Frequency of Antibiotic Resistance in a Swine Facility 2.5 Years After a Ban on Antibiotics
- First Online:
- Cite this article as:
- Pakpour, S., Jabaji, S. & Chénier, M.R. Microb Ecol (2012) 63: 41. doi:10.1007/s00248-011-9954-0
- 553 Views
The addition of antibiotics to livestock feed has contributed to the selection of antibiotic-resistant bacteria in concentrated animal feeding operations and agricultural ecosystems. The objective of this study was to assess the occurrence of resistance to chlortetracycline and tylosin among bacterial populations at the Swine Complex of McGill University (Province of Quebec, Canada) in the absence of antibiotic administration to pigs for 2.5 years prior to the beginning of this study. Feces from ten pigs born from the same sow and provided feed without antibiotic were sampled during suckling (n = 6 for enumerations, n = 10 for PCR), weanling (n = 10 both for PCR and enumerations), growing (n = 10 both for PCR and enumerations), and finishing (n = 10 both for PCR and enumerations). The percentage of chlortetracycline-resistant anaerobic bacterial populations (TetR) was higher than that of tylosin-resistant anaerobic bacterial populations (TylR) at weanling, growing, and finishing. Prior to the transportation of animals to the slaughterhouse, resistant populations varied between 6.5 and 9.4 Log colony-forming units g humid feces−1. In all pigs, tet(L), tet(O), and erm(B) were detected at suckling and weanling, whereas only tet(O) was detected at growing and finishing. The abundance of tet(O) was similar between males and females at weanling and growing and reached 5.1 × 105 and 5.6 × 105 copies of tet(O)/ng of total DNA in males and females, respectively, at finishing. Results showed high abundances and proportions of TetR and TylR anaerobic bacterial populations, as well as the occurrence of tet and erm resistance genes within these populations despite the absence of antibiotic administration to pigs at this swine production facility since January 2007, i.e., 2.5 years prior to the beginning of this study. This work showed that the occurrence of bacterial resistance to chlortetracycline and tylosin is high at the Swine Complex of McGill University.
Antibiotics have been utilized broadly in the last 50 years in food animals to treat, prevent, or control infectious illness or to enhance efficiency of feed utilization and weight gain . In Canada, swine production is an important economic activity and a major source of environmental problems because of the large volume of swine wastes that it generates . In this country, tetracyclines and macrolides are the first and third most abundantly distributed antibiotics for use in animals, respectively , and are also important for the treatment of human infections. Intensive and extensive use of these antibiotics creates a pressure for the selection and dissemination of both pathogenic and commensal antimicrobial-resistant bacteria in swine husbandry and subsequently across the food processing chain . An investigation derived from different steps of the production chain of pork meat industries illustrates the wide distribution of multidrug-resistant staphylococci . Moreover, resistant bacteria and resistance genes can be transferred from food animals, their waste, and their meat to humans via the food processing chain or the environment . Studies show that human diseases and deaths caused by strains of multidrug-resistant pathogenic Salmonella enterica serotype Typhimurium DT104 in Denmark  and Enterococcus faecium in China  originated from swine herds. The selection of antibiotic-resistant bacteria in swine husbandry and the subsequent dissemination of such bacteria from animal reservoirs to humans increase the occurrence of infectious diseases that become difficult to treat with currently available antibiotics. This represents a major food safety and human/animal health concern because of the increasing emergence of antibiotic resistance phenotypes both in strains of clinical/veterinary significance and in usual commensal bacteria .
In the mammalian gut, pathogens are greatly outnumbered by commensal bacteria that can harbor the same resistance determinants as their disease-producing counterparts. Since the oxygen level in the swine large intestine is below 1%, anaerobic bacteria, for which the culturable fraction can reach 109 to 1010 colony-forming units (CFU) g humid feces−1, dominate over aerobic bacteria by one to over two orders of magnitude . Since anaerobic commensal bacteria constitute the numerically and ecologically dominant subpopulation in the swine large intestine, we hypothesized that they serve as a diversified and highly abundant reservoir of resistance genes that may be transferred to pathogenic bacteria. This implies that not only pathogenic bacteria, but mainly commensal bacteria, especially anaerobes, must be targeted when applying a relevant, whole-population approach as described in this work.
The objective of this study was to assess the occurrence of antibiotic resistance among bacterial populations at the Swine Complex of McGill University (Province of Quebec, Canada), where the addition of antibiotics to swine feed has been discontinued since January 2007. Specific objectives were (1) to determine the abundance of enteric anaerobic bacterial populations (total, TetR, and TylR) in the swine intestine; (2) to detect the resistance genes present among enteric bacterial populations; and (3) to determine the abundance of the most frequently detected resistance gene among enteric bacterial populations.
Swine Rearing and Sampling
Swine rearing and sampling
1 to 3
June 17–July 8
4 to 11
July 9–September 1
12 to 16
September 2–October 7
17 to 21
October 8–November 12
Anaerobic phosphate buffer [APB: 313 μM KH2PO4, 2 mM MgCl2·6H2O, 0.001 g resazurin l−1 (redox potential indicator), pH 7.2] was boiled for 30 min to reduce the oxygen concentration, transferred into an anaerobic jar, vacuumed for 1 min, and flushed for 5 min with a 10% H2, 10% CO2, and 80% N2 gas mixture (Megs, St-Laurent, QC, Canada). The last two steps (vacuuming and flushing) were repeated three times. APB was distributed in glass test tubes with screw caps inside the anaerobic chamber. Glass test tubes were screwed, autoclaved, and sterile Na2S·9H2O (reducing agent to maintain anaerobic conditions) was added at a final concentration of 0.2 mmol l−1 .
Bacterial enumerations were completed in an anaerobic chamber (Coy Laboratory Products, Grass Lake, MI, USA). Serial dilutions (10−1 to 10−10) of fresh fecal samples kept on ice were performed in APB. For each sample, anaerobic bacterial populations were determined using a spread-plate procedure (three replicate Petri plates per dilution for each sample). Brain heart infusion agar (BHIA; Becton Dickinson, Sparks, NJ, USA) was utilized as it allowed the highest quantitative bacterial growth from the swine fecal samples in comparison to tryptic soy agar (Becton Dickinson) (data not shown). Total bacterial populations were enumerated using BHIA, chlortetracycline-resistant bacterial populations (TetR) were enumerated using BHIA supplemented with 20 mg chlortetracycline HCl l−1 (Sigma-Aldrich, St-Louis, MO, USA), and tylosin-resistant bacterial populations (TylR) were enumerated using BHIA supplemented with 10 mg tylosin tartrate l−1 (Sigma-Aldrich). In order to use antibiotic concentrations similar to those reported in the literature, chlortetracycline and tylosin were added at concentrations 25% higher than minimum inhibitory concentration (MIC) breakpoints for tetracycline-resistant (≥16 mg l−1) and macrolide-resistant (≥8 mg l−1) bacteria isolated from animals . Petri plates were incubated in an anaerobic incubator in the dark at 37°C. Growth was observed periodically in an anaerobic chamber until no additional colonies could be visually detected (10 days of incubation). Triplicate Petri plates (from a specific dilution) containing 30 to 300 CFU were selected to determine the abundance, evolution (fold change), and percentage of total, TetR, and TylR anaerobic bacterial populations in swine feces.
DNA was extracted from fresh fecal samples and from samples preserved either at −20°C or −80°C. The Qiagen QIAamp DNA Stool Mini Kit (QIAgen, Mississauga, ON, Canada) provided better total DNA recovery than the Ultra Clean DNA Kit (MoBio, Carlsbad, CA, USA), especially for samples preserved at −80°C, and relatively little PCR inhibition occurred (data not shown). Hence, the DNA extracts analyzed were those obtained with the QIAamp DNA Stool Mini Kit and preserved at −80°C in 1 mmol l−1 Tris–HCl pH 8.0. DNA was checked for integrity by agarose gel electrophoresis with Lambda DNA HindIII Digest standards (New England BioLabs, Ipswich, MA, USA), and DNA concentration of each extract was determined using the NanoDrop ND-1000 UV-visible spectrophotometer (Thermo Fisher Scientific, Marietta, OH, USA). The total DNA from pure cultures was also extracted, purified, quantified, and preserved in the same manner and used as positive controls in PCR and real-time PCR experiments (see below).
Oligonucleotide primers, annealing temperatures, and positive control strains used for PCR amplification of bacterial genes
Annealing temperature (°C)
Amplicon size (bp)
CCT ACG GGA GGC AGC AG
Pseudomonas aeruginosa ATCC 27853
ATT ACC GCG GCT GCT GG
GCG CGA TCT GGT TCA CTC G
AGT CGA CAG YRG CGC CGG C
TAC GTG AAT TTA TTG CTT CGG
Strain Ct4afooB (Tn10)
ATA CAG CAT CCA AAG CGC AC
GCG GGA TAT CGT CCA TTC CG
GCG TAG AGG ATC CAC AGG ACG
GGA ATA TCT CCC GGA AGC GG
CAC ATT GGA CAG TGC CAG CAG
GTT ATT ACG GGA GTT TGT TGG
AAT ACA ACA CCC ACA CTA CGC
TTA TGG TGG TTG TAG CTA GAA A
AAA GGG TTA GAA ACT CTT GAA A
GTM GTT GCG CGC TAT ATT CC
GTG AAM GRW AGC CCA CCT AA
ACA GAA AGC TTA TTA TAT AAC
TGG CGT GTC TAT GAT GTT CAC
ACG GAR AGT TTA TTG TAT ACC
TGG CGT ATC TAT AAT GTT GAC
GAA AGC TTA CTA TAC AGT AGC
AGG AGT ATC TAC AAT ATT TAC
ATT TGT ACC GGC AGA GCA AAC
GGC GCT GCC GCC ATT ATG C
TCT AAA AAG CAT GTA AAA GAA
CTT CGA TAG TTT ATT AAT ATT AGT
GAA AAG GTA CTC AAC CAA ATA
AGT AAC GGT ACT TAA ATT GTT TAC
TCA AAA CAT AAT ATA GAT AAA
GCT AAT ATT GTT TAA ATC GTC AAT
The PCR conditions for antibiotic resistance genes started with an initial DNA denaturation (94°C for 5 min), followed by 30 cycles of 1 min at 94°C (denaturing), 1 min of annealing at the temperatures specified in Table 2, and 1 min at 72°C (extension), followed by a final extension of 7 min at 72°C. For the 16S rDNA gene, the V3 region was targeted with the Bacteria universal primers 341-F and 518-R . The PCR conditions for the 16S rDNA gene were 5 min of denaturation at 99°C, followed by 10 min at 80°C during which the Taq DNA polymerase was added (hot start), two cycles of 5 min at 94°C, 5 min at 55°C, 2 min at 72°C, then 28 cycles of 1 min at 94°C, 1 min at 55°C, 2 min at 72°C, and finally, an extension period of 10 min at 72°C. The size (Table 2), specificity (unique band), and abundance of PCR products were determined by comparison with DNA standards (GeneRuler 100 bp DNA Ladder, MBI Fermentas, Burlington, ON, Canada) after agarose gel electrophoresis . PCR products obtained with positive controls described in Table 2 were also sequenced (McGill University and Genome Quebec Innovation Centre, Montreal, QC, Canada) and compared with sequences available in the GenBank database to confirm their identity (data not shown).
Standard for Real-Time PCR
Since tet(O) was detected in all pigs at all production stages and as it can be horizontally transferred among bacterial populations via different mechanisms, this gene was selected for further quantitative analysis by real-time PCR based on protocols described by Smith et al. . The tet(O)-bearing plasmid pU0A1 was extracted and purified from Escherichia coli using the QIAquick PCR Purification Kit (QIAgen) and used as standard to determine the number of copies of tet(O) in the samples by real-time PCR. Following extraction and purification of pU0A1, tet(O) was amplified by conventional PCR, using the reaction mixture and conditions described above, and the concentration of the amplified product (~171 bp) was determined using the NanoDrop ND-1000 UV-visible spectrophotometer (Thermo Fisher Scientific). Subsequently, serial tenfold dilutions of the tet(O) PCR products were prepared in triplicate in 1 mmol l−1 Tris–HCl pH 8. Dilutions ranged from 3 × 103 copies/ng total DNA (lower precision limit) to 3 × 107 copies/ng total DNA (upper precision limit) and included a negative control (containing no target DNA).
Optimization of Real-Time PCR
Real-time PCR amplifications were carried out in an Mx3000P (Stratagene, Cedar Creek, TX, USA). For optimization of the real-time PCR mixture and conditions, different concentrations of forward primer (5′AAGAAAACAGGAGATTCCAAAACG; 600 to 900 nmol l−1), reverse primer (5′CGAGTCCCCAGATTGTTTTTAGC; 600 to 900 nmol l−1), TaqMan fluorogenic probe [5′FAM-ACGTTATTTCCCGTTTATCACGG-Tamra, labeled with 6-carboxyfluorescein (FAM) at its 5′ end and with 6-carboxytetramethyl-rhodamine (TAMRA) at its 3′ end; 450 and 900 nmol l−1], and tet(O) DNA standard (3 × 103 to 3 × 107tet(O) copies/ng total DNA), as well as a range of annealing temperatures (53°C to 59°C) were assessed.
The optimal real-time PCR mixture (25 μl) contained 10 ng of DNA extracted from individual pig’s fecal sample as template, 800 nmol l−1 of each forward and reverse primers, 450 nmol l−1 of TaqMan fluorogenic probe, and the TaqMan Universal PCR Master Mix (Stratagene, La Jolla, CA, USA) . For each DNA extract, duplicate PCR tubes were analyzed for the quantification of tet(O). A negative control, consisting of the reaction mixture without DNA, was used in each PCR run. The optimal real-time PCR conditions for tet(O) started with an initial DNA denaturation (95°C for 10 min), followed by 40 cycles of 30 s at 95°C (denaturing), 1 min at 55°C (annealing), and 30 s at 72°C (extension), followed by a final extension of 7 min at 72°C. PCR products were analyzed by the Mx3000P software and checked for nonspecific PCR amplification by comparison with DNA standards (GeneRuler 100 bp DNA Ladder, MBI Fermentas, Burlington, ON, Canada) after agarose gel electrophoresis .
The upper and lower precision limits of each standard curve were determined by the highest and lowest (respectively) tet(O) concentration at which the relative standard deviation on the average of triplicate Ct (threshold cycle) was <5%. A standard curve was considered linear if R2 was >0.98. The percentage efficiency of the real-time PCR amplification for a standard curve was considered satisfactory if higher than 90% and lower than 105% . Finally, successive standard curves were considered accurate if the relative standard deviation on the average of their R2, efficiency, and slope were all <5%. These quantitative criteria were applied to the optimization experiments to determine the optimal real-time PCR mixture and conditions, as described above, and to each standard curve used thereafter for the determination of tet(O) concentrations in samples.
At the finishing stage, i.e., prior to the transportation of animals to the slaughterhouse, resistant populations varied between 6.5 and 9.4 Log CFU g humid feces−1 (Fig. 1), TetR populations averaged at 69% and approximately 100% of bacterial populations in males and females, respectively, whereas TylR populations accounted for only 2% and 5% of the communities in males and females, respectively (Fig. 2C).
Antibiotic Resistance Genes
DNA concentrations in extracts from swine feces
DNA concentration (μg total DNA/g feces)
Optimization and Standard Curve for Real-Time PCR
Accuracy of standard curves for the real-time PCR assay for tet(O) quantification
Relative standard deviation (%)
Real-Time PCR for tet(O)
Our starting hypothesis was that enteric commensal bacteria, especially anaerobes, constitute a diversified and highly abundant reservoir of antibiotic resistance genes in the swine intestine. Despite the discontinuation of subtherapeutic applications of antibiotics at the Swine Complex 2.5 years prior to the beginning of the study and even in the absence of therapeutic applications to the sow and the experimental pigs prior to and during this study, high abundances of TetR and TylR anaerobic bacterial populations were observed in swine feces throughout the rearing period (Fig. 1). This is of environmental and public/animal health significance since antibiotic-resistant bacteria can be transmitted among pigs reared in intensive production systems, as well as to farm employees  and facilities . Also, the spreading of swine manure to fertilize agricultural fields can introduce resistant bacteria, whether commensal or pathogenic, into farmlands [2, 3, 39] and crops  produced for animal feed or human consumption, field equipment, buildings, and workers , as well as surface  and subsurface [9, 29] water resources used for drinking, irrigation, aquaculture, or recreation.
Our findings support previous observations regarding a higher abundance of TetR anaerobic populations than TylR in a swine farm effluent . Accordingly, it was reported that 71% of Enterococcus faecalis isolates from a farrowing house effluent were resistant to tetracycline , whereas 4–32% of bacteria isolated from swine feces and manure storage pits were resistant to tylosin . Similarly, in a previous report, fluorescence in situ hybridization analyses revealed that potential resistance to macrolides by ribosomal methylation was present in manure samples from two swine farms where no antibiotic use was reported (37.6 ± 6.3% and 40.5 ± 5.4%, respectively) . However, in a recent investigation of the impact of land application of swine manure on antibiotic resistance levels in soils, spreading of manure from farms with tetracycline use did not result in increased levels of TetR and TylR in the soil microbial communities in comparison to soils fertilized with swine manure from organic farms (i.e., without antibiotic applications) .
Antibiotic Resistance Genes
Despite the absence of antibiotic use prior to and during this study, 3 out of 14 selected resistance genes [tet(O), tet(L), and erm(B)] were detected in swine feces at suckling and weanling, although only tet(O) was detected at growing and finishing. In a survey done by Jindal et al., several tetracycline resistance determinants, including tet(O), were present in the swine waste of four conventional farms and one organic farm . A metagenomic approach investigating the tetracycline resistome of the organic farm pig gut revealed that tet(C), tet(W), tet(40), and novel genes encoding resistance to the tetracyclines minocycline and doxycycline were present in swine gut . Although tylosin was not used prior to or during the present study, erm(B) was detected at suckling and weanling. This gene is one of the most frequently detected macrolide resistance genes in bacteria of animal and human origin [21, 34]. The occurrence of antibiotic resistance genes in the absence of usage of a specific antibiotic can result from the colocalization of resistance determinants within DNA segments, such as the linkage of genes conferring resistance to macrolides, streptogramins, and glycopeptides [21, 23]. Since no antibiotics have been used subtherapeutically and medicinally in the pigs selected for the present study, two hypotheses can be presented regarding the detection of resistance genes in swine feces. First, it is possible that 2.5 years is not a long enough period of time to observe a decrease in the diversity of resistance genes. We monitored the short-term temporal evolution (~20 weeks, i.e., within the lifetime of pigs at the farm) of antibiotic resistance shortly after (2.5 years) the discontinuation of antibiotic applications. Only few investigations reported a lower occurrence of antibiotic resistance in swine farms without antibiotic applications in comparison to farms where antibiotics are used [21, 33]. Further work is required to investigate the medium-term (i.e., for successive herds of pigs) and long-term (over several years) temporal evolution of the diversity and abundance of resistance genes in swine production. Second, it may be that some selective pressure maintains antibiotic resistance genes among these enteric bacterial populations. One possible selective pressure originates from metals included in the feeds given to pigs, which contain 125 mg of copper/kg and between 80 and 200 mg of zinc/kg, depending on pig age (Agribrands Purina information documents: http://www.agripurina.ca/Screens/pigtech.aspx). Since it is known that antibiotic and metal resistance genes can be colocated on mobile genetic elements , the metal concentrations mentioned above may exert a selective pressure for maintaining both antibiotic and metal resistance genes in swine husbandry despite the absence of any antibiotic application [16, 32].
Until now, at least 42 genes are known to confer resistance to tetracyclines and at least 67 genes were reported to confer resistance to macrolides [35, 36]. These genes are widely distributed among Gram-positive and Gram-negative bacteria, pathogenic and commensal bacteria, aerobes and anaerobes, and are found in several agricultural, environmental, and industrial ecosystems [35, 36]. It is likely that tetracycline and macrolide resistance genes other than those detected by PCR are present in the microflora of pigs at the Swine Complex, which may explain the high abundance of resistant populations (Figs. 1 and 2).
PCR analysis revealed that erm(B) was detected in the experimental pigs but not in the sow. Such a discrepancy could be the result of differences in intestinal microenvironmental conditions such as pH, temperature, and mucin composition or concentration, which supply a variety of niches and affect the types of bacteria (and their resistance genes) that can successfully compete for nutrients and space in the intestine .
Abundance of tet(O)
Despite the absence of antibiotic additions to swine feed, the abundance of tet(O) remained at high levels throughout the lifetime of the pigs (Fig. 3). tet(O), which confers resistance against tetracyclines through ribosomal protection , was first identified on plasmids of Campylobacter spp. [43, 47] and was later found in ruminal bacteria . It is now distributed in at least 10 Gram-negative genera, including 3 anaerobic, and 12 Gram-positive genera, including 5 anaerobic . This gene, as well as other resistance determinants, is likely present in other species in agricultural ecosystems like swine production . A wide species distribution of tet(O) in the microflora of pigs monitored in the present study is another potential explanation for the high abundance of TetR populations (Fig. 1) and tet(O) (Fig. 3).
Mobile genetic elements, such as plasmids, transposons, and integrons, contribute to the horizontal transfer of resistance genes among related or unrelated species through transformation, bacteriophage-mediated transduction, and cell-to-cell conjugation . Since tet(O) can be located on plasmids, it is subject to be horizontally transferred among bacterial species in various environments and hosts, including swine farms [6, 9, 48]. Interestingly, neither tet(O) nor any of the other 13 resistance genes monitored by PCR were detected in feces of the 2 employees working at the Swine Complex (results not shown).
This study showed high abundances and proportions of TetR and TylR enteric populations, as well as the occurrence of tet and erm resistance genes within these populations despite the absence of antibiotic administration to pigs at the Swine Complex of McGill University. This is of environmental and public/animal health significance since antibiotic-resistant bacteria can be disseminated among livestock, humans, and the environment. This dissemination is one of the several factors contributing to the dynamic evolution of antibiotic resistance in agricultural, environmental, and industrial ecosystems, as well as in clinical settings and the community. This work showed that, despite the absence of antibiotic usage for 2.5 years prior to the beginning of this study, the occurrence of bacterial resistance to chlortetracycline and tylosin was high at the Swine Complex of McGill University.
This work was supported by the National Sciences and Engineering Research Council of Canada (NSERC) and the Fonds Québécois de Recherche sur la Nature and les Technologies (FQRNT) to M.R.C. S.P. benefited from the Alexander Graham Bell Canada Graduate Scholarship M (NSERC), the Principal’s Graduate Fellowship (McGill), and the Macdonald Class of '44 Rowles Graduate Bursary (McGill). We acknowledge Dr. Huilan Chen for generously sharing her expertise about real-time PCR. We thank Dr. Josée Harel (Faculty of Veterinary Medicine, Université de Montréal), Dr. Luke Masson (Biotechnology Research Institute, National Research Council Canada, Montreal), and Dr. Marilyn C. Roberts (University of Washington, USA) for providing positive control strains for PCR.