Vancomycin-Resistant Enterococcus spp. Isolated from Wastewater and Chicken Feces in the United States

Abstract

Vancomycin-resistant Enterococcus spp. (VRE) were isolated from sewage and chicken feces but not from other animal fecal sources (dog, cow, and pig) or from surface waters tested. VRE from hospital wastewater were resistant to ≥20 μg of vancomycin/ml and possessed the vanA gene. VRE from residential wastewater and chicken feces were resistant to 3 to 5 μg of vancomycin/ml and possessed the vanC gene.

Vancomycin resistance in Enterococcus species is becoming a major concern in clinical settings as the rate of occurrence of vancomycin-resistant Enterococcus spp. (VRE) implicated in disease increases. For example, by 1999 the incidence of VRE-mediated nosocomial infections in intensive care units had increased 43% from that of the period of 1994 to 1998 (23). Enterococcus faecalis and E. faecium are reported more frequently as etiological agents of disease than are other enterococci (14, 15), but other species, such as E. avium (25), occasionally cause disease.

Several operons that mediate vancomycin resistance in Enterococcus spp. have been identified. Perhaps the most significant from an epidemiological standpoint is vanA-mediated vancomycin resistance, as these genes are carried on transposon Tn1546 (4) and confer inducible high-level resistance to vancomycin and teicoplanin (19). The chromosomally encoded phenotype mediated by the vanC gene is marked by low-level resistance to vancomycin (20) and is an intrinsic characteristic of E. gallinarum (11), E. casseliflavus, and E. flavescens (24).

VRE that are resistant to high levels of vancomycin can be readily isolated from the feces of domestic animals in Europe (2, 10) and from humans with no exposure to hospitals (17, 31). There has been no report of high-level vancomycin resistance (>32 μg/ml) in Enterococcus spp. from animal feces or from humans without hospital exposure in the United States. In spite of a 1997 call for the investigation of sources of VRE outside the health care setting in the United States (21), there are remarkably few publications containing such data (8). VRE that are intrinsically resistant to low levels of vancomycin, such as E. gallinarum, E. casseliflavus, and E. flavescens, have been isolated from bird feces (30), and E. gallinarum containing the vanC gene has been isolated from chickens and farm lagoons (8).

As part of another study (18), our laboratory isolated thousands of fecal streptococci (a group that includes Enterococcus spp. and other group D Streptococcus spp.) from animal feces, wastewater, and surface waters. Some of these isolates were resistant to high levels (>32 μg/ml) of vancomycin. In order to investigate the presence of VRE in wastewater, animal feces, and surface waters, all VRE were identified to the genus and species level.

Isolation of vancomycin-resistant fecal streptococci.

Fecal streptococcus isolates were obtained from the feces of cattle, chickens, dogs, pigs, and wild animals (birds and raccoons). Fecal streptococci were also isolated from wastewater samples collected at a central sewer lift station (designated LF) serving residential neighborhoods (sample LF1) (Table 1) and from a line connecting a hospital to the main sewer line (hospital wastewater samples 1 and 2) (Table 1) in Tampa, Fla. Surface water samples were collected from three major tributaries of the St. Johns River in Jacksonville, Fla, and from the Hillsborough River in Tampa, Fla.

TABLE 1.

Screening for VRE

Source (sample)	No. of isolates screened	Vancomycin concn (μg/ml)	No. of colonies on vancomycin-amended media	No. of isolates:		Identified organism(s)
				Preliminarly identified as Enterococcus	Identified as PYR+ and LAP+
Dog	481	10	9	3	0	Weissella, Leuconostoc
Cow	474	10	0	NAa	NA	NA
Chicken	470	10	2	2	2	E. gallinarum
Pig	362	10	9	0	NA	NA
Residential wastewater (LF1)	96	10	0	NA	NA	NA
Residential wastewater (LF2)	10,000	10	36	0	NA	NA
Residential wastewater (TP)	10,000	3	20	14	14	E. gallinarum
Residential wastewater (LF3)	10,000	3	192	2	2	E. gallinarum
Hospital wastewater (1)	96	10	6	6	6	E. avium
Hospital wastewater (2)	96	10	5	5	5	E. faecium
Surface water	2,112	10	7	0	NA	Weissella, Leuconostoc

^aNA, not applicable. 

Fecal streptococci were isolated by membrane filtration followed by incubation with selective-differential media. For stool samples, 1.0 g of feces was dissolved in 100 ml of phosphate-buffered saline (PBS) (pH 7.2) followed by thorough homogenization by vortexing. Subsequent dilutions over several orders of magnitude were made from the homogenate. Wastewater samples were similarly vortexed and diluted in PBS. A volume of 0.1 to 100 ml of each surface water sample, depending upon the contamination level at the site, was analyzed. Each sample was filtered through a 0.45-μm-pore-size membrane, which was placed on a sterile filter paper pad soaked with Enterococcosel broth (BBL) and incubated at 37°C for 24 h. The colonies were individually transferred with sterile toothpicks to a 96-well microtiter plate whose wells each contained 200 μl of Enterococcosel broth and were incubated at 37°C for 24 h. Isolates in wells that turned black or dark brown, indicating esculin hydrolysis, were transferred by a replica plating device to Trypticase soy agar plates amended with 10 μg of vancomycin/ml. All sources except the cattle feces yielded some colonies with the ability to grow on vancomycin-amended plates (Table 1). Isolates that displayed growth on the vancomycin-amended plates were subjected to further biochemical testing.

Fecal streptococci that grew on vancomycin-amended media were subjected to preliminary tests to determine their membership in the genus Enterococcus, including Gram staining and growth in brain heart infusion broth (Becton Dickinson, Cockeysville, Md.) with 6.5% NaCl (35°C), at pH 9.6 (35°C), and at 45°C (3). Most isolates from surface waters and animal feces were determined not to belong to the genus Enterococcus (Table 1). Isolates that passed the preliminary Enterococcus tests were assayed for leucine aminopeptidase (LAP) and pyroglutamyl aminopeptidase (PYR) activity using the BactiCard Strep test kit (Remel, Inc., Lenexa, Kans.). These tests differentiate Enterococcus species from members of the intrinsically vancomycin-resistant genera Leuconostoc, Pediococcus, and Weissella (13). Isolates from dog feces were PYR⁻ and were identified as members of the genus Weissella (Table 1). All of the isolates from chicken feces and hospital wastewater and some of the isolates from residential wastewater were PYR⁺ and LAP⁺, confirming their membership in the genus Enterococcus. Confirmed VRE isolates were identified to the species level with the API Strep biochemical test system (BioMerieux, St. Louis, Mo.)

Detection of VRE by primary isolation on vancomycin-amended media.

Presumptive VRE were also isolated from wastewater by plating directly on vancomycin-amended mEI agar (34) (3 or 10 μg/ml; Table 1). The LF lift station was described above, and samples LF2 and LF3 were taken on separate dates at the LF lift station. Wastewater was also sampled at a second residential lift station (designated TP), which was not connected by flow to the LF lift station. The sampling events for each of the residential wastewater samples were approximately 2 months apart.

One milliliter of wastewater was suspended in 50 ml of PBS. Each sample was filtered onto a membrane filter (pore size, 0.45 μm), which was transferred to mEI agar amended with vancomycin (10 μg/ml) and incubated at 41°C for 24 h. No vancomycin-resistant Enterococcus isolates were obtained from residential wastewater screened against 10 μg of vancomycin/ml (isolate LF2, Table 1). Two samples (LF3 and TP) were therefore filtered and cultured on mEI agar amended with 3 μg of vancomycin/ml at 41°C for 24 h. The subsequent processing steps were the same for all isolates. Colonies were transferred with sterile toothpicks to microtiter dish wells containing 180 μl of Enterococcosel broth. These cultures were incubated at 41°C for 24 h. Growth from the microtiter dishes was transferred with a replicator to Mueller-Hinton agar plates amended with 3 (samples LF3 and TP) or 10 μg of vancomycin/ml, and the plates were incubated for 24 h at 41°C.

Presumptive VRE isolates were confirmed as described above and were identified to the species level with the API Strep system. PCR was used to assess the genotype of each VRE isolate. The vanA, vanB, and vanC genes were targeted with specific primers (7) in separate PCRs. The primer sequences for vanA were (5′→3′) CAT GAA TAG AAT AAA AGT TGC AAT A and CCC CTT TAA CGC TAA TAC GAT CAA; for vanB, they were GTG ACA AAC CGG AGG CGA GGA and CCG CCA TCC TCC TGC AAA AAA; and for vanC, they were GAA AGA CAA CAG GAA GAC CGC and ATC GCA TCA CAA GCA CCA ATC. Positive controls were E. faecalis A256 for VanA, E. faecalis V583 for VanB, and E. gallinarum VR-42 for VanC. These combinations of primer sets and control strains produced PCR products of ca. 1,030, 433, and 796 bp, respectively, as expected based on the literature (7). PCR products of the correct size were obtained from hospital wastewater isolates using the vanA-specific primers and from the chicken feces and residential wastewater isolates using the vanC-specific primers. DNA from each VRE could be amplified with only one primer set (data not shown).

To confirm the specificity of PCR, amplicons were sequenced and aligned with DNA sequences in GenBank using BLAST 2.1 (National Center for Biotechnology Information). The PCR product from each strain was purified using the QIAquick DNA purification kit (Qiagen, Valencia, Calif.). The first primer of each set for PCR (see above) was used for DNA sequencing of the PCR product. Cycle sequencing was carried out with an ABI model 310 automated sequencer (Perkin-Elmer, Boston Mass.). The DNA sequences of the vanA PCR amplicon and the vanC PCR amplicon demonstrated over 90% identity with representative sequences in GenBank. The vanA amplicon was aligned with the E. faecium plasmid pIP816 vanA gene sequence (GenBank accession no. X56895) (12), and the vanC amplicon was aligned with the E. gallinarum vanC gene sequence (GenBank accession no. AF162694) (11).

The MICs of vancomycin, ampicillin, erythromycin, and tetracycline for VRE isolates were determined by the broth microdilution method as described in National Committee for Clinical Laboratory Standards guidelines. The end points of the antibiotic series used in these experiments were defined by previously published breakpoints corresponding to full clinical resistance or susceptibility for Enterococcus (22). The three VRE isolates from hospital wastewater that were identified as E. avium (Table 2) displayed resistance to erythromycin (>10 μg/ml) as well as to vancomycin (20 to 32 μg/ml). Coupled resistance to vancomycin (≥32 μg/ml) and erythromycin (≥8 μg/ml) has been previously noted in E. faecium isolates harboring the vanA gene (1) but not, to the best of our knowledge, in E. avium. Two of the vancomycin-resistant E. avium strains displayed unstable resistance in that they were resistant to 64 μg of vancomycin/ml when first isolated but displayed successively less resistance when subcultured in the presence of vancomycin. Their resistance to 20 μg of vancomycin/ml, however, was stable.

TABLE 2.

MICs of antibiotics for VRE isolates from residential and hospital wastewater

Source (isolate)	Species	Genotype	MIC (μg/ml)
			Ampicillin	Erythromycin	Tetracycline	Vancomycin
Hospital (2A1)	E. avium	vanA	5	>10	<2	32
Hospital (2F5)	E. avium	vanA	5	>10	<2	20
Hospital (2H2)	E. avium	vanA	2.5	>10	<2	20
Hospital (2B29)	E. faecalis	vanA	<0.5	>10	<2	>100
Hospital (1K4)	E. faecalis	vanA	<0.5	>10	<2	>100
Residential (8B3)	E. gallinarum	vanC	<0.5	<0.1	22	3
Residential (1C4)	E. gallinarum	vanC	<0.5	<0.1	<2	3
Residential (3H1)	E. gallinarum	vanC	<0.5	<0.1	>25	3

The two E. faecalis isolates from hospital wastewater were also resistant to erythromycin. The three residential wastewater isolates, which were identified as E. gallinarum, were resistant to low levels of vancomycin, and two of the E. gallinarum isolates were resistant to tetracycline. The E. gallinarum isolates from chicken feces were lost in a laboratory accident before they were tested for resistance to other antibiotics.

Enterococcus species that were resistant to high levels of vancomycin could be readily isolated from hospital wastewater even without the use of vancomycin during the initial screening (Table 1). Enterococci resistant to low levels of vancomycin were isolated in the same manner from chicken feces, demonstrating their relatively high prevalence. It was necessary to screen residential wastewater isolates on vancomycin-amended media in order to isolate low-level VRE, and no high-level VRE were found in residential wastewater. It should be noted that this screening was not exhaustive and serves mainly to illustrate the vastly higher frequency of occurrence of VRE in hospital wastewater than in residential wastewater in this Florida community. For comparison, a study in Spain noted a much lower frequency of high-level VRE (0.4%) in wastewater entering a main treatment plant (33) than demonstrated here for hospital wastewater (6.25 and 5.21% for two samples). However, VRE were present at a higher frequency than was measured for residential wastewater in this study. In France, 3.7% of enterococci isolated from fecal samples of hematology patients and 1.8% of enterococci isolated from fecal samples of a control outpatient group were resistant to high levels (≥16 μg/ml) of vancomycin (17).

Although one tends to think of wastewater that has entered a central sewer collection system as well-contained, breaches in the integrity of such systems can contribute to the contamination of natural waters by pathogens and indicator organisms. Infiltration, or the flow of groundwater into wastewater collection systems when soils are saturated with water, is a major problem in wastewater management, increasing the cost and decreasing the efficiency of wastewater treatment (32, 35). The converse of infiltration can also occur; that is, sewage may leak out of cracked pipes when groundwater levels are low, and microorganisms can be transported to groundwater or surface water under appropriate hydrological conditions. Combined sewer overflow systems, which handle both sewage and stormwater, can allow the release of microorganisms to surface water during storm events.

VRE that are resistant to high levels of vancomycin may well still be uncommon outside health care settings in the United States, as demonstrated by the failure to isolate high-level VRE from residential wastewater in this study. However, each of the vancomycin-resistant Enterococcus species isolated during this study, including E. gallinarum, has been implicated in disease. E. faecalis is one of the major pathogens of the genus, while E. avium (9, 25–27) and E. gallinarum (28, 36) are documented etiological agents of endocarditis and bacteremia. Intrinsically resistant VRE such as E. gallinarum can be particularly troublesome, as in vitro tests may indicate vancomycin susceptibility in E. gallinarum isolates that are resistant to treatment in vivo (28).

Epidemiological evidence from Europe suggests that VRE are horizontally transmitted from animals to humans (5, 6, 31). The transmissibility of VRE by nonnosocomial routes, coupled with the difficulties encountered in treating infections caused by VRE, indicates that great care should be taken to avoid introducing these organisms into the environment. Such contamination can occur during sewage spills or as the result of other, less obvious failures of wastewater systems, such as leaky collection lines or ineffective on-site (septic) systems. Public health may be threatened by VRE release, particularly if the organisms reach groundwater, which may be consumed without treatment, or if they impact recreational waters. Evidence exists that Enterococcus spp. can proliferate in subtropical and tropical soils and waters (16, 29); therefore, introduction of VRE into such environments may be especially problematic.

This study found that high-level VRE could be isolated without enrichment or screening on antibiotics from hospital wastewater but not from residential wastewater. Much less is known about the distribution of VRE in healthy humans in the United States than in Europe, and almost no information exists on the survival and growth of VRE that are released into natural waters. Accurate assessment of the magnitude of the public health threat represented by VRE in wastewater depends upon further investigation of these questions.