Populations of antibiotic-resistant coliform bacteria change rapidly in a wastewater effluent dominated stream

Abstract

Incomplete elimination of bacteria and pharmaceutical drugs during wastewater treatment results in the entry of antibiotics and antibiotic-resistant bacteria into receiving streams with effluent inputs. In Mud Creek in Fayetteville, AR, ofloxacin, trimethoprim, and sulfamethoxazole have been detected in water and sediment, and tetracycline has been detected in sediment downstream of treated effluent input. These antibiotics have been measured repeatedly, but at low concentrations (< 1 μg/L) in the stream. To determine if effluent input results in detectable and stable changes in antibiotic resistances downstream of effluent input, antibiotic resistance in Escherichia coli and total coliform bacteria in Mud Creek stream water and sediment were determined using a culture-based method. Isolated E. coli colonies were characterized for multiple antibiotic resistance (MAR) patterns on solid media and to evaluate E. coli isolate richness by amplification of a partial uidA gene followed by denaturant gradient gel electrophoresis (DGGE). Despite temporal variability, proportions of antibiotic-resistant E. coli were generally high in effluent and 640 m downstream. The MAR pattern ampicillin–trimethoprim–sulfamethoxazole was associated with a DGGE profile that was detected in effluent and downstream E. coli isolates, but not upstream. Percent resistance among coliform bacteria to trimethoprim and sulfamethoxazole was higher 640 m downstream compared to upstream sediment and water (with one exception). Resistance to ofloxacin was too low to analyze statistically and tetracycline resistance was fairly constant across sites. Resistances changed from 640 m to 2000 m downstream, although dissolved nutrient concentrations within that stream stretch resembled effluent. Antibiotic resistant bacteria are entering the stream, but resistances change within a short distance of effluent inputs, more quickly than indicated based on chemical water properties. Results illustrate the difficulty in tracking the input and fate of antibiotic resistance and in relating the presence of low antibiotic concentrations to selection or persistence of antibiotic resistances.

Introduction

Antibiotics and antibiotic-resistant bacteria are being detected in many streams (Ash et al., 2002, Kolpin et al., 2002). One of the sources for antibiotic resistance in streams is wastewater treatment plant (WWTP) effluent. Approximately 16,000 wastewater treatment facilities are located in the United States alone, including municipal WWTPs that treat less than 3.8 million liters per day of wastewater (Bitton, 2005, U.S. Environmental Protection Agency (U.S. EPA), 2004). Recent studies indicate incomplete elimination and release of antibiotic-resistant bacteria to effluent-receiving streams (Ferreira da Silva et al., 2006, Reinthaler et al., 2003, Watkinson et al., 2007). Antibiotic resistance genes carried on genetic elements that can be transferred between distantly related bacteria have been isolated from activated sludge and final effluent of WWTPs (reviewed in Schlüter et al., 2007). If antibiotic-resistant bacteria persist, dissemination of antibiotic resistance among bacteria may occur in effluent-receiving streams.

In addition, antibiotics are released to streams from WWTPs at low concentrations. Mud Creek, the stream in this study, has been found consistently to have antibiotics at concentrations less than 1 μg L^− 1 in water (Galloway et al., 2005). The presence of antibiotics at low concentrations may influence the levels of antibiotic resistance in stream ecosystems. Antibiotics may act as signaling compounds at sub-inhibitory concentrations and select for antibiotic-resistant bacteria. For example, efficiency of transfer of a Bacteroides conjugative transposon can be increased in the presence of tetracycline at low concentration (50 μg L^− 1) (Song et al., 2009). A lab-scale experiment on the impacts of sub-inhibitory concentrations (20 to 250 μg L^− 1) of oxytetracycline on aquatic mesocosms indicated an increase in the rate of resistance gene selection relative to the 16S rRNA gene pool and an increase of the resistance genes (mostly tet(M) and tet(W)) (Knapp et al., 2008). Although inputs of antibiotic-resistant bacteria and antibiotics to streams have been reported, it is not clear if or to what extent inputs of low-levels of antibiotics from WWTPs (< 1 μg L^− 1) affect antibiotic resistance among effluent bacteria after entry into the receiving stream.

The input of WWTP effluent may also influence antibiotic resistance of bacteria by impacting antibiotic concentrations and resistance in sediment. Bacteria create highly complex microbial structures called biofilms on or in sediment where individuals are protected from stressors, including antibiotics (Anderl et al., 2000, De Beer et al., 1994, Elasri and Miller, 1999). Some antibiotics such as ofloxacin and sulfonamides can be sorbed to soils (Thiele-Bruhn et al., 2004, Tolls, 2001); thus, antibiotics introduced from WWTPs may persist in stream sediment. In fact, Massey et al. (2010) reported that antibiotics introduced from WWTP such as ofloxacin and sulfamethoxazole are retained and can accumulate in Mud Creek sediment (1 ng g^− 1 to > 1 μg g^− 1).

E. coli are utilized as indicators of fecal contamination because they are intestinal bacteria that are not supposed to persist and grow in the environment, but yet E. coli are known to persist in stream sediment up to several weeks (Jamieson et al., 2005). Because bacteria live closely together in biofilms, bacteria may be more likely to exchange their genes. With the possibility of antibiotics influencing phenotypic and genotypic traits of bacteria, biofilms potentially are important reservoirs for the development and horizontal gene transfer of antibiotic resistance in aquatic ecosystems (Davey and O'Toole, 2000, Molin and Tolker-Nielsen, 2003).

While low concentrations of antibiotics can affect behavior and antibiotic resistance gene selection in bacteria, the question remains as to how quickly and to what extent resistance is changed in bacteria after effluent enters freshwater streams and these streams have continual exposure to antibiotic inputs at low concentrations. Additionally, because of multiple antibiotic resistances, increased resistances may occur for number of antibiotics because of the presence of other antibiotics in water or sediment. In this study, levels of resistance in E. coli and total coliform bacteria were determined seven times in effluent and stream water, upstream (20 m from effluent) and at two locations downstream of effluent input (640 and 2000 m). In addition, E. coli were isolated after incubation of water samples to confirm antibiotic resistance, characterize multiple antibiotic resistance (MAR) patterns, and to determine richness of E. coli isolates using PCR amplification-denaturant gradient gel electrophoresis (DGGE).

Generally, greater numbers of bacteria can be found in stream sediment compared to the water column (Schmidt et al., 2000). Bacterial numbers in the water column can increase by resuspension of biofilms, sediments, and inputs from adjacent land when stream flow increases (McDonald et al., 1982, Jamieson et al., 2005). If streambed sediments are seeding water with antibiotic-resistant bacteria, resistance may be greater downstream. Antibiotic resistance was therefore measured three times in stream sediment upstream (20 m) and downstream (640 m) of effluent input.

We hypothesized that levels of antibiotic resistance in E. coli and total coliform bacteria, and composition of antibiotic-resistant E. coli isolate populations, would be similar to effluent and higher downstream compared to upstream of effluent inputs in water and sediment. Additionally, we hypothesized that levels of antibiotic resistance and multiple antibiotic resistance (MAR) patterns would be related to antibiotics present in the stream.