1. Introduction
Antibiotic resistant bacteria (ARB) are having profound effects on the treatment of human diseases. In the United States, the number of ARB-related hospitalizations continues to increase [
1] with an estimated cost of up to
$30 billion annually [
2]. In recent years the importance of ARB has come to the forefront of many scientific disciplines including environmental microbiology. Antibiotic resistant genes in the environment have the potential to spread into the human population presenting a possible public health problem. Although the full impact of environmental transfer of antibiotic resistance genes (ARGs) on public health is not currently known, several reports have highlighted the presence of ARB and ARGs in wastewater treatment plants [
3,
4,
5,
6,
7,
8], agricultural feedlots [
9,
10,
11] soils [
12], rivers and lakes [
9,
13,
14,
15,
16,
17,
18] raising concerns about potential public health impacts from these sources. As the number and cost of ARB-related illnesses continues to grow, multifaceted efforts are needed to control ARB in the clinic along with further investigation into potential impacts arising from environmental sources.
The presence of antibiotics and ARB in wastewater treatment plants (WWTPs) and water reclamation plant (WRP) effluents are of specific interest. The primary reasons for concern include the presence of clinically relevant ARB and ARGs in raw sewage entering these facilities, the potential for ineffective removal and/or selection of ARG/ARB by WWTPs and, the possibility of human contact with treated waters [
13,
17,
19]. Additionally, some studies have revealed the presence of ARB in treated wastewater effluents and have shown that ARGs and ARB are more prevalent downstream from WWTPs [
3,
15,
20,
21,
22,
23,
24,
25] which suggest WWTPs have the potential to influence the ARB population in receiving waters. Almost all of these reports, however, focused on the impact of facilities using traditional primary and secondary stage treatments with or without disinfection. Currently, the effects of tertiary wastewater treatment processes, such as media filtration, on ARB and ARGs have not been well characterized.
Conventional activated sludge wastewater treatment is divided into distinct stages. The first or primary stage consists of physical removal of oils and greases along with sedimentation of large particles. The secondary phase of treatment utilizes microbial organisms to reduce the amount of organic matter in the wastewater. Typically, processes in addition to traditional secondary treatment which are employed to further improve water quality are referred to as tertiary-stage treatment [
26]. Filtration, a commonly used type of tertiary treatment, has been shown to be effective in reducing the concentrations of viable indicator bacteria and viruses such as bacteriophage, substantially decreasing the number of hazardous microbes in final effluent waters [
27]. Furthermore, tertiary filtration yields additional reductions in suspended solids and biological oxygen demand producing a higher quality effluent. Consequently, disinfection of tertiary effluents can be more efficient due to lower chlorine demand and higher transmittance of UV light compared to secondary effluent [
28].
The public health and environmental hazards associated with ARGs are related to their capacity to be transferred between bacteria coupled with positive selective pressure occurring from the pervasive use of antibiotics both clinically and agriculturally. Specifically, the transfer of ARGs between bacteria horizontally is one of the main factors that have led to the rapid spread of antibiotic resistance across the globe [
29,
30]. The three main mechanisms by which ARGs are acquired by bacteria include: the direct exchange of genetic material between two viable bacteria (conjugation), through phage infections (transduction) and by uptake of extracellular, free DNA (transformation). Data has been presented indicating that horizontal gene transfer (HGT) of resistant genes occurs at a higher frequency among more closely related bacteria [
31].
When considering the potential environmental impact of ARGs related to anthropogenic activities, it is important to recognize that ARGs occur naturally and can readily be found in environmental matrices unaffected by human activities [
32,
33,
34,
35]. Antibiotic resistance genes and ARB have been found in 30,000 year old permafrost as well as in remote caves isolated from humans for over four million years [
36,
37] signifying that these genes are commonly found in the environment and have evolved over thousands to millions of years prior to the therapeutic use of antibiotics by humans. Moreover, research has shown that most of the functional ARGs found in WWTPs are specific to those matrices and are not widely disseminated into human populations or the environment [
38,
39]. Therefore, it is likely that, at any given time, only a small proportion of the total wastewater treatment plant (WWTP) resistome may be associated with potential adverse public health effects however; the possibility exists for the sudden mobilization and horizontal transfer of new ARGs from within the wastewater resistome which could result in additional ARGs entering the human population. It stands to reason, that assessing wastewater matrices not only for the presence of ARGs but their ability to be transferred would result in data that is more relevant to public health. Subsequently, WWTPs that reduce the fraction of clinically relevant ARGs that can be horizontally transferred would lower the probability of treated effluent waters contributing to this public health crisis. Unfortunately, obtaining a direct measurement of HGT events in wastewater has been difficult to achieve making the evaluation of treatment options more challenging.
In water-based matrices, HGT mechanisms can be associated with either the cellular or extra-cellular fractions. The cellular fraction consists of ARGs existing within intact and/or viable bacterial cells or attached to sediments and particulates. Conversely, extracellular ARGs include bacterial DNA that has been excreted or released upon death of the organism and genes contained within viral particles. In aqueous solutions, the extracellular ARGs exist dissolved in solution whereas cell-associated ARGs can settle with other solids. The separation of cellular and extra-cellular ARGs in water can easily be achieved using centrifugation that partitions the components necessary for different HGT pathways into either the pellet or the supernatant. Specifically, HGT by conjugation would be primarily associated with viable bacteria partitioned to the solids fraction (pellet) whereas the dissolved phase (supernatant) would contain the components necessary for transformation (extra-cellular DNA) and transduction (phage particles). Quantifying the amount of ARGs in different fractions would identify where the majority of ARGs reside and consequently, which HGT pathways are likely to be involved in the transfer of ARGs at a particular treatment stage. Furthermore, such information could also be used to identify processes that are better at removing ARGs from each of the particular fractions.
This study was designed to determine the fate of dissolved and solids-associated ARGs through different stages of a tertiary WRP and to identify treatment processes that result in the best removal of ARGs. A qPCR approach was utilized to quantify two ARGs throughout wastewater treatment: the first, targets the sulfonamide resistance gene sul1, and a second targets SHV and TEM-type β-lactamase (blaSHV/TEM) resistance genes. Furthermore, a pilot-scale solid media filter was used to provide a more detailed characterization of tertiary filtration on ARGs.
2. Materials and Methods
2.1. Sample Collection and Processing
Samples were obtained from a tertiary WRP designed for an average flow of 235 million L/day of raw sewage (daily dry weather ranges from 94 to 340 million L/day). The treatment scheme included primary sedimentation, activated sludge with nitrification and denitrification (NDN), secondary clarification (flocculation/settling), tertiary media filtration (anthracite coal/sand/gravel) and, sequential chlorine disinfection (free chlorine with a minimum required residual of 1.0 mg/L followed by chloramine with typical residual concentrations from 1.0 to 3.0 mg/L).
Unconcentrated samples were collected in sterile one-liter Nalgene bottles. Sodium thiosulfate (one milliliter of a 10% weight/volume solution) was added to dechlorinate the water. Samples were collected at the WRP from incoming raw sewage, activated sludge, clarified secondary effluent, and disinfected final effluent. Samples were collected for each water type on at least three separate occasions.
2.2. Collection of Hollow Fiber Filtration (HFF) Concentrated Wastewater Samples
Antibiotic resistance genes from secondary and final effluent waters were concentrated using a one-step HFF procedure. Briefly, approximately 10
9 copies of AdvIPC:pSMART plasmid were added to 100 L of each matrix (secondary or final effluent) then pumped through an Asahi Kasei REXEED
TM-25S filter (Asahi Kasei, Oita, Japan) in a recirculation configuration with pressure within the filter maintained between four and eight psi. Bacteria and nucleic acids were eluted by recirculation of 0.05 M Glycine containing 0.01% Antifoam A (pH 7.0) for 10 to 30 min. New tubing was used with each sample to prevent interferences associated with reused tubing (biofilm, DNA, etc.). The original samples were collected in 50 L plastic carboys that had been pre-washed with hypochlorite (to remove DNA), dechlorinated, and then autoclaved. Secondary effluent was collected directly from the WRP secondary settling tank and tertiary-treated disinfected effluent (final effluent) was collected from a channel immediately after dechlorination but prior to entering the distribution system. Sodium thiosulfate was added to each carboy after collection and total chlorine concentrations were measured using the Hach DPD (
N,
N-diethyl-p-phenylenediamine) colorimetric procedure (Hach Co., Loveland, CO, USA) according to method SM4500-Cl [
40].
2.3. Separation of Solids and Dissolved Fractions
Each sample collected was separated into a solids-associated and dissolved fraction via centrifugation. Specifically, one milliliter of each sample was placed in a sterile, DNA-free tube and subjected to centrifugal force (9600× g) for five minutes. Larger volume samples (50 mL) collected during the pilot-scale filtration experiments were separated by centrifugal force (6000× g) for ten minutes. The supernatants (dissolved fraction) were decanted into a sterile DNA-free tube with the pellets (solids-associated fraction) remaining in the original tube and stored at −80 °C until the DNA was extracted.
2.4. DNA Extraction
The DNA extraction of the fractionated wastewater samples was performed with the QIAPrep® spin miniprep (solids fraction) and the QIAquick® gel extraction kit (dissolved fraction) according to the manufacturer’s instructions (Qiagen Inc., Hilden, Germany). Elution volumes for the full-scale plant samples were 40 µL and 30 µL for the pilot-scale filtration experiments. Samples collected from the full-scale WRP were extracted on the same day that the samples were collected. Samples collected from the pilot-scale dual-media filter experiments were extracted within one week of collection. The DNA was divided into individual single-use aliquots and stored at −80 °C.
2.5. AdvIPC:pGEM-T and AdvIPC:pSMART Vectors
Two plasmids were constructed to contain a DNA sequence not found in nature [
41]. The synthetic DNA sequence (termed Adv IPC; synthesized by Integrated DNA Technologies Inc., Coralville, IA, USA) was inserted into the pGEM-Teasy vector (Promega Co., Madison, WI, USA) and pSMART
® GC LK vector systems (Lucigen Co., Middleton, WI, USA) using standard molecular biological techniques including restriction digestion and ligation. The pGEM-Teasy vector was used with the pilot-scale filtration experiments whereas the pSMART
® vector, which did not contain a β-lactamase gene, was used as an internal control for the HFF concentration of secondary and final effluents. See
supplementary materials for additional sequence information and plasmid map.
2.6. Quantitative PCR Primers/Probes and Plasmids
Quantitative PCR primers and probes targeting the β-lactamase and sulfonamide resistance genes were used for the analysis of wastewater matrices. Additionally, qPCR assays were utilized to assess total bacterial biomass (
16S rRNA gene) and to determine recovery/loss of DNA (AdvIPC) during sample concentration by HFF. A sample processing control (SPC) was used to monitor DNA extraction and qPCR inhibition in all samples analyzed. The SPC consisted of approximately 120 ng of salmon sperm DNA (Ambion, Waltham, MA, USA) added into each sample prior to DNA extraction which was subsequently quantified by qPCR. If inhibition was identified, the sample was diluted and re-analyzed. Primer and probe sequences for all qPCRs are listed in
Table 1. Each
sul1 qPCR consisted of 1× SsoAdvanced™ Universal SYBR
® Green Supermix (BioRad, Hercules, CA, USA), 120 nM primer and 1 µL template in a final volume of 25 µL. The
blaSHV/TEM and
16S qPCRs had a final concentration of 1× SsoAdvanced™ Universal Probes Supermix (BioRad, Hercules, CA, USA), 320 nM each primer and probe, 3 µL template in a total volume of 25 µL. The AdvIPC qPCR contained the same reaction constituents as the
blaSHV/TEM but the final probe concentration was 200 nM. The SPC qPCR consisted of 1× SsoAdvanced™ Universal Probes Supermix, 1 µM each primer, 800 nM probe, 2 µL template in a final volume of 25 µL. The qPCR reactions were performed and analyzed using either a RotorGene (Qiagen Inc., Hilden, Germany) or LightCycler
® 480 (Roche, Basel, Switzerland) real-time thermocycler. The qPCR efficiency, limit of detection (LOD), and regression statistics (mean square error for the LightCycler
® 480 software and r
2 for the RotorGene software) were determined for each reaction (see
Supplementary Materials). The specificity of the
blaSHV/TEM PCR was confirmed by DNA sequence analysis of the amplified products. All primers and probes were procured from Integrated DNA Technologies (Coralville, IA, USA). Data analyses were performed using the software associated with each respective thermocycler (LightCycler
® 480 software version 1.5.1.62, Roche Diagnostics Corporation, Indianapolis, IN, USA, and RotorGene Q software version 2.1.0, Qiagen Inc., Hilden, Germany). Full-scale WRP samples were extracted in triplicate and each extract analyzed singularly (RotorGene) or in triplicate (LightCycler
® 480 analyses) by qPCR. Samples were only deemed positive if at least two of the three extractions resulted in a positive signal. Pilot-scale dual-media filtration samples were extracted and analyzed singularly (RotorGene). Turbidity readings were also taken for the full-scale WRP samples and are provided in the
Supplementary Materials. The qPCR cycling conditions along with the average efficiencies and linearity data are provided in the
Supplementary Materials along with additional information on the development of the qPCRs.
2.7. Pilot-Scale Dual-Media Filtration
A pilot-scale dual-media filter was constructed using a PVC tube (15.24 cm diameter) and the same depth of sand (31 cm) and anthracite coal (61 cm) as is used in the full-scale WRP filter beds. The flow was controlled via an electric pump and monitored with a flow meter. The filter was operated at hydraulic loading rate of 3.0 to 3.8 liters per minute (L/min), (166 L/(min m2) to 209 L/(min m2)) similar to full-scale filters used at the WRP (122–204 L/(min m2)).
The pilot-scale filtration experiments were conducted using a laboratory modified plasmid (AdvIPC:pGEM-T) added to secondary effluent. Briefly, 379 L of clarified secondary effluent from a WRP was added to a 2271 L polyethylene tank and dosed with the AdvIPC:pGEM-T plasmid to achieve a final concentration of 1.95 × 108 copies/mL in the secondary effluent. Samples were collected at a total of six locations/time points: (1) prior to the addition of plasmid, (2) after addition but before filtration, (3–5) filtrate at 20, 90 and 100 min. from the start of the filtration and (6) after backwashing the filter. Dissolved and solids-associated fractions were evaluated for exogenous plasmid concentrations in each of the samples as described above. Additionally, the filtration experiments were performed under both “dirty” and “clean” conditions: (1) using a filter that had been in service for over four weeks to simulate a bio-fouled filter, and (2) using a filter treated with high-dose chlorine to simulate a clean filter condition.
The concentration of the exogenous plasmid was determined by qPCR in each of the samples upon collection and again after chlorine treatment (10 mg/L for 45 min followed by dechlorination). Liquid hypochlorite was prepared using Clorox as a stock solution (5.84% available chlorine) and diluted with ultrapure water (double reverse osmosis, carbon filtered, UV disinfected). The Clorox was stored in a light impenetrable container at room temperature. Chlorine was diluted and used within two days. Total chlorine concentrations were measured prior to treatment via the Hach DPD colorimetric method. Treatment doses were based on the total chlorine concentrations measured using the DPD method.
2.8. Statistical Analyses
Statistical analyses were performed using either Microsoft Excel 2010 (version 14.0.7190.5000, Microsoft Corporation, Redmond, WA, USA) or SigmaPlot version 11.0 software (Systat Software, Inc., San Jose, CA, USA).
4. Discussion
Antibiotic resistance genes have been identified as contaminants of emerging concern in environmental matrices and could be subject to regulatory monitoring requirements in the future. However, ARGs have been found in every environment on earth and their presence alone does not necessarily signify a major public health concern. The primary factors which have led to the current public health crisis with respect to antibiotic resistance are the horizontal transfer of resistance genes to antibiotic susceptible bacteria coupled with the prodigious use of antibiotics which provided a positive selective pressure resulting in the rapid dissemination of AR across the globe. Therefore, it becomes important to not only identify clinically relevant ARGs in the environment but to also characterize their ability to be transferred between bacteria.
The ARGs chosen for this study represent both clinically relevant and transmissible genes suggesting they are appropriate indicators of ARGs in wastewater. In particular, β-lactams are some of the most prescribed antibiotics worldwide [
46]; moreover,
bla ARGs are prevalent in human pathogens as well as in wastewater [
47,
48] and have shown the ability to undergo HGT [
49]. The
sul1 gene is known to be associated with class I integrons on conjugative plasmids [
50,
51] and is one of the most abundant ARGs in wastewater [
52].
Previous data combined from several reports show that concentrations of ARGs can vary throughout wastewater treatment depending on the ARG studied, the extent of treatment, geographical location, and operational parameters (reviewed in [
52]). The results presented herein show that the concentrations of
blaSHV/TEM and
sul1 were similar to what has been reported by others for raw wastewater, activated sludge, secondary and tertiary effluent although the
sul1 concentrations were slightly higher in the activated sludge than what has been reported previously [
52]. The tertiary WRP examined in this study produced ARG reductions of approximately four-log
10 in the final disinfected effluent compared to the raw sewage concentrations which are some of the largest overall reductions reported for activated sludge–type WWTPs [
53,
54,
55,
56,
57]. The full extent of the β-lactamase reductions in the disinfected final effluent from the full-scale WRP, could not be precisely quantified because the values for all samples analyzed were below the detectable limit of the qPCR assay.
Although
sul1 and
blaSHV/TEM concentrations were greatly reduced after activated sludge treatment, the ARGs were still present in both fractions in the secondary effluent. In contrast, the
blaSHV/TEM genes could not be detected in tertiary-treated final effluent (after chlorination) even upon concentrating the water samples over 200-fold. The
sul1 gene was detected in higher numbers throughout each of the treatment stages compared to the
blaSHV/TEM yet, was only detected in three of six concentrated final effluent samples for the solids fraction and five of six dissolved fraction samples. This indicated that substantial removal occurred during the tertiary filtration and disinfection processes (solids: >2-log
10 and dissolved: >3-log
10 reduction). Pilot-scale filtration experiments were incorporated to further study the media filtration and disinfection processes. The results indicated that tertiary filtration coupled with chlorine disinfection provided a synergistic benefit with respect to removal of ARGs compared to chlorination alone. More specifically, chlorine mediated reduction of ARGs was more effective on filtered effluents compared to non-filtered effluents. The increased effectiveness of the chlorine in the filtered water may reflect the removal of substances that increase chlorine demand during filtration. As a consequence, more chlorine would be available in the filtered water compared to the unfiltered secondary effluent resulting in greater degradation of the ARGs. A previous report found that disinfection of tertiary-treated effluent waters provided less than a one-log
10 reduction in ARG concentration [
55]. Several factors may explain the discrepancy between the results presented here and Munir et al. (2011): the WRP studied here incorporated biological ammonia removal through a nitrification and denitrification process lowering ammonia levels in the effluent resulting in less chloramine formation. Therefore, the removal rates documented here may differ for WWTPs that do not incorporate ammonia removal processes. Specifically, other investigators have reported that chloramines are less effective in reducing ARG concentrations in wastewater compared to free chlorine [
58]. It should be noted that the effects of ammonia on ARG removal were not specifically evaluated as part of this project. Additionally, the geographic locations of the WWTPs differed in the two studies and the operational parameters were not reported; both factors could have contributed to differences in the presence and removal of ARGs. Taken together, these data indicate that WWTPs that include the use of tertiary filtration with disinfection can provide additional ARG reductions and thereby further minimize any potential public health and environmental impacts compared to those without filtration.
The pilot-scale tertiary filtration and chlorine disinfection experiments resulted in a removal of at least five log10 of dissolved ARG plasmid. Very little ARG removal was attributed to the filtration process itself (less than 0.9 log10) whereas the chlorinated, filtered effluent decreased the ARG concentration to undetectable limits, signifying a substantial role for chlorine in the overall removal. Treatment of secondary effluents with chlorine successfully reduced ARG concentrations however; tertiary filtration was shown to enhance the chlorine mediated removal of ARGs and provided at least an additional order of magnitude more reduction compared to chlorinated secondary effluent alone. The full-scale WRP data showed an average removal of greater than two log10 between secondary and tertiary filtered, disinfected effluents indicating that further optimization of the filtration and disinfection processes could produce additional ARG reductions. The use of pilot-scale processes offered some experimental advantages not available at the full-scale level. In particular, the pilot-scale filtration permitted the study of ARG removal using a single homogenous plasmid added into the secondary effluent immediately prior to filtration whereas, the full-scale plant experiments detected ARGs from multiple sources of DNA (plasmids, genomic, phage, etc.) that could interact with elements present throughout the treatment process. Furthermore, chlorination of the pilot-scale samples was performed under controlled conditions in the laboratory where the precise concentration and contact time were monitored along with the use of fresh liquid chlorine thus, maximizing its oxidative potential. Future studies focused on the optimization of ARG removal by full-scale filtration and chlorination processes would provide additional information on how to further reduce these constituents.
Currently, it is difficult to directly assess the total HGT activity in wastewater ecosystems. While there are culture-based methods available that can approximate transfer in particular model organisms or by certain mechanisms, they have drawbacks [
59]. The use of non-culture-based methods such as PCR offer the ability to identify ARGs and transmissible elements from essentially all sources in a sample and because HGT mechanisms are associated with cellular or extracellular fractions, PCR analysis of ARGs in each fraction can provide information concerning which HGT pathways have a greater likelihood of leading to ARG transfer. This approach assumes that higher concentrations of ARGs in a particular fraction correlate with an increased chance of HGT occurring by the pathways associated with each fraction. In support of this, transformation rates in water have been shown to increase with increasing gene concentrations [
60] and the same logic would be expected to apply with transduction-mediated HGT. With respect to the solids-associated fraction, conjugation rates in water have been shown to be dependent on the number of cells containing the transferrable gene [
61].
While not a direct assessment of HGT, analyzing the cellular and extracellular fractions for ARGs provided some insight into where these genes reside during different treatment stages as well as which HGT mechanisms may have a higher probability of occurring in each water type. For example, the ratio of solids to dissolved ARGs was at least 100-fold higher in the activated sludge compared to the tertiary-treated, disinfected effluent, which would suggest that HGT via conjugation would be more likely to occur in the activated sludge compared to the final effluent. Note that there are multiple factors that can affect the horizontal transfer of genes in the environment and while the fractionation experiments give some insight into the probability of one transfer pathway occurring over another it is not a direct analysis of HGT. Given the importance of HGT with regard to spread of AR, a more thorough study targeting HGT in these waters would be beneficial.
The concentrations of dissolved and solids-associated ARGs differed between the treatment processes. The dissolved concentrations of both ARGs were highest in the raw sewage and showed reductions at each successive stage of treatment (raw > activated sludge > secondary effluent > tertiary effluent). In contrast, the concentrations of both ARGs and 16S DNA in the solids fraction increased in the activated sludge compared to the raw sewage. The sul1 and 16S genes increased in a similar fashion while the blaSHV/TEM demonstrated a smaller increase than the other two genes (statistically significant) suggesting the genes may be present in separate microbial populations. This trend was not observed in subsequent stages as both ARGs were reduced substantially in the solids fraction of the secondary and final effluents (raw < activated sludge > secondary effluent > final effluent).
These results would be consistent with a model in which ARG concentrations increase in conjunction with the expansion of microbial populations during the biological phase of treatment (activated sludge) followed by a decrease in the amount of solids associated ARGs in the secondary effluent resulting from floc settling and solids removal. Finally, an increase in the percentage of dissolved ARGs observed after tertiary filtration and disinfection could result from the chlorine-mediated destruction of bacterial cells releasing additional DNA in the aqueous fraction. Reductions in viable indicator bacterial concentrations after tertiary filtration and disinfection have been well documented and, coupled with data that demonstrate the presence of extracellular
Bacteroides DNA in the absence of the viable bacteria after wastewater treatment [
62], lend support to this theory. Additionally, a recent disinfection study using enterococci containing the
vanA resistance gene demonstrated that
vanA DNA could be detected after chlorination of secondary effluent when the concentration of viable enterococci was reduced to below detectable limits [
63]. What effect, if any, the release of dissolved DNA has on the presence and dissemination of ARGs in waters that receive treated effluents is currently unknown.
Presently there is no consensus as to whether ARGs are selected for or against during wastewater treatment with published reports illustrating both scenarios [
17,
20,
24,
54,
64,
65,
66,
67,
68,
69]. However, the activated sludge treatment stage in particular, has been proposed as a potential source for ARG transfer [
56]. This study examined samples from each stage of treatment and evaluated the data for positive and negative selection by comparing the two target ARGs to the number of
16S rDNA genes (an indicator of the total bacterial biomass). The ratio of
sul1 to
16S did not change significantly throughout each stage of treatment (for both fractions) suggesting this ARG was neither selected for or against. The
blaSHV/TEM:
16S ratio in the solids fraction decreased between the raw sewage and activated sludge matrices in a statistically relevant manner indicating that the number of
blaSHV/TEM genes in relation to the total bacterial biomass went down compared to the raw sewage. The solids fraction ratio increased slightly in the secondary effluent compared to the activated sludge (not statistically significant) but was still below the ratio observed in the raw sewage. A comparison of the dissolved fractions showed little change between the raw and activated sludge
blaSHV/TEM:
16S but decreased significantly in the secondary effluent (statistically significant). The fact that the ARGs did not increase at a significantly larger rate than the total bacterial population suggests that they were not selected for during the treatment processes with respect to the overall bacterial populations present at each stage. In total, the data showed no evidence that the two ARGs analyzed in this project were being selected for during wastewater treatment. It should be noted that not all ARGs were evaluated in this study. While the two resistant genes selected for this project represent highly prevalent and clinically relevant genes, the possibility cannot be ruled out that other genes might be selected for during wastewater treatment. In addition, operational parameters have been shown to affect ARG concentrations [
19,
70] thus ARG reductions may differ between facilities operating under different conditions.