Antibiotic resistance profiles in cultivable microbiota isolated from some romanian natural fishery lakes included in Natura 2000 network

Background The present study aims the characterization of antibiotic resistance phenotypes and encoding genes in bacterial strains isolated from some Romanian aquatic fishery lowland salted lakes. Material/Methods This study was conducted on 44 bacterial strains, mainly belonging to species used as microbiological indicators of fecal pollution isolated from four natural fishery lakes. All strains were tested for their antibiotic susceptibility by disk diffusion method. Simplex and multiplex PCR were performed to identify the β-lactams antibiotic resistance genes (blaNMD, blaOXA−48, blaVIM, blaIMP, blaCTX−M, blaTEM), sulfonamides (Sul1, Sul2), tetracyclines (TetA, TetB, TetC, TetD, TetM), aminoglycosides (aac3Ia), vancomycin (VanA, VanB, VanC), macrolides (ermA, ermB, ermC) as well as the plasmid-mediated quinolone resistance (PMQR) markers (QnrA, QnrB, QnrS), and class 1 integrons (Int1, drfA1-aadA1). Results The Enterococcus spp. isolates exhibited phenotypic resistance to vancomycin (35 %) and macrolides (erythromycin) (75 %); from the vancomycin – resistant strains, 5 % harboured VanA (E. faecalis), while the erythromycin resistant isolates were positive for the ermA gene (E. faecalis − 10 %, E. faecium − 5 %). The Gram- negative rods (GNR) exhibited a high level of resistance to β-lactams: cefuroxime (63 %), cefazolin (42 %), ceftriaxone (8 %), ceftazidime and aztreonam (4 % each). The genetic determinants for beta-lactam resistance were represented by blaCTX−M−like (33 %), blaNDM−like and blaIMP−like (8.33 %) genes. The resistance to non-β-lactam antibiotics was ascertained to the following genes: quinolones (QnrS − 4.16 %); sulfonamides (Sul1–75 %, Sul2–4.16 %); aminoglycosides (aac3Ia − 4.16 %); tetracyclines (tetA – 25 %, tetC − 15 %). The integrase gene was found in more than 50 % of the studied strains (58.33 %). Conclusions The cultivable aquatic microbiota from fishery lakes is dominated by enterococci and Enterobacterales strains. The GNR strains exhibited high levels of β-lactam resistance mediated by extended spectrum beta-lactamases and metallo-β-lactamases. The Enterococcus sp. isolates were highly resistant to macrolides and vancomycin. The high level and diversity of resistance markers, correlated with a high frequency of integrons is suggesting that this environment could act as an important reservoir of antibiotic resistance genes with a great probability to be horizontally transmitted to other associated species from the aquatic sediments microbiota, raising the potential zoonotic risk for fish consumers.


Background
Antimicrobial resistance (AR) is an increasing worldwide concern, as antibiotics still represent a very important option for human and animal health protection [1]. It is expected that the mortality rates due to AR will reach as many as 10 million people/year by 2050, if appropriate action is not taken [2]. AR is now considered without any doubt a typical One Health problem, the environment having a crucial role as a reservoir and transmission route to humans. Consequently, one of the main priorities of the Joint Programme Initiative on Antimicrobial Resistance (JPIAMR) is to foster the research of environmental AR reservoirs, which are very poorly known, in comparison with clinical ones [3][4][5][6]. In 2014, Romania reported for the first time to ESVAC the consumption of veterinary antimicrobial agents [7]. At the top of the most-solded classes were tetracyclines, penicillins and aminoglycosides. Similarly, in 2016, the most consumed classes for food-producing animals in Romania according ESVAC were tetracyclines, penicillins, macrolides and aminoglycosides [8]. As antibiotics are one of the most popular pharmaceuticals used in medicine, veterinary care, and farming [9,10], they can be released, most of them unchanged, into the environment. They appear as contaminants of the water bodies, especially those frequently impacted by anthropogenic activities (e.g., wastewater, municipal sewage, influents and effluents of wastewater treatment plants) [11], which have been suggested to be ideal reservoirs and vectors for the AR origin, evolution and spread [12][13][14]. In wastewater effluents, lakes, rivers or streams, bacteria from different sources, probably selected by intensive antibiotic usage, are collected and mixed with environmental strains, which on their turn, could introduce the newly acquired antibiotic resistance genes (ARGs) into the clinics [15][16][17].
Enterococci and Enterobacterales species were found in high concentrations in human and animal faeces, so their presence in any type of waters indicate the faecal contamination of the respective water body. The presence of Enterobacterales (which do not survive in the water for a long time) indicates a recent faecal water pollution with potentially pathogenic microorganisms originating from anthropogenic systems (household wastewater, livestock), but also with allochtonous species originating from natural ecosystems; enterococci can survive longer in water, due to their high resistance to physical, chemical and biological agents, indicating a chronic faecal contamination of the water [18].
These commensal intestinal bacteria could often act as reservoirs for AR determinants [19]. Therefore, the evaluation of AR profiles of these microbiological indicators of water quality could predict for the ARGs disemination from humans and animals into the environment and the risk of horizontal gene transfer (HGT) to native species.
On the other hand, AR has been detected in different aquatic environments and some resistance determinants, such as the recently emerging plasmid-mediated quinolone resistance determinants (PMQR) from the Qnr family [20] and CTX-M [21] from aquatic Kluyvera sp. have been found to originate from aquatic bacteria.
In aquaculture, antibiotics are given as a component of their food, and occasionally in baths and injections [22]. The unconsumed food, as well as the fish faeces, containing antibiotics reach the sediment and the antibiotics could be ingested by wild fish and other organisms, consequently altering the composition of the acquatic microbiota, selecting antibiotic-resistant bacteria (ARB) and favoring the HGT of ARGs [22,23]. In this study we have investigated the Balta Albă-Amara-Jirlău-Lacul Sărat Câineni lakes, whose vulnerability resides in the spillage of household waste, significant floods and consecutive natural (Buzău river's flood) and artificial (due to fish farming activities) changes of water composition [24,25]. These aquatic environments were scarcely studied, the few previous studies focusing on the study of halophilic microorganisms (bacteria and arhaea), phytoplankton and zooplankton species, as well as on the organic compounds hydrolythic activity of the acquatic microbiota isolated from Amara, Balta Albă, Movila Miresei salted lakes and from Ocnele Mari area, in relation with the physical-chemical parameters (pH, density and chloride content) of the respective waters [26,27].
Our group has previously evaluated the physicochemical and microbiological characteristics of Balta Albă-Amara-Jirlău-Lacul Sărat Câineni site included in Natura 2000 Network for establishing the degree of organic pollution and generating the knowledge required for the design and implementation of appropriate measures for maintaining the balance between the water protection and the sustainable use of these protected ecosystems [28]. By this present work, we will complete the knowledge with data regarding the most prevalent ARB and ARGs present in these natural lakes. In this context, the aim of this study was to isolate and characterize the AR phenotypes (disk diffusion method) and ARGs (PCR method) in Enterococcus sp. and Gramnegative rods (mainly Enterobacterales) strains isolated in 2016 from Romanian aquatic fishery lowland salted lakes included in Natura 2000 Network [28], in the first attempt to evaluate their contribution to the AR aquatic reservoir.
In our study, the disk diffusion susceptibility assay results revealed different resistance rates among the isolated Enterococcus spp. strains (Fig. 1).
The molecular support of the AR for all the strains isolated from fishery lakes was determined using simplex and multiplex PCR performed on genomic DNA in order to identify the genes encoding for resistance to βlactams (bla NMD , bla OXA−48 , bla VIM−2 , bla IMP , bla CTX−M , bla TEM ), PMQR markers (QnrA, QnrB, QnrS), sulfonamides (Sul1, Sul2), tetracyclines (TetA, TetB, TetC, Table 1 The antibiotic resistance genes and antibiotic susceptibility profiles in the analyzed Enterococcus strains distributed by the isolation sources
Although some strains were resistant to antibiotics from different classes, they still remained sensitive to cephalosporins. This does not apply in case of the isolated Acinetobacter calcoaceticus strain, that proved to be resistant to all tested beta-lactams (data not shown).

Discussion
This study investigated for the first time the AR features of cultivable microbiota of four lowland salted lakes of natural origin located in two counties from the southern region of Romania, namely Buzău and Brăila.
We observed an ubiquitary presence of Enterococcus spp. isolates, indicating the presence of chronic fecal pollution in all analyzed lakes (Table 1). Moreover, the E. faecalis strains isolated from Salty lake Câineni proved to be resistant to vancomycin (Van A) and erythromycin (ermA), exhibiting ARGs mostly found in clinical isolates. Enterococci are known to acquire relatively easy AR and to diseminate the ARGs to other species [29]. To date, nine distinct vancomycin resistance clusters have been described in enterococci (vanA, vanB, vanC, vanD, vanE, vanG, vanL, vanM and vanN), the vanA cluster being the most common mediator of vancomycin resistance in enterococci. Multidrug-resistant (MDR) and vancomycin-resistant Enterococcus (VRE) are now commonly isolated from clinical samples, sewage, aquatic enviroments, agricultural run-off and animal sources, which indicates their ability to enter the human food chain. VRE threatens to compromise the effective treatment of infections caused by these MDR bacteria, particularly in seriously ill patients who may need Fig. 1 The antibiotic resistance profile of the Enterococcus spp. isolated strains revealed that 75 % of the strains were resistant to erythromycin, followed by vancomycin (35 %) and tetracycline (15 %), and 5 % of them to doxycycline and linezolid. Ampicillin and penicillin, the most active βlactam antibiotics against enterococci were found to be 100 % efficient against isolated strains, as well as the tested fluoroquinolone (levofloxacin) treatment with vancomycin where other antibiotics have failed [30]. The VRE conjugative transposons can transfer vancomycin resistance to Staphylococcus aureus, streptococci and lactobacilli. Historically, the vancomycin use was also coupled with the emergence and spread of methicillin-resistant S. aureus (MRSA) in the 1960s [31]. E. faecalis has been reported to transfer plasmids harbouring AR traits to other enterococci and to Listeria monocytogenes in wastewater treatment plants [32]. Also, E. faecium conjugative transposons can be transferred from animal to human microbiota.
The increased levels of erythromycin resistance observed in this study is also alarming, taking into account that macrolide-lincosamide-streptogramin (MLS) antibiotics constitute an alternative therapy for the treatment of insidious VRE infections [33]. Macrolide-resistant Enterococcus sp. strains have been often isolated in humans and animals [34,35]. Two major resistance mechanisms that cause macrolide resistance in enterococci isolates are the target modification due to the ribosomal methylase encoded by erm genes (MLSB phenotype) and the expression of efflux pumps encoded by mef(A/E) and msr genes (M phenotype) [36]. Our isolates have shown the ermA genotype, affecting both macrolides and lincosamides.
Regarding the isolated GNR strains, our experimental data have shown that E. coli strains were isolated in a higher proportion in Jirlău and Câineni lakes compared to the other species ( Table 2).
The emergence of GNR resistance in the aquatic environments can be associated with an increased risk to human and animal health [37][38][39]. Nowadays carbapenem-resistant Enterobacteriaceae are very difficult to treat especially in immunocompromised patients and these strains have also been reported in long-term care facilities [40]. Carbapenemases represent the most IMP-1 (imipenemase) first reported in S. marcescens in Japan in 1991; the most frequently carbapenemases in non-fermentative GNB are the VIM type (Verona imipenemase) firstly described in Pseudomonas aeruginosa isolates, which has been identified also in Enterobacteriaceae strains [41][42][43].
ESBLs are a group of plasmid-mediated enzymes that are posing a major therapeutic challenge today in the treatment of hospitalized and community patients. Most of ESBLs are members of TEM, CTX-M, OXA and SHV β-lactamase families [44]. CTX-M15 (the most commonly worldwide) was predominantly revealed in humans, while CTX-M1 in farming animals. The global spread of CTX-M type enzymes also associated with quinolones resistance has become the main concern nowdays. The trimetoprim-sulphamethoxazole resistance is most frequently encoded on plasmids coexpressing ESBLs [45,46]. The ESBLs and carbapenemases have been reported as the major type of beta-lactams resistance mechanisms in both human and veterinary medicine among GNR. However, such resistance phenotypes were increasingly reported in the environment, which both receives and further disseminates MDR bacteria [47].
In the aquatic environments there have been demonstrated that efflux genes (tet A, B, C, D and E) are frequently detected in Enterobacterales species found in nature, humans, and animals [48]. The tet(A) gene is commonly associated with plasmids, and can be easily transferred between Gram negative bacteria [49].
In our study the integrase gene was found in more than 50 % of the investigated strains (58.33 %) exhibiting a MDR phenotype. Class I integrons are capable of transferring genes responsible for resistance to β-lactam, aminoglycoside, sulfonamide and quaternary ammonium salts [50]. Several studies have correlated high class 1 integron abundance in the environmnent to anthropogenic activity. In one study, concentrations were found to be higher in aquatic environments contaminated with metals and antibiotics compared to unexposed environments. Another study showed that class 1 integrons were more abundant in detergent and antibiotic-contaminated sewage sludge and pig slurry compared to unexposed agricultural soils. The capacities of mobile integrons to disseminate among bacteria, to confer adaptive advantages to continously changing environmental conditions, and to use the gene cassettes of the environmental metagenome make them likely facilitators of environmental AR dissemination [51].
Another study performed on 108 bacterial strains isolated from eels and aquaculture ponds, revealed the presence of some ARGs (blaTEM, tetC, sull, aadA, floR and qnrB) in a high percentage and also detected the class I and II integrons [52].

Conclusions
The cultivable aquatic microbiota from fishery lakes is mainly represented by enterococci and Enterobacterales strains. The source of contamination of all these lakes is Fig. 2 The antibiotic resistance profiles of the isolated GNRs. The analyzed GNR proved to be 100 % susceptible to aminoglycosides (excepting kanamycin), carbapenems, and trimethoprim/sulfamethoxazole. The GNR strains have shown 100 % resistance to ticarcilin, 21 % to tetracycline, 4 % to aztreonam and expressed different resistance levels to cefems (between 4 % and 63 %) probably the neighboring urban sewage or factories. The GNR strains exhibited high levels of β-lactam resistance mediated by ESBL and metallo-β-lactamases, accompanied with a lower resistance to the majority of all other important classes of non-β-lactam antibiotics. The majority of Enterococcus spp. isolates were resistant to macrolides and vancomycin. The high level and diversity of resistance markers, correlated with a high frequency of class 1 integrons is suggesting that this environment could act as an important reservoir of ARGs with a great probability to be horizontally transmitted to other associated species in sediments microbiota, but also to animal and human pathogens raising the potential zoonotic risk for fish consumers. Thus, the findings of this study may assist in developing strategies to avoid the spread of ARB. The limitations of the article are represented by the small number of analyzed strains, the lack of details regarding the sources of pollution with ARB or other possible characteristics that might have explained the presence of resistant microorganisms. Further studies, some of them being already underway through ongoing projects will address these limitations in the future. For now, this article represents a red flag to warn about the resistant microorganisms found in Romanian natural salted lakes included in Natura 2000 network.

Sampling site and isolation of the strains
The water samples were taken in sept. 2016 from four lowland salted lakes belonging to Natura 2000 network.
The sampling points have been represented by four lakes circumscribed to a plain area of approx. 6500 ha of water from Buzău and Brăila counties, namely: ROSCI0005 Balta Albă-Amara-Lacul Sărat Câineni-Jirlău, ROSPA0004 Balta Albă-Amara-Jirlău and the area of reservations 2.271-Balta Albă, 2.272-Balta Amara, 2.260-Lacul Jirlău (Fig. 3). The area is polluted in particular by the actions and activities of the residents such as: discharging into the lakes of the domestic residues, natural and artificial modifications of the water composition, changing the methods of cultivation of land from the traditional ones in intensive agriculture and the excessive use of fertilizers. The lakes on this site are populated with fish of different species (carp, flax, chickpea, saltpeter and slumber), which are fished and consumed, and can be carriers of ARGs. The strains were isolated from samples on a selective (EMB, McConkey) and differential culture media (blood agar) as described previously [8]. A total of 44 lowland salted lakes strains were recovered from the selective and differential culture media and identified using the MALDI-TOF-MS Bruker system: 9 from the Balta Albă lake, 11 from the Balta Amara lake, 13 from the Jirlău lake and 11 from the salty lake Câineni.

Antimicrobial sensitivity test for identified strains
The antibiotic susceptibility profiles were checked using disk diffusion method, according to Clinical and Laboratory Standards Institute (CLSI 2016 and 2017) using   Table 3), using a reaction mix of 20 µl (PCR Master Mix 2x, Thermo Scientific) containing 1 µl of bacterial DNA extracted using an adapted alkaline extraction method. In this purpose 1-5 colonies of each strain was resuspended in 20 µl NaOH (0.05 M) + SDS (0.25 %), the suspensions were heated at 94°C for 15 minutes followed by the addition of 180 µl of TE 1x buffer solution and centrifugation at 13,000 rpm, 3 minutes. The supernatant was checked by electrophoretic migration in a 1 % agarose gel and 1, 5 %, 45 minutes at 90 V and stained with 3.5 µg/ml ethidium bromide. Authors' contributions VL, IG, CC, MIP and MCC designed the study and corrected the manuscript; DD facilitated the collection of water samples from four fishery lakes from Buzău and Brăila counties; VCC, IS and FM performed the isolation, identification and phenotypic susceptibilities of the isolated strains, IG and CC performed the molecular analyses, GLP has contributed to the design of the study, analysis of the results, drafting and the correction of the manuscript; VL, IG, MCC and CC drafted the paper and have equally contributed to this paper as main authors.

Funding
This study was supported by the Romanian Executive Agency for Higher Education, Research, Development and Innovation (https://uefiscdi.gov.ro/) research project PN-III-P4-ID-PCCF-2016-0114 POSCCE (RADAR) and COFUND-JPI-EC-AMR-ARMIS. The funding had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Availability of data and materials All data analyzed or generated during this study are included in this published article and its Supporting Information files. Any additional information is available from the corresponding author on reasonable request.