Antibiotic resistant and extended-spectrum β-lactamase producing faecal coliforms in wastewater treatment plant effluent

Wastewater treatment plants (WWTPs) provide optimal conditions for the maintenance and spread of antibiotic resistant bacteria (ARB) and antibiotic resistance genes (ARGs). In this work we describe the occurrence of antibiotic resistant faecal coliforms and their mechanisms of antibiotic resistance in the effluent of two urban WWTPs in Ireland. Effluent samples were collected from two WWTPs in Spring and Autumn of 2015 and 2016. The bacterial susceptibility patterns to 13 antibiotics were determined. The phenotypic tests were carried out to identify AmpC or extended-spectrum β-lactamase (ESBL) producers. The presence of ESBL genes were detected by PCR. Plasmids carrying ESBL genes were transformed into Escherichia coli DH5α recipient and underwent plasmid replicon typing to identify incompatibility groups. More than 90% of isolated faecal coliforms were resistant to amoxicillin and ampicillin, followed by tetracycline (up to 39.82%), ciprofloxacin (up to 31.42%) and trimethoprim (up to 37.61%). Faecal coliforms resistant to colistin and imipenem were detected in all effluent samples. Up to 53.98% of isolated faecal coliforms expressed a multi-drug resistance (MRD) phenotype. AmpC production was confirmed in 5.22% of isolates. The ESBL genes were confirmed for 11.84% of isolates (9.2% of isolates carried blaTEM, 1.4% blaSHV-12, 0.2% blaCTX-M-1 and 1% blaCTX-M-15). Plasmids extracted from 52 ESBL isolates were successfully transformed into recipient E. coli. The detected plasmid incompatibility groups included the IncF group, IncI1, IncHI1/2 and IncA/C. These results provide evidence that treated wastewater is polluted with ARB and MDR faecal coliforms and are sources of ESBL-producing, carbapenem and colistin resistant Enterobacteriaceae. Importance Antibiotic resistant bacteria (ARB) are an emerging environmental concern with a potential impact on human health. The results provide the evidence that treated wastewater is polluted with antibiotic resistant bacteria containing mobile resistance mechanisms of importance to clinical treatment of pathogens and multi-drug resistant (MDR) faecal coliforms. They are sources of relatively high proportions of ESBL-producing Enterobacteriaceae, and include carbapenem and colistin resistant Enterobacteriaceae. The significance of this study is the identification of the role of WWTPs as a potential control point to reduce or stop the movement of ESBL, MDR and colistin resistant bacteria into the environment from further upstream sources, such as human or animal waste.


83
The dissemination of antimicrobial resistance within bacterial communities and the selection 84 of new resistance mechanisms are due to the large-scale use of antibiotics in agricultural, 85 veterinary and human clinical applications (1-8). The emergence of antibiotic resistant 86 bacteria (ARB) is a major public health issue which poses a serious therapeutic challenge 87 worldwide (6). Wastewater treatment plants (WWTPs) are considered potential sources of 88 ARB and antibiotic resistance genes (ARGs), which play an important role in the spread of 89 antibiotic resistance into the environment (9, 10). The urban WWTPs receive wastewater 90 from human communities, which contains high concentrations of chemical matter, including 91 antibiotics and microorganisms, including ARB. Therefore, WWTPs are favourable 92 environments with optimal conditions for the development and spread of ARB and ARGs 93 (11,12). Both ARB and ARGs were detected in wastewater samples globally (13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24). 94 However, little is known of the fate of these bacteria; and the role of WWTPs in releasing 95 ABR and ARGs into the environment through treated effluent (13). A recent report by Flach 96 et al shows no evidence for the selection of antibiotic resistance in WWTPs (25); however, 97 large amounts of resistant bacteria were identified throughout the wastewater treatment 98 process (7,26). The conventional wastewater treatment process can remove some ARB (27,99 28), but ARB were still found in large proportions in the effluent (18,29). In some cases, 100 ARB were detected at higher rates in WWTP effluent than in the influent (30,31). 101 AmpC cephalosporinases and extended-spectrum β-lactamases (ESBLs) are some of the most 102 clinically important antibiotic resistance mechanisms (32). The dissemination of AmpC or 103 ESBL producing Escherichia coli were identified in different types of aquatic environments, 104 particularly in wastewater (33)(34)(35)(36). The prevalence of AmpC or ESBL producing bacteria 105 pose a global health problem due to limitations of therapeutic options for the treatment of 106 infections caused by these bacteria (37). The ESBL genes are frequently located on mobile 107 genetic elements (38). Among more than 300 subtypes of ESBL genes, bla TEM and bla SHV 108 groups were the most common ESBL genes identified in human pathogens until the late 109 1990s. These groups were replaced by bla CTX-M genes since the beginning of the 2000s and 110 Escherichia coli became the most prevalent ESBL producing bacteria among clinical 111 Enterobacteriaceae (39). 112 The monitoring of antibiotic resistance from WWTPs provides the information required to 113 track the dissemination of ARB and ARGs into the environment (40, 41). Moreover, the 114 analysis of ARB and ARGs in urban WWTPs is considered as an alternative method for the 115 indirect study of antibiotic resistance in human populations from which the WWTPs receive 116 wastewater (42). Indeed, the resistance rates of indicator bacteria in wastewater may give 117 useful information to identify the changes in resistance in the human populations (43). The 118 objectives of this study were to characterise the faecal coliforms resistome leaving urban 119 WWTPs via the effluent. This was achieved through i) assessment of the prevalence of 120 antibiotic resistant faecal coliforms in the effluent from two Irish urban WWTPs, ii) 121 characterisation of the antibiotic resistance profiles of these bacteria, iii) identification of the 122 occurrence of AmpC or ESBL producing faecal coliforms and iv) identification of the 123 resistance mechanisms and their potential mobility. 124 125

Isolation of total faecal coliforms 127
Final effluent samples were collected from two urban WWTPs in Ireland during early Spring 128 and late Autumn in 2015 and 2016. These WWTPs were representative of medium sized 129 WWTPs with 100% urban agglomerations, include tertiary treatment, and the distance 130 between them was less than 100km. Faecal coliforms were isolated using the membrane 131 filtration method (44). The effluent samples (1 mL and 10 mL) were filtered through 132 nitrocellulose membranes (Sigma Aldrich/Merck). The filters were then incubated on mFC 133 agar at 37 o C for 24 hours in the presence or absence of antibiotics: amoxicillin (32 mg/L), 134 ciprofloxacin (1 mg/L) or tetracycline (16 mg/L). All procedures were performed in triplicate. 135

Antibiotic susceptibility test using agar dilution and disk diffusion methods 136
Bacterial isolates were subjected to antibiotic susceptibility testing using the agar dilution 137 method following the CLSI recommendations (45). The antibiotics tested were tetracycline, 138 amoxicillin, ampicillin, ciprofloxacin, kanamycin, gentamicin, colistin, chloramphenicol and 139 trimethoprim. The imipenem, meropenem, cefotaxime and ceftazidime susceptibilities were 140 determined using the disk diffusion method (45). The minimum inhibitory concentration 141 (MIC) breakpoints for Enterobacteriaceae in the CLSI guidelines (45) were used to identify 142 ARB. The EUCAST MIC breakpoint for colistin was used (46). Bacterial isolates resistant to 143 three or more different antibiotic classes of antibiotics were defined as multidrug resistant. 144 The resistance percentages of bacteria were calculated as: percentage (%) = [(Number of 145 resistant faecal coliforms to an antibiotic/ total number of tested faecal coliforms) X 100]. 146

Phenotypic identification of the production of Metallo-beta-lactamase (MBL), ESBL 147
and AmpC enzymes 148 Bacteria resistant to imipenem and/or meropenem were subjected to the Imipenem-EDTA 149 double-disk synergy test as described previously (47). Isolates resistant to cefotaxime and/or 150 ceftazidime were subjected to ESBL testing following the CLSI guidelines and AmpC testing 151 using phenylboronic acid (45,48). 152

Identification of antibiotic resistance genes and bacterial species 153
Putative MBL producing carbapenem resistant isolates (identified using the imipenem-EDTA 154 double-disk synergy test) were subjected to multiplex PCR to identify the carbapenem 155 resistance genes. The primer sets included bla GES , bla GIM , bla IMI , bla IMP , bla KPC

Antibiotic susceptibility patterns 184
In total, 498 faecal coliforms were isolated from all WWTP effluent samples, comprising 226 185 isolates from WWTP A and 272 from WWTP B. These isolates were subjected to antibiotic 186 susceptibility testing. Among the tested β-lactam antibiotics, more than 90% of bacteria 187 isolated from the two WWTPs were resistant to amoxicillin and ampicillin (Table 2, Figure  188 1); greater than 20% were resistant to cefotaxime or ceftazidime. All ceftazidime resistant 189 isolates were also resistant to cefotaxime. Carbapenem resistance was detected at relatively 190 lower levels (Table 2). Colistin resistance was found at a higher percentage in WWTP B 191 effluent than in WWTP A. We also identified that there were no differences in the 192 identification of colistin resistant isolates in antibiotic susceptibility testing by the agar 193 dilution method compared to the microbroth dilution method. Multi-drug resistant (MDR) 194 faecal coliforms were detected at approximately 50 % of the total isolates tested (Table 2). 195 The resistance prevalence to other antibiotics were found at similar levels between the two 196 WWTPs. 197

Detection of antibiotic resistance genes and speciation 198
Faecal coliforms with resistance to cefotaxime and ceftazidime were considered putative 199 ESBL-producers (n = 157). AmpC production was confirmed in 26 of these isolates (13 from  In total, 79 isolates (36 from WWTP A and 43 from WWTP B) resistant to imipenem were 210 screened for MBL production. Metallo-beta-lactamase production was identified in 36 211 isolates. However, all isolates were negative for known MBL genes, using the MBL 212 multiplex PCR. More than 100 colistin resistant isolates were detected. However, none of 213 these isolates were positive for the mcr-genes using the mcr-targeted PCR analysis. 214

Plasmid transformation, plasmid resistance profile and replicon typing 215
The plasmids were extracted from 62 ESBL faecal coliform isolates (30 from WWTP A and 216 32 from WWTP B). Plasmids extracted from 52 ESBL isolates (46 plasmids carried bla TEM , 2 217 plasmids bla CTX-M15 , 1 bla CTX-M-1 and 3 with bla SHV-12 ) were successfully transferred into E. 218 coli Dh5α recipients. The plasmids extracted from the 10 ESBL isolates (3 with bla TEM , 3 219 with bla CTX-M15 and 4 with bla SHV-12 ) could not be transferred, suggesting that the ESBL genes 220 identified in these isolates are located on the bacterial chromosomes. 221 All transformants were sensitive to the tested carbapenem antibiotics (imipenem, meropenem 222 and ertapenem) and gentamicin, 16 were resistant to chloramphenicol, 16 resistant to 223 tetracycline, 10 resistant to trimethoprim, 3 resistant to colistin, 2 resistant to kanamycin, 1 224 resistant to amikacin and 1 resistant to ciprofloxacin. Nine transformants show a resistance 225 phenotype to two drugs from different classes and eight show a multidrug resistance 226 phenotype. The plasmids from 21 transformants could be typed using PCR replicon typing. 227 Of these, 11 plasmids were IncF replicon type (3 IncFIA, 1 IncFIIA, 1 IncFIB and 6 non-228 specific IncF group), 5 were IncI1 replicon type, 2 were IncHI1 replicon type, 2 were IncHI2 229 replicon type and 1 was IncA/C replicon type. The IncF group of replicons was the most 230 prevalent across plasmids from both WWTPs. The IncI1 type was detected only in bacteria 231 isolated from the WWTP B effluent samples. 232 233

234
Our work presents the antibiotic resistance patterns of faecal coliforms in the effluent of two 235 WWTPs in Ireland. Wastewaters with faecal contamination are considered reservoirs for 236 ARB and ARGs in the environment (57, 58). Among all tested antibiotics, resistance to 237 amoxicillin and ampicillin was most prevalent, followed by tetracycline and, cefotaxime and 238 ciprofloxacin. High levels of β-lactam resistance were previously detected in 239 Enterobacteriaceae in an urban WWTP (59), and resistance to tetracycline and 240 fluoroquinolones was found at lower rates. The higher resistance rate of E. coli to ampicillin 241 and tetracycline as well as lower rates of resistance to ciprofloxacin were detected in WWTPs 242 in Portugal (60). A study of raw sewage in Brazil identified 100% sensitivity of E. coli to 243 ciprofloxacin and amoxicillin and tetracycline resistance levels in the range of 50 -75% (61). 244 The proportions of ciprofloxacin resistant faecal coliforms were 31.42% in WWTP A effluent 245 and 26.47% in WWTP B. This resistance rate is higher in comparison to reported levels of 246 ciprofloxacin resistance in the E. coli isolated from other WWTPs (59,62,63). In general, 247 differences in antibiotic resistance percentages were observed between the two WWTPs, 248 particularly for colistin, trimethoprim and kanamycin ( Figure 1). As these two WWTPs are 249 using the same treatment process, this difference may be associated with their location or 250 other external factors. Multi-drug resistant faecal coliforms were retrieved at high rates from all effluent samples. In 270 the study of Lefkowitz and Durán (78), 60% of E. coli in WWTP effluent were resistant to 271 two or more antibiotics and 25% to four or more antibiotics. The study of Garcia et al (79) in 272 WWTP effluent showed no more than 12% of E. coli were resistant to two antibiotics and 273 less than 10% to three or more antibiotics. Escherichia coli (34.3%) were found to be 274 resistant to two or more antibiotics and 8.8% to four or more antibiotics in treated wastewater 275 in Portugal (59). The MDR faecal coliforms in our study were found in the same range of the 276 study of Lefkowitz and Durán, but at a higher percentage than in others. 277 The ESBLor AmpC producing faecal coliforms were recovered from all WWTP effluent 278 samples. The rate of ESBL producing Enterobacteriacea in our work were within the range 279 of previous studies. It was considerably high in comparison to some studies of WWTP 280 effluent (0.5-9.8%) (34, 80-83). However it was lower than those studies in wastewater (45-281 100%) (84, 85). The high load of bacteria and rich nutrient environment in WWTPs 282 facilitates the transfer of ARGs among bacteria (11,12). These may explain the relatively 283 high rates of ESBL producers in WWTP effluent. 284 The bla TEM were the most prevalent beta-lactamase in this study, which is similar to a study 285 on hospital wastewater in Brazil (85) to antibiotics of last line of defence. The ability of these bacteria to survive in water has been 310 demonstrated for many years. The significance of this study is the identification of the role of 311 WWTPs as a potential control point to reduce or stop the movement of resistant bacteria and 312 genes into the environment from further upstream sources, such as human or animal waste. In 313 addition, this enables the use of additional treatment technologies to be added to WWTPs to 314 stop or reduce such ARB and ARGs entering the water environments globally.     extended-spectrum beta-lactamase-and carbapenemase-producing Enterobacteriaceae 454 Isolates from rivers and lakes in Switzerland. Appl Environ Microbiol 79:3021-6. 455