Presence of Antimicrobial-Resistant Bacteria and Resistance Genes in Soil Exposed to Wastewater Treatment Plant Effluent

Abstract

Antimicrobial resistance (AMR) has become a world-wide health issue, and anthropogenic antibiotics entering the environment is cause for concern with regard to impacts on environmental bacteria. As water resources have become scarcer, reuse of wastewater treatment plant (WWTP) effluent has increased, creating a conduit for environmental antibiotic pollution. The aim of this study was to determine the impact of spray-irrigating effluent on the incidence of AMR in soil organisms in agricultural lands (Astronomy Site, Pennsylvania State University). This study performed culture work to assess resistance of Gram-negative and Gram-positive soil bacteria to four antibiotics (sulfamethoxazole, trimethoprim, ciprofloxacin, and ampicillin) and molecular work (qPCR) to quantify genes associated with AMR (sulI, sulII, ermB, and intI1). Compared to a control site, Gram-negative bacteria at the Astronomy Site appeared to have increased resistance to sulfamethoxazole and trimethoprim. Higher numbers of resistance genes by depth (below 35 cm) and by location were consistently observed at the Astronomy Site with copy numbers of some genes up to 10⁶-fold higher than the control site. Increased quantities of sulI and intI1 in the top 0–5 cm of the soil profile appeared to be dependent upon the amount of effluent irrigation received, whereas the presence of sulII and ermB showed the opposite patterns. Overall, long-term reuse of WWTP effluent to spray irrigate cropped lands does appear to alter and possibly increase AMR in soil environments; however, additional work is necessary to determine potential impacts on human, wildlife, plant, and soil health.

1. Introduction

Since the late 1990s, low levels of pharmaceuticals and personal care products (PPCPs) have been detected in waterways and soil systems throughout the world [1,2,3]. One of the main pathways for PPCPs to enter into the environment is through discharge of wastewater treatment plant (WWTP) effluent, either directly via disposal into natural waterways or as spray irrigation over agricultural land [4]. Although concern of acute toxicity is low, pharmaceuticals are designed to elicit highly specific physiological responses [5,6]. Therefore, release of sub-clinical concentrations of PPCPs into the environment may result in unexpected biological and ecological impacts in target and non-target organisms that alter the dynamics within an ecological system [7].

Antibiotics are an area of rising concern, especially for human and animal health, due to selection and development of drug-resistant bacteria that limit the drugs available to fight infections [8]. As cases of resistance increase, the integrity and efficacy of antibiotics come into question, especially with the development of pathogens with multidrug resistance [9]. While in the past, antimicrobial resistance genes (ARGs) in the environment and clinical health settings have been viewed as discrete entities, recently, ARGs in the environment have been linked to human pathogens [10,11]. Possible scenarios for shared resistance between human and animal pathogens and the environment are influxes of anthropogenic antibiotics into the environment creating selective pressures for environmental bacteria to evolve cross-resistance to commonly prescribed antibiotics [12]. Then, due to natural water cycling, bacteria and/or ARGs may be transported throughout the environment, human populations, and animal husbandry operations [13].

Sewage treatment systems are not currently equipped to fully remove antibiotics [14]. Even though total bacterial numbers are significantly reduced by typical wastewater treatments, complete removal of antibiotics has been determined to be incomplete [15,16,17]. A main concern about this uncontrolled release of anthropogenic antibiotics into waterways is the conduit created for intimate interactions of environmental bacteria with synthetic antibiotics. In addition, ARGs in the wastewater may alter microbial communities and negatively impact ecological systems. Due to the nature of wastewater reuse, the presence of anthropogenic antibiotics in effluent and, consequently, in aquatic environments, may allow antibiotics, ARGs, and antimicrobial-resistant bacteria (ARB) to cycle through ecosystems and eventually impact ecological, human, and animal health. One possible approach to mediate the release of anthropogenic antibiotics into waterways is the use of soils to complete the tertiary treatment of effluent. Tertiary treatment of wastewater utilizes both chemical and physical processes to remove harmful contaminants. Utilizing soil for tertiary treatment will not only potentially remove contaminants but will also conserve water resources [18,19]; however, by exposing natural systems to trace levels of PPCPs, this reuse practice may impact the presence of AMR in soil bacteria.

Recent research has shown that anthropogenic and agricultural antibiotics released into the environment have variable impacts on AMR in bacteria depending on the system: water, soil, or sediment [20,21,22,23,24]. The increased presence of ARGs and ARB in waterways downstream of WWTPs has been well documented [13,25,26]. The impact of effluent on AMR when reused for the purpose of irrigating soil systems appears to be significantly lower than what has been observed in waterways and sediments [27,28]. However, the number of studies analyzing the impact of effluent reuse on AMR in soil systems is limited and given the variability in climates and soil types as well as the potential impacts in human and animal populations, additional studies are necessary to fully understand the impact of effluent reuse on the presence of ARGs and ARB in soil systems [29].

The main goal of this study was to assess the levels and types of AMR in a soil system impacted by the introduction of anthropogenic antibiotics and other PPCPs due to irrigation with WWTP effluent and to compare those results with natural levels of AMR in areas not impacted by WWTP effluent. A preliminary study with a limited number of samples was performed to assess possible alterations in AMR at a site receiving effluent for the purpose of irrigating croplands. This initial study analyzed the resistance of cultivated soil bacteria to four antibiotics (sulfamethoxazole, trimethoprim, ciprofloxacin, and ampicillin) and quantified the copy numbers of two ARGs (sulI and sulII). A full-scale study was later performed to quantify four antibiotics (sulfamethoxazole, trimethoprim, ofloxacin, and lincomycin) and a suite of ARGs (sulI, sulII, and ermB) and a class 1 integron-integrase (intI1). The ARGs were selected based on the importance of the antibiotics to which they confer resistance in human health. IntI1 was selected because this gene has been associated with the presence of ARGs, and it has been suggested as a proxy for anthropogenic pollution [30].

2. Materials and Methods

2.1. Study Site

Pennsylvania State University (University Park, PA, USA), which is located in central Pennsylvania at 40.7982° N, 77.8599° W, has a unique water renovation site called the Living Filter, where the University Park WWTP effluent is spray irrigated on grassed, cropped, and forested lands year-round to allow tertiary treatment in the soil profile. The site has been in full-scale operation since 1982. The equivalent of 5 cm of effluent is permitted to be applied to the land weekly and passes through a soil profile of ≈30.5 m. Groundwater wells are located around the site to monitor the efficacy of the tertiary treatment in the soil (Figure 1). Maintaining the quality of the Living Filter groundwater is imperative since this system feeds into the Spring Creek watershed, a premier cold-water stream for Pennsylvania trout and a tributary of the Chesapeake Bay Watershed.

During the time frame of this study, the land was planted with wheat (Triticum aestivum). Soil samples were collected prior to planting and after harvest [31]. The soil at the site has been mapped primarily as a Hagerstown silty clay loam; a fine, mixed, semiactive, mesic Typic Hapludalf; and Hublersburg silt loam, a clayey, illitic, mesic Typic Hapludult with approximate pH of 6.5. For additional information on soil properties, see the Supplementary Material (Table S1).

2.2. Sample Collection

For the initial work, soil cores were collected from two sites: a control-cropped site at Rock Springs that had not received irrigation and a cropped site at the Astronomy Site of the Living Filter. All samples were collected using a clean 25 mm diameter metal corer to obtain 15 cm long cores. A depth of 15 cm was selected due to preliminary work demonstrating that this was the expected maximum depth for antibiotic movement in the soil profile with downward water movement through the soil pores during each wetting event. For the full-scale assessment of AMR, soil cores were taken at a control-cropped site at Rock Springs that had not received irrigation and a cropped site at the Astronomy Site of the Living Filter. For both the preliminary work and full-scale assessment, the control and Astronomy Site locations had similar soil types and crop rotations.

At the Astronomy Site during the full-scale assessment, soil samples were collected two times after 1 week and 10 weeks of irrigation, in triplicate, from summit and depression locations (Figure 1). Sampling locations were 15 m from a solid set spray irrigation head (direct irrigation), which is approximately the midpoint between two laterals of a solid set sprinkler system, and 75 m from the last irrigation head (indirect irrigation to capture any impacts due to drift from the irrigation system’s last irrigation lines). At the 15 m distance, soil samples were taken in between two irrigation lines and on the side of an irrigation line that was the last one before a wooded area (i.e., location would only receive effluent irrigation from one irrigation line). At the 15 m distance, soil would receive the most direct spray and largest quantity of effluent during an irrigation event [32] and the soil sampling location in between two irrigation lines would receive even more effluent, which was confirmed by placing rain meters and monitoring quantity of water received during one irrigation event. Drift cards, which are sensitive to water, were placed at 75 m to provide visual confirmation that irrigation drift occurred at that distance from the spray irrigation lines. Furthermore, whether a location was at a summit or depression would also impact the amount of effluent received since depression locations could receive run off from summit locations.

The agricultural site where irrigated soil samples were collected had not received effluent irrigation for 7 months over the fall/winter. The first sampling (Sampling #1) occurred one day following the first 12 h period of effluent irrigation. Soil core samples were taken with a hydraulic Giddings probe (Giddings Machine Company, Windsor, CO, USA) down to an 80 cm depth at summit and depression locations to account for possible run off from summits to depressional areas. The second sampling period (Sampling #2) occurred approximately 20 weeks after Sampling #1. The site had received a total of 10 weeks of effluent irrigation events (12 h interval with ≈5 cm effluent/week) over the summer into early September, with the final irrigation event occurring the day prior to sampling. Summit and depression locations were sampled using a hydraulic soil sampler down to an 80 cm depth.

Samples for preliminary analysis were stored in sterile, sealed bags and kept on ice until reaching the laboratory, while full-assessment samples were kept in plastic soil sampling sleeves with end caps until reaching the laboratory. A HOBO device (UA-001-64 Pendant Temperature/Alarm (Waterproof) 64K Data Logger) (Onset: Bourne, MA, USA) was utilized to monitor the ambient temperature of soil samples during the sampling period. Soil samples for preliminary analysis were then segmented into 5 cm increments, and full-assessment samples were divided into 0–5 cm, 5–15 cm, 15–30 cm, and 30–80 cm segments. All samples were stored at 4 °C in sterile, sealed bags to await further analysis. Samples were analyzed for ARB and/or DNA was extracted within one week following collection.

2.3. Chemicals

Sulfamethoxazole (SMX) (99.9%), trimethoprim (TMP) (99.7%), ciprofloxacin (CIP) (99.8%), and ampicillin (AMP) (≥95.0%) were used (Sigma-Aldrich Inc., St. Louis, MO, USA) for making serial dilutions. Water used during this study was deionized (18 MΩ) for making general stocks and reagents and for culturing. SYBR Green I nucleic acid gel stain (96%) was used for qPCR work and obtained from Molecular Probes (Eugene, OR, USA). MacConkey agar and Mueller–Hinton agar were used for isolation and culturing soil bacteria (ThermoFisher Scientific, Waltham, MA, USA).

2.4. Analysis of Antimicrobial-Resistant Bacteria (Culture-Based Analysis)

Bacterial isolates were cultured from soil samples during the preliminary analysis work to determine if bacteria were present that had resistance to one or more of the four antibiotics (SMX, TMP, CIP, and AMP) at break point concentrations. Bacteria were selectively enriched for Gram-positive and Gram-negative bacteria in Mueller–Hinton (MH) broth with added Staph/Strep selective supplement (Thermo Scientific) and Escherichia coli (EC) broth, respectively. After incubation for 16–20 h at 35 ± 2 °C, cultures for Gram-positive and Gram-negative bacteria were streaked on MH and MacConkey agar, respectively. These plates were incubated for 12–16 h at 35 ± 2 °C. Isolates were selected from each plate and streaked on the appropriate agar and incubated for 12–16 h at 35 ± 2 °C. Five colonies were then collected from the plates of selectively cultured Gram-negative and Gram-positive bacteria and added to 5 mL of molecular-grade water, from which 50 μL were added to 10 mL of double strength MH and EC broth to achieve a turbidity of 0.5 McFarland standard.

Serial dilutions of SMX, TMP, CIP, and AMP (μg mL⁻¹) were made based on the expected breakpoint threshold of susceptibility for Gram-negative and Gram-positive bacteria for each antibiotic, individually, according to CLSI guidelines [33]. Breakpoints for SMX, TMP, CIP, and AMP were 32 μg mL⁻¹, 16 μg mL⁻¹, 4 μg mL⁻¹, and 32 μg mL⁻¹ (Gram-negative)/4 μg mL⁻¹ (Gram-positive), respectively. An aliquot (50 mL) of each pure colony suspension in double strength broth was added to an equal aliquot (50 mL) of each antibiotic dilution in a 96 well plate with blank broth as controls. Plates were incubated for 16–20 h at 35 ± 2 °C. The plates were then read both visually and on a spectrophotometer at 600 nM to obtain an absorbance reading that correlates with cell growth. The minimum inhibitory concentrations (MICs) of each antibiotic against the soil isolates were determined by these serial dilutions.

2.5. Isolation of DNA from Soil

For preliminary analysis, DNA was extracted from soil (0.25 ± 0.015 g, field moist) using a MO BIO PowerSoil DNA Isolation Kit (Carlsbad, CA, USA—now DNeasy, Germantown, MD, USA). Full assessment analyses utilized the MO BIO PowerLyzer PowerSoil DNA Isolation Kit (Carlsbad, CA, USA—now DNeasy, Germantown, MD, USA). The steps were followed according to the directions of the manufacturer. DNA concentrations per sample were quantified using a Nanodrop 2000 UV–VIS spectrophotometer (Thermo Scientific, Wilmington, DE, USA) with averages per sample and standard deviations recorded. Extracting nucleic acids from soil and removing contaminants is quite challenging, but this procedure recovered DNA concentrations of 10–20 μg/g of soil (field moist) with average purities of 1.9 (260/280 nm).

2.6. Analysis of Resistance Gene Copy Numbers in Soil Bacteria (Non-Culture-Based Analysis)

Resistance gene copy numbers were determined for each sample utilizing quantitative real-time PCR (qPCR). Gene-specific primers (Table S2) and intercalating dye SYBR Green I (Molecular Probes, Inc., Eugene, OR, USA) were used to quantify three ARGs (sulI, sulII, and ermB) commonly found in human pathogens and environmental bacteria, an integrase (intI1) commonly associated with the presence of ARGs in environmental bacteria, and the 16S rRNA gene. qPCR reactions for preliminary work looking only at sulI and sulII were performed using a BioRad iCycler (Primer Design, Inc., Southampton, UK). Analyses for preliminary work were performed using the BioRad iQ5 Real Time PCR Detection System and software (version: 2.1.1022.0523) (BioRad, Hercules, CA, USA). For the full-scale study looking at sulI, sulII, ermB, and intI1, an Applied BiosystemTM 7500 Fast Real Time PCR System and software (version: 2.3) (Life Technologies, Fort City, CA, USA) were used for qPCR reactions and analyses. For each gene, the reaction mixture consisted of 1× SYBR Green PCR Master Mix (Qiagen, Valencia, CA, USA); 0.2 μM of each primer; 1 μL or 2.5 μL of template for preliminary work or full scale study, respectively; and sterile DNase/pyrogen free water to create a final volume of 25 μL. A specific temperature and cycling program was optimized for each gene analyzed (Table S2) [34].

Primers, annealing temperatures, reaction protocols, amplicon lengths, and sequences can be found in Table S2. The qPCR runs included a standard curve that covered six orders of magnitude, and each sample was analyzed in triplicate. Standards were created using gBlocks Gene Fragments (Integrated DNA Technologies, Coralville, IA, USA). The r² values for all standard curves were >0.98, and efficiencies for amplification ranged from 73 to 100%. The ARG and intI1 gene copy numbers were normalized to 16s rRNA gene copy numbers to generate relative gene abundances. Relative gene abundances of target genes were also normalized to 1 g of soil (defined as relative concentrations: copies per gram). Approximately 20% of the samples were spiked with a GFD gene marker (1 μL—300,000 gene copies) before extraction of total DNA to account for the DNA extraction efficiency. The GFD gene marker was avian specific (gull, geese, chickens, and ducks) and was not expected to be present in the cropped lands at the Living Filter Astronomy Site [35]. Extraction efficiencies were 80% or higher for all samples analyzed.

2.7. Statistical Analysis

Mean values and standard deviations were calculated for each bacterial group and treatment. One-way analysis of variance (ANOVA) was performed on plate count data (CFU g⁻¹ dry soil) analyzing culture medium and soil type. A repeated-measures ANOVA model was used to determine if treatment was significantly related to abundance of ARGs. Correlation analysis (PROC CORR) was used to calculate Pearson correlation coefficients (r) and p-values between ARGs and ARGs and antibiotic concentrations. Highly correlated variables were defined as r > 0.34. Statements of statistical significance refer to p < 0.05. The software utilized for computation of statistics were SAS (version 9.2, SAS Institute, Cary, NC, USA) and Excel (version 16.30, Microsoft, Redmond, WA, USA). Limit of blank was determined first (mean of blank + 1.645 × standard deviation of blank). Limit of detection (LOD) and limit of quantification (LOQ) were determined by mean of blank + 3.3 × standard deviation of blank and mean of blank + 10 × standard deviation of blank, respectively (Table S2) [36].

3. Results and Discussion

Throughout the year, the soil at the Astronomy Site of Penn State’s Living Filter receives, on average, weekly WWTP effluent irrigation events that result in the soil bacteria being consistently exposed to varying concentrations of antibiotic compounds as well as other PPCPs. The effluent that is applied to the soil is known to contain concentrations of four antibiotics that range from <LOD for LIN to 22 μg L⁻¹ for SMX (Figure 2) with higher concentrations during the spring and fall and lower concentrations during the summer [19]. This seasonal variation could be due to various reasons, including lowered removal of antibiotic compounds during colder months; changes in the population density of University Park, a college town, where the population is higher during fall and spring; and/or the number of antibiotic prescriptions rising during seasons of higher bacterial infections (colder months) [37,38]. Concentrations of these human antibiotics have also been detected in the soil profile to a depth of 80 cm (Figure 3). At the first sampling event (Sampling #1), which occurred one day after the first 12 h period of effluent irrigation in seven months, SMX and OFL were detected throughout the soil profile, but TMP and LIN were below the limit of detection. After 10 weeks of effluent irrigation events (Sampling #2), all four antibiotics were detected within the soil profile. Concentrations of the antibiotics were always higher at the soil surface (0–5 cm), with maximum average concentrations of 650 ± 204 ng kg⁻¹ for OFL, 730 ± 360 ng kg⁻¹ for SMX, 190 ± 71 ng kg⁻¹ for TMP, and 13 ± 13 ng kg⁻¹ for LIN [19].

3.1. Preliminary Analysis of AMR

Preliminary AMR work at the Astronomy Site resulted in notable findings with regard to resistance at the site. Antimicrobial resistant bacteria (Gram-negative) cultivated from the Astronomy Site demonstrated resistance to TMP and SMX above MICs, whereas this level of resistance was absent at the control site (Figure 4A). Resistance above MIC levels to CIP (a commonly prescribed fluoroquinolone similar to ofloxacin) was absent at both the Astronomy Site and the control site. Quantification of a sulfonamide resistance gene, sulI, showed that while the presence of this resistance gene near the surface (0–10 cm) at the Astronomy Site did not differ from a control site, the quantity of sulI further down the profile (10–15 cm) was higher than at the control site (Figure 4B).

These findings from the preliminary work suggest that the long-term irrigation with wastewater effluent may be impacting AMR at the Astronomy Site. However, this preliminary work was not conclusive given the natural variability within a soil system across a specific landscape combined with the limited number of samples taken and analyses performed. As a result, a larger scale assessment of AMR was developed and performed to confirm the validity of these findings. The focus of the full-scale study was quantification of select ARGs and a gene for a class 1 integron-integrase associated with AMR. Culture work was not included in the subsequent full-scale study due to various factors, including the time intensiveness of culturing for ARB and that >99% of soil bacteria are not cultivable [39,40,41].

3.2. Full-Scale Analysis of a Suite of Three ARGs and intI1

The full-scale analysis of AMR at the Astronomy Site of the Living Filter analyzed four genes: three ARGs (sulI, sulII, and ermB) and an integron-integrase gene (intI1) that has been associated with AMR. This full-scale study showed that alterations in AMR were present at the Astronomy Site when compared to a control site. Such alterations varied by the depth in the soil profile, the class of gene, the amount of effluent irrigation received, and/or the location at the site (Figure 5 and Figure 6). Unlike the preliminary work, samples were taken down to a depth of 80 cm, which provided greater insight into the long-term implications of spray irrigating effluent for nearly 40 years. Not only were antibiotic concentrations found at a depth of 80 cm (Figure 3) [19], but quantities of ARGs and intI1 were also present (Figure 5 and Figure 6). While the control site did not receive any irrigation, locations sampled at the Astronomy Site received (i) direct irrigation of effluent from the spray heads from one lateral (cropped 15 m), (ii) direct irrigation of effluent from two laterals (cropped 15 m between laterals), or (iii) indirect irrigation due to drift that is known to occur at the site due to weather and wind patterns (cropped 75 m) (Figure 1). At each of these sampling locations, samples were taken after one week of irrigation (1 week irrigation) and 10 weeks of irrigation (10 weeks irrigation). For certain genes, the amount of irrigation received made a difference in the number of gene copies present, whether it was direct vs. indirect irrigation (sulI and intI1), and/or 1 week vs. 10 weeks of irrigation (intI1, sulI, sulII). In addition, two different types of elevation levels were sampled at the Astronomy Site: cropped summits and depressions due to the possibility of runoff from summits into depressional areas leading to a greater influx of effluent and antibiotic load in depression locations (Figure 3), and, subsequently, higher copy numbers in depression locations for certain genes analyzed (sulI, ermB). For the quantities (gene copies/16S rRNA in 1 g of soil) of the three ARGs (sulI, sulII, and ermB) and an integron-integrase gene (intI1) at each sampling location and sampling event, see Table S3.

3.2.1. Comparison of Gene Abundance by Soil Profile Depth

A major trend that was consistent for each of the genes analyzed and helped confirm the results from the preliminary work with sulI was that by depth, the quantities of resistance genes and intI1 found at the Astronomy Site were higher than the control site. In fact, the quantities of sulI, intI1, and ermB at the Astronomy Site at a depth of 35–80 cm were significantly higher than the control site at the same depth (Figure 5 and Figure 6). This trend held true, regardless of location at the Astronomy Site and amount of effluent irrigation received. For example, with sulII and ermB, the quantities of these resistance genes in the upper 0–15 cm of the soil profiles at all locations of the Astronomy Site were not significantly different (higher) than at the control site at the same depth (Figure 5 and Figure 6). However, below 35 cm, the quantities of these two ARGs increased at the Astronomy Site so that all locations had higher numbers of the ARGs than the control site. In fact, for ermB, the gene was not detected below 15 cm at the control site.

Given this trend of higher numbers of ARGs and intI1 by depth at the Astronomy Site, the long-term application of WWTP effluent for the purpose of irrigating crops appears to have implications for AMR development and maintenance, especially at depths where increased resistance may not be expected in the soil environment. Most interesting are the results for the macrolide ARG, ermB, which is completely absent at the control site below a depth of 15 cm but found at the Astronomy Site down to 80 m. Soil is a naturally limiting environment with most of the nutrients for bacterial growth and survival near the surface where plants are abundant and organic matter is present. Typically, the size and diversity of the bacterial community will diminish by depth in the soil profile as organic matter decreases. Concurrently, the proportion of ARGs should either remain similar or decrease by depth. One theory as to why this is not occurring at the Living Filter could be due to the presence of antibiotics reaching the depths of 80 cm in the soil profile, which would not only create selective pressure for bacteria with resistance to those compounds but may also provide a carbon source in an otherwise carbon-limiting environment for those bacteria that are resistant to those antibiotics [42].

3.2.2. Gene Abundance by Type

The actual quantities of each type of gene analyzed did vary from one another at the Astronomy and control sites. Notably, the overall abundance of the macrolide resistance gene, ermB, was about 1,000,000-fold higher than any other gene analyzed, regardless of location, including the control site. Macrolide resistance genes are some of the most common ARGs in the environment and are found even in pristine soils [43], most likely due to the natural origin of macrolide antibiotic compounds in soil environments and environmental bacteria requiring a defense mechanism to survive [44,45]. Therefore, this higher overall abundance of ermB compared to the other ARGs and intI1 is most likely driven by the levels that would naturally occur in the soil environment, which is evidenced by the control site having similar quantities of ermB in the top 15 cm of the soil profile compared to sampling locations at the Astronomy Site (Figure 5 and Figure 6).

However, as noted previously, ermB persists down through the soil profile at the Astronomy Site, while it disappears below 15 cm at the control site. The presence of this ARG, especially by depth at the Astronomy Site, is important since ermB has been recommended as an indicator to evaluate overall levels of Macrolide–Lincosamide–Streptogramin (MLS) resistance genes in the environment [46]. Furthermore, macrolides are an important medicine for human use since they are broad-spectrum antibiotics effective against Gram-positive bacteria and Mycoplasma and can be substituted for penicillin clinically [47]. An increase in this resistance gene in the environment may be of concern for human health due to its clinical importance [48].

While the higher levels of ermB near the surface may be partly due to resistance naturally found in the soil environment, the persistence of this resistance through the soil profile at the Living Filter site is most likely due to long term irrigation with WWTP effluent, which is confirmed by correlation analysis of the quantities of this gene with antibiotic concentrations. However, ermB is the only gene analyzed in this study that significantly correlated (p < 0.05) with the presence of antibiotic compounds and was positively correlated with OFL, TMP, and LIN at the Astronomy Site locations. The fact that only ermB strongly correlated with antibiotic concentrations helps demonstrate that the presence of antibiotics in the soil environment is not the only driving factor in the alterations in AMR resulting from wastewater reuse at this site.

3.2.3. Gene Correlations

Only two of the genes analyzed during this study were positively correlated with one another (intI1 and sulI), while the two other genes (sulII and ermB) had different patterns with no correlation with other genes. This finding is of note, because intI1 has been recommended as a proxy for anthropogenic pollution and the presence of AMR, since it is a class 1 integron-integrase and has been found to be associated and strongly correlated with ARGs [30]. However, that correlation with other ARGs did not hold true in this study, since sulI was the only ARG with which intI1 was significantly correlated (p < 0.05). Positive correlations of intI1 with sulI, sulII, and ermB have been reported previously [49,50,51,52]; however, these studies were performed in either surface waters, a different matrix, or soil environments impacted by manure and/or dairy wastewater, which would have higher concentrations of antibiotics than treated effluent from a wastewater treatment plant. Therefore, based on this research, utilizing only intI1 to determine anthropogenic impacts in soil environments impacted by wastewater reuse may not provide an accurate assessment of environment pollution in those systems.

3.2.4. Impacts of Amount of Effluent Irrigation and Location on Gene Abundance

During this study, the total amount of effluent irrigation that a site received varied based on the type of irrigation (direct from one or two laterals or indirect irrigation at the Astronomy Site versus no irrigation at the control site), whether it was located in a depression or on a summit, and number of weeks of irrigation received (1 week versus 10 weeks). The impact of the amount of effluent received by a particular location on the three ARGs (sulI, sulII, and ermB) and intI1 were most clearly observed in the surface of the soil profile from 0 to 5 cm. For sulI, the total amount of effluent received made a clear difference in the quantity of the gene present in the soil profile at the surface (Figure 5 and Figure 6). On the other hand, the presence and quantities of intI1, sulII, and ermB in the upper 0–5 cm of the soil profiles had varying patterns with some locations that received fewer effluent applications having higher quantities of the gene than the locations that received the greatest amount of direct effluent.

The highest quantities of the sulI gene were present in the depression locations after 10 weeks of irrigation with 1.76 × 10⁻³ ± 2.27 × 10⁻³ and 1.50 × 10⁻³ ± 2.41 × 10⁻³ gene copes/16S rRNA in 1 g of soil at sites receiving irrigation from one lateral and two laterals, respectively (Figure 5). The lowest quantities of the gene were found at the 75 m Astronomy Site location (indirect irrigation via drift) and the 15 m cropped summit location receiving 10 weeks of direct effluent from one irrigation lateral with 8.41 × 10⁻⁵ ± 5.49 × 10⁻⁵ and 1.58 × 10⁻⁴ ± 6.33 × 10⁻⁵ gene copies/16S rRNA in 1 g of soil, respectively (Figure 5). Both locations had lower concentrations than background levels of sulI present at the control site (2.39 × 10⁻⁴ ± 3.23 × 10⁻⁵ gene copes/16S rRNA in 1 g of soil) with the 75 m location being significantly lower than the control site (p < 0.05). Of note, after one week of irrigation, the 15 m depression locations had quantities of sulI that were similar to the control site; however, after 10 weeks of irrigation, these locations had significantly higher quantities of this resistance gene compared to the control site (Figure 6). These trends for this resistance gene mirror what occurred in the soil with concentrations of SMX, the antibiotic compound to which sulI confers resistance. Quantities of SMX were higher in the depression sites compared to the summit sites, most likely due to effluent runoff from the summits into depressions. Additionally, while SMX was present in the soil surface (0–5 cm) after one irrigation event, its concentration increased in the profile after 10 weeks of effluent irrigation.

While the intI1 gene was positively correlated with sulI when taking into account all sample locations and depths of the soil profile, the presence of intI1 near the surface did not follow as clear of a pattern as sulI with regard to sites receiving the most effluent having the highest quantities of the resistance gene. The location that had the highest quantities of this gene was the 15 m summit site after 10 weeks of irrigation from two laterals with 1.94 × 10⁻⁵ ± 2.90 × 10⁻⁵ gene copies/16S rRNA in 1 g of soil (Figure 5). However, the location with the second highest quantities of intI1 was the 15 m depression location after only one week of effluent from one lateral with 1.64 × 10⁻⁵ ± 8.8 × 10⁻⁶ gene copies/16S rRNA in 1 g of soil (Figure 6). While this location received the least amount of direct effluent irrigation (one week from one lateral), it had higher quantities of the gene than the 15 m depression location that received the highest amount of direct irrigation over 10 weeks of irrigation from two irrigation laterals (6.96 × 10⁻⁶ ± 9.55 × 10⁻⁶ gene copies/16S rRNA in 1 g of soil). Furthermore, when comparing the 15 m summit versus 15 m depression locations between two laterals after 10 weeks of irrigation, the 15 m summit site had the highest quantities of the intI1 gene, even though it most likely received less effluent infiltration than the depression site due to run off from summits to depressions. The location that had the lowest quantity of intI1 was the 75 m site that received indirect irrigation (8.60 × 10⁻⁷ ± 3.24 × 10⁻⁷ gene copies/16S rRNA in 1 g of soil) and would have received the least amount of effluent irrigation (Figure 5). All of the locations that received direct irrigation had higher quantities of intI1 than the control site (1.09 × 10⁻⁶ ± 3.78 × 10⁻⁷ gene copies/16S rRNA in 1 g of soil); however, they were not significantly higher (p > 0.05). Overall, for intI1 within the 0–5 cm range of the soil profile, the quantities of this gene were typically higher in sites that received the most effluent irrigation and lowest in sites that received the least; however, a few aberrations do exist such as a location receiving only one week of irrigation from one lateral having higher or similar quantities of intI1 as sites receiving 10 weeks of effluent from two laterals.

The presence of the resistance gene sulII within the 0–5 cm of the soil profile had the opposite trend of sulI, even though both genes confer resistance to the same class of antibiotics (sulfonamides). The locations that had the highest quantities of the resistance gene were the 15 m depression sites that received the least amount of direct irrigation with one week of irrigation from one lateral and two laterals (1.44 × 10⁻⁴ ± 1.86 × 10⁻⁴ and 4.09 × 10⁻⁴ ± 2.59 × 10⁻⁴ gene copies/16S rRNA in 1 g of soil, respectively) (Figure 6). Both of these locations had significantly higher quantities of sulII than the control site (1.49 × 10⁻⁵ ± 2.95 × 10⁻⁶ gene copies/16S rRNA in 1 g of soil) (p < 0.05). The rest of the locations (summits and depressions) even after 10 weeks of irrigation applications had lower quantities of sulII than the depression locations that only received one week of irrigation. Of the locations receiving 10 weeks of irrigation, only the 15 m summit location between two laterals had significantly higher quantities of this resistance gene (2.84 × 10⁻⁵ ± 1.23 × 10⁻⁵ gene copies/16S rRNA in 1 g of soil) than the control site (p < 0.05). Even though the 15 m summit and depression locations had received 10 weeks of irrigation from one lateral, they had the lowest quantities of sulII with 1.03 × 10⁻⁵ ± 6.45 × 10⁻⁶ and 1.00 × 10⁻⁵ ± 4.34 × 10⁻⁶ gene copies/16S rRNA in 1 g of soil, respectively.

Similar to sulII in the top 0–5 cm of the soil profile, ermB was most abundant at the locations that received the lowest amount of effluent irrigation throughout the study and least abundant at the sites that received the most effluent. The 15 m depression sites after one week of effluent irrigation from one lateral and two laterals had the highest quantities of ermB with 8.98 × 10² and 6.39 × 10² ± 1.64 × 10² gene copies/16S rRNA in 1 g of soil, respectively, and were significantly higher than the control site (1.72 gene copies/16s rRNA in 1 g of soil) (p < 0.05) (Figure 6). The locations with the lowest quantities of ermB were the 15 m depression and summit locations between two laterals after receiving effluent applications for 10 weeks (8.07 × 10–2 ± 1.12 × 10⁻¹ and 3.4 × 10⁻² ± 3.83 × 10⁻² gene copies/16S rRNA in 1 g of soil, respectively). These sites even had gene copy numbers lower than the control site (Figure 5).

While nothing conclusive can be stated about the exact reason for these patterns in gene copy numbers for intI1, sulII, and ermB at the Astronomy Site based on this study, higher copy numbers in samples that received less direct effluent may be the result of what is present in the effluent water (e.g., a specific antibiotic compound or contaminant) instead of the quantity of effluent received or due to other environmental factors (physical, chemical, and/or biological) not taken into account during this study. For intI1, all sites that received direct effluent applications had higher quantities than the control site, which infers that effluent itself is having an impact. However, for sulII and ermB, not all of the sites receiving direct effluent irrigation had higher copy numbers of those resistance genes than the control site, which could mean that other environmental factors could be altering and/or impacting the presence of the genes.

Given that out of the four genes analyzed during this study, only sulI followed a clear pattern of increasing quantities of the gene as more effluent was applied, this demonstrates the complexity of AMR within soil systems impacted by WWTP effluent. Typically, a frequent concern with regard to anthropogenic antibiotics entering soil environments is the creation of selective pressures for bacteria to harbor certain ARGs or other genes, specifically genes that they normally would not possess in non-impacted environments and that may impact human and/or animal health [53]. This concern is based on the assumption that as more anthropogenic antibiotics or other selective agents enter the soil environment, then the presence of ARGs and/or other genes related to resistance will increase. While this phenomenon has been well documented within soil systems impacted by manure, which contains higher overall concentrations of antibiotics [54,55,56,57], this research analyzing the reuse of wastewater shows that the presence of antibiotics (and other emerging contaminants, such as PPCPs) at relatively low concentrations does not always result in increased presence of ARB and ARGs.

4. Conclusions

Overall, based on preliminary work and a larger scale study analyzing four different genes associated with AMR (sulI, sulII, ermB, and intI1), long-term reuse of wastewater that contains low levels of antibiotics does appear to impact the presence of AMR in the soil even down to a depth of 80 cm. The most significant trend at the Astronomy Site where effluent is used for irrigation year-round was that the presence of these resistance genes persisted down through the soil profile and in many cases either remained at the same level (quantities) as the surface or increased down through the soil profile. In fact, at a depth of 35–80 cm, the presence of all four genes at the Astronomy Site was higher at all sampling locations and time points compared to the control site. Also, of note, while ermB was absent from the control site below a depth of 15 cm, this gene was still present at the Astronomy Site. The application of effluent does appear to alter the level of resistance near the surface (0–5 cm) so that some locations at the Astronomy Site had significantly higher quantities of certain genes compared to the control site; however, a clear pattern of more effluent applications and higher quantities of a gene was only observed with sulI, which demonstrates the complexity of AMR within soil systems impacted by WWTP effluent. While this work does provide new information about the potential of long-term impacts on AMR in soils receiving effluent irrigation, additional work is necessary to determine other factors that may be altering AMR within soil systems impacted by WWTP effluent and if AMR in groundwater and crops are potentially impacted, both of which could more directly impact human and animal populations.