1. (a) Spatial and temporal distribution of detected antibiotics in PCA; (b) Shifts of ARGs of three sampling sites between two months; it was not potentially linked to the changes of antibiotics levels in MSW processing chain.
Figure 2. Relative abundances of six antibiotic resistance genes (ARGs) detected in during the MSW treatment process in Shanghai. All values are normalized (Log ARGs/16SrDNA). Box plots represent median and range values (n = 3).
D
DOI: 10.1021/es506081z Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology Table 2. Bivariate and Partial Correlations with Target ARGsa sul1 parameters SAs TCs MLs Ni As b Cd b Cr
BC nd
0.852 0.824 0.887 0.860
sul2 PC nd
nd nd
BC nd
0.864 0.846 0.899 0.869
tetM PC
tetQ
ermB
BC
PC
BC
PC
0.887
nd
0.814
nd
0.883 0.875 .922 0.881
nd nd
.949 .948 .941 0.915
nd nd
mefA
BC
PC
BC
PC
n.d. 0.905 0.903 .926 0.896
nd nd nd
nd nd nd 0.815 0.784
nd nd nd
nd
nd nd
a
nd Means significant correlation was not detected (values in bold p
