OUP user menu

Disappearance of oxytetracycline resistance genes in aquatic systems

Christina A. Engemann, Laura Adams, Charles W. Knapp, David W. Graham
DOI: http://dx.doi.org/10.1111/j.1574-6968.2006.00419.x 176-182 First published online: 1 October 2006

Abstract

The disappearance of selected tetracycline resistance genes was investigated in different simulated receiving waters to determine conditions that maximize resistance gene loss after release. Wastewater from an operating cattle feedlot lagoon was provided to four pairs of duplicate 3-L flasks, and tet(O), tet(W), tet(M), tet(Q), and 16S rRNA gene levels were monitored over 29 days using real-time PCR. Treatments included simulated sunlight with 0, 25, and 250 µg L−1 nominal oxytetracycline (OTC) levels, respectively, and ‘dark’ conditions. Gene disappearance rates were always highest when light was present, regardless of OTC level. First-order loss coefficients (kd) for the sum of resistance genes were 0.84, 0.75, and 0.81 day−1 for 0.0, 25, and 250 µg L−1 OTC treatments over the first 7 days after release, respectively, whereas kd was 0.49 day−1 under dark conditions, which is significantly lower (P<0.10). kd varied fourfold among the four individual genes, although disappearance patterns were similar among genes. Results suggest that light exposure should be maximized in receiving waters in order to maximize resistance gene loss rate after release.

Keywords
  • antibiotic resistance
  • oxytetracycline
  • real-time PCR
  • gene disappearance kinetics
  • aquatic systems

Introduction

Antibiotics are one of the few classes of drugs that are used in large quantities in both agriculture and medicine (Chopra & Roberts, 2001). As such, agricultural use of antibiotics, especially for nontherapeutic applications, has become a major concern due to the development of antibiotic resistance in commensal bacteria and pathogens that might enter the food and water supply (Mølbak, 2004). Although links have been made between the food and the transmission of resistant organisms to humans (Witte, 1998; Angulo et al., 2000), similar links between water and human exposure are less established. A key question related to predicting risks associated with resistance gene migration via water to humans (and/or their microbial host) is determining the retention time of such genes after release into the environment. No quantitative data currently exist on resistance gene disappearance rates in aquatic systems, which was the focus of this study.

Antibiotics are used for many applications in agriculture, although they are most often used in cattle concentrated animal feedlot operations (CAFO) for the treatment of sick animals and/or prevention prophylaxis (NCCLS, 1999). Regardless of application, the antibiotic is fed to the animals where it temporarily resides in the animal and is then released through feces and urine to the environment. Although active agents are often detected in aquatic systems, levels are usually low, and a possibly greater concern via water is the release of resistant genes and/or microbial hosts to such systems. In fact, resistance gene levels can be high in receiving waters near CAFOs (Peak et al., in press; unpublished results), presumably due to gene or host migration via the hydrologic cycle.

To predict the impact of resistance gene release in surface waters, the fate kinetics of such genes must be determined. It is suspected that enteric organisms released in feces or urine are not likely to be ecologically competitive in natural aquatic systems, although ecological success or horizontal resistance gene transfer to environmental organisms are possible. Therefore, our goal here was to develop quantitative data that describe the fate of resistance genes (and host organisms) after release into different receiving waters. With such data, models might be developed to predict environmental levels of ‘resistance’ after release and to optimize waste handling systems.

Two factors related to receiving waters were assessed in this report; ambient oxytetracycline (OTC) levels and light conditions. Four water conditions (treatments) were simulated including light-exposed and dark conditions with high and low OTC loadings. Each treatment was provided lagoon wastewater from a CAFO, and tet(O), tet(Q), tet(W), tet(M), tet(B), and 16S rRNA gene levels were quantified over time using established real-time PCR methods (Smith et al., 2004; Peak et al., in press). OTC was chosen because it is used in both clinical and veterinary applications, and the genetic basis of resistance is well established permitting quantitative PCR methods. The five genes chosen for study were selected because they have been detected in many resistant genera (Roberts, 2005), they have been observed in large-animal operations (Aminov et al., 2002), and their sum (tetR) correlates with plate count data on the same samples. It was hypothesized that as OTC levels increased, resistance gene and organism retention times would increase due to selective enrichment, and that light supply might amplify this effect.

Materials and methods

Experimental design and conditions

Four aquatic treatments were monitored in duplicate using 3-L glass flasks under temperature- and light-controlled conditions. Treatments included simulated sunlight with 0, 25, and 250 µg L−1 OTC, respectively, and a dark condition where the flasks were enclosed in foil (receiving no OTC). In the light treatments, light was provided according to a 14/10 (on/off) hour diel cycle by a bank of ‘reptile’ lamps that simulate sunlight with a ‘day’ light intensity of 320 µE m−2 s−1(Cardoza et al., 2005). Reptile lamps provide 10–1000 times greater light intensities in the 400–650 nm range than normal ‘cool white’ indoor lights and, as such, are ideal for supporting photosynthetic activity in indoor systems. All flasks were mixed continuously at 200 r.p.m., provided Saran™ wrap lids to minimize evaporation, and maintained at 30°C. OTC (Sigma Aldrich, St Louis, MO) was provided via aqueous slurry every third evening to the OTC-amended flasks to balance photodegradative losses and provide a pseudo-steady load of OTC to those units.

The source water for the experiment was a protected surface reservoir at the Nelson Environmental Study Area located north of Lawrence, Kansas that had no previous contact with surface contamination (Graham et al., 1999; Cardoza et al., 2005). Ten liters of water was collected from the reservoir and 1.1 L aliquots were provided to the eight flasks. Mixing and the diel light cycle were commenced immediately, and the flasks were maintained to allow the waters to adjust to room conditions. After 2 weeks, flask waters were combined, mixed, and redistributed back into the flasks to create ‘uniform’ conditions before individual treatments. After 10 more days, specific treatments were created and the feedlot waste added. Chemical and microbiological analyses were performed throughout the preparation period to establish background water conditions.

The experiment commenced the evening before the waste was added by supplying OTC to four flasks and placing foil around the two dark flasks. The next morning (Day 0), the lagoon wastewater (initial OTC level ~3 µg L−1), which was collected the previous afternoon from the Kansas State University (KSU) Beef Cattle Research Center wastewater lagoon was added to all eight flasks at a 1 : 50 dilution relative to flask-water volume. The KSU Research Center is a controlled research facility where OTC is only used for sick animals (i.e., no growth promotion use). The dilution ratio was chosen based on measured resistance gene levels in the wastewater and was designed to achieve similar gene numbers seen in receiving waters during our parallel field-monitoring program (Peak et al., in press).

Sample processing and analysis routine

Samples for molecular, chemical, and biological analyses were collected on Days 0, 1, 4 (gene only), 7, 14, 21, and 29 after waste addition. For molecular work, samples were collected 100 mL per flask, which were sterilely filtered (0.2 µm; Nalgene, Rochester, NY), and then stored at −20°C before extraction and analysis. One milliliter subsamples were serially diluted in triplicate, and plated on Difco™ Plate Count-Agar media amended with 0.0 or 16 µg mL−1 OTC to quantify ‘total’ vs. ‘resistant’ culturable organisms. The plates were incubated at 25°C for 7 days in the dark before counting. Plate count data were used to verify that quantified gene levels correlated with culturable organisms and to also confirm that the gene data reflect viable cells. The 16 µg mL−1 concentration was based on observed minimum inhibitory levels for most organisms (NCCLS, 1999).

Water chemistry included temperature, dissolved oxygen (DO), pH, total phosphorus (TP), total nitrogen (TN), total organic carbon (TOC), and dissolved organic carbon (DOC), all of these were processed and analyzed according to previous methods (Graham et al., 1999; APHA, 2001; Cardoza et al., 2005). Additional samples were collected on Days 1, 8, 15, and 25 for OTC analysis, which were stored immediately in the dark at 4°C until further processing (Smith et al., 2004). Free OTC (i.e., soluble) was quantified using the RIDASCREEN® ELISA tetracycline detection kit (R-Biopharm, Darmstadt, Germany) according to the manufacturer's instructions (Aga et al., 2003), which was shown previously to significantly correlate with liquid chromatography electrospray ionization mass spectrometry data on similar samples (Peak et al., in press). Typically, 2.0 mL of water were collected per treatment (1 mL per flask), centrifuged at 3000 g for 1 min, decanted, and then diluted at a 1 : 1 (125 to 125 µL) or 1 : 24 (20 to 480 µL) ratio with sample buffer for 25 and 250 µg L−1 OTC treatment analysis, respectively. Duplicate analyses and the analysis of standards were performed to verify the accuracy of the method.

Molecular sample processing and real-time PCR

DNA was extracted using the MoBio UltraClean Soil DNA kit (Solona Beach, CA) with some modifications (Mo-Bio-Laboratories, 2004). The filters, beads, and extraction buffer were combined and agitated using a FastPrep (Qbiogene, Irvine, CA) cell disruptor for 30 s (6.0 speed) to pulp the filters. The samples were then incubated at 70°C for 10 min to enhance lysis of Gram-positive bacteria, and reagitated for 15 s (4.5 speed). The remaining purification steps followed the manufacturer's protocols.

Real-time PCR was used to quantify 16S rRNA genes and the five resistance genes using methods, probes/primers, and plasmid standards described elsewhere (Harms et al., 2003; Smith et al., 2004; Peak et al., in press). In summary, each DNA template (2 µL), and appropriate primers and probes were combined with iQ Supermix PCR reagent (BioRad, Hercules, CA). Reactions were performed using a BioRad iCycler with an iCycler iQ fluorescence detector and software version 2.3 (BioRad).

Data analysis

Mean physical/chemical water conditions for each treatment were estimated based on five samples per flask collected in duplicate (n=20), and confidence intervals (CI) were estimated from standard deviations. Mean gene copy numbers and CI were estimated like the water chemistry parameters except for n=12. The Kruskal–Wallis (K–W) test was used to assess differences among treatments for the molecular and chemistry data.

Gene disappearance rate coefficients (kd) for individual genes and tetR per treatment were estimated by fitting disappearance curves using a first-order model. Standard errors of the rate coefficients were calculated using ln-transformed linear regression. Given that gene disappearance was clearly biphasic, two kd estimates were made for treatment. Individual analysis of covariance (ancova) used a General Linear Model with time as a covariant to assess whether disappearance rate coefficients differed significantly among treatments. All statistical analysis was performed with Minitab® (Minitab, 2005).

Results

Estimates of microbial community size using real-time PCR and plate count data

Real-time PCR gene copy numbers vs. plate count data are summarized in Fig. 1, which show that 16S rRNA gene levels significantly correlate with total plate counts (Fig. 1a) and percent-resistance genes correlate with percent resistant culturable organisms (Fig. 1b), respectively (P<0.05). It should be noted that similar patterns of ‘resistance’ disappearance were noted using plate count data; however, molecular data are emphasized because real-time PCR detects genes in unculturable species, permits greater sample throughput, more rapid results, and higher precision than culturing techniques.

Figure 1

(a) Correlation between percent resistance genes relative to total 16S rRNA genes (i.e., tetR/16S rRNA) compared with percent resistance organisms relative to total plate counts (r2=0.67, P<0.01). (b) Correlation between total 16S rRNA gene copy number with total plate counts for equivalent samples from the experiment (r2=0.79, P<0.01).

Water conditions before and around the initial treatments

Mean physical and chemical water conditions among the eight flasks before the lagoon waste was added were pH=8.91, DO=6.6 mg L−1, DOC=6.81 mg L−1, TOC=13.9 mg L−1, TN=1.26 mg L−1, and TP=36 µg L−1, and no resistance genes nor OTC were detected in flask waters (<0.25 copies mL−1 and 0.5 µg L−1 detection limits, respectively). However, 1 h after waste addition (on Day 0), water conditions changed dramatically to pH=7.74, DO=8.1 mg L−1, DOC=20.1 mg L−1, and TOC=37.5 mg L−1. TN and TP levels were out of detection range and were not subsequently monitored. Mean initial gene copy numbers averaged across treatments were 7930, 48 900, 15 500, 29 000, and 101 000 copies mL−1 for tet(O), tet(W), tet(M), tet(Q), and 16S rRNA genes, respectively. tet(B) was not detected (10 copies mL−1 detection limit). Therefore, it was not monitored further. These initial gene levels were roughly proportional to measured levels in original feedlot lagoon samples.

Physical, chemical, OTC, and biological conditions as a function of treatment

Mean water conditions for each treatment are summarized in Table 1. Conditions were statistically the same among treatments (K–W test; P>0.05) with the exception of pH in the dark treatment, which was significantly lower than the other treatments (P<0.01). Measured OTC levels were below nominal values, although values differed significantly among the 0, 25, and 250 µg L−1 OTC light treatments (P<0.05). In contrast, 16S rRNA gene levels were significantly higher in the control and 25 µg L−1 OTC treatments relative to the dark and 250 µg L−1 OTC treatments (K–W test; P<0.05), presumably due to reduced primary production and increased antibiotic impact, respectively.

View this table:
Table 1

Physiochemical and biological conditions in each treatment

TreatmentPHDO (mg L−1)Temperature (°C)DOC (mg L−1)TOC (mg L−1)OTC level (μg L−1)16S rRNA genes (copies mL−1)
Control9.418.0229.020.437.1NA4.8 × 107
(1.19)(0.30)(2.01)(9.53)(47.1)(2.7 × 107)
Dark7.657.4129.713.524.2NA2.9 × 107
(0.21)(1.11)(1.81)(19.0)(17.0)(1.6 × 107)
25 μg L−1 OTC9.557.9929.718.049.54.354.7 × 107
(1.23)(1.21)(1.55)(11.4)(15.9)(3.45)(1.6 × 107)
250 μg L−1 OTC9.508.2630.219.850.5121.03.0 × 107
(1.17)(1.26)(1.47)(7.73)(15.6)(89.5)(1.0 × 107)
  • pH, dissolved oxygen, and temperature are arithmetric mean values based on ten separate samples.

  • DOC is mean values based on six separate samples.

  • TOC is mean values based on four separate samples.

  • 95% confidence intervals provided in brackets.

  • DO, dissolved oxygen; DOC, dissolved organic carbon; TOC, total organic carbon; OTC, oxytetracycline; NA, not applicable.

tetR and individual gene disappearance rate coefficients as a function of treatment

Figure 2 presents gene disappearance curves for tetR and tetR/16S rRNA gene ratio for the four treatments. Both nonnormalized and normalized data are presented because each provides useful information relative to resistance gene disappearance. tetR describes the absolute decline in resistance genes over time, whereas tetR/16S rRNA gene ratio describes the proportional decline in resistance genes relative to other genes in the system. Overall, resistance gene disappearance patterns were similar for both tetR and tetR/16S rRNA gene data. Gene disappearance was always initially rapid, but then slowed after about 5–7 days to a more moderate pace through Day 29. Owing to this biphasic disappearance pattern, separate rate coefficients were estimated for Days 0–7 and Days 7–29, respectively, to describe rates immediately after waste addition and over the longer term because it was suspected that both parameters may be of practical importance.

Figure 2

Total resistance (tetR) and 16S rRNA gene copy numbers over time for the four treatments. Gene abundances are presented as (○) absolute resistance gene numbers, (•) resistance gene numbers normalized to ambient 16S rRNA gene level, and (Δ) 16S rRNA gene level for each treatment over time. Error bars show the range of measured values at each point. Trendlines indicate the set of points used for estimating decay rate coefficients for resistance genes.

Table 2 summarizes all estimated kd values for tetR and the four individual genes for Days 0–7. Table 2 and Fig. 2 show that the presence or absence of light was the dominant factor to resistance gene disappearance rate in the systems. Except for tet(O) when not normalized (P=0.25), rate coefficients were significantly lower in the dark treatment relative to light treatments for all measures (P<0.10, ancova). Unexpectedly, no significant differences in kd were noted among the different OTC treatments for Days 0–7 data. Table 3 presents similar data for Days 7–29. In contrast to Days 0–7 data, no statistically significant differences in kd were observed among treatments using tetR data; however, significantly lower kd values were seen in the dark treatment vs. the other treatments for tetR/16S rRNA gene data (P<0.10). Figure 2 indicates that 16S rRNA gene levels were relatively constant over time for the three light treatments, whereas 16S rRNA gene levels declined significantly over time in the dark treatment, further suggesting reduced primary production (kd=−0.043 day−1; ancova, P<0.05).

View this table:
Table 2

Rate coefficients (kd) for (a) nonnormalized and (b) normalized resistance gene decay under the four treatments for Days 0–7

TreatmenttetRtet(O)tet(W)tet(M)tet(Q)
(a) Nonnormalized first-order decay coefficients (in day−1)
        Control−0.84−0.27−0.66−0.99−1.2
(0.10)(0.06)(0.09)(0.17)(0.09)
        Dark−0.49−0.18−0.38−0.48−0.66
(0.10)(0.09)(0.09)(0.10)(0.13)
        25 μg L−1 OTC−0.75−0.27−0.67−0.72−1.1
(0.17)(0.10)(0.13)(0.15)(0.19)
        250 μg L−1 OTC−0.82−0.30−0.69−0.81−1.2
(0.12)(0.05)(0.13)(0.09)(0.09)
(b) Normalized first-order decay coefficients to 16S rRNA gene level
        Control−0.98−0.41−0.80−1.1−1.4
(0.20)(0.19)(0.15)(0.22)(0.19)
        Dark−0.42−0.11−0.31−0.40−0.59
(0.10)(0.10)(0.07)(0.10)(0.14)
        25 μg L−1 OTC−0.82−0.33−0.73−0.78−1.1
(0.20)(0.13)(0.15)(0.19)(0.22)
        250 μg L−1 OTC−0.83−0.31−0.70−0.82−1.2
(0.14)(0.08)(0.16)(0.12)(0.11)
  • Standard errors of coefficient provided in brackets.

  • Significantly different coefficients provided in bold text (ancova, P<0.10).

  • Rate coefficients are provided for the sum of the four genes (tetR) as well as the four individual resistance genes.

  • OTC, oxytetracycline.

View this table:
Table 3

Rate coefficients (kd) for (a) nonnormalized and (b) normalized resistance gene decay under the four treatments for Days 7–29

TreatmenttetRtet(O)tet(W)tet(M)tet(Q)
(a) Nonnormalized first-order decay coefficients (in day−1)
        Control−0.09−0.11−0.12−0.03−0.15
(0.03)(0.03)(0.03)(0.06)(0.03)
        Dark−0.11−0.10−0.12−0.10−0.16
(0.02)(0.16)(0.02)(0.02)(0.04)
        25 μg L−1 OTC−0.12−0.13−0.18−0.09−0.21
(0.01)(0.02)(0.03)(0.01)(0.04)
        250 μg L−1 OTC−0.12−0.12−0.16−0.08−0.16
(0.01)(0.02)(0.03)(0.02)(0.05)
(b) Normalized first-order decay coefficients to 16S rRNA gene level
        Control−0.12−0.13−0.15−0.06−0.17
(0.04)(0.03)(0.03)(0.06)(0.02)
        Dark−0.07−0.07−0.08−0.06−0.13
(0.01)(0.01)(0.02)(0.02)(0.05)
        25 μg L−1 OTC−0.11−0.11−0.16−0.08−0.19
(0.02)(0.03)(0.03)(0.02)(0.03)
        250 μg L−1 OTC−0.14−0.14−0.19−0.11−0.19
(0.01)(0.02)(0.03)(0.02)(0.05)
  • Standard errors of coefficient provided in brackets.

  • Significantly different coefficients provided in bold text (ancova, P<0.10).

  • Rate coefficients are provided for the sum of the four genes (tetR) as well as the four individual resistance genes.

  • OTC, oxytetracycline.

Disappearance curves and rate coefficients for the individual genes followed the same general trends seen for total resistance gene loss in both Days 0–7 and 7–29 data. Table 2 shows that dark conditions resulted in the lowest kd values for all resistance genes for Days 0–7, regardless of normalization to 16S rRNA gene level. However, actual disappearance rates for each individual gene differed slightly with tet(O) kd values, typically the lowest, and tet(Q) values, typically the highest.

Finally, Fig. 3 presents relative abundances of the four resistance genes over time in the four treatments. Tet(M) and tet(Q) were consistently the dominant genes in all systems [i.e., 80.0% <tet(M) plus tet(Q) <87.2%], whereas tet(O) was an order magnitude lower than the other three genes [0.8% <tet(O) <1.1%]. Interestingly, tet(Q) was dominant in the original waste, whereas dominance switched to tet(M) by Day 29 in all treatments, indicating possible selective enrichment of organisms bearing this gene in the simulated receiving waters.

Figure 3

The proportional abundance of each gene over the 29-day sampling period for the four genes and corresponding disappearance curves for tetR/16S rRNA gene data for the four treatments.

Discussion

The possibility of bacteria gaining resistance to antibiotics is not a new concern and was, in fact, considered in the 1920s when penicillin was discovered (Roberts, 2000). It is generally agreed that overuse of antibiotics is resulting in more resistant species, which is of particular concern relative to pathogens in the food or water supply (Witte, 1998; Aarestrup & Wegener, 1999; Angulo et al., 2000; Chopra & Roberts, 2001; Mølbak, 2004). However, although resistance genes in the environment are a concern, minimal quantitative data exists on resistance gene and organism fate after release into aquatic systems, which is crucial for future risk assessments.

Results summarized in Tables 2 and 3, and Fig. 2 indicate that all four detected resistance genes were retained longer in the dark compared with well-lit conditions, and that OTC level had minimal influence on gene retention under the timeframe of this experiment. Specifically, all four individual genes and tetR disappeared two to four times more slowly in dark vs. lit systems, even when the lit systems were provided supplemental OTC. The lack of influence of OTC on resistance gene retention is very surprising because elevated OTC clearly reduced total bacterial community size (Table 1). The practical question is why was light so important, and OTC was not. Although our results are not conclusive, speculation is possible.

Darkness might impact gene fate in two general ways, either by reducing photolytic affects on genes or the host cells, or by altering community ecology due to reduced photosynthesis. Although reduced cell photolysis is possible, our results suggest that an altered ecology is likely the more important factor for two reasons. First, photosynthesis and primary production were clearly reduced in the dark treatment, as indicated by lower pHs and 16S rRNA gene levels. In theory, reduced primary production will result in less bioavailable organic carbon for ‘environmental’ nonphotosynthetic heterotrophs (i.e., competitors), which provides them less of a competitive advantage over resistance gene bearing organisms that are likely of enteric origin.

Evidence of this possibility is seen in biphasic gene disappearance in our systems and how gene disappearance rates differ between phases in lit vs. dark systems. We suggest that biphasic gene disappearance results from the ‘carryover’ of competitors, predators, and nutrients from the original lagoon waste, creating initial microbial ‘over-population’ in the systems. Further, the impact of overpopulation was greatest when fueled by greater photosynthesis, which is why initial disappearance rates were much higher in well-lit vs. dark systems. In theory, the second (slow) phase of gene disappearance, where differences in kd are less pronounced among treatments, results from reduced ecological pressure as competitors and predators from the original waste senesced or shifted dominance.

The second reason why ecological interactions are suggested over cell photolysis is that tetR/16S rRNA gene ratio kd data were consistently positive and correlated well with light/dark effects. tetR/16S rRNA gene ratio describes the relative abundance of resistance genes to total genes. Therefore positive values indicate that resistance genes are disappearing more rapidly than ‘other’ genes in the system. This is consistent with the possibility that resistance genes are hosted by less competitive (in receiving water), probably enteric organisms, which is most evident when primary production is greater and competition is more intense.

One interesting other result from this study was that the four detected resistance genes followed the same general disappearance pattern, although rates differed among genes. It had been expected that some genes would disappear more rapidly than other genes due to differ levels of gene mobility among organisms in the aquatic community (Clewell et al., 1995; Salyers et al., 1995a, b; Luna & Roberts, 1998; Roberts, 2000; Billington et al., 2002). However, although rates differed among genes, no statistically significant differences were noted even in the presence of elevated OTC. This result is further evidence that gene disappearance rate (over the short term) is primarily driven by ecological interactions that affect most organisms and are not necessarily gene specific. Despite these comments, Fig. 3 does hint that tet(M) was slightly enriched over 29 days, especially in high OTC water, although data are not statistically significant and require verification.

The practical implication of this study is that lagoon wastes, which contain organisms harboring resistance genes, should be provided maximum light after release into receiving water in the environment. Conversely, light-reduced habitats, such as bottom sediments or poorly mixed lagoons, will likely result in longer in situ retention of resistance genes and organisms. Therefore, if elevated light does increase resistance gene disappearance rates, possible remediation strategies are suggested. For example, we suggest that CAFO wastewater lagoons might be mixed to enhance light exposure effects, thus increasing resistance gene disappearance rates in such systems.

Acknowledgements

This research was supported by U.S. Environmental Protection Agency STAR Grant #R82900801-0. The authors would like to thank Patricia Keen, Teresa Lane, and Susan Okalebo who assisted in sampling and analysis, and Jim Droulliard of Kansas State University for providing lagoon waste for the work and general advice related to antibiotic use at cattle feedlots.

Footnotes

  • Editor: Stefan Schwarz

References

View Abstract