Detection of beta-lactam antibiotic resistance in aquatic Enterobacteriaceae isolates Open Access

Selection pressure on the bacterial populations occurring as a consequence of inappropriate antibiotic use has a crucial role in the selection and dissemination of antibiotic-resistant bacteria. The overuse of antibiotics leads to the emergence of many different resistance determinants involving highly complex processes in bacterial cells (Kummerer 2004; Ocal 2012).

Antibiotic-resistant bacteria (ARB) and antibiotic-resistance genes (ARG) are detected in household and hospital wastewater, sewage sludge and coastal lakes, as well as in streams contaminated with groundwater and wastewater. However, research has also been focused on revealing the spreading rates of ARB and ARG in some disconcerting water environments like tap water, recycled water, drinking water pipe biofilms, recreational lakes, etc. in the last few years (Rocha et al. 2019; Kimbell et al. 2021; Li et al. 2022). ARB and ARG may enter the natural water sources by the direct discharge of unprocessed effluents or sewage sludges discharged from wastewater pretreatment plants. They are also transferred to soil from treated bio-waste of sewage treatment plants and animal manure farms, then mixed with groundwater and transferred to surface water by runoff or erosion (Prado et al. 2008; Zhang et al. 2009).

It is considered that the greatest environmental sources of antibiotic-resistant bacteria are hospital effluent and poultry manure. Recent studies have mentioned that enteric bacteria, which are transmitted from hospital effluents and pose a serious threat to human health, are at higher levels in wastewater effluents than those originating from other sources (Khan et al. 2019; Asghari et al. 2021). Studies have also reported that antibiotics, which are used intensively due to their therapeutic and growth-promoting effects, cause ARB and ARG to occur in poultry manure and the wastewater of poultry slaughterhouses and contaminate the aquatic environments (Savin et al. 2021; Tian et al. 2021).

Beta-lactam antibiotics that are used several ways such as veterinary and agricultural activities, aquacultures etc., are constantly preferred as they are more reliable agents with lower toxicity and bactericidal properties compared with other antibiotic groups. In addition, these antibiotics are prominent because of their functions which provide variation by changing their chemical configurations (Henriques et al. 2006a; Livermore & Woodford 2006). Beta-lactam resistance often consists of more than one resistance mechanism that work together; so the bacterial infection control and the treatment processes are often complicated (Livermore & Woodford 2006).

Bacterial species from Enterobacteriaceae family are Gram-negative microorganisms that live in human and animal intestines (Schwartz et al. 2003; Prado et al. 2008). Important pathogenic microorganisms of Enterobacteriaceae are frequent in clinical environments and able to adapt to different environments and habitats. They rapidly spread resistance factors, which cause serious problems in terms of public health. Production of beta-lactamase enzymes is the primary resistance mechanism in Enterobacteriaceae (Zhang et al. 2009). Beta-lactamases, which maintain their functionality by causing hydrolysis of beta-lactam antibiotics, are encoded by both chromosomal DNA regions and genes found on different mobile genetic elements such as plasmids (Yang et al. 2018).

Surface waters have also been the subject of worldwide investigations about various ARB strains and resistance genes (Tong et al. 2020). Since Turkey is a country surrounded by seas on three sides and harbors a large number of surface waters, it undoubtedly constitutes an important reservoir for similar studies. There are dams for purposes such as drinking water, irrigation, and electricity generation in the western Black Sea reservoir of Turkey, which is located very close to the organized industrial zones. According to the report ‘Wastewater Treatment Action Plan (2008–2012) of the Turkish Ministry of Environment and Forestry’, (Report for the Wastewater Treatment Action Plan of The Ministry of Environment and Forestry General Directorate of Environmental Management of the Turkish Republic, 2008-2012) the Western Black Sea is one of the basins with the highest increase in domestic wastewater pollution in Turkey in recent years and most companies that discharge wastewater into this basin do not have any treatment plants. Lake Manyas, also known as Manyas Bird Paradise, which hosts many bird species, is polluted through the flowing streams with factory wastes and the releases of egg poultry activities for years (Karafistan & Arık-Colakoglu 2005). The waters of Lake Iznik, where irrigation, fishing, and tent tourism are common, are used for wound treatment and as drinking water. All these activities make it worthwhile to investigate the three separate surface waters located in different geographical areas of Turkey in terms of their potential to contain ARB and ARGs. That is the reason we believe that this study will provide novel and helpful notions about resistance to beta-lactam antibiotics acquired by beta-lactamase enzymes produced by Enterobacteriaceae isolated from coastal seawaters of the western Black Sea, Lake Manyas and Lake Iznik in Turkey, which are contaminated by industrial activities and anthropogenic factors. In this context, using both phenotypic and genotypic methods to search for beta-lactamases also allowed us to test the reliability of phenotypic methods by comparing them with gene expression results.

The water sources, from which 27 Enterobacteriaceae strains used in this study were isolated are shown in Table 1. The western Black Sea and Lake Manyas strains, which were isolated in our previous studies (Arslan Aydogdu et al. 2017; Kimiran Erdem et al. 2017) and randomly selected from the Enterobacteriaceae family, were included in this study. Escherichia coli ATCC 25922 (American Type Culture Collection, Rockville, MD), E. coli ATCC 35218 (American Type Culture Collection, Rockville, MD) and Klebsiella pneumoniae ATCC 700603 (American Type Culture Collection, Rockville, MD) were used as standard bacterial strains. Water samples from Lake Iznik were filtered through 0.45-μm pore size membrane filters and incubated in Endo-NPS medium (Sartorius, Germany) at 37 °C for 24 hours. Identification and species verification of standard bacteria and Lake Iznik isolates were performed with API 20E system (Biomerieux, France).

Beta-lactam antibiotic susceptibility of 27 Enterobacteriaceae isolates obtained from three different water sources was determined by the Kirby-Bauer disc diffusion test (Bauer et al. 1966). The involved beta-lactam antibiotic discs (Becton Dickinson, USA) were: ampicillin (AM; 10 μg), carbenicillin (PY; 100 μg), amoxicillin/clavulanic acid (AMC/CLA; 20/10 μg), cephalothin (CF; 30 μg), cefoxitin (FOX; 30 μg), ceftazidime (CAZ; 30 μg), cefotaxime (CTX; 30 μg), ceftriaxone (CRO; 30 μg), moxalactam (MOX; 30 μg), cefoperazone (CFP; 75 μg), cefepime (FEP; 30 μg), imipenem (IMP; 10 μg) and aztreonam (ATM; 30 μg). Test results were evaluated according to Clinical and Laboratory Standards Institute (CLSI) guidelines and E. coli ATCC 25922 (American Type Culture Collection, Rockville, MD) was used as a control (CLSI 2016).

Based on CLSI (2016) guidelines, a combined disc test was applied to E. coli, K. pneumoniae and K. oxytoca strains to determine if clavulanic acid could inhibit ESBLs. Accordingly, strains were identified as positive for ESBL production when the diameter of the inhibition zone around ceftazidime/clavulanic acid (CAZ/CLA; 30/10 μg) or cefotaxime/clavulanic acid (CTX/CLA; 30/10 μg) disc was at least 5 mm broader than the inhibition zone around the CAZ or CTX disc only. K. pneumoniae ATCC 700603 (American Type Culture Collection, Rockville, MD) was used as the standard strain.

The combined disc test was modified by making use of the ability of boronic acid to inhibit AmpC beta-lactamases (Coskun & Altanlar 2012). The test involved ceftazidime, cefotaxime, ceftriaxone or cefoxitin antibiotic discs and their modified forms. The discs were prepared by combining with 3-aminophenylboronic acid (APBA; Sigma-Aldrich, Turkey). According to this confirmation test, strains were deemed positive for AmpC production when the diameter of the inhibition zone around cephalosporin/APBA disc was at least 5 mm broader than the inhibition zone around the cephalosporin disc only (Yagi et al. 2005).

The plasmid DNA isolation procedure from the Zyppy™ Plasmid Miniprep Kit (Zymo Research, USA) was applied to the 27 Enterobacteriaceae isolates, while the chromosomal DNA isolation procedure of IDPURE™ Spin Column Bacterial Genomic DNA Isolation Kit (ID Labs, Canada) was applied to three E. cloacae from Lake Iznik. Plasmid and chromosomal DNA products were stored at −20 °C for later use.

The classical PCR method for detecting the genes encoding plasmid-mediated penicillinase, cephalosporinase and carbapenemase enzymes and semi-nested PCR to detect the gene encoding chromosomally mediated AmpC cephalosporinase were performed.

PCR mixtures were primarily prepared to amplify the plasmid and chromosomally mediated beta-lactamase genes. For each PCR mixture, 12.5 μl of Taq PCR Mastermix (Biomatik, USA), 1 μl each of forward and reverse primers (Biomatik, USA; 10 μM), 1.5 μl of DNA (plasmid/chromosomal) and 1.5 μl of DMSO were added to microtubes, then the mixtures were completed to 25 μl with bidistilled water. Amplification of the plasmid and chromosomally mediated beta-lactamase genes was carried out in a thermal cycler (Bio-Rad, USA). Primer sequences and annealing temperatures of PCR reactions are shown in Table 2. PCR conditions for bla_TEM, bla_SHV and bla_CTX-M target genes were as follows: initial denaturation (9 min at 94 °C); 30 cycles of denaturation (30 s at 94 °C), annealing (30 s at temperatures indicated in Table 2) and extension (1 min at 72 °C); and final extension (10 min at 72 °C). PCR program for the bla_IMP target gene consisted of initial denaturation (2 min at 94 °C); 30 cycles of denaturation (1 min at 94 °C), annealing (1 min at 55 °C) and extension (90 s at 72 °C); and final extension (8 min at 72 °C). PCR processes for the CIT, MOX and FOX gene groups included initial denaturation (10 min at 95 °C); 35 cycles of denaturation (40 s at 94 °C), annealing (30 s at 63 °C) and extension (2 min at 72 °C); and final extension (10 min at 72 °C). Lastly, the first and second steps for the semi-nested PCR conditions of the chromosomal bla_AmpC target gene of E. cloacae were as follows: initial denaturation (3 min at 94 °C); 35 cycles of denaturation (1 min at 94 °C), annealing (1 min at temperatures indicated in Table 2) and extension (2 min at 72 °C); and final extension (8 min at 72 °C).

PCR products were prepared on a 1.2% agarose gel including GelRed™ DNA stain in water (Biotium, USA), operated for 60/75 minutes at 100 volts with gel electrophoresis and visualized using a UV transilluminator (Kodak GL 1500). The band sizes of PCR products were determined via a 1 kb DNA ladder (Biomatik, USA).

The disc diffusion test, including 13 total antibiotic discs from five different groups of beta-lactams, indicated that 25 (92.5%, n = 27) of the bacterial strains were resistant to AM, 20 (74.1%, n = 27) were resistant to PY, 17 (62.9%, n = 27) to CF and 10 (37%, n = 27) to AMC/CLA, which presented high-level resistance profiles. All the strains were susceptible to MOX and IMP.

It was determined that three E. coli isolated from the western Black Sea (B236, B237 and B243) were resistant to third-generation cephalosporins including CTX, CRO and CFP; two of the strains (B237 and B243) were also resistant to CAZ, FEP and ATM. It was shown that three E. cloacae strains isolated from Lake Iznik exhibited resistance to AMC/CLA, CF and FOX and two strains of Lake Manyas were not resistant to any beta-lactam antibiotics except penicillins including AM and PY (Table 3). Apart from three E. coli isolates (B57, B90 and B109) from the western Black Sea, all 24 (88.8%, n = 27) of the studied strains were resistant to more than one antibiotic used in the disc diffusion test. Moreover, 19 (79.1%, n = 24) of the isolates were found to be resistant to three or more beta-lactam antibiotics. In total, 17 (77.2%, n = 22) of the western Black Sea isolates were resistant to at least two beta-lactam antibiotic groups. From all isolates, 25 (92.5%, n = 27) were resistant to penicillin, 17 (62.9%, n = 27) to cephalosporin, 2 (7.4%, n = 27) to monobactam and 10 (37%, n = 27) to beta-lactam/beta-lactamase inhibitor. All the isolates (n = 27) were susceptible or moderately susceptible to imipenem and moxalactam. In total, 17 (62.9%, n = 27) strains were found to be resistant to first generation cephalosporin (CF), three isolates (11.1%, n = 27) were resistant to second generation cephalosporin (FOX), two isolates (7.4%, n = 27) were resistant to fourth-generation cephalosporin (FEP) and three isolates (11.1%, n = 27) were resistant to at least one of the five different types of third-generation cephalosporins.

The highest frequency of resistance was observed in two E. coli isolates (B237 and B243), which were resistant to nine of the beta-lactam antibiotics, while the other E. coli strain (B236) showed resistance to seven of the antibiotic discs. It was determined that these three isolates provided resistance to CTX, CRO and CFP, which are third-generation cephalosporins and only two E. coli strains (B237 and B243) showed resistance to fourth-generation cephalosporin (FEP) (Table 3).

Four plasmid-mediated beta-lactamase genes (bla_TEM, bla_SHV, bla_CTX-M and bla_IMP) and three plasmid-mediated beta-lactamase gene groups (CIT, MOX and FOX) were targeted by PCR in this study. Beta-lactamase encoding gene profiles and resistance phenotypes of 27 Enterobacteriaceae isolates were presented in Table 3.

DNA amplification of the chromosomal bla_AmpC gene for E. cloacae was detected through semi-nested PCR in all three Lake Iznik isolates.

One bla_TEM + bla_SHV, one bla_CTX-M + CIT, and one bla_SHV + CIT gene profiles in three different E. coli isolates from the western Black Sea were detected (Table 4).

The antibiotic resistance mechanism provided by beta-lactamases, owing to frequent antibiotic use at human–animal–environment triangle, is on course for being the main reason for ecological pollution along with threatening public health. Thus, the investigations about this mechanism, which have been focusing on clinical bacterial strains for years, need to be directed to environmental bacteria.

Beta-lactamase-producing bacteria from aquatic environments could transfer their resistance determinants both to each other and the pathogenic bacteria sharing the same habitats via horizontal gene transfer from mobile genetic elements including plasmids, transposons and integrons (Matyar et al. 2008; Zhang et al. 2009). As a result of becoming antibiotic resistant, pathogenic bacteria are exposed to selection pressure, which could complicate the treatment of infections that they cause by constituting the dominant population in animal and human tissues (Santos & Ramos 2018).

Conversely, resistance determinants including bacteria, resistance genes and antibiotic residues passing through the effluents of healthcare facilities and wastewater treatment plants are other important factors increasing the resistance rates in natural aquatic habitats (Henriques et al. 2006b; Yang et al. 2018).

Although the latest global research have been raising awareness about the spread of antimicrobial resistance caused by microbial contamination in different aquatic environments, there are still very few and noninclusive studies on ARB and ARG in the cycle of water reservoirs of Turkey. This study, in which the common aquatic species of Enterobacteriaceae family from surface waters were included, dealing with the extent of resistance acquired by a specific beta-lactamase production mechanism. In addition, this study questions the phenotypic reflections of beta-lactamase-encoding genes and the reliability and sufficiency of the phenotypic susceptibility test in detecting plasmid-mediated resistance.

In this study, 27 Enterobacteriaceae isolates including four different species from surface waters of western Black Sea, Lake Manyas and Lake Iznik were inspected. In antibiotic susceptibility tests, 20 (100%, n = 20) E. coli strains were resistant to AM, 16 (80%, n = 20) to PY, and 14 (70%, n = 20) to CF. This resistance pattern of E. coli isolates to different groups of beta-lactams appeared higher than some similar studies previously performed on E. coli isolates from surface waters (Rasheed et al. 2014; Blaak et al. 2015). This study presented a lower rate of resistance compared with a study that previously marked as resistance to third-generation cephalosporins in Enterobacteriaceae strains from various aquatic environments (Amador et al. 2015). In addition, the highest resistance frequency was shown against AM, PY and CF antibiotic discs, respectively. Our findings were still lower than the resistance rates seen in previous similar studies on clinical Enterobacteriaceae isolates (Niranjan & Malini 2014; Sabir et al. 2014). However, it is not difficult to predict that antimicrobial contamination in environments will reach undeniable proportions in the coming years. Recent studies have revealed that there is a great deal of similarity between the antibiotic-resistance genes of bacteria in aquatic cultures and terrestrial bacteria that cause disease in humans and animals (Santos & Ramos 2018).

The increasing use of broad-spectrum beta-lactam antibiotics in various fields has led to the emergence of many different derivatives of TEM, SHV and CTX-M beta-lactamases. Therefore, methods such as isoelectric point detection, agar diffusion test based on clavulanic acid inhibition are no longer sufficient options alone for determining the ESBL phenotype (Tasli & Bahar 2005; Pitout & Laupland 2008). Instead, a recent molecular diagnostic technique, PCR, which is performed to amplify beta-lactamase-encoding genes such as bla_TEM, bla_SHV and bla_CTX-M by using specific primer sequences, is preferred.

In this study, the presence of TEM and SHV enzymes, the most common plasmid-mediated penicillinase providing resistance to penicillins and first generation cephalosporins in Enterobacteriaceae (Bradford 2001), were investigated. The TEM encoding bla_TEM gene in three (11.1%, n = 27) of the strains and SHV encoding bla_SHV gene in nine (33.3%, n = 27) of the strains were detected by PCR. The three E. coli isolates (B89, B91 and B120) carrying bla_TEM gene were found to be multiresistant to AM, PY and CF. This apparently confirms that the resistance was encoded by the bla_TEM gene in these isolates. Among nine strains (B85, B120, B155, B176, B180, B185, B233, B235 and M55) carrying bla_SHV gene; six were E. coli, one was K. pneumoniae and two were K. oxytoca. The fact that three out of four strains from the genus Klebsiella in this study expressed the blaSHV gene supports the previous studies of Nordmann (1998) and Pitout & Laupland (2008), who emphasized that the gene is frequently reported in Klebsiella spp. In a previous study, Harada et al. (2016) demonstrated ESBLs and/or AmpCs in 31 (34.8%, n = 89) K. pneumoniae isolates, but not in any of 14 K. oxytoca isolates by phenotypic testing. They also demonstrated by DNA sequencing that 5 of 34 (14.7%) cephalosporin-resistant K. pneumoniae isolates produced bla_SHV-type ESBL. In this study, bla_SHV-producing Klebsiella spp. strains showed neither cephalosporin-resistance nor ESBL and/or AmpCs in antibiotic susceptibility tests.

It has been known that high levels of plasmid-mediated narrow-spectrum TEM-1, TEM-2 and SHV-1 enzymes may induce resistance to penicillins combined with beta-lactamase inhibitors (Farsak Orak 2005). Hence, this may be the reason why AMC resistance was detected in 5 of the 11 strains, in which bla_TEM and/or bla_SHV genes were amplified. It is considered that there is a slight possibility that these genes are ESBL, since all strains which were confirmed to carry these bla_TEM and/or bla_SHV were susceptible to cephalosporins and ATM. Yet, for precise information on the differentiation of types of enzymes, sequence analysis is required for these genes. Two (I3 and I5) of another 16 non-bla_TEM or bla_SHV-carrying strains presented intermediate resistance to AM, which could be due to method variation or decreased levels of resistance expression. Nevertheless, five of these non-bla_TEM- or bla_SHV-carrying strains were resistant to AM and nine were resistant to both AM and PY, which indicates the presence of plasmid or chromosomally mediated beta-lactamase or resistance mechanisms other than beta-lactamases.

The most convenient method for phenotypic confirmation of ESBLs is the inhibitor-based combined disc test suggested by CLSI (Thomson 2001). According to the combined disc test, three (12.5%, n = 24) E. coli strains belonging to E. coli, K. pneumoniae and K. oxytoca were ESBL positive. Two of these ESBL(+) E. coli strains were B237 and B243, which showed resistance to CAZ, CTX, CRO, FEP and ATM, while the remaining 21 ESBL(−) strains were all susceptible to oxy-imino-cephalosporins and ATM. It was remarkable that the ESBL(+) B236 strain was resistant to CTX, moderately susceptible to CAZ and ATM and susceptible to FEP. As the three ESBL(+) strains were all resistant to CFP, this study proves that additional research is needed on the relationship between this resistance pattern and ESBL production. It has been previously reported by Pitout & Laupland (2008) that the presence of the bla_CTX-M gene mostly provides sufficient evidence that an ESBL-type CTX-M enzyme is responsible for the cephalosporin-resistance phenotype. Accordingly, bla_CTX-M genes amplified by PCR in the three E. coli isolates (B236, B237 and B243) show that the high-level cephalosporin-resistance pattern of these strains is probably provided by a CTX-M-type ESBL.

IMP resistance was not observed in any of the strains in our disc diffusion tests, and no strains were amplified to carry the bla_IMP gene in PCR reactions. As the low IMP susceptibility was possibly associated with chromosomal bla_AmpC expression, it was not surprising that the bla_IMP gene was not found in E. cloacae strains. In addition, the most common carbapenem resistance mechanism in Enterobacteriaceae is the reduction of outer membrane permeability (Nordmann 1998). Accordingly, it will be necessary to investigate different gene regions or resistance mechanisms in the other three isolates (B54, B58, B65) with low imipenem susceptibility in our study. Carbapenem resistance is rarely seen in Enterobacteriaceae because the increased dissemination of resistant isolates is generally more associated with the coexistence of impermeability-induced resistance with CTX-M or AmpC beta-lactamases (Livermore & Woodford 2006). Considering the assumption that the presence of plasmid-mediated ESBL and AmpC in bacterial isolates will lead to more frequent use of carbapenems for the treatment of infections in the coming years, the lack of detection of carbapenemases does not guarantee that carbapenem resistance followed by mutations will not increase in the future.

No phenotypic screening test has yet been standardized for the detection of plasmid-mediated AmpC beta-lactamases, which have been determined to originate from the beta-lactamases encoded by inducible chromosomal DNA of the bacterial species including E. cloacae, Serratia marcescens, Citrobacter freundii etc. (Philippon et al. 2002). Interpretation of results in the AmpC detection tests is not simple enough for routine use in clinical microbiology laboratories. In addition, these tests could be misleading in strains that simultaneously produce an ESBL or a metallo-beta-lactamase (Perez-Perez & Hanson 2002; Yagi et al. 2005).

A modified combined disc test with boronic acid, which has been known as an AmpC inhibitor for many years (Jacoby 2009), has already been successfully practiced by certain researchers. Some previous reports on clinical Enterobacteriaceae strains demonstrated that this method is reliable for detecting AmpC beta-lactamases (Manoharan et al. 2012). Yagi et al. (2005) observed that applying 300 micrograms of 3-aminophenylboronic acid could sensitively detect AmpC beta-lactamases by inhibiting the growth around CAZ or CTX antibiotic discs.

This study was intended to confirm the presence of AmpC beta-lactamases that hydrolyse the 7-α-methoxy cephalosporins (cephamycins) (Philippon et al. 2002) using a modified combined disc test. 3-Aminophenylboronic acid was found to inhibit the growth of four (B243, İ3, İ5, İ8) of the six bacterial strains, which were resistant to at least one of the third- and fourth-generation cephalosporins. Three different plasmid-mediated gene groups (CIT, MOX, FOX) were also investigated in all bacterial strains involved in the study and 25 of them, including three E. cloacae strains, gave negative results in PCR. CIT gene group expression was detected in one E. coli strain (B243), which was confirmed to produce AmpC in modified combined disc tests. Some studies have mentioned that the phenotypic susceptibility tests usually fail to detect ESBL in some AmpC producing bacterial strains that cause high mortality rates (Pai et al. 2004). In relation to this, after detecting the presence of both the CIT gene group and bla_SHV gene in an E. coli strain (B185), which phenotypically showed low levels of resistance and gave negative results in the ESBL and AmpC disc tests, this study revealed once again that phenotypic detection tests alone are not reliable.

E. cloacae species possess inducible chromosomal AmpC beta-lactamases (Thomson 2001; Philippon et al. 2002). Studies in recent years have reported that these types of AmpCs cause natural resistance to broad-spectrum cephalosporins in E. cloacae and reduce IMP susceptibility (Peter-Getzlaff et al. 2011). In this study, the presence of an inducible chromosomal bla_AmpC gene was revealed by semi-nested PCR analysis in three strains (I3, I5, I8) isolated from the waters of Lake Iznik and identified as E. cloacae. FOX resistance accompanied by low IMP susceptibility, observed through the effect of boronic acid in the phenotypic AmpC confirmation test, demonstrated a strong induction of chromosomal bla_AmpC gene expression by these antibiotic agents in E. cloacae isolates. AmpC-type enzymes produced in E. cloacae can sometimes mask the presence of ESBL and cause false-negative evaluations (Pitout & Laupland 2008). However, in our study, a resistance phenotype that would be compatible with the presence of ESBL was not observed in E. cloacae isolates, and no amplification was achieved for bla_TEM, bla_SHV and bla_CTX-M genes by PCR.

In our study, more than one plasmid-mediated gene was detected in three different strains isolated from the western Black Sea water samples. According to this, both bla_TEM and bla_SHV genes were detected in an E. coli strain (B120). Yet another E. coli isolate (B243) was positive for both ESBL and AmpC in phenotypic confirmatory tests. In addition, PCR amplification of both the plasmid-mediated bla_CTX-M gene and the CIT gene group proved that the strain had both types of enzymes. Another E. coli strain (B185) showed positive results for both bla_SHV and the CIT gene group in PCR reactions, contradicting the resistance phenotype.

As far as we are concerned, this is the first study to investigate antibiotic-resistance genes in Lake Manyas and Lake Iznik. While the bla_SHV plasmid-mediated resistance gene was found in a K. pneumoniae strain (M55) from Lake Manyas, only three E. cloacae strains were isolated from Lake Iznik and none of the strains expressed the plasmid-mediated genes included in this study. At least one plasmid-mediated gene was observed in 13 (48.1%, n = 27) strains selected from the western Black Sea. Although this study was conducted with a small number of Enterobacteriaceae strains, it is suggested that the coasts of the western Black Sea, which are polluted by legal or illegal means, are exposed to more antibiotic residues than the enclosed environments of lakes.

Antibiotic-resistant bacteria and resistance genes contaminate surface waters widely used in agricultural irrigation activities, as well as ground water, drinking water and tap water through soil and the industrial wastes. This is definitely a major concern from the point of both human health and ecological balance. It would be a critical step towards reducing selection pressure to construct risk assessments and implement measures to control the spread of antibiotic use.

All data obtained in this study prove that resistance caused by beta-lactamases in bacterial strains evolves rapidly not only in clinical settings but also in natural aquatic environments. This study clearly reveals that the western Black Sea, Lake Manyas and Lake Iznik, which are important surface waters open to activities that directly affect human health, are contaminated with bacteria resistant to beta-lactam antibiotics. As far as we know, no other study has investigated these three important reservoirs of Turkey in terms of the most widespread plasmid-mediated beta-lactam resistance genes in Enterobacteriaceae. ESBL and AmpC, encoded by plasmid-mediated genes, have a high potential for intra- and/or inter-species spread in bacteria; therefore, these are the enzymes that make it difficult to keep resistance under control. These resistance factors, which should be considered by both clinical microbiology and microbiological research laboratories, will cause us to encounter major problems in terms of public health in the near future, unless necessary worldwide precautions are taken.

Research studies on resistant bacteria and resistance mechanisms will contribute to canalize the decisions into the right direction for developing efficient control methods. The obtained results throughout our study show that plasmid-mediated beta-lactamase production because of transmission of resistance by horizontal gene transfer, continues actively in bacteria isolated from aquatic environments. With this study, the necessity of selecting appropriate methods to determine the resistance phenotype and investigate the basic genetic factors of resistance in order to detect and interfere the antibiotic resistance provided by beta-lactamases in a reliable way, has been demonstrated once again.

This work was supported by the ‘Research Fund of the Istanbul University’ (grant number 36585).

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.