Article pubs.acs.org/est
Air-Drying Beds Reduce the Quantities of Antibiotic Resistance Genes and Class 1 Integrons in Residual Municipal Wastewater Solids Tucker R. Burch,† Michael J. Sadowsky,‡,§ and Timothy M. LaPara*,†,‡ †
Department of Civil Engineering, University of Minnesota, Minneapolis, Minnesota 55455, United States BioTechnology Institute, University of Minnesota, St. Paul, Minnesota 55108, United States § Department of Soil, Water, and Climate, University of Minnesota, St. Paul, Minnesota 55108, United States ‡
S Supporting Information *
ABSTRACT: This study investigated whether air-drying beds reduce antibiotic resistance gene (ARG) concentrations in residual municipal wastewater solids. Three laboratory-scale drying beds were operated for a period of nearly 100 days. Real-time PCR was used to quantify 16S rRNA genes, 16S rRNA genes specific to fecal bacteria (AllBac) and human fecal bacteria (HF183), the integrase gene of class 1 integrons (intI1), and five ARGs representing a cross-section of antibiotic classes and resistance mechanisms (erm(B), sul1, tet(A), tet(W), and tet(X)). Air-drying beds were capable of reducing all gene target concentrations by 1 to 5 orders of magnitude, and the nature of this reduction was consistent with both a net decrease in the number of bacterial cells and a lack of selection within the microbial community. Half-lives varied between 1.5 d (HF183) and 5.4 d (tet(X)) during the first 20 d of treatment. After the first 20 d of treatment, however, half-lives varied between 8.6 d (tet(X)) and 19.3 d (AllBac), and 16S rRNA gene, intI1, and sul1 concentrations did not change (P > 0.05). These results demonstrate that air-drying beds can reduce ARG and intI1 concentrations in residual municipal wastewater solids within timeframes typical of operating practices.
■
discharged from the treatment process.11 Thus, residual municipal wastewater solids treatment could be targeted as a key point in the treatment process to implement strategies to reduce the quantity of ARGs discharged from municipal wastewater treatment plants. Previous work has demonstrated the potential for some existing treatment technologies to reduce concentrations of ARGs in residual municipal wastewater solids. For instance, full-scale thermophilic anaerobic digestion has been demonstrated to remove substantial fractions (≥75%) of intI1, tet(A), tet(O), and tet(X) relative to 16S rRNA genes.7 Furthermore, removal of 80% to 90% of intI1, sul1, sul2, tet(G), and tet(X) genes has been demonstrated in laboratory-scale mesophilic anaerobic digestors, while similar removal efficiencies have been demonstrated for intI1, sul1, sul2, erm(B), erm(F), and tet(W) in laboratory-scale thermophilic anaerobic digestors.17 Halflives for intI1, tet(A), tet(L), tet(O), tet(W), and tet(X) in anaerobic digestors appear to be on the order of days and may decrease with increasing temperature.18 However, the effect that other existing residual solids treatment technologies may have on ARG removal is poorly understood. These alternative technologies (e.g., aerobic digestion, air-drying, composting,
INTRODUCTION Antibiotics are important to the success of modern medicine. Bacterial resistance to antibiotics, however, is believed to be an inevitable consequence of their use that eventually negates their benefits.1 Two solutions to the resistance problem are typically employed by medical practitioners. The first is to prescribe newer antibiotics for which there has not been sufficient time for resistance to develop.2 The second is to prescribe a combination of antibiotics to be taken simultaneously; this combination is typically lethal to the infectious microorganisms because multiple resistance is not as frequent as resistance to a single antibiotic.2 Unfortunately, both approaches are becoming less effective, which has led to a search for new solutions for managing antibiotic resistance. As part of these efforts, antibiotic resistance genes (ARGs) have been identified as an emerging pollutant of concern.3 As with all pollutants, a critical component of elucidating and managing the fate of ARGs is identifying the environmental reservoirs from which they are currently discharged. One of the most important environmental reservoirs of ARGs appears to be municipal wastewater. ARGs have been found in relatively high concentrations at almost every point in the municipal wastewater treatment process, including the raw influent, primary effluent, aeration tanks, secondary effluent, and residual solids.3−16 The residual solids are of particular interest because they contain the vast majority of prokaryotic biomass in the treatment process and, as a result, the vast majority of ARGs © 2013 American Chemical Society
Received: Revised: Accepted: Published: 9965
June 3, 2013 August 1, 2013 August 2, 2013 August 2, 2013 dx.doi.org/10.1021/es4024749 | Environ. Sci. Technol. 2013, 47, 9965−9971
Environmental Science & Technology
Article
FastDNA Spin Kit (MP Biomedicals LLC, Solon, OH). Each dry sample was stored at −20 °C until being mixed with 500 μL of lysis buffer (CLS-TC, MP Biomedicals LLC, Solon, OH) in bead-beating tubes (Lysis Matrix E, MP Biomedicals LLC, Solon, OH). Each tube was then subjected to bead-beating for 30 s in a Bio101 Savant FastPrep instrument followed by genomic DNA extraction using a FastDNA Spin Kit for Soil (MP Biomedicals LLC, Solon, OH). Quantitative PCR. Real-time quantitative PCR (qPCR) was used to quantify the concentrations of three tetracycline resistance genes (tet(A), tet(W), and tet(X)), an erythromycin resistance gene (erm(B)), a sulfonamide resistance gene (sul1), and the integrase gene of class 1 integrons (intI1).7,22−26 The tetracycline resistance genes represent each of the three known tetracycline resistance mechanisms: efflux pumps, ribosomal protection proteins, and enzymatic modification systems, respectively.27 The erythromycin resistance gene encodes an rRNA methyltransferase that confers resistance to macrolides, lincosamides, and streptogramin B.28 These drugs are among the most frequently prescribed antibiotics in human medicine. Class 1 integrons were quantified due to their association with multiple antibiotic resistance; they enable bacteria to collect multiple, exogenous ARGs and modulate their expression.29 Real-time PCR was also used to determine the concentrations of 16S rRNA genes (a measure of total bacterial biomass), all Bacteroides spp. 16S rRNA genes (AllBac, a measure of total fecal material), and human-specific Bacteroides spp. 16S rRNA genes (HF183, a measure of human fecal material).30−33 The primer sequences, expected amplicon size, and annealing temperature of each gene target can be found in Table S1. Real-time PCR was carried out using an Eppendorf Mastercycler EP Realplex thermal cycler (Eppendorf, Westbury, NY). PCR assays were optimized to reduce or eliminate the formation of primer-dimers and other nonspecific products. Typical qPCR assays began with a 1 min initial denaturation at 95 °C, followed by 40 cycles of denaturation at 95 °C for 15 s and combined annealing and extension at the primer-specific annealing temperature for 1 min. Reaction volumes of 25 μL consisted of 12.5 μL of BioRad iTaq SYBR Green Supermix with ROX (Life Science Research, Hercules, CA), 25 μg of bovine serum albumin, optimized quantities of forward and reverse primers, and approximately 1 ng of template genomic DNA. Standards were made from PCR products selected from municipal wastewater solids or well-described bacterial isolates. PCR products were ligated into a pGEM-T Easy cloning vector, transformed into JM109 competent cells, and extracted from cell cultures using an alkaline lysis procedure.34 The DNA concentrations of qPCR standards were quantified using a TD700 fluorometer and Hoechst 33258 dye. The standard curve for each real-time PCR assay consisted of a 10-fold dilution series of the qPCR standard containing at least 7 points (r2 ≥ 0.99). Amplification efficiencies were 100% ± 12% (maximum deviation from 100%). Data Analysis. Simple linear regression (obtained using Arc 1.06) was used to determine the goodness of fit of the data to both a monophasic first-order kinetic model and a biphasic first-order kinetic model. The monophasic first-order kinetic model was initially hypothesized based on previous empirical observations that it tends to fit this type of data well for a variety of gene targets in several different environmental conditions. 18,35−37 The model also provides a useful interpretive tool and basis of comparison to other studies that may consider different gene targets and environmental
lime stabilization, etc.) might also offer opportunities for reducing ARG concentrations in residual solids. Furthermore, approximately 80% of all treatment plants in the country employ technologies other than anaerobic digestion for treatment of residual solids.19 Many of these treatment plants serve small municipalities with limited financial and technical resources and, as a result, use less sophisticated types of residual solids treatment technology. Because of this, there is a need to assess the treatment potential of these technologies for their ability to remove ARGs. Our overall hypothesis is that existing technologies for treating residual municipal wastewater solids can be used to reduce ARG concentrations during solids treatment. The goal of the work presented here was to assess the potential of airdrying beds to remove ARGs from municipal wastewater solids. Air drying is designated in the U.S. as a “Process to Significantly Reduce Pathogens” (PSRP). It is used to produce Class B treated residual solids at more than 400 U.S. wastewater treatment plants and is also used to dewater treated solids at approximately 30% of treatment plants in the United States.19,20 Air drying is accomplished by loading wastewater solids to a relatively shallow depth onto an outdoor drying bed, typically constructed of gravel, sand, and concrete or wood.21 The solids are left to dry for a minimum of three months, with an average ambient daily temperature above 0 °C for at least two of those three months.20 Air-drying beds produce strong odors and tend to require a large physical footprint relative to alternative technologies. However, they are also simple to operate and are characterized by relatively low capital costs. As a result, they may represent a tractable strategy for removing ARGs from municipal wastewater solids in smaller treatment plants and sparsely populated municipalities if they are found to reduce ARG concentrations effectively.
■
MATERIALS AND METHODS Experimental Design. Three drying beds were constructed outdoors at the University of Minnesota. Each drying bed was approximately 0.6 m wide, 0.6 m long, and 0.6 m deep with two 8 cm (diameter) holes drilled in the bottom to allow for water drainage. A 15 cm layer of gravel, covered by a 15 cm layer of sand, was placed into each drying bed to support the residual solids and allow for drainage. Residual solids, which consisted of a mixture of primary and secondary solids, were collected from a full-scale treatment plant in southern Minnesota. Precipitation and temperature data were obtained from the Minnesota Climatology Working Group Web site for National Weather Service Station 214884 (44° 55′ N, 93° 11′ W) located near St. Anthony Falls, which is approximately 2 km west of where the drying beds were located. Moisture content was quantified periodically by collecting and weighing approximately 15 g triplicate samples from each drying bed. Solids were dried overnight at 103 °C, allowed to cool in a desiccator, and weighed again. Moisture content was calculated as the fraction of the total mass lost during the drying process. Sample Collection and Genomic DNA Extraction. Three 200 μL replicates (for “liquid” samples) or 500 mg replicates (for “dry” samples) were collected directly from the surface of each drying bed following thorough horizontal and vertical mixing of the bed contents. Each liquid sample was diluted with 500 μL of lysis buffer (120 mM sodium phosphate buffer, 5% dodecyl sulfate, pH 8.0 ± 0.1) and subjected to three consecutive freeze−thaw cycles followed by incubation at 70 °C for 90 min. Genomic DNA was then extracted using a 9966
dx.doi.org/10.1021/es4024749 | Environ. Sci. Technol. 2013, 47, 9965−9971
Environmental Science & Technology
Article
conditions. However, a formal lack-of-fit test provided evidence that several time series of gene target concentrations varied significantly (P ≤ 0.05) from the fitted monophasic first-order kinetic models.38 A biphasic first-order model was subsequently hypothesized based on the observation that the time series trends of drying bed moisture content and several gene targets appeared to be biphasic, with the major switch in rates occurring at approximately 20 d. A biphasic pattern is also consistent with empirical observations of the behavior of gene targets representing ARGs in surface waters and fecal indicators in manure-amended soils.36,39,40 The biphasic first-order model was constructed by splitting each time series of gene target concentrations into two phases and determining independent first-order kinetic fits for each phase using simple linear regression (obtained using Arc 1.06). The first phase of each biphasic model contains gene target concentrations for times less than 20 d, while the second phase contains gene target concentrations for times greater than 20 d.
■
RESULTS Residual municipal wastewater solids in each bed dried substantially over the course of the experiment despite several relatively large precipitation events (>2 cm) during the first 50 d (Figure S1). Drying took place over the course of more than 3 months, during which the ambient air temperature remained above 0 °C. The moisture content of the drying solids was approximately 70% after 10 d, which is consistent with expectations for full-scale drying beds operated under favorable conditions.21 Total bacterial biomass and the concentrations of fecal bacteria decreased over time, although they followed separate patterns (Figure 1). The concentration of 16S rRNA genes decreased by an order of magnitude during the first 10 d and then remained constant for the duration of the experiment. In contrast, the concentration of all Bacteroides spp. 16S rRNA genes decreased rapidly by 4 orders of magnitude during the first 30 d and continued to decrease for the remainder of the experiment. Similarly, the concentration of human-specific Bacteroides spp. 16S rRNA genes decreased rapidly during the first 10 d of the experiment, after which they were below the detection limit (5.7 × 107 copies g−1 dry weight). Concentrations of all ARGs and intI1 decreased, although the rate and extent depended on the specific gene target (Figure 2). Two targets, intI1 and sul1, exhibited patterns similar to that of the 16S rRNA gene; they decreased by an order of magnitude within the first 10 d and then remained constant for the remainder of the experiment. In contrast, concentrations of erm(B), tet(A), and tet(W) all decreased by 4 to 5 orders of magnitude during the course of the experiment. The concentration of tet(X), however, exhibited a unique pattern among the gene targets examined in this study. The tex(X) concentration initially decreased by an order of magnitude within the first 10 d but then increased by nearly 2 orders of magnitude by 27 d, after which it again decreased for the remainder of the experiment. The ratios of ARGs and intI1 to the 16S rRNA gene concentration exhibited distinct patterns compared to absolute concentrations (Figure 3). The ratios of intI1 and sul1 to the 16S rRNA gene remained constant throughout the experiment, while the ratios of tet(A) and tet(W) to the 16S rRNA gene remained constant for time less than 20 d, after which they decreased. Only the ratio of erm(B) decreased throughout the course of the experiment relative to 16S rRNA genes, and the
Figure 1. The quantities of (A) 16S rRNA genes, (B) fecal indicator bacteria as measured by 16S rRNA genes of all Bacteroides spp. (AllBac), and (C) fecal indicator bacteria as measured by 16S rRNA genes of human-specific Bacteroides spp. (HF183) in residual solids applied to 3 replicate drying beds (closed circles, open circles, and closed triangles represent unique experimental replicates). Values are the arithmetic mean of triplicate samples; error bars represent one standard deviation.
ratio of tet(X) to 16S rRNA genes initially increased for the first 20 d, after which it decreased. A biphasic first-order kinetic model fits the data better than a monophasic first-order kinetic model for concentrations of ARGs and intI1 (Tables 1 and 2). When the concentrations of each gene target were modeled as monophasic first-order kinetic processes, half-lives varied between 1.5 d (16S rRNA gene of human-specific Bacteroides spp.) and 36.7 d (16S rRNA gene). However, according to lack-of-fit P values, only the concentrations of tet(A) and the 16S rRNA gene from humanspecific Bacteroides spp. fit a monophasic first-order model well (P > 0.05). In contrast, 15 of 17 time series for the biphasic models are characterized by favorable (P > 0.05) lack-of-fit P values. Half-lives for concentrations of gene targets in the first 20 d of the experiment varied between 1.5 d for the humanspecific Bacteroides spp. 16S rRNA gene and 5.4 d for tet(X). Half-lives for concentrations of gene targets after the first 20 d varied between 8.6 d for tet(X) and 19.3 d for all Bacteroides spp. 16S rRNA genes, while concentrations of the 16S rRNA gene, intI1, and sul1 did not change (P > 0.05). A biphasic first-order kinetic model also fit well for the ratios of ARGs and intI1 to the 16S rRNA gene (Table 3). During the first 20 d of the experiment, the half-life for the ratio of erm(B) to 16S rRNA genes was 3.3 d, and the ratio of tet(A) to 16S rRNA genes did not change (P > 0.05). Ratios of the remaining ARGs and intI1 to 16S rRNA genes actually increased (P ≤ 0.05) during the first 20 d, with doubling times varying between 9967
dx.doi.org/10.1021/es4024749 | Environ. Sci. Technol. 2013, 47, 9965−9971
Environmental Science & Technology
Article
Figure 3. The ratios of intI1, sul1, erm(B), tet(A), tet(W), and tet(X) to the 16S rRNA gene in residual solids applied to 3 replicate drying beds (closed circles, open circles, and closed triangles represent unique experimental replicates). Values are the arithmetic mean of triplicate samples; error bars represent one standard deviation.
Figure 2. The quantities of intI1, sul1, erm(B), tet(A), tet(W), and tet(X) in residual solids applied to 3 replicate drying beds (closed circles, open circles, and closed triangles represent unique experimental replicates). Values are the arithmetic mean of triplicate samples; error bars represent one standard deviation.
Table 1. First-Order Kinetic Coefficients (k)a, Half-Lives (t1/2), r2, and Lack-of-Fit P from Monophasic First-Order Kinetic Models of Gene Target Concentrations in 3 Replicate Drying Beds
7.9 d for tet(X) and 16.9 d for tet(W). After the first 20 d, the ratios of intI1 and sul1 to 16S rRNA genes did not change (P > 0.05), while half-lives for ratios of the remaining ARGs to 16S rRNA genes varied between 9.0 d for tet(X) and 15.9 d for tet(W).
■
gene target
DISCUSSION Air-drying beds can be used to reduce ARG concentrations in residual municipal wastewater solids. In this study, the concentrations of gene targets for erm(B), tet(A), tet(W), and tet(X) were reduced by 2 to 5 orders of magnitude over the course of 100 d, while the concentrations of the most persistent gene targets in the group, intI1 and sul1, were reduced by approximately an order of magnitude. The extents of gene loss found here are comparable to those found for the same genes in other treatment technologies, including mesophilic anaerobic digestion, thermophilic anaerobic digestion, and aerobic digestion.7,17,18,37 Thus, small municipalities with limited resources, which are likely to find drying beds appealing for residual solids treatment due to their low capital and operating costs, may be able to achieve significant reduction of ARG concentrations with air-drying beds. Furthermore, municipalities that use drying beds for dewatering treated residual solids (approximately 30% in the U.S.) may be able to use air-drying beds to provide further ARG removal in addition to that achieved by upstream treatment units.19 Drying beds, however, do not reduce ARG concentrations in residual municipal wastewater solids as quickly as other treatment technologies. The half-lives determined here for erm(B), intI1, sul1, tet(A), tet(W), and tet(X) within the first 20
16S rRNA gene all Bacteroides spp. human-specific Bacteroides spp. erm(B) intI1 sul1 tet(A) tet(W) tet(X)
k (d‑1) ± std. error (d‑1)
t1/2 (d)
r2
lack-of-fit P
10−2 −3
±
36.7
0.29
