Christopher K. Yost, Department of Biology, University of Regina, 3737 Wascana Parkway, Regina, SK S4S0A2, Canada. E-mail: email@example.com
Aims: To isolate and characterize multiple antibiotic resistance plasmids found in swine manure and test for plasmid-associated genetic markers in soil following manure application to an agricultural field.
Methods and Results: Plasmids were isolated from an erythromycin enrichment culture that used liquid swine manure as an inoculant. Plasmids were transformed into Escherichia coli DH10β for subsequent characterization. We isolated and DNA sequenced a 22 102-bp plasmid (pMC2) that confers macrolide, and tetracycline resistances, and carries genes predicted to code for mercury and chromium resistance. Conjugation experiments using an pRP4 derivative as a helper plasmid confirm that pMC2 has a functional mobilization unit. PCR was used to detect genetic elements found on pMC2 in DNA extracted from manure amended soil.
Conclusions: The pMC2 plasmid has a tetracycline-resistant core and has acquired additional resistance genes by insertion of an accessory region (12 762 bp) containing macrolide, mercury and chromium resistance genes, which was inserted between the truncated DDE motifs within the Tn903/IS102 mobile element.
Significance and Impact of the Study: Liquid swine manure used for manure spreading contains multiple antibiotic resistance plasmids that can be detected in soil following manure application.
The observed increase in antibiotic resistance in clinical isolates of bacterial pathogens is undermining physicians’ ability to control invasive bacterial infections leading to serious consequences for patient health (Nicolau 2011). The increasing use of antibiotics in clinical and agricultural settings is a possible factor contributing to the increase in antibiotic-resistant bacterial populations (Aminov 2009; Martinez 2009a,b). Animal husbandry practices have contributed to the intensive use of antibiotics in the livestock industry for both therapeutic and nontherapeutic purposes such as for growth promotion (Winckler and Grafe 2001; Peak et al. 2007; Chee-Sanford et al. 2009; Kazimierczak et al. 2009). Various studies have documented abundance of antibiotic-resistant bacteria (ARB) associated with livestock such as swine, cattle and chickens (Wassenaar 2005; Aarestrup et al. 2008; Alexander et al. 2008; da Costa et al. 2011). ARB and associated antibiotic resistance genes (ARGs) have also been isolated in agricultural environments such as wastewater lagoons at animal feedlots and in the manure of antibiotic-fed livestock (Peak et al. 2007; Binh et al. 2008; Alexander et al. 2009, 2011; Heuer et al. 2009; Aslam et al. 2010; Cessna et al. 2011). Increasing abundance of the ARGs poses a risk for the proliferation and dissemination of these ARGs to human and environmental bacteria (Ghosh and LaPara 2007). For example, swine manure has been shown to contain high diversity of bacterial communities carrying ARGs encoding resistance to a variety of clinically relevant antibiotic classes (Binh et al. 2008; Heuer et al. 2009; Kazimierczak et al. 2009). A standard waste management practice is to spread large quantities of swine manure onto cultivated fields for soil fertilization. Plasmids and other mobile genetic elements (MGEs) such as transposons, insertion sequence (IS) elements and integrons may play a role in facilitating transfer of ARGs to various bacterial communities in the environment (Bennett 2008; Heuer et al. 2009). Therefore, the potential release of antibiotic resistance plasmids (ARPs) due spreading of swine manure may increase the proliferation of ARB in the environment. Furthermore, transport into ground water and streams may be detrimental if the ARPs are ultimately mobilized to opportunistic human pathogens.
Given the potential threats of manure application in releasing ARPs into the environment, further studies are required to measure the diversity of ARPs in manure and quantify their fate following application to agricultural fields. In this study, we isolated and characterized a unique multiple antibiotic resistance plasmid from swine manure and have demonstrated its detection and persistence in soil following application of the manure to an agricultural field.
Materials and Methods
Sample preparation, plasmid DNA extraction and resistance characterization
Plasmids were isolated from swine manure slurry obtained from a southern Saskatchewan pork producer; this slurry was later applied as fertilizer onto a research field at the Canada-Saskatchewan Irrigation Diversification Centre located in Outlook, Saskatchewan, Canada. Fifty millilitres of the liquid manure was re-suspended in 500 ml sterile water, solid particles were allowed to settle for 30–45 min and the liquid was filtered through a 0·45-μm membrane filter (Millipore Corp., Billerica, MA) using vacuum filtration. Filters were placed in a flask containing 250 ml Luria-Bertani (LB) medium (Sambrook et al. 1989) supplemented with erythromycin (400 μg ml−1) and cultured overnight at 37°C with agitation. Total plasmid DNA from the culture media was isolated and purified using a NucleoBond® Xtra Midi prep kit (Macherey Nagel, Duren, Germany) according to manufacturer’s instructions. DNA yield was estimated using a NanoDrop 1000 spectrophotometer (Thermo Fischer Scientific, Wilmington, DE). Five microlitres of plasmid DNA (50 ng μl−1) was used to transform high efficiency DH10βEscherichia coli competent cells (Invitrogen, Carlsbad, CA, USA) with selection on LB agar plates supplemented with 400 μg ml−1 erythromycin. The transformed colonies were subsequently tested for antibiotic resistance in LB plates supplemented with various antibiotic concentrations: erythromycin (400 μg ml−1), tetracycline (10 μg ml−1), gentamycin (15 μg ml−1), streptomycin (100 μg ml−1), kanamycin (50 μg ml−1), neomycin (20 μg ml−1) ampicillin (100 μg ml−1), streptomycin (100 μg ml−1), spectinomycin (100 μg ml−1) and rifampicin (30 μg ml−1). Transformed strains displaying multiple resistances were selected and stored at −80°C in LB broth containing 400 μg ml−1 erythromycin and 10% glycerol. The size of the plasmids was estimated by Eckhardt gel electrophoresis as described by Hynes et al. (1985). The pMC2 clone was selected from the library for further characterization.
Plasmid DNA sequencing and bioinformatics analysis
The plasmid DNA (25 ng μl−1) was sequenced using a Roche GS-FLX sequencer and 454 technology at the Public Health Ontario Laboratories (ON, Canada). The sequence data obtained were imported into Sequencher computer software (GeneCodes® Corp., Ann Arbor, MI, USA). Sequence reads and contigs were analysed using the basic local alignment search tool (Blast) (http://blast.ncbi.nml.nih.gov/Blast.cgi). A complete consensus sequence was imported into an online computer program GeneMark heuristic model 2.0 ver. 2.8 (http://opal.biology.gatech.edu/genemark/) for gene predictions (Besemer and Borodovsky 1999; Besemer et al. 2001). Amino acid translated sequences were compared against the conserved domain database using position specific iterative Blast search tool (Altschul et al. 1997). Primers for polymerase chain reaction (PCR) and mapping of the plasmid were designed using Primer 3 (ver. 0.4.0) online program (http://frodo.wi.mit.edu/primer3/) (Rozen and Skaletsky 2000), and oligos were obtained from Sigma-genosys (https://row.sigma-genosys.eu.com/). Visual gene annotations and a visual map were generated using vector NTI 10.3.0 computer software (Invitrogen Corp., Carlsbad, CA). Additional protein annotation and domain predictions were analysed using Simple Modular Architecture Research Tool (Smart) program ver. 3.4, genomic MODE (http://smart.embl-heidelberg.de/) (Letunic et al. 2009). Complete DNA sequence comparisons with other closely related or similar plasmid sequences were performed by Mauve multiple genome alignment software ver. 2.0 (http://gel.ahabs.wisc.edu/mauve/) (Darling et al. 2010). The complete annotated nucleotide sequence of the plasmid pMC2 is available in GenBank database under accession number JN704639.
Detection of pMC2 in manure applied soil
Liquid swine manure was applied to a nine-hectare research field site at the Canada-Saskatchewan Irrigation Diversification Center (Outlook, SK, Canada). To the best of our knowledge, the field has never received manure or any organic fertilizers in the last decade (T. Hogg, personal communication). The liquid manure was applied via injection from tanker trucks at a rate of c. 20 000 l ha−1 on 20–21 May 2010. The field was subsequently cultivated with corn. Soil samples were obtained on November 20, 2009, prior to swine manure application, and at intervals following application until October 2010. Two hundred and fifty grams of soil was collected in triplicates at soil depths of 0–10, 10–20 and 20–30 cm, from the four different sites of the field (North East, NE; North West, NW; South East, SE; South West, SW). The soil samples collected before and following manure application were from varying depths of 0–10, 10–20 and 20–30 cm. DNA was extracted from the soil using the PowerSoil DNA Isolation Kit and the manufacturer’s protocol (MoBio Laboratories Inc., Carlsbad, CA, USA). DNA concentration was determined using a Nanodrop spectrophotometer (Thermo Scientific, ON, Canada), and the DNA samples were stored at −20°C. DNA samples used for PCR detection of various pMC2 markers were selected from samples where total DNA concentrations were above 5 ng μl−1. In total, 15 samples after manure application and five samples before manure application were selected for PCR analysis. Several primer pairs were used targeting the pMC2 repA, tetA(C) as well as intergenic regions within the tnpA-IS102/merR (transposon/mercury resistance region) and chrA area (chromium resistance region) (Table 1, Fig. 1). A duplex PCR reaction was performed using primer pairs; Rep3-F, Rep4-R, tetA1-F, tetA2-R and primer pairs IS102A-F, merC2-R and chr1-F, chr3-R were used in single PCR reactions. A total of 25 μl reaction master mix was prepared containing 4 μl of template DNA (5 ng μl−1), 2·5 μl of primers (3 μM), 2·5 μl of Mg2SO4 (20 mM), 2·5 μl of 10× reaction buffer, 0·2 μl of Taq DNA polymerase (50 U μl−1) and 8·8 μl of de-ionized sterile water. The following conditions were used in both duplex and single PCR reaction cycle: 94°C for 5 min initial denaturing, followed by 30 cycles [of denaturing at 94°C; 50°C for 30 s for duplex PCR, annealing; 58°C for 30 s for single PCR, extension; 72°C for 2 min] and final extension at 72°C for 5 min. Cloned PCR products were sequenced to confirm correct amplification of target sequences and compared with other sequences in the NCBI database using the BLASTn. The frequency of detection (FOD) of the repA and tetA(C) genes and intergenic regions surrounding tnpA-IS102/merR and chrA in the soil was determined as described by Storteboom et al. (2010).
Table 1. Description of PCR primer pairs designed for this study
Amplicon size (bp)
tnpA-IS102/merR mercury resistance
Plasmid mobilization determination
To test whether the predicted mobilization genes annotated in the pMC2 DNA sequence are functional, mobilization of the plasmid was performed. pMC2 was transformed into chemically competent cells of the genetically engineered mobilizer E. coli strain S17-1 that has the pRP4 plasmid tra region integrated into its chromosome (Simon et al. 1983). Donor S17-1 E. coli strain is auxotrophic for proline and does not grow on minimal media. Recipient strains included DH5αE. coli carrying a kanamycin-resistant nonmobilizable plasmid, pUCP20tk (West et al. 1994) and an environmental isolate of Pantoea agglomerans. Escherichia coli was selected as a host with the highest probability for successful conjugation while P. agglomerans was selected as another member of the Enterobacteriaceae that is an opportunistic pathogen and is often isolated from agricultural environments. Both recipients are erythromycin sensitive. For conjugation, one millilitre of overnight donor and recipient cultures was centrifuged at 4500 g for 3 min, and the pellet was resuspended in 100 μl of LB broth. One hundred microlitres of donor culture was mixed with 100 μl of recipient culture, and the mixture including 100 μl of controls was spot plated on LB agar plates and incubated overnight at 37°C. Colonies were then scraped from LB plates and resuspended in 900 μl sterile water; 100 μl of serial dilutions were plated on LB plates with selectable antibiotic markers. DH5αE. coli transconjugants were selected on LB supplemented with erythromycin (400 μg ml−1) and kanamycin (50 μg ml−1), and P. agglomerans transconjugants were selected on Vincent’s minimal medium with 1% mannitol (Vincent 1970) supplemented with erythromycin (400 μg ml−1). All the transconjugants were confirmed for plasmid carriage by DNA isolation and subsequent PCR targeting the pMC2 replication and tetracycline resistance genes. Conjugation transfer efficiency/frequency was calculated as the number of transconjugants per recipient cell (Phornphisutthimas et al. 2007; Soda et al. 2008).
DNA analysis of pMC2
Roche 454 DNA sequencing and subsequent assembly yielded a single contig with an average depth of coverage of 25×. PCR reactions were used to confirm the correct contig assembly and confirm that pMC2 is a circular plasmid. The plasmid length is 22 102 bp and annotation prediction lists at least 30 genes with functions involved in antibiotic resistance, heavy metal resistance, replication, conjugative mobility as well as hypothetical genes of unknown function (Fig. 1, Table S1). A 9256-bp pMC2 region shares high similarity to pSC101 (X01654), a 9239-bp naturally occurring low-copy-number plasmid originally described by Cohen and Chang (1977), and DNA sequenced by Bernardi and Bernardi (1984). pMC2 also shares close sequence identity to a recently described 9323-bp mini-plasmid (FJ158001) isolated from a tetracycline-resistant bacterial community sampled from a porcine gut (Kazimierczak et al. 2009). The accessory region of pMC2 containing mercury and macrolide resistance modules has a high similarity to a region in a large 120 592-bp multiple resistance plasmid pRSB107 isolated from a wastewater treatment plant in Germany (Szczepanowski et al. 2005) (Fig. 2).
Organization of resistance genes on pMC2
The resistance genes are clustered together with several transposable elements. The tetR(C), coding for repressor protein, and tetA(C), coding for tetracycline efflux protein (class C) genes responsible for tetracycline resistance (McNicholas et al. 1992), are found downstream of the tnpA-IS102; a Tn903 transposon associated with insertion sequence IS102. The macrolide resistance gene cluster comprised of mph(A), mrx(A) and mphR(A) and appears between insertion sequence elements IS26 and IS6100. The putative heavy metal resistance genes are also found organized in a similar fashion in the pMC2 backbone. The mercury resistance operon consists of a set of genes that code for mercuric ion regulatory (merR and merD), mercuric ion transport (merP, merT and merC), mercuric ion reductase (merA) and unknown function (merE) proteins responsible for narrow-spectrum resistance to inorganic mercury (Schluter et al. 2003; Rojas et al. 2011). Downstream of the mercury resistance operon is a conserved hypothetical gene (urf2) associated with the Tn402 transposon. Another heavy metal resistance gene, chrA, codes for chromate ion transporter proteins responsible for chromium resistance (Cervantes et al. 1990; Cervantes and Silver 1992). This gene is found upstream of the Tn903/IS102 element, associated with another conserved hypothetical gene (orfC) and insertion sequence IS6100.
The Tn903/IS102 element is located downstream of repA gene and hypothetical orfA gene carrying the helix-turn-helix (HTH) DNA-binding domain of truncated merR-like proteins, and upstream of the tetR(C). The Tn903/IS102 element carries a DDE motif coding for a putative DDE and IS102 transposase. The putative mercury resistance operon, macrolide resistance gene cluster and the putative chromate resistance gene are all located within the Tn903/IS102 element; the mercury and macrolide resistance modules are both associated with IS26 and show high similarity at nucleotide level to a Tn21 transposon region in pRSB107 (Fig. S1a).
pMC2 replication, and mobilization
Plasmid pMC2 has a single replication gene repA and conjugative mobilization genes mobA, mobX and traD (also trwB) that represent the core mobilization (MOB) unit of the pMC2 backbone (Fig. S1b). The repA replicon is homologous to the proteins classified into the rep3 super family of proteins and serves as replication initiator (Bertini et al. 2010). The mobA gene has been described as an endonuclease that serves a function as virD2-like DNA relaxase of the type IV secretion system, responsible for cleaving a specific site at the origin of transfer (oriT) initiating plasmid DNA transfer during conjugation (Smillie et al. 2010). The oriT is predicted to be located downstream of the mobA gene and upstream of the traD gene, which encodes the type IV secretion system coupling protein (T4CP). In vitro conjugation experiments indicated that the pMC2 plasmid could be mobilized into E. coli DH5α (pUCP20tk) at a very high frequency of transfer of 6·2 × 10−1 transconjugants per recipient cell, confirming the functionality of the predicted mob region. The pMC2 plasmid was also mobilized into P. agglomerans at a frequency of transfer of 2·63 × 10−5 transconjugants per recipient cell. Plasmid mobilization required tra genes provided by the broad host-range helper plasmid pRP4.
pMC2 detection in agricultural soil following spread of swine manure
PCR amplicons corresponding to internal regions of the pMC2 repA gene, tetA(C) gene and intergenic regions spanning the tnpA-IS102/merR, and the tnpA-IS102/chrA, were detected in soil samples following manure spreading (Table 2, Fig. 3). PCR using DNA isolated from five control soil samples, taken before manure application, did not yield detectable amplicons (Table 2, Fig. 3). Analysis of the sequenced PCR products confirmed correct amplification of target sequences. The majority of detections occurred in the 0–10 cm soil fractions, and in fact, only the tetA(C) marker detected in three of the 10–20 cm fraction soil samples (data not shown). The tetA (C) gene and repA gene amplicons were detected most frequently in the soil samples (Table 2, Fig. 3). Notably, amplicons were detected in DNA from soil sampled in both June and October of 2010 (Table 2).
Table 2. PCR amplification of target sequences in selected soil DNA samples collected at different times and locations following addition of the swine manure
Soil DNA samples
*The gene marker detection from these samples are shown in Fig. 3.
NW, North West; SW, South West; NE, North East; +, positive amplification of target gene/region by PCR.
C – control soil sample, sampled before manure addition.
20 November 2009
4 June 2010
10 June 2010
21 June 2010
21 June 2010
21 June 2010
19 October 2010
Based on DNA sequence analysis, pMC2 may have acquired multiple resistance genes through insertion of a mobile genetic element. Salmonella plasmid pSC101 and uncultured bacterium clone 2 from the tetracycline resistome of the pig intestine may represent a common ancestral backbone with the replication, mobilization and tetracycline resistance genes. In addition, both plasmids also have a similar Tn903/IS102 element with a DDE motif. The nomenclature is based on the relatedness of previously described Tn903 kanamycin resistance transposon and pSC101-IS102 element (Bernardi and Bemardi 1981; Oka et al. 1981). The DDE domain within the Tn903/IS102 element in pMC2 has been disrupted by insertion of an additional 12 762-bp resistance region containing mercury, macrolide and chromium resistance genes (Fig. 4). The insertion event resulted in simultaneous acquisition of both the macrolide resistance module and heavy metal resistance genes. The presence of the transposable elements (IS elements and transposons) in pMC2 is notable, and given the relatively small size of the plasmid, it can potentially insert itself into bacterial chromosomes or other plasmids by insertion and transposition events, which can lead to the development of larger multiple resistance gene plasmids (Bennett 2008).
Many plasmids conferring resistance to antibiotics have been frequently found in association with genes coding for resistance to heavy metal (Tennstedt et al. 2003; Stokes et al. 2006; Moura et al. 2007). This association is not well understood, some studies have suggested the presence of both antibiotic and metal resistance genes may help in plasmid maintenance in environments with no antibiotic selection pressure (Schluter et al. 2005, 2008). For example in environments polluted with mercury, antibiotic-resistant bacterial communities carrying such plasmids will be co-selected. Other heavy metals that have been detected in pig manure slurry and agricultural soils include As, Cu and Zn and have been shown to co-select and enhance the spread of antibiotic resistance genes in microbial populations found in the soil (Bolan et al. 2004; Marcato et al. 2009; Berg et al. 2010; Heuer et al. 2011). Furthermore, co-selection may arise from the use of macrolides in swine husbandry, for example, tylosin use as a growth promoter in swine has been implicated with the development of macrolide resistance in bacteria (Alban et al. 2008; Chen et al. 2010; Juntunen et al. 2011).
Plasmid pMC2 is a mobilizable plasmid with a MOB region that is 99% identical, at the nucleotide level, to the Salmonella plasmid pSC101 and uncultured bacterium clone 2 plasmid. The presented experimental data confirmed that pMC2 can be mobilized in the presence of a conjugative helper plasmid. The mobA and mobX region of pMC2 has been described in plasmid pSC101; mutations in either gene affect mobilization (Meyer 2000, 2009; Becker and Meyer 2003). The mobA domain shares similar functions with the incQ type relaxases (Grohmann et al. 2003), such as those found in broad host-range incQ mobilizable plasmid R1162 (M13380), RS1010 (M28829). The MOB unit of these incQ plasmids consists of mobA, mobB and mobC (see (Meyer (2000); Becker and Meyer (2003) for comparisons). Plasmid pMC2, like pSC101, R1162, RSF1010, is classified into MOBQ group based on related sequences and functions of the MOB units; this group is comprised of relatively small mobilizable plasmids ranging up to 30 kb in size (Garcillán-Barcia et al. 2011). The presence of a functional MOB unit implies that pMC2 could contribute to the transfer and spread of antibiotic resistance genes in the environment.
The detection of pMC2 sequences in soil samples taken in October 2010 is suggestive that pMC2 can be maintained in bacterial populations within the soil for a significant period following its introduction in the soil by manure application in May 2010. The tetracycline resistance gene amplicon was detected at high frequencies and below the top soil. This is likely due to the fact that the PCR primers are capable of amplifying tetA(C) genes from other sources found in the manure in addition to pMC2. Notably, the tetA(C) marker was not detected in the selected DNA samples obtained before manure application. Therefore, its higher rate of detection relative to the pMC2-specific amplicons related to intergenic regions suggests a greater abundance of sources of the tetA(C) gene in the manure treated soil. This observation is consistent with results from other studies, for example, several tetracycline resistance genes have been detected by PCR from lagoon and groundwater close to the pig production facility during a period of over one year (Koike et al. 2007). Similarly, Storteboom et al. (2010) have observed high FOD of sulphonamide and tetracycline resistance genes in various environments associated with agricultural practices.
Application of manure to soil has been suggested to contribute significantly to the release of ARB populations into the soil (Binh et al. 2008; Chee-Sanford et al. 2009; Heuer et al. 2011). Our results further demonstrate that multiple antibiotic resistance plasmids are a likely component to these populations and have the potential to persist and potentially mobilize to native soil bacteria. The fate of multiple resistance plasmids in the soil after manure application in the field may depend on a variety of factors. Conditions such as temperature and pH of the soil may play an important role in survival of plasmid-carrying hosts in the soil. Soil contamination by other antimicrobials substances including erythromycin, tetracycline and mercury residues may also play a role in the maintenance of pMC2 in soil bacterial communities.
Our study has shown that swine manure is a source of multiple antibiotic resistance plasmids such as pMC2, these plasmids can be mobilized and potentially transfer resistance genes to other bacterial species through conjugation mechanisms, and the presence of mobile elements may further disseminate the genes through transposition. In addition, multiple resistance plasmids can persist in the soil for a significant time following introduction into a previously untreated soil environment. This persistence may allow further environmental spread of bacteria carrying antibiotic resistance plasmids through transport into water sources from run-off events. Future research will monitor transport of plasmids like pMC2 from manure spread soils into water and measure their persistence in aquatic ecosystems using quantitative molecular methods. Accurate quantification of pMC2 in the soil could help determine the significance of multiple resistance plasmids as pollutants (Rahube and Yost 2010) released into the environment and have the potential to spread to human pathogenic bacteria.
Acknowledgement of a graduate scholarship to TOR is extended to the Botswana International University of Science and Technology (BIUST) task force, sponsorship awarded through the Ministry of Education and Skills Development (MoESD) in Botswana. We gratefully acknowledge Dr David Alexander for providing the resources for 454 sequencing of the pMC2 plasmid. We also acknowledge the technical support provided by Tyler Boa and helpful discussions from Dr Bastien Fremaux. This work is supported by the Canada Research Chairs program and Natural Sciences and Engineering Research Council of Canada funding to CKY. We thank Terry Hogg and technical staff at the Canada-Saskatchewan Irrigation Diversification Center for assistance with soil sample collection.