Identification of Antibiotic-Resistance-Gene Molecular Signatures

Feb 16, 2010 - Animal feeding operations (AFOs) and wastewater treatment plants (WWTPs) are potential sources of antibiotic resistance genes (ARGs) in...
17 downloads 15 Views 966KB Size
Download PDF
Environ. Sci. Technol. 2010, 44, 1947–1953

Identification of Antibiotic-Resistance-Gene Molecular Signatures Suitable as Tracers of Pristine River, Urban, and Agricultural Sources H. STORTEBOOM,† M. ARABI,† J . G . D A V I S , ‡ B . C R I M I , †,§ A N D A . P R U D E N * ,†,| Departments of Civil and Environmental Engineering and of Soil and Crop Sciences, Colorado State University, Fort Collins, Colorado 80523

Received September 23, 2009. Revised manuscript received January 18, 2010. Accepted January 29, 2010.

Animal feeding operations (AFOs) and wastewater treatment plants (WWTPs) are potential sources of antibiotic resistance genes (ARGs) in rivers and/or antibiotics that may select for ARGs in native river bacteria. This study aimed to identify ARG distribution patterns that unambiguously distinguish putative sources of ARG from a native river environment. Such molecular signatures may then be used as tracers of specific anthropogenic sources. Three WWTPs, six AFO lagoons, and three sites along a pristine region of the Cache la Poudre (Poudre) River were compared with respect to the frequency of detection (FOD) of 11 sulfonamide and tetracycline ARGs. Principlecomponent and correspondence analyses aided in identifying the association of tet(H), tet(Q), tet(S), and tet(T) (tet group HQST) with AFO environments and tet(C), tet(E), and tet(O) (tet group CEO) with WWTPs. Discriminant analysis indicated that both tet group HQST and tet group CEO correctly classified the environments, but only the tet group HQST provided a significant difference in FOD among the environments (p < 0.05). Sul(I) was detected in 100% of the source environments but just once in the pristine Poudre River, which was dominated by tet(M) and tet(W). Tet(W) libraries generated from the pristine Poudre River, WWTPs, and AFO lagoons were also discernible based on restriction fragment length polymorphism and phylogenetic analysis. Thus, a novel approach was developed and demonstrated to be effective for the model river system, taking an important step in advancing the fundamental understanding of ARG transport in the environment.

Introduction Antibiotic resistance is recognized by the World Health Organization and the Center for Disease Control as a critical * Corresponding author phone: (540) 231-3980; email: [email protected]. † Department of Civil and Environmental Engineering, Colorado State University. ‡ Department of Soil and Crop Sciences, Colorado State University. § Current address: Department of Biotechnology, Delft University of Technology, Julianalaan 67, 2628 BC Delft, The Netherlands. | Current address: Department of Civil and Environmental Engineering, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061. 10.1021/es902893f

 2010 American Chemical Society

Published on Web 02/16/2010

public health challenge of our time (1, 2). Because new drug discovery is no longer capable of keeping pace with emerging antibiotic-resistant infections (3), strategies for prolonging the efficacy of available antibiotics by reducing the spread of antibiotic resistance must be developed (1–3). It has been demonstrated that the diversity and concentration of antibiotic resistance genes (ARGs) can be influenced by landuse and anthropogenic activity (4–8). Still, the primary mechanism(s) involved in the proliferation and dissemination of ARGs via environmental pathways has not been determined. Up to 90% of consumed antibiotics pass through the body unmetabolized into manure (9). Excreted antibiotics can then enter river environments through a variety of pathways, including point discharges from wastewater treatment plants (WWTPs), animal feeding operations (AFOs), and fish hatcheries or via nonpoint sources such as overland flow from fields where manure or biosolids have been applied. Antibiotics have been detected typically at part per trillion concentrations in many natural and engineered environments, such as surface waters and sediments (10, 11), municipal WWTPs (12), AFO lagoon systems (13), and fields fertilized with animal manures or biosolids (14). The presence of antibiotics, even at subinhibitory concentrations, can stimulate bacterial metabolism and thus contribute to the selection and maintenance of ARGs (15, 16). Thus, the presence of antibiotics could drive the elevated levels of ARGs that have been observed previously in anthropogenically impacted river environments (4, 5, 17). Alternatively, we propose that non-native bacteria possessing ARGs, or ARGs as free DNA, may themselves be transported to surface waters via similar pathways as antibiotics. Once present in rivers, ARGs are capable of being transferred among bacteria, including pathogens, through horizontal gene transfer. The fate and transport of ARGs is expected to be unique relative to other contaminants, considering that they may be amplified in the presence of selecting agents such as antibiotics, be transferred between diverse types of bacteria, and exist as either intra- or extracellular entities. Direct molecular analysis provides a means to access ARGs present in uncultured bacteria, explore the extent of their diversity, and also directly quantify the total pool of resistance determinants via quantitative polymerase chain reaction (QPCR). Previous investigations of ARGs in source environments have generally been limited to the culturing of one organism (e.g., Hu et al. (8)) or the molecular characterization of only one source (human or animal; e.g., Koike et al. (7)). The need for improved characterization of the background biogeography of ARGs has been duly noted (18), and a comprehensive molecular-level analysis of ARGs in pristine river environments is particularly lacking. Additionally, differentiation of human and animal sources of ARGs can shed light on areas where intervention may be most effective in helping to reduce the spread of ARG contaminants through environmental matrixes. The overall objective of this study was to identify patterns of ARG distribution that are suitable for discrimination between sources of ARGs and native river environments. In support of the objective, the frequency of detection (FOD) of 11 tetracycline and 2 sulfonamide ARGs, along with tet(W) restriction fragment length polymorphism (RFLP) phylotype and phylogenetics, were determined at pristine sampling sites at the headwaters of the Cache La Poudre (Poudre) River and compared with putative urban (WWTPs) and agricultural (AFO lagoons) sources using exploratory statistics to identify the characteristic distributions of ARGs. The overall approach is intended to have broader value for the identiVOL. 44, NO. 6, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1947

TABLE 1. Description of Environments and Sites environment

AFO lagoons

WWTPs wildlife trout-rearing unit pristine Cache La Poudre (Poudre) River

site

size/location

beef feedlot 1 beef feedlot 2 large organic dairy large conventional dairy small organic dairy small conventional dairy WWTP1 WWTP2 WWTP3 trout-rearing unit Poudre River-0a Poudre River-0b Poudre River-1

18 000 head cattle 28 000 head cattle 1800 cows 6000 cows 150 cows 800 cows 0.5 million gallons per day 14.3 million gallons per day 1.5 million gallons per day 50 000 trout Elephant Rock, Roosevelt National Forest Profile Rock, Roosevelt National Forest Greyrock trailhead, Roosevelt National Forest

fication of appropriate molecular signatures suitable as tracers of anthropogenic sources in other watersheds, taking a significant step in advancing the fundamental understanding of ARG fate and transport in the environment.

Experimental Section In the text, the term “site” refers to all samples from a specific location at all time points. The term “environment” refers to all sites of a particular classification (i.e., pristine Poudre River, AFO lagoon, WWTP, or trout-rearing unit; Table 1). Pristine Environment: Upstream Poudre River Water and Sediment Samples. The upstream portion of the Poudre River has a pristine source arising from snowmelt in the Rocky Mountains with few tributaries or anthropogenic influences (Figure S1 in the Supporting Information). The Poudre River watershed has been analyzed with respect to both antibiotics and ARGs in previous works (4, 5, 11). To date, no antibiotics have been detected at Poudre River-1, though ARGs were found to be present (4). Thus, Poudre River-1 and two sites further upstream located at the Elephant Rock landmark (Poudre River-0a) and the Profile Rock landmark (Poudre River-0b) were chosen for analysis in this study. River water and sediment samples were collected three times during lowflow conditionssOctober 2006, February 2007, and October 2007sand once during high-flow conditionssMay 2007. Water was collected in sterile 1 L Nalgene bottles from the approximate midpoint of the river. A metal spade was used to collect ∼0.5 kg of sediment from the top 5 cm of the riverbed. The sample was mixed, and a portion was collected in sterile 50 mL centrifuge tubes. Samples were preserved on ice for transport to the laboratory (maximum 12 h). Source Environments: AFO Lagoons, Municipal WWTPs, and Trout-Rearing Unit. Water and settled solids from AFO lagoons were collected in coordination with river sampling events. The AFO lagoons (small organic dairy, large organic dairy, small conventional dairy, large conventional dairy, and beef feedlots) analyzed in this work have been described previously (19). In April 2008 and April 2009, 250 mL of mixed-liquor activated sludge from three municipal WWTPs and 50 mL of anaerobically digested and dewatered biosolids from WWTP1 were collected. In April 2009, 500 mL of effluent was also collected at each WWTP. All samples were stored on ice for transport. Two of the WWTPs, WWTP1 and WWTP3, treat nearly 100% municipal wastewater, whereas WWTP2 also receives 15% commercial and 20% light industrial wastewater (Table 1). A wildlife trout-rearing unit was the only known potential point source of ARGs in the pristine region of the Poudre River. Therefore, water from the discharge stream and the underlying sediment were collected from two separate raceway discharges in October 2008. 1948

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 6, 2010

Sample Preparation and DNA Extraction. Suspended sediments from water were collected via filtration through presterilized membranes [0.45 µm mixed-cellulose esters (Millipore)] placed in sterile 10 mL Petri dishes and frozen prior to DNA extraction. Filters were cut into small pieces and added directly to extraction tubes. A total of 1 g of sediment was extracted in duplicate using the UltraClean Soil DNA Kit (MoBio Laboratories, Inc.) according to the manufacturer’s instructions. Lagoon and activated sludge samples were centrifuged at 6000 rpm for 10 min to obtain pellets (0.25 g), which were extracted using the PowerSoil DNA Kit (MoBio Laboratories, Inc.) according to the manufacturer’s instructions. The DNA quality was verified by agarose gel electrophoresis and diluted 1:10 with sterile water for PCR assays. PCR Assays. A total of 47 AFO lagoon, 11 WWTP, 24 pristine Poudre River, and 4 hatchery samples were screened by PCR for the presence of 4 tetracycline efflux genes [tet(B), tet(C), tet(E), tet(H)], seven ribosomal protein protection genes [tet(M), tet(O), tetB(P), tet(Q), tet(S), tet(T), and tet(W)], and two sulfonamide resistance genes [sul(I) and sul(II); Table S2a,b in the Supporting Information]. Primers and PCR assays have been described previously (4, 7, 20, 21). A typical reaction mixture (25 µL total volume) for PCR of tet genes contained 1X ExTaq Buffer (TaKaRa Bio Inc., Pittsburgh, PA), 0.5 µM each primer, 0.2 µM each dNTP, 1.25 units of TaKaRa ExTaq polymerase, and 1 µL of template. The presence of two sul genes, sul(I) and sul(II), was determined using PCR assays described in Pei et al. (4). The reaction mixture (25 µL total volume) contained 5 µL of 5X buffer, 2.5 µL of 10X buffer, 0.2 µM dNTPs, 0.2 µM each primer, 1.75 units of Taq polymerase, and 1 µL f template. The thermoprofile for all of the genes [except tet(H)] was as follows: initial denaturing at 94 °C, followed by 40 cycles of 15 s at 94 °C, 30 s at the annealing temperature (Table S1 in the Supporting Information), 30 s at 72 °C, and a final extension step for 7 min at 72 °C. The thermoprofile for tet(H) has been described previously (7). Aliquots of PCR products of 10 µL were mixed with 1X SYBR Green DNA gel stain (Invitrogen, Carlsbad, CA) and loading dye and analyzed by electrophoresis on 2% agarose gels (w/ v). Negative and positive controls (PCR products with their identity verified by sequencing or cloned plasmids donated by K. McMahon) were included with all PCR reactions. Samples in which the ARG of interest was not detected were randomly subjected to a spiked matrix test to determine if PCR inhibition may have contributed to false negatives. Spiked controls were 2-3 orders of magnitude greater than the limit of detection (1000 genes), defined as the last serial dilution of controls that produced a visible band on the gel. Spiked controls were not inhibited in any of the matrixes examined (all sample matrixes were tested for inhibition of at least one gene).

TABLE 2. FODs of 13 ARGs in AFO, WWTP, Trout-Rearing-Unit, and Pristine River Environments tetracycline efflux ARGs

tet ribosomal protection ARGs

sul ARGs

environment

tet(B)

tet(C)

tet(E)

tet(H)

tetB(P)

tet(M)

tet(O)

tet(Q)

tet(S)

tet(T)

tet(W)

sul(I)

sul(II)

AFO lagoons WWTPs trout-rearing unit pristine river

0.04 0.00 0.00 0.00

0.77 0.91 0.75 0.08

0.28 0.45 0.00 0.00

0.89 0.36 0.75 0.04

0.04 0.00 0.00 0.00

1.00 0.91 0.00 0.25

0.85 0.91 0.00 0.08

0.87 0.45 0.00 0.00

0.49 0.00 0.00 0.00

0.96 0.45 0.00 0.04

0.96 1.00 0.25 0.33

1.00 1.00 1.00 0.04

0.94 0.82 0.25 0.17

Shot-Gun Cloning and RFLP Analysis of tet(W). A 1152bp region of the tet(W) gene was amplified in six AFO samples and three WWTP samples with 1.25 units of high-fidelity TaKaRa ExTaq as described by Koike et al. (7) with primer (2 pM) and Mg2+ (4 mM) concentrations optimized for the DNA matrixes of all samples. PCR products were cloned using the TOPO-TA cloning kit (Invitrogen), and libraries were generated directly from clones by standard PCR incorporating M13 primers specific to the pCR4-TOPO vector. A restriction digest assay for the amplicons was designed using NEBcutter V2.0 (New England Biolabs, Inc.) based on sequences with the accession numbers DQ309636, DQ309637, DQ309647, DQ309651, DQ309659, DQ309659, DQ309667, DQ309687, DQ309688, and DQ309691 and five sequences cloned from organic dairy sludge. Clones generated from AFO lagoon samples (n ) 311) and WWTP activated sludge (n ) 162) were subjected to a BstUI restriction digest incorporating 2 µL of M13 PCR product, 18 µL of New England Biolabs Buffer 2, and 1 unit of restriction enzyme. Digests were carried out at 60 °C for 1 h and visualized on 2% agarose gel. Selected clones were sequenced in both directions by Colorado State University’s Proteomics and Metabolomics facility using T3 and T7 primers. Large amplicons were assembled using ChromasPro 1.41 software (Technelysium Pty Ltd.). Sequence alignments and neighbor-joining tree constructions were carried out using MEGA, version 4 (22). tet(W) trees were rooted to the Aquifex aeolicus fusA gene (AE000669) (23) and used to perform similarity analysis between environments using the online environmental analysis software UniFrac (available at http:// bmf2.colorado.edu/unifrac/index.psp) (24). UniFrac significance tests and jackknife cluster analysis were performed using abundance weights, determined as the number of identical sequences for each environment. Jackknife resampling was set at 100 permutations, with the smallest sample size as the minimum number of sequences kept in the analysis. Sequences were deposited in GenBank under accession numbers GU116770-GU117056. PCR amplification of the 1152-bp region of tet(W) ARGs from pristine sites was inhibited for unknown reasons. The 167-bp region generated in the PCR assay was sequenced (accession numbers: GU596412-GU596413) and compared to the existing clone library of 1152-bp tet(W) amplicons. To reduce bias that could have been introduced by the use of a different reverse primer, sequences that did not match 100% identity with the reverse priming region (148-167 bp) were eliminated from analysis (n ) 201 sequences). The remaining sequences (n ) 129) were aligned, and trees were created as described above. PCR Detection Assays. The FOD of each gene at a sampling site for all time points was calculated as the total number of detections (ARG+) of that ARG divided by the total number of samples at the site (n) (eq 1). FOD ) ARG+/n

(1)

Correspondence analysis, a descriptive statistical approach, was utilized as a means to capture dependencies between the presence, persistence, and FOD of analyzed

ARGs in the source and pristine environments and also to identify combinations of ARGs that could potentially discriminate between environments. Raw categorical responses (FOD) of each ARG were analyzed in SAS 9.1 (SAS Institute Inc., Cary, NC) with the PROC CORRESP method. The Statistical Toolbox in Matlab R2008a (The MathWorks Inc., Natick, MA) was used to implement analysis of variance (ANOVA) to test the statistical significance between the FOD of groups of correlated ARGs hypothesized to be able to correctly classify environments.

Results and Discussion This study represents a comprehensive direct molecular characterization and comparison of ARGs occurring in urban (WWTPs), agricultural (AFOs), and pristine river environments within the same watershed. Through principlecomponent and correspondence analyses of ARG FODs and tet(W) RFLP phylotypes, characteristic patterns of ARG distribution were identified that were distinct among the two putative sources and the native river environment. The unique distribution of ARGs identified among the environments was further corroborated by phylogenetic analysis of tet(W). FOD of ARG Determined by PCR. The overall FOD of ARGs was observed to be the highest in the AFO lagoon environment, with each ARG being detected in at least two samples. Because tet(B) and tetB(P) were each found at only one site (beef feedlot 2 and the small conventional dairy, respectively), they were unlikely to be suitable indicator genes and thus were excluded from downstream analysis. The FOD of the 11 remaining ARGs was highest in AFO lagoons (0.82), followed by WWTPs (0.66), the trout-rearing unit (0.30), and the native Poudre River (0.09) (Table 2). The sul(I) gene was detected in 100% of AFO lagoons, WWTPs, and trout-rearingunit samples but only once in the upstream river at site Poudre River-0b, immediately downstream of the trout-rearing unit. Thus, the presence of sul(I) may be an indicator of influence from a non-native environment. The results of the correspondence analysis of 11 ARGs are presented in Figure 1, providing a comparison of the distribution of ARGs at each site in the pristine Poudre River, AFO lagoon, and WWTP environments. It was observed that sites representing the pristine Poudre River, WWTP, and AFO lagoon environments were each clustered together, respectively, indicating the uniqueness of the ARG FOD to an environment. The loading of tet(S) was clearly associated with the AFO lagoons, as indicated by both the direction and magnitude of the corresponding vector (Figure 1). In fact, tet(S) was detected at least once in each AFO lagoon, but not in any other environment. All other ARGs were detected in at least 75% of the AFO samples. The ribosomal protection protein genes, tet(Q) and tet(T), were positively correlated and also were potential predictors of AFO influence, being found in 87% and 96% of AFO samples, respectively, but only 45% of WWTP samples. With respect to WWTPs, tet(E) was also a major contributor to the variation in the data and was loaded in the direction of all three WWTPs with an overall FOD of 0.46. The FOD of VOL. 44, NO. 6, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1949

FIGURE 2. Comparison of the total FODs of tet(C), tet(H), tet(M), tet(O), tet(T), tet(W), sul(I), and sul(II) in upstream Poudre River sites and the trout-rearing unit. The total FOD was calculated as the sum of all detections of each ARG divided by the total number of possible detections (n × 11 ARGs total).

FIGURE 1. Correspondence analysis of ARG probabilities. (A) Number of positive detections of each ARG within the AFO, WWTP, and pristine Poudre River (PR) environments subject to correspondence analysis. The x axis represents the dimension (dimension 1) explaining 47.89% of the total variation; the y axis represents the second principal component (dimension 2), explaining 24.78% of the total variation for 72.67% cumulative explained variability. Vectors from the origin to the ARG points indicate the loading of the presence/absence of 11 ARGs [tet(B) and tetB(P) were excluded from further analysis because each ARG was detected in only one AFO lagoon]. (B) Plot of ARG variable FODs of tet group HQST [tet(H), tet(Q), tet(S), and tet(T)] vs FODs of tet group CEO [tet(C), tet(E), and tet(O)]. tet(E) in the AFO environment was 0.28, and while tet(E) was detected in each dairy lagoon, it was not detected in beef feedlot lagoons. Similarly, tet(C) was loaded in the same direction as tet(E), and its overall FOD was 0.91 in WWTP compared to 0.77 in AFOs. On the basis of correspondence analysis, it would follow that tet(H), tet(Q), tet(S), and tet(T) (tet group HQST) and tet(C), tet(E), and tet(O) (tet group CEO) could be used as indicators of AFO and WWTP sources, respectively. To explore this possibility, the average normalized FODs of tet group HQST and tet group CEO were plotted against each other for pristine Poudre River values, AFOs, and WWTPs (Figure 1B). It was found that the three environments could clearly be discriminated in this manner (Figure 1B). While both tet groups HQST and CEO appear to have the potential for source discrimination, ANOVA revealed that only tet(HQST) provided 1950

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 6, 2010

a significant difference in the FODs among the environments in this study (p < 0.05; Figure S2 in the Supporting Information). It is important to note that detection/nondetection of a particular ARG by PCR is not exactly equivalent to its presence/absence because an ARG may be present below the detection limit. The detection limit is largely influenced by the nature of the matrix, including the presence of PCR inhibitors and nontarget DNA. In this study, the background matrix was variable among the three environments, which could arguably lead to variable detectability of the ARG. Therefore, particular attention was given in this study to verify negative signals through a randomized spiked matrix inhibition test, which yielded no false negatives. Also, the AFO lagoon matrix, which would be expected to contain the highest levels of PCR inhibitors, actually yielded the highest FODs. False-positive results are less of a concern to this study, given that the primers employed are highly specific, have been widely tested in the literature (4–8, 20, 21), and have been verified to amplify only target sequences when products amplified from complex environments have been sequenced (7, 20, 21). Comparison of Upstream Poudre River to the TroutRearing Unit. A lower ARG FOD was expected in the upstream region of the Poudre River because of the low level of anthropogenic influence. In total, all but tet(E), tet(Q), and tet(S) were detected. Because of the location of the troutrearing unit within the upstream Poudre River watershed, its possible impact on the river was also explored. Commercial aquaculture facilities are well-known to have an intense impact on the surrounding environment because of the release of ammonia and other pollutants (25), and, interestingly, even antibiotic-free feeds employed have been noted to harbor ARGs (26). However, wildlife fish-rearing units are likely to have stricter regulations concerning antibiotic use. Still, the trout-rearing-unit environment was distinct from the upstream Poudre River in both its overall detection and distribution of ARGs (Figure 2). Whereas the upstream Poudre River was dominated by tet(M) and tet(W), tet(M) was detected just once in the trout-rearing unit and no other tet ribosomal protection protein ARGs were detected (FOD 0.04). In contrast, sul(I), tet(C), and tet(H) were the most commonly detected ARGs in the trout-rearing unit. It is noteworthy that although these ARGs were also detected in the pristine Poudre River, they were only found at site Poudre River-0b, located immediately downstream of the trout-rearing unit, suggesting influence at this site. Interestingly, upstream (Poudre River0a) and downstream (Poudre River-1) sites exhibited greater

similarity to each other than to Poudre River-0b. Thus, it is likely that any impact of the trout-rearing unit attenuates as the river flows downstream through pristine wilderness to the mouth of the canyon. These observations are in agreement with published literature regarding the predominance of tetracycline efflux ARGs in similar aquatic environments (27, 28). tet(C), which was predominant in the trout-rearing unit, has also been isolated from the fish pathogen Aeromonas hydrophila (29). RFLP Analysis of tet(W). tet(M) and tet(W) were frequently detected in the pristine Poudre River, AFO, and WWTP environments. tet(M) has been reported to be found in 69 genera and tet(W) in 24 genera, including aerobic, anaerobic, Gram-positive, and Gram-negative genera (28, 30). Both ARGs are located on conjugative transposons, mobile genetic elements that code for their own transfer, can integrate themselves into a host’s chromosomal DNA, and can pass from organism to organism (28). These factors may contribute to their widespread distribution across environments, both with and without anthropogenic disturbances, as described here and elsewhere (28), which makes them prime candidates for obtaining higher resolution source information via sequence analysis. In a previous work, tet(W) phylogenetics was successfully applied for linking groundwater ARGs to AFO lagoon sources, providing contrast with the “native” upgradient environment (7). tet(W) was thus selected in this study for further comparison of environments via RFLP phylotyping and phylogenetics. A total of 95% of all clones generated four RFLP patterns (patterns 1-4) that were easily identifiable between libraries. Additional unique patterns were compared for further classification. In Figure 3A, correspondence analysis of the number of the conserved restriction patterns normalized to total clones from each library is presented. A clear distinction between the AFO and the WWTP tet(W) clone libraries was observed, suggesting greater diversity within the AFO environment. In terms of loadings, pattern 2 appeared to be an indicator of AFO lagoons and accounted for 17% of AFO clones but only 1% of WWTP clones. Pattern 1 was loaded in the direction of the WWTP libraries and represented 9% of all WWTP clones versus 3% of AFO clones. RFLP patterns 3 and 4 exhibited a split loading between the two environments and were negatively correlated (loaded in opposite directions). Thus, these patterns were not strong indicators of either environment and were less useful for the classification of AFOs and WWTPs. On the basis of these results, higher proportions of pattern 1 versus pattern 2 were predicted for environments similar to WWTPs versus AFOs, respectively. Discrimination between source environments was achieved as expected by plotting the percent distribution of RFLP patterns 1 and 2 (Figure 3B). Phylogenetic Analysis of the tet(W) ARG. In Figure 4A, a phylogenetic tree representing 100% of the clones exhibiting patterns 1, 2, and 4 and 50% of pattern 3 clones is presented (n ) 223 AFO + 107 WWTP clones). The number of clones selected from each pattern for sequencing and phylogenetic analysis was based on the relative sequence diversity observed from preliminary sequencing. The presence of tet(W) in pristine environments was demonstrated in this study and others (4, 5) by amplification of a small (167-bp) region of the gene; however, despite several attempts under various conditions, amplifying the 1152-bp region of the gene in upstream Poudre River samples was hindered for unknown reasons. When all six AFO lagoons were considered as one environment, the AFO lagoon environment was significantly different (p < 0.01) from the WWTP environment (activated sludge from all three WWTPs) according to the UniFrac Significance test. While the WWTP libraries appeared highly similar, AFO lagoon libraries displayed notable divergence

FIGURE 3. Correspondence analysis of RFLP patterns of tet(W) clones from AFO lagoons and WWTP activated sludge. (A) RFLP pattern counts of clones from the AFO and WWTP environments subjected to correspondence analysis. The x axis represents the first dimension, explaining 48.63% of the total variation; the y axis represents the second dimension, explaining 37.73% of the total variation for 86.36% cumulative explained variability. Vectors indicate the loading of the presence/absence of the four conserved RFLP patterns. (B) Discrimination of environments based on the percent occurrence of patterns 1 and 2. from each other, as indicated by the jackknife cluster analysis of the clone libraries (Figure 4A) and the RFLP patterns (Figure S3 in the Supporting Information), which produced similar clusters. In both trees, all three WWTPs form a cluster, with the greatest similarity observed between WWTP1 and WWTP3. Also, the small and large conventional dairies were observed to be similar to each other based on both RFLP and phylogenetic analysis. Because phylogenetic analysis supported the results of RFLP analysis, the latter may have potential for application in the absence of sequence information as a more practical means to discriminate between sources and classify impacted environments. Overall, the libraries derived from the small and large conventional dairies and beef feedlot 1 possessed many unique sequences, resulting in a greater separation from the other clone libraries (Figure 4A). The striking divergence of AFO-derived ARG from other sequences in the clone libraries and from reference sequences has been noted previously VOL. 44, NO. 6, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1951

FIGURE 4. Jackknife cluster tree of the AFO and WWTP environments based on sequencing of the tet(W) ARG. (A) Jackknife cluster tree of the 1152-bp tet(W) clone libraries generated from the AFO lagoons (water solids ) ws; settled solids ) ss) and WWTP activated sludge (as) environments. Percentages at nodes represent the jackknife cluster support for each node. (B) Jackknife cluster tree of the pristine river, AFO, and WWTP environments based on sequencing of the 167-bp amplicon of the tet(W) ARG. There was >99% jackknife cluster support for tree nodes. (7, 31). A clone derived from settled solids in the beef feedlot 1 lagoon and a clone from the water of the small conventional dairy lagoon were particularly unique from other clones. These clones shared 99% and 96% identity, respectively, to the tet(W) ARG of a Gram-negative animal pathogen, Lawsonia intracellularis (AM180252), that is typically transmitted through environmental contamination with feces from infected animals (32). Other divergent clones shared closest identity with various uncultured bacterial clones isolated from swine feedlots (7) and AFOs (31). The clone libraries of the organic dairies exhibited greater similarity to the other environments (Figure 4A). This could be explained by the existence of greater selection pressures on conventional dairies and feedlots compared to organic farms due to the more prevalent use of antibiotics. In fact, this is consistent with the trends observed in the tetracycline concentration in the lagoons (19). Tetracyclines were detected in organic dairy lagoons ([total tetracycline] ) 5-116 µg kg of dry weight-1) but at lower concentrations than in conventional dairy lagoons ([total tetracycline] ) 213-2300 µg kg of dry weight-1). The total tetracycline concentration in the beef feedlot 1 lagoon was between that of the organic and conventional dairies ([total tetracycline] ) 97-599 µg kg of dry weight-1) (19). Two groups of consensus sequences were identified. An identical group of 15 clones (3 WWTPs and 12 AFOs) shared 100% homology with the tet(W) reference gene of Bifidobacterium sp. ISO3519 (AF202986), a high-GC Gram-positive bacteria. Another identical group of 14 clones (8 WWTPs and 6 AFOs) shared complete identity with the tet(W) ARG found in the Gram-positive rumen anaerobe Butyrivibrio fibrisolvens (AJ222769), the Gram-negative commensal human-colonic anaerobe Mitsuokella multacida (AY603069), and the Gramnegative rumen anaerobe Megasphaera elsdenii (AY485125). Two unique sequences were identified from the 10 167bp tet(W) clones obtained from the Poudre River-0b sediment sample. The UniFrac tree obtained from the alignment with the existing clone library of 1152-bp sequences is presented in Figure 4B. Despite the small sample size of the Poudre River-0b tet(W) clone library, the nodes of the tree were recovered more than 99% of the time, indicating that there was sufficient sample size to determine the relationships among environments. The WWTP and AFO environments were more similar to each other than to the pristine Poudre River-0b clone library. When the Poudre River-0b sequences were compared to the entire clone library developed in this study and to the nucleotide database using the Basic Alignment Search Tool (available online http://blast.ncbi.nlm.nih.gov/Blast.cgi), no sequences sharing 100% identity were found. Though arguably Poudre River-0b was influenced by the trout-rearing unit, according to the FOD studies, it is unlikely that it is the source of tet(W) at this site. This is considering that tet(W) was never detected at the rearing unit in the FOD studies. tet(W) was detected twice at the 1952

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 6, 2010

trout-rearing unit by Q-PCR, but this method has greater sensitivity than traditional PCR, and tet(W) was found to be very close to the detection limit (data not shown). Potential Application to Other Watersheds. While the approach developed and applied was demonstrated to be effective for the Poudre River as a model system, it is expected that it will be of broad value to other watersheds as well. Though the precise molecular signatures identified as tracers of putative urban and agricultural sources identified in this study may not be directly applicable, the overall method can be applied to identify ARG distributions specific to the watershed of interest. Specific ARG may then be quantified by Q-PCR, depending on the intended application. Particularly encouraging was that FOD, tet(W) RFLP, and tet(W) phylogenetics all led to highly congruent classifications of the environments, indicating that it may not be necessary to apply all three methods in some situations. Overall, the study advances recognition of ARG sources to impacted environments, taking an important step in the identification of dominant processes of ARG proliferation and transport.

Acknowledgments Funding for this study was provided by the Colorado Water Resources Research Institute in addition to the National Science Foundation CAREER CBET Award 0547342 and the USDA Agricultural Experiment Station at Colorado State University. The findings of this study do not necessarily reflect the views of either agency. We also thank Kathy Doesken and Chad McKinney for assistance in sampling, K. McMahon for the donation of positive controls, and the reviewers for critical input to the manuscript.

Supporting Information Available PCR primers used in this study (Table S1), presence of 13 ARG in pristine Poudre River and trout-rearing unit (Table S2a), presence of 13 ARG in WWTPs and AFO lagoons (Table S2b), map of pristine Poudre River sampling locations and surrounding land use (Figure S1), comparison of tet groups CEO and HQST FOD across environments (Figure S2), and cluster analysis of tet(W) RFLP patterns (Figure S3). This material is available free of charge via the Internet at http:// pubs.acs.org.

Literature Cited (1) Overcoming antimicrobial resistance; WHO Annual Report on Infectious Disease; WHO, World Health Organization: Geneva, Switzerland, 2000 [online]. (2) CDC, Centers for Disease Control. National Nosocomial Infections Surveillance (NNIS) system report, data summary from Jan 1992-June 2001, Issued Aug 2001. Am. J. Infect. Control 2001. 29, 404-421. (3) Talbot, G. H.; Bradley, J.; Edwards, J. E.; Gilbert, D.; Scheld, M.; Bartlett, J. G. Bad bugs need drugs: an update on the development pipeline from the antimicrobial availability task force of

(4)

(5) (6) (7)

(8)

(9)

(10)

(11) (12)

(13)

(14)

(15) (16)

(17)

the infectious diseases society of America. Clin. Infect. Dis. 2006, 42 (5), 657–668. Pei, R.; Kim, S. C.; Carlson, K. H.; Pruden, A. Effect of river landscape on the sediment concentrations of antibiotics and corresponding antibiotic resistance genes (ARG). Water Res. 2006, 40, 2427–2435. Pruden, A.; Pei, R.; Storteboom, H.; Carlson, K. H. Antibiotic resistance genes as emerging contaminants: Studies in northern Colorado. Environ. Sci. Technol. 2006, 40, 7445–7450. Auerbach, E. A.; Seyfried, E. E.; McMahon, K. D. Tetracycline resistance genes in activated sludge wastewater treatment plants. Water Res. 2007, 41, 1143–1151. Koike, S.; Krapac, I. G.; Oliver, H. D.; Yannarell, A. C.; CheeSanford, J. C.; Aminov, R. I.; Mackie, R. I. Monitoring and source tracking of tetracycline resistance genes in lagoons and groundwater adjacent to swine production facilities over a 3-year period. Appl. Environ. Microbiol. 2007, 73 (15), 4813–4823. Hu, J. Y.; Shi, J. C.; Chang, H.; Li, D.; Yang, M.; Kamagata, Y. C. Phenotyping and genotyping of antibiotic-resistant Escherichia coli isolated from a natural river basin. Environ. Sci. Technol. 2008, 42 (9), 3415–3420. Sukul, P.; Lamshoft, M.; Kusari, S.; Zuhlke, S.; Spiteller, M. Metabolism and excretion kinetics of 14C-labeled and nonlabeled difloxacin in pigs after oral administration, and antimicrobial activity of manure containing difloxacin and its metabolites. Environ. Res. 2009, 109, 225–231. Kolpin, D. W.; Furlong, E. T.; Meyer, M. T.; Thurman, E. M.; Zaugg, S. D.; Barber, L. B.; Buxton, H. T. Pharmaceuticals, hormones, and other organic wastewater contaminants in US streams, 1999-2000: A national reconnaissance. Environ. Sci. Technol. 2002, 36, 1202–1211. Yang, S.; Carlson, K. Evolution of antibiotic occurrence in a river through pristine, urban and agricultural landscapes. Water Res. 2003, 37, 4645–4656. Batt, A. L.; Bruce, I. B.; Aga, D. S. Evaluating the vulnerability of surface waters to antibiotic contamination from varying wastewater treatment plant discharges. Environ. Pollut. 2006, 142, 295–302. Chee-Sanford, J. C.; Aminov, R. I.; Krapac, I. J.; GarriguesJeanjean, N.; Mackie, R. I. Occurrence and diversity of tetracycline resistance genes in lagoons and groundwater underlying two swine production facilities. Appl. Environ. Microbiol. 2001, 67, 1494–1502. Thorsten, C.; Schneider, R. J.; Fa¨rber, H. A.; Skutlarek, D.; Meyer, M. T.; Goldbach, H. E. Determination of antibiotic residues in manure, soil, and surface waters. Acta Hydrochim. Hydrobiol. 2003, 31, 36–44. Martinez, J. L. Antibiotics and antibiotic resistance genes in natural environments. Science 2008, 321 (5887), 365–367. Knapp, C. W.; Engemann, C. A.; Hanson, M. L.; Keen, P. L.; Hall, K. J.; Graham, D. W. Indirect evidence of transposon-mediated selection of antibiotic resistance genes in aquatic systems at low-level oxytetracycline exposures. Environ. Sci. Technol. 2008, 42 (14), 5348–5353. Munoz-Aguayo, J.; Lang, K. S.; LaPara, T. M.; Gonzalez, G.; Singer, R. S. Evaluating the effects of chlortetracycline on the proliferation of antibiotic resistant bacteria in a simulated river water

ecosystem. Appl. Environ. Microbiol. 2007, 73 (17), 5421–5425. (18) Singer, R. S.; Ward, M. P.; Maldonado, G. Can landscape ecology untangle the complexity of antibiotic resistance. Nat. Rev. Microbiol. 2006, 4 (12), 943–952. (19) McKinney, C. W.; Loftin, K. A.; Meyer, M. T.; Davis, J. G.; Pruden, A. tet and sul antibiotic resistance genes in livestock lagoons of various operation type, configuration, and antibiotic occurrence. Environ. Sci. Technol. 2009. submitted. (20) Aminov, R. I.; Garrigues-Jeanjean, N.; Mackie, R. I. Molecular ecology of tetracycline resistance: development and validation of primers for detection of tetracycline resistance genes encoding ribosomal protection proteins. Appl. Environ. Microbiol. 2001, 67, 22–32. (21) Aminov, R. I.; Chee-Sanford, J. C.; Garrigues-Jeanjean, N.; Teferedegne, B.; Krapac, I. J.; White, B. A.; Mackie, R. I. Development, validation, and application of PCR primers for detection of tetracycline efflux genes of gram-negative bacteria. Appl. Environ. Microbiol. 2002, 68, 1786–93. (22) Tamura, K.; Dudley, J.; Nei, M.; Kumar, S. MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol. Biol. Evol. 2007, 24, 1596–1599. (23) Aminov, R. I.; Mackie, R. I. Evolution and ecology of antibiotic resistance genes. FEMS Microbiol. Lett. 2007, 271, 147–161. (24) Lozupone, C.; Hamady, M.; Kelley, S. T.; Knight, R. Quantitative and qualitative (beta) diversity measures lead to different insights into factors that structure microbial communities. Appl. Environ. Microbiol. 2007, 73, 1576–1585. (25) Cabello, F. C. Heavy use of prophylactic antibiotics in aquaculture: a growing problem for human and animal health and for the environment. Environ. Microbiol. 2006, 8 (7), 1137–1144. (26) Kerry, J.; Hiney, M.; Coyne, R.; Nicgabhainn, S.; Gilroy, D.; Cazabon, D.; Smith, P. Fish feed as a source of oxytetracyclineresistant bacteria in the sediments under fish farms. Aquaculture 1995, 131 (1-2), 101–113. (27) Furushita, M.; Shiba, T.; Maeda, T.; Yahata, M.; Kaneoka, A.; Takahashi, Y.; Torii, K.; Hasegawa, T.; Ohta, M. Similarity of Tetracycline Resistance Genes Isolated from Fish Farm Bacteria to Those from Clinical Isolates. Appl. Environ. Microbiol. 2003, 69, 5336–5342. (28) Roberts, M. C. Update on acquired tetracycline resistance genes. FEMS Microbiol. Lett. 2005, 245, 195–203. (29) Henriques, I. S.; Fonseca, F.; Alves, A.; Saavedra, M. J.; Correia, A. Tetracycline-resistance genes in Gram-negative isolates from estuarine waters. Lett. Appl. Microbiol. 2008, 47, 526–533. (30) Roberts, M. C. Tetracycline and MLS Nomenclature. http:// faculty.washington.edu/marilynr (accessed Jan 14, 2010). (31) Yu, Z. T.; Michel, F. C.; Hansen, G.; Wittum, T.; Morrison, M. Development and application of real-time PCR assays for quantification of genes encoding tetracycline resistance. Appl. Environ. Microbiol. 2005, 71, 6926–6933. (32) Collins, A.; Love, R. J.; Pozo, J.; Smith, S. H.; McOrist, S. Studies on the ex vivo survival of Lawsonia intracellularis. Swine Health Prod. 2000, 8, 211–125.

ES902893F

VOL. 44, NO. 6, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1953