Occurrence of the Transferable Copper Resistance Gene tcrB among Fecal Enterococci of U.S. Feedlot Cattle Fed Copper-Supplemented Diets

Abstract

Copper, an essential micronutrient, is supplemented in the diet at elevated levels to reduce morbidity and mortality and to promote growth in feedlot cattle. Gut bacteria exposed to copper can acquire resistance, which among enterococci is conferred by a transferable copper resistance gene (tcrB) borne on a plasmid. The present study was undertaken to investigate whether the feeding of copper at levels sufficient to promote growth increases the prevalence of the tcrB gene among the fecal enterococci of feedlot cattle. The study was performed with 261 crossbred yearling heifers housed in 24 pens, with pens assigned randomly to a 2×2 factorial arrangement of treatments consisting of dietary copper and a commercial linseed meal-based energy protein supplement. A total of 22 isolates, each identified as Enterococcus faecium, were positive for tcrB with an overall prevalence of 3.8% (22/576). The prevalence was higher among the cattle fed diets supplemented with copper (6.9%) compared to normal copper levels (0.7%). The tcrB-positive isolates always contained both erm(B) and tet(M) genes. Median copper MICs for tcrB-positive and tcrB-negative enterococci were 22 and 4 mM, respectively. The transferability of the tcrB gene was demonstrated via a filter-mating assay. Multilocus variable number tandem repeat analysis revealed a genetically diverse population of enterococci. The finding of a strong association between the copper resistance gene and other antibiotic (tetracycline and tylosin) resistance determinants is significant because enterococci remain potential pathogens and have the propensity to transfer resistance genes to other bacteria in the gut.

INTRODUCTION

Copper (Cu) is an essential micronutrient required for various biochemical functions in prokaryotic and eukaryotic cells (1). In livestock, copper requirements are met by the inclusion of copper, generally as sulfate, as part of a multimineral supplement in the diet. The inclusion of copper in diets at elevated concentrations, particularly in cattle and swine, has growth-promotional effects (2). In cattle, copper is included at concentrations of 10 to 100 mg/kg of diet to reduce mortality and morbidity in calves and for growth promotion in feedlot cattle (3). Copper in excess is very reactive and toxic to cells; therefore, intracellular Cu²⁺ concentration is highly regulated (4). In biological systems, copper homeostasis is mediated by the expression of copper chaperone proteins, Cu²⁺-transporting proteins, Cu²⁺-transporting ATPases, and a Cu²⁺ chaperone for Cu-Zn superoxide dismutase, which help in the regulation of Cu uptake, intracellular transport, and export (1). A group of membrane-spanning proteins called CPx-ATPases, encoded by the four-gene operon copYZAB, regulate copper homeostasis in Gram-positive bacteria, such as Enterococcus hirae (5). Acquired copper resistance has been reported both in Gram-positive (6, 7) and in Gram-negative (8) bacteria. A transferable copper resistance gene, tcrB, in Enterococcus faecium, E. faecalis, E. gallinarum, E. casseliflavus, and E. mundtii (6, 9) was first reported in Denmark. In Europe, the tcrB gene often is found on a plasmid, which also carries erm(B) and vanA genes encoding resistance to macrolides and glycopeptides, respectively (6, 9). We have previously reported the occurrence and prevalence of the tcrB gene among fecal E. faecium and E. faecalis isolates of piglets fed elevated concentrations of copper in the diet (10, 11). The tcrB gene in fecal E. faecium and E. faecalis of piglets in the United States was on a plasmid which carried macrolide [erm(B)] and tetracycline [tet(M)] resistance genes, but no vanA genes. The potential genetic link between copper and antibiotic resistance is of interest because tylosin, a macrolide, and tetracyclines are widely used as feed additives in feedlot cattle (12). The present study was conducted to determine the occurrence of the tcrB gene in fecal enterococcal isolates of cattle supplemented with elevated levels (i.e., 100 mg/kg) of copper in the diet and to determine phenotypic susceptibilities to copper, tetracycline, tylosin, and vancomycin. Because enterococci, a common gut commensal of animals and humans (13, 14), are a frequent cause of nosocomial infections in humans, as well as sporadic infections in animals (15, 16), we also explored the presence of major virulence genes. Multilocus variable number tandem repeat analysis (MLVA) was performed to determine the clonal relationship of the enterococcal isolates to the MLVA types (MTs) reported in the global epidemiological database. In addition, Southern blot hybridization was performed to show that tcrB was located on a plasmid and a conjugation assay was conducted to demonstrate the rate of transfer and cotransfer of tcrB, erm(B), and tet(M) genes.

(Portions of this study were presented at the Annual Meeting of the Missouri Valley Branch of the American Society for Microbiology, Manhattan, Kansas, USA, 12 to 14 April 2012, and at the Third American Society for Microbiology Conference on Antimicrobial Resistance in Zoonotic Bacteria and Food-borne Pathogens, Aix-en-Provence, France, 26 to 29 June 2012.)

MATERIALS AND METHODS

The use of animals and the procedures employed were approved by the Kansas State University Animal Care and Use Committee.

Animals, study design, and sampling schedule.

Crossbred yearling heifers (n = 261) were blocked by initial body weight into heavy (330 ± 11 kg)- and light (280 ± 16 kg)-weight groups and assigned randomly to 24 pens with 10 or 11 heifers per pen. The pens were assigned, in a randomized complete block design, to a 2×2 factorial arrangement of treatments of dietary copper, normal (10 mg/kg of feed) or elevated (100 mg/kg of feed) and a commercial nutritional supplement (Linpro; Oleet Processing, Ltd., Regina, Saskatchewan, Canada) at 0 or 10% of the diet on a dry matter (DM) basis. The nutritional supplement was an extruded product of full-fat flaxseed and split field peas with added vitamins and minerals (22% crude protein; 23% fat) used as an energy and protein supplement in cattle diets. The basal diet included (on a DM basis) 35% dry-rolled corn grain, 35% wet corn gluten feed, 15% pelleted soybean hulls, 10% corn silage, and 5% minerals (1,000 mg of sodium, 1,500 mg of chlorine, 7,000 mg of calcium, 7,000 mg of potassium, 0.10 mg of cobalt, 0.6 mg of iodine, 0.25 mg of selenium, 60 mg of manganese, and 60 mg of zinc per kg of the feed), vitamins (2,200 IU of vitamin A and 22 IU of vitamin E per kg of feed), and antibiotics (300 mg of monensin [Rumensin; Elanco Animal Health, Greenfield, IN]/animal/day and 90 mg of tylosin [Tylan; Elanco Animal Health]/animal/day). Cattle were fed once daily and had ad libitum access to feed and water. Eight pen-floor fecal samples were collected with new plastic spoons from freshly defecated fecal pats from each pen (n = 24 pens) on day 116 (a day before heifers in the heavy-weight group were shipped for harvest) and from the remaining 12 pens housing light-weight heifers on day 131, which was a day before the shipment of the remaining lighter cattle for harvest. Spoons with feces were placed in individual Whirl-Pak (Nasco, Ft. Atkinson, WI) bags and transported in a cooler with ice packs to the Molecular Epidemiology and Microbial Ecology laboratory of Kansas State University. A schematic representation of the study design and sampling schedule is shown in Fig. 1.

Fig 1.

Schematic representation of the study design. Linpro is a commercial feed supplement containing an extruded blend of flax seed and field peas with added vitamins and minerals.

Isolation and speciation of Enterococcus.

The culture media used in the study were from Becton Dickinson (Sparks, MD). Approximately 1 g of each fecal sample was mixed in 10 ml of sterile phosphate-buffered saline, and 50 μl of the suspension was spread plated onto M-Enterococcus agar and incubated at 42°C for 24 h. Two putative colonies (pin-point red, pink, or metallic red) were picked from each plate, streaked onto blood agar plates, and incubated overnight at 37°C for 24 h. An esculin hydrolysis test was conducted, for presumptive genus identification, by inoculating a single colony into 100 μl of Enterococcosel broth in a 96-well microtiter plate (Becton Dickinson) and incubating at 37°C for 4 h. Esculin-positive isolates were stored in protect beads (Cryo-Vac, Round Rock, TX) at −80°C until further use. Species identification of enterococcal isolates was performed by multiplex PCR and superoxide dismutase (sodA) gene sequence analysis. The multiplex PCR identifies E. faecium, E. faecalis, E. gallinarum, and E. casseliflavus (17). DNA extraction was performed by suspending a single colony from the blood agar plate in nuclease-free water with Chelex 100 resin (Bio-Rad Laboratories, Hercules, CA) and boiling for 10 min. The ATCC strains of E. faecium (ATCC 19434), E. faecalis (ATCC 19433), E. gallinarum (ATCC 49579), and E. casseliflavus (ATCC 25788) were used as positive controls. Superoxide dismutase (sodA) gene PCR and gene sequence analysis were carried out for further confirmation of species (18).

PCR detection of tcrB, erm(B), tet(M), vanA, vanB, and virulence genes.

The DNAs of enterococcal isolates were extracted as described above. All primers used in the study were supplied by Invitrogen Life Technologies (Carlsbad, CA). The PCR assay for tcrB was as described by Hasman et al. (9), and a tcrB-positive E. faecium isolate (7430275-4) obtained from those authors' laboratory in Denmark served as the positive control. The primers and PCR conditions for the detection of erm(B) and tet(M) genes were as described previously by Amachawadi et al. (10). E. faecalis MMH 594 (Lynn Hancock, Division of Biology, Kansas State University) and an Escherichia coli strain harboring plasmid pFD 310 (19) served as positive controls for erm(B) and tet(M) genes, respectively. The vanA and vanB gene primers and PCR conditions were based on procedures described by Kariyama et al. (20). E. faecium (ATCC 51559) and E. faecalis (V583) were used as positive controls for vanA and vanB genes, respectively. A multiplex PCR was performed to identify asa1 (aggregation substance), gelE (gelatinase), cylA (cytolysin), esp (enterococcal surface protein), and hyl (hyaluronidase) genes (Vankerckhoven et al. [21]). E. faecalis MMH594 was again used as a positive control. A subset of tcrB-negative isolates, equal to the number of tcrB-positive isolates (n = 22; matched by pen, treatment, and date of sampling) was speciated and included for detection and statistical analysis of antibiotic resistance and virulence genes.

Copper and antibiotic susceptibility determinations.

Copper susceptibilities of enterococcal isolates were determined by the agar dilution method (9). Two tcrB-positive strains from our previous study (BAA-2127 and BAA-2128 [10, 11]) served as positive controls. The copper sulfate–Mueller-Hinton agar plates were prepared with concentrations of copper sulfate (Fischer Scientific, Fair Lawn, NJ) of 0, 2, 4, 8, 12, 16, 20, 24, 28, 32, 36, or 40 mM. Bacterial inocula were prepared by growing isolates in Mueller-Hinton II broth for 5 to 6 h, and the turbidity was adjusted to a McFarland standard of 0.5. The plates were spot inoculated with 20 μl of bacterial inoculum and incubated for 36 to 48 h at 37°C to determine growth or no growth. The MICs of tetracycline, tylosin, and vancomycin (Sigma-Aldrich, St. Louis, MO) were determined by microbroth dilution method according to Clinical and Laboratory Standards Institute (CLSI) guidelines (22). Antibiotic concentrations tested were based on the potency of the antibiotics. Stock solutions were prepared with sterile distilled water to obtain an initial concentration of 1,000 μg/ml. Antibiotics were tested at concentrations of 100, 50, 25, 12.5, 6.25, 3.125, 1.56, 0.78, 0.39, 0.195, and 0.098 μg/ml. Bacterial cultures were grown in 10 ml of Mueller-Hinton II broth for 6 h, and the inoculum concentrations were adjusted to 0.5 McFarland turbidity standards. The susceptibility determinations for antibiotics were carried out using 96-well microtiter plates (Becton Dickson) by incubating the plates with the bacterial inocula at 37°C for 24 h, and the results were recorded as growth or no growth. MIC determinations were carried out on both tcrB-positive (n = 22) and tcrB-negative (n = 22) isolates.

Transferability of tcrB gene.

The transferability of the tcrB gene from tcrB-positive to tcrB-negative strains via conjugation was determined by a filter-mating procedure (23). E. faecium strain TX5034, resistant to tetracycline [positive for tet(M); MIC > 100 μg/ml] and susceptible to spectinomycin (MIC = 12.5 μg/ml), and E. faecalis strain OG1SSp, resistant to spectinomycin (MIC > 100 μg/ml) and susceptible to tetracycline [negative for tet(M); MIC = 0.78 μg/ml], were used as controls (these strains were provided by Barbara E. Murray, University of Texas-Houston Medical School). The tcrB-positive donor strains were resistant to tetracycline [positive for tet(M)] and susceptible to spectinomycin. The tcrB-negative recipient strains were resistant to spectinomycin and susceptible to tetracycline [negative for tet(M)]. Filter mating was carried out with donors and recipients in a ratio of 1:10. Overnight cultures of donors (0.5 ml) and recipients (4.5 ml) were mixed, followed by cell collection using a 0.22-μm-pore-size nitrocellulose membrane filter. The strains were selected on brain heart infusion (BHI) agar for the tcrB-positive strains (donor, 40 μg of tetracycline/ml), tcrB-negative strains (recipient, 500 μg of spectinomycin/ml), and the transconjugants using both tetracycline and spectinomycin at these concentrations. The donor and recipient strains were grown on BHI agar plates containing 40 μg of tetracycline/ml and 500 μg of spectinomycin/ml, respectively. The transfer frequency was expressed as the number of CFU transconjugants per recipient CFU. The resultant transconjugants were also tested for tcrB, erm(B), and tet(M) genes by PCR and their susceptibilities to copper by using the procedures described above.

Southern blot hybridization.

Southern blot hybridization was performed on three tcrB-positive E. faecium strains according to the procedure described by Amachawadi et al. (11). The procedure was performed using digoxigenin High Prime labeling and detection starter kit II (Roche Diagnostics, Indianapolis, IN).

MLVA of tcrB-positive enterococci.

MLVA, a typing scheme based on six different tandem repeat loci for E. faecium, was performed as described by Top et al. (24). Six variable number tandem loci (VNTR)—VNTR-1, VNTR-2, VNTR-7, VNTR-8, VNTR-9, and VNTR-10—were used for genotyping tcrB-positive enterococcal isolates. The alleles were defined based on the amplicon size difference at each locus. The MLVA types (MTs) were assigned by using the MLVA database (http://www.umcutrecht.nl/subsite/MLVA/Method/). The population snapshot for MLVA types were performed using eBURST analysis (http://eburst.mlst.net).

Statistical analyses.

The statistical analysis was performed using STATA v12.1 (StataCorp, College Station, TX). Bivariate descriptive statistics of gene or phenotype prevalence by diet type were assessed prior to multivariable analysis. The likelihood-ratio chi-square test was performed to statistically assess unadjusted differences in prevalence proportion among two variables or between two genotypes [tcrB versus erm(B), tcrB versus tet(M), and erm(B) versus tet(M)]. Thereafter, the multivariable adjusted effects of important variables such as block, copper supplementation, commercial nutritional supplement (Linpro), and sampling days on the prevalence of tcrB, erm(B), and tet(M) genes in enterococci also were assessed. The association of each variable with prevalence was tested by using mixed (random and fixed)-effects logistic regression. Pen was considered as the experimental unit. The data were considered multilevel and longitudinal in nature since some of the pens were sampled twice, animals were clustered within each pen, multiple isolates within each fecal sample were analyzed, and there existed also the potential presence of within-animal dependencies because of the possibility of repeated sampling of the same animal's feces throughout the study period. The data were analyzed for each of the main effects of copper supplementation, commercial nutritional supplement (Linpro), blocking (by body weight), and the sampling days on the prevalence of tcrB gene first and then the interaction terms among remaining main effects. Copper and antibiotic MIC susceptibility data were analyzed using a log₂MIC-transformed and adapted survival analysis, and the MIC was expressed and compared as the median (MIC₅₀) after a log-rank assumption for normality (25). The differences in transfer frequency (number of transconjugants per recipient) of tcrB gene for both intra- and interspecies conjugation were assessed using the two-sample t test according to a normality assessment using the Shapiro-Wilk normality test. The transfer frequencies were also tested for equality of variance (variance comparison test), and independence was assumed.

Nucleotide sequence accession number.

The partial nucleotide sequence of one tcrB-positive E. faecium (strain 11A) was deposited in the NCBI GenBank database (accession no. JN660793).

RESULTS

Prevalence of tcrB, erm(B), tet(M), and virulence genes among fecal enterococci.

A total of 576 enterococcal isolates, comprising 288 isolates in each of the two main trial groups, control or copper supplemented, from fecal samples collected on days 116 and 131, were obtained. Twenty-two (3.8%) of the 576 isolates were positive for the tcrB gene. Of the 22 tcrB-positive isolates, 2 (2/288; 0.7%) were from cattle fed the basal diet with a normal concentration of copper (10 mg/kg), while 20 (20/288; 6.9%) were from cattle fed the basal diet supplemented with a higher concentration (100 mg/kg) of copper (Table 1). The overall prevalence of tcrB-positive enterococcal isolates was higher (P < 0.001) in the copper-supplemented group than in the control group (i.e., 6.9 versus 0.7%). The prevalence of tcrB gene also was affected by sampling period (P = 0.03), and the initial blocking of heifers by body weight (P = 0.03). The Linpro supplementation had no effect (P = 1.0) on the overall prevalence of tcrB gene. Based on multiplex PCR and sodA gene sequence analyses, all tcrB-positive (n = 22) and tcrB-negative (n = 22) isolates (selected by matching pen, date of sampling, and treatments) were E. faecium. All tcrB-positive (n = 22) and tcrB-negative (n = 22) isolates were positive for erm(B) and tet(M) and negative for vanA and vanB genes. The prevalence of erm(B) and tet(M) genes or multigene types [tcrB, erm(B) and tet(M) alone or in combination] is shown in Table 1. Forty-one of the 288 isolates (14.2%) of the control group and 56 of the 288 isolates (19.4%) of the copper group were negative for tcrB, erm(B), and tet(M) genes. The overall prevalence (control and copper groups) of erm(B), tet(M), and erm(B)-tet(M) combinations in enterococcal isolates were 53.3, 62.3, and 36.2%, respectively (Table 1). Both tcrB-positive (n = 22) and tcrB-negative (n = 22) isolates were negative for asa1 (aggregation substance), gelE (gelatinase), cylA (cytolysin), esp (enterococcal surface protein), and hyl (hyaluronidase) genes.

Table 1.

Prevalence of copper resistance gene tcrB and antibiotic resistance genes erm(B) and tet(M) among fecal enterococci in cattle fed diets supplemented with or without copper at elevated concentrations

Treatment group and copper concn (mg/kg) in the diet	Sampling period (days)	No. of enterococcal isolates	No. of enterococcal isolates (%) positive for:
			None	tcrB	erm(B)	tet(M)	erm(B) + tet(M)	tcrB + erm(B) + tet(M)
Control (10)	116	192	31	2	116	134	91	2
	132	96	10	0	65	62	41	0
Total		288	41 (14.2)	2 (0.7)	181 (62.8)	196 (68)	132 (45.8)	2 (0.7)
Copper (100)	116	192	41	14	82	106	51	14
	132	96	15	6	44	57	26	6
Total		288	56 (19.4)	20 (6.9)	126 (43.7)	163 (56.6)	77 (26.7)	20 (6.9)

Phenotypic susceptibilities to copper, tetracycline, tylosin, and vancomycin.

All tcrB-positive isolates grew on MH agar containing copper at a concentration of 20 or 24 mM. However, the tcrB-negative isolates did not grow on MH agar containing copper at a concentration of >8 mM. The two positive reference strains BAA-2127 and BAA-2128 had MICs of 20 and 24 mM, respectively. The median copper MIC of tcrB-positive isolates (22 mM) was higher (P < 0.001) than that of the tcrB-negative (4 mM) isolates (Fig. 2). The failure-function graph (Fig. 2) represents the cumulative susceptibility of both tcrB-positive and -negative isolates. Both tcrB-positive and -negative isolates were resistant (based on the CLSI breakpoint; ≥16 μg/ml) to tetracycline with median MICs of >100 and 37.5 μg/ml, respectively. The median tetracycline MIC of tcrB-positive isolates was significantly higher (P < 0.001) than that of the tcrB-negative isolates. All tcrB-positive isolates were resistant (based on a CLSI breakpoint of ≥32 μg/ml) to tylosin, with a median MIC of 50 μg/ml, and all tcrB-negative isolates were susceptible (based on a CLSI breakpoint of ≤8 μg/ml) to tylosin, with a median MIC of 1.6 μg/ml, respectively (P < 0.001). Both tcrB-positive and -negative isolates were susceptible (based on a CLSI breakpoint of ≤4 μg/ml) to vancomycin though the median MIC of tcrB-positive (0.39 μg/ml) was significantly higher (P < 0.001) than that of tcrB-negative (0.2 μg/ml) isolates. The median MIC results are shown in Table 2.

Fig 2.

Failure function graph showing cumulative susceptibilities of tcrB-positive and tcrB-negative isolates to increasing copper concentrations.

Table 2.

Median MICs of antimicrobial compounds for fecal enterococcal isolates with or without tcrB^a

Enterococcal isolate group (n)	Copper (mM)	MIC50 (μg/ml)
		Tetracycline	Tylosin	Vancomycin
tcrB-positive isolates (22)	22*	100*	50*	0.39*
tcrB-negative isolates (22)	4	37.5	1.6	0.20

^a*, significantly different from tcrB-negative isolates at P < 0.05.

Transferability of tcrB gene.

All 22 tcrB-positive E. faecium isolates were used to demonstrate both intraspecies (E. faecium) and interspecies (E. faecalis) transferability of tcrB gene by conjugation. The mean transfer frequencies of 22 isolates for both intra- and interspecies conjugations were 1.7 × 10⁻⁵ and 2.0 × 10⁻⁵, respectively. The conjugation frequency was higher in interspecies conjugation compared to intraspecies (P = 0.02). The transconjugants were positive for tcrB, erm(B), and tet(M) genes and phenotypically resistant to copper, tylosin, and tetracycline. The median copper MIC of the transconjugants was 20 mM.

Southern hybridization.

Hybridization of DNA from tcrB-positive E. faecium isolates with tcrB or erm(B) gene probes yielded distinct bands of ca. 175 to 180 kbp (Fig. 3). The tcrB-negative isolates showed bands ranging from 150 to 170 kbp. However, probing of tcrB-negative isolates with tcrB gene showed no hybridization.

Fig 3.

Southern blot hybridization of tcrB probe with S1 nuclease-digested genomic DNA of tcrB-positive enterococcal isolates. Lane 1, midrange PFG marker II; lanes 2, 3, and 4, S1 nuclease-digested genomic DNA of E. faecium isolate; lanes 5, 6, and 7, hybridization with tcrB probe.

MLVA.

The MLVA procedure revealed 20 novel MTs as shown in Fig. 4. Among 22 tcrB-positive isolates, two isolates were clonally related to the existing MT 7 and MT 190 in the database. The other 20 tcrB-positive isolates with unassigned MTs were submitted to the curator (J. Top, University of Utrecht, Utrecht, Netherlands) for sequence types and cluster assignment and also to update the global epidemiological database of E. faecium. The MT 449 isolate was closely related to the predominant hospital-adapted clone MT 12, which had been isolated from blood and feces of hospitalized patients. On eBURST, clustering of eight of our isolates suggests that they share one or more alleles at loci with MT 12 and MT 1, which were from hospitalized patients or clinics in the European Union. Fourteen of our tcrB-positive isolates were clonally different (437, 438, 440, 441, 442, 444, 445, 446, 447, 450, 452, 454, 455, and 456) and did not share alleles at any loci with existing MTs.

Fig 4.

eBURST-based population snapshot of E. faecium on 455 available MTs in the E. faecium MLVA database (http://www.umcutrecht.nl/subsite/MLVA). eBURST clustering of 20 two MTs of E. faecium isolates from feedlot cattle is indicated in boldface and enclosed in ovals. Each MT represents one node with a difference in one locus. The dotted circles indicate MTs that are closely related to the existing predominant clones in the database (solid circles).

DISCUSSION

This study, designed primarily to evaluate the effects of feeding an elevated concentration of copper (125 mg/kg of the feed) and a commercial energy and protein supplement (Linpro) on cattle performance, was utilized to investigate the effects of feeding elevated concentrations of copper on the prevalence of tcrB-mediated, copper-resistant enterococci in feedlot cattle feces. In the United States, cattle diets are supplemented with in-feed antimicrobials for growth promotion, to increase feed efficiency, to decrease respiratory infections, or to reduce the prevalence of liver abscesses at slaughter (26). Most in-feed antimicrobials are known to alter the gut microbial flora, thereby reducing fermentation loss of nutrients and suppressing pathogens (27). There is concern that gut bacterial commensals and pathogens in food-producing animals develop and propagate resistance to antibiotics and thereafter contaminate food products, providing a direct pathway for transmission to humans. Therefore, alternatives to antibiotics, including heavy metals and particularly copper and zinc, has been used as feed additives in food animal production (2).

In the present study, cattle diets were supplemented with a commercial energy and protein supplement (Linpro) to evaluate its performance benefits. This factor was a nuisance parameter for our study purposes and, as expected, the commercial supplement had absolutely no effect on the prevalence of tcrB-mediated copper-resistant fecal enterococci (P = 1.0). In cattle fed a higher concentration of copper, only a small proportion (6.9% [20/288]) of fecal enterococcal isolates harbored the tcrB gene and these isolates also were phenotypically resistant to copper. However, the higher prevalence of tcrB gene in cattle supplemented with copper compared to the control group (6.9% versus 0.7%) suggests the exertion of selection pressures of feeding elevated level of copper. Similar low prevalence (4.6 and 11.9%) were reported for enterococcal isolates from piglets fed an elevated level (125 mg/kg) of copper in two studies (10, 11). In contrast, Danish studies have reported a prevalence of 76% of copper-resistant enterococci in pig feces obtained at the time of slaughter (6), perhaps because of higher levels (175 to 250 mg/kg) of copper and supplementation for a longer period of time during their production cycle, starting from birth to slaughter.

All tcrB-positive enterococcal isolates in our study were E. faecium, which is in contrast to studies from Denmark and Australia that have reported the occurrence of the tcrB gene not only in E. faecium but also in E. faecalis, E. gallinarum, E. durans, E. casseliflavus, and E. mundtii (9, 28). Among Enterococcus species, E. faecium and E. faecalis most commonly colonize the gut in humans and animals (29). In our study, there were no E. faecalis isolates among the 22 tcrB-positive enterococci. Because only a limited number (n = 22) of tcrB-negative isolates were speciated, the occurrence of other species in cattle feces cannot be ruled out; in fact, it is more likely than not. Fluckey et al. (30), in a study on the species distribution of Enterococcus in cattle in the United States, reported E. faecium, E. faecalis, and E. durans as the dominant species. Factors such as age of the animal, diet, type, and level of supplement used in the diet, in-feed antibiotics, the geographical location and even isolation methods used are likely to have an impact on species distribution (17).

In Denmark, the copper resistance in enterococci was shown to coselect both macrolide and glycopeptide resistance (6). In our study, all 22 tcrB-positive isolates contained the erm(B) gene and were phenotypically resistant to tylosin; this is not surprising since the diet included tylosin, a macrolide that is routinely included in feedlot cattle diets in the United States for the prevention of liver abscesses (12). The erm(B) gene is the most common macrolide resistance genetic determinant present in animal and human enterococcal isolates (31). Because all tcrB-negative isolates were also positive for erm(B) gene, an assessment of coselection could not be made in our study (12). Perhaps the prevalence of erm(B) in all of the isolates tested is reflective of the selection pressure applied over a long period with the wide spread use of tylosin, a macrolide widely used to control liver abscesses in feedlot cattle (12). All tcrB-positive isolates from our study also contained the tet(M) gene, which is in agreement with our earlier studies that showed all tcrB-positive E. faecium and E. faecalis isolates of piglets to contain the tet(M) gene (10, 11). Unlike for erm(B), there were tcrB-negative isolates that also were negative for tet(M). Both erm(B) and tet(M) genes are carried on conjugative transposons, Tn916 and Tn916/Tn1545, which have a broad bacterial host range and a strong capacity for horizontal gene transfer among gut commensals (32). Our tcrB-positive isolates did not contain either vanA or vanB genes and were phenotypically susceptible to vancomycin. Although susceptible based on the CLSI breakpoint, the vancomycin MIC values were significantly (P < 0.001) higher for tcrB-positive isolates than the tcrB-negative isolates. This suggests a nonspecific function, such as an efflux pump used for copper homeostasis, might also aid in ridding bacterial cells of vancomycin, although not at an MIC sufficient to confer resistance. The prevalence of vancomycin-resistant enterococci in Europe is likely because of the historical use of avoparcin, a glycopeptide. Avoparcin was once applied extensively in livestock feed in Europe for growth promotion in broiler chickens, pigs, calves, and beef cattle (33). The absence of vanA or vanB genes in our study is not surprising since avoparcin was never approved for use in animal agriculture in the US (34) and a U.S. Food and Drug Administration prohibition on extra-label use of glycopeptides in animal agriculture has been in place since 1997 (35).

Southern blot hybridization revealed that the tcrB gene is located on an ∼180-kb plasmid. In our previous study in nursery piglets, the tcrB gene was on a conjugative mobilizable plasmid of size ∼194 kb (11). Earlier studies from Europe have reported the presence of tcrB, erm(B), and vanA genes on a 175-kb plasmid in an E. faecium isolate from a pig (6). Generally, megaplasmids are known to harbor antibiotic and heavy metal resistance genes (36). Conjugation systems involving plasmids and transposons are abundant in the genus Enterococcus and contribute to the dissemination of both antibiotic resistance determinants as well as virulence factors. The intra- and interspecies transferability of tcrB gene in our study is in agreement with our earlier studies with pig isolates (10, 11). Interestingly, the conjugation frequency was higher in the interspecies assay, possibly because of the production of pheromones (peptide signal) by the recipient E. faecalis isolate strains, which are specific for a particular plasmid or a group of related plasmids (37, 38). In this pheromone-inducible plasmid transfer, the recipient bacterium excretes the peptide pheromone molecule into the medium, where it can diffuse to a donor cell, thus facilitating transfer of antibiotic resistance determinants. The sex pheromones have also been described in E. faecium; more often, they are associated with vancomycin resistance (38). However, we detected a lower intraspecies conjugation frequency rate compared to interspecies conjugation, which may be due to the absence of vancomycin resistance gene determinants in our tcrB-positive isolates.

Enterococci have become major nosocomial pathogens in humans and often are associated with multiple antibiotic-resistant infections of bloodstream, the urinary tract, and surgical wounds (39). Virulence in enterococci is associated with several genes, particularly those that encode for aggregation substance (asa1), gelatinase (gelE), cytolysin (cylA), enterococcal surface protein (esp), and hyaluronidase (hyl) (40, 21). Studies have shown that esp is the major virulence gene in E. faecium, followed by asa1, gelE, and hyl genes (41, 21). None of the tcrB-positive (n = 22) or tcrB-negative (n = 22) isolates in our study carried any of the virulence genes tested. However, the evolution and propagation of antibiotic resistance involve a complex network consisting of multilevel populations and variable selection pressures. Both ecology and evolution play a vital role in linking antibiotic resistance and bacterial virulence (42, 43).

All 22 tcrB-positive isolates possessed diverse MTs, suggesting a genetically diverse and heterogeneous population in cattle. The majority of newly assigned MTs were clonally related in single and or double alleles at loci with clinically isolated human isolates. Note that this does not imply direction or causation. The order of entry into libraries is typically driven by various biases, such as the more intensively sampled clinical isolates. Twenty of the 22 strains were novel MTs. The MT 451 isolate was closely related to the predominant clone MT 112, which was isolated from blood and feces of hospitalized patients in Germany and in feces of chickens in the Netherlands. The MT 190 isolate was closely related to MT 325 (as per the MLVA database). The genetic diversity among the E. faecium isolates may be attributable to their propensity for genetic exchange coupled with a high rate of recombination (44). No doubt, as broader studies into populations of bacteria outside of clinical realms enter databases, a clearer picture will emerge.

In summary, the feeding of elevated dietary copper levels resulted in an increased prevalence of tcrB-mediated copper-resistant fecal enterococci in feedlot cattle. The tcrB gene was on a plasmid that can be transferred by conjugation, which is strongly suggestive of potential transferability of the gene under selection pressure. The finding of strong associations between copper resistance and other antibiotic (tetracycline and tylosin) resistance determinants may be significant because of the propensity of enterococci to transfer resistance genes to other bacteria in the gut when placed under selection pressure. Further studies are needed to investigate the effects of feeding antibiotics and their association with a copper resistance gene in enterococci to better understand the potential coselection and molecular epidemiology of multidrug-resistant enterococci in cattle.