IncF Plasmids Are Commonly Carried by Antibiotic Resistant Escherichia coli Isolated from Drinking Water Sources in Northern Tanzania

The aim of this study was to identify the replicon types of plasmids, conjugation efficiencies, and the complement of antibiotic resistance genes for a panel of multidrug resistant E. coli isolates from surface waters in northern Tanzania. Standard membrane filtration was used to isolate and uidA PCR was used to confirm the identity of strains as E. coli. Antibiotic susceptibility was determined by breakpoint assay and plasmid conjugation was determined by filter-mating experiments. PCR and sequencing were used to identify resistance genes and PCR-based replicon typing was used to determine plasmid types. Filter mating experiments indicated conjugation efficiencies ranged from 10−1 to 10−7. Over 80% of the donor cells successfully passed their resistance traits and eleven different replicon types were detected (IncI1, FIC, P, FIIA, A/C, FIB, FIA, H12, K/B B/O, and N). IncF plasmids were most commonly detected (49% of isolates), followed by types IncI1 and IncA/C. Detection of these public health-relevant conjugative plasmids and antibiotic resistant traits in Tanzanian water suggests the possible pollution of these water sources from human, livestock, and wild animal wastes and also shows the potential of these water sources in the maintenance and transmission of these resistance traits between environments, animals, and people.


Introduction
Increased mortality and morbidity due to antibiotic treatment failure make antimicrobial resistance (AMR) one of the 21st century's major global public health challenges [1]. Overuse and misuse of antibiotics are considered major reasons for the emergence of resistant bacteria in many lowincome countries [2,3]. Antibiotic resistance has been documented for enteric bacteria from various water sources and these water sources could facilitate dissemination of resistant bacteria to a wider community of people and animals [4]. This is particularly true for low-income countries like Tanzania where water sources are frequently shared between animals and people [5,6]. For example, a report from Kenya reported a high prevalence of antibiotic resistance E. coli from water and fish in Lake Victoria [ampicillin (64%), tetracycline (76%), and cotrimoxazole (80%)] [7] where untreated water is consumed routinely.
Tanzanian hospitals have reported a high proportion (80%-90%) of clinical E. coli isolates that are resistant to antibiotics such as ampicillin, cotrimoxazole, tetracycline, gentamicin, and amoxicillin/clavulanic acid. These bacteria infect people within a healthcare system where, in most cases, there are no laboratory diagnostics to guide antibiotic treatment [8][9][10]. Another study reported a high number of antibiotic resistant E. coli, possessing resistance to cephalosporins, from free-range buffalo, zebra, and wildebeest [11]. These animals were located in mixed grazing areas with potential contact with people and livestock. Contaminated water was suspected as the source of resistant bacteria found in these 2 International Journal of Microbiology wild animals [11]. Contaminated water most likely plays a role in the dissemination of antibiotic resistant bacteria and the probability of transmission likely increases when people and animals use that water. Antimicrobial resistance (AMR) genes are transferred to other bacteria, sometimes at the species level, by horizontal gene transfer (transduction, transformation, and conjugation). Plasmid-mediated horizontal transfer of multidrug resistance between different bacteria is a major concern because this contributes to the evolution and emergence of antibiotic resistant bacteria in the environment [12,13].
In northern Tanzania, surface water such as rivers and ponds is often shared between animals and people on daily basis. Consequently, these water sources become polluted with human and animal excreta and might harbor antibiotic resistant enteric bacteria. Consumption of water containing these bacteria is likely to increase the risk that antibiotic resistant and pathogenic bacteria will be transmitted. Nevertheless, to date, no studies have been conducted in Tanzania to determine if drinking water represents a risk factor for transmission of antibiotic resistant bacteria to people and animals. The objective of this study was to characterize the replicon types of plasmids that harbor drug resistant traits, their conjugation efficiencies, and the complement of antibiotic resistance genes for a panel of multidrug resistant E. coli isolates that were obtained from drinking water sources in northern Tanzania.

Study Design.
Convenience sampling was used to collect water samples between March and August 2014. Each source was visited twice and one sample was collected from each source per visit (in Tanzania, March is the rainy season and August is during dry season). Sample locations included the Kilimanjaro Region (Moshi Municipal, Moshi Rural, and Hai Districts), the Arusha Region (Arusha City, Arumeru, Longido, and Monduli Districts), and the Manyara Region (Simanjiro and Babati Districts). A convenience approach was used to select sampling sites with appropriate permission from local authorities. Water samples from ponds were collected from localities near Maasai villages. All sites, including streams, may have been impacted by people and wildlife. We also collected opportunistic samples from taps and wells.

Isolation and Identification of E. coli.
Water samples were collected in 500 mL sterile bottles and were transported in cooler boxes with ice packs to the laboratory for processing within 6 h of collection. Out of 500 mL, 100 mL water samples were analyzed using a standard membrane filtration technique with minor modifications [20]. Following filtration, each filter membrane was placed on a chromogenic selective agar plate (HiCrome E. coli Agar, HiMedia Laboratories Prt. Ltd., Mumbai, India). The agar plates were initially incubated at 37 ∘ C for 4 h, followed by incubation for 16-22 h at 44 ∘ C. Plates produced 1-200 CFU of which individual colonies were subcultured to ensure purity for further characterization. The identity of E. coli isolates was confirmed using a PCR genotyping test that detects the presence of the uidA gene [21].
Antibiotic break point assays were used to determine the resistance profile of each E. coli isolate against a panel of important antibiotics. MacConkey (MAC) (Thermo Oxoid Remel) agar plates with each antibiotic at their CLSI recommended minimum inhibitory concentrations [22] (Table 1) were used to perform the break point assays [23]. E. coli strains K-12 (negative control; susceptible to all antibiotics tested) and H4H E. coli (positive control; resistant to all antibiotics tested) were used as reference strains for antibiotic susceptibility testing.

Plasmid Characterization.
A set of 31 E. coli isolates that were susceptible to nalidixic acid and resistant to more than 2 antibiotics tested were chosen for this study. Filter-mating experiments were performed to determine the conjugation rates of plasmids with the nalidixic acid susceptible MDR (wild-type) E. coli isolates as donors and a plasmid-free recipient strain (E. coli K-12, nalidixic acid resistant, Nal r ) as recipient as described earlier [23,24] with minor modifications. Briefly, single colonies of E. coli K-12 and potential donor strains were grown separately overnight in LB medium (Luria-Bertani medium; Difco6 LB Broth Lennox, Sparks, MD, USA) at 37 ∘ C. Equal quantities (10 L) of overnight cultures of donor and recipient strains were added on top of a nitrocellulose (∼1 cm 2 ) membrane overlaid on LB agar with no antibiotics. After 24 h of incubation at 37 ∘ C, cells from the membrane were suspended in 500 L of sterile phosphate-buffered saline (PBS, pH 7.0) and spread onto LB agar plates containing 32 g/mL nalidixic acid (Sigma-Aldrich) and another antibiotic to which the donor cells were resistant ( Table 2). Colonies that grew on these selective agar plates were considered transconjugants. The conjugation efficiency of plasmid was calculated by dividing the number of transconjugants by the number of donor cells. Transconjugants were screened for their donor's antibiotic resistance phenotypes and presence of tet(A), tet(B), TEM-1 , SHV , and CTX-M genes [25,26]. CTX-M grouping (group 1, group 2, and group 9) was further evaluated for all CTX-M positive isolates. All CTX-M E. coli isolates were positive for CTX-M group 1 and the PCR products were subsequently sequenced by Functional Bioscience (Madison, WI). Sequencher (ver 5.0) software was used to process sequence traces, and the final sequences were analyzed with CLC Genomic Workbench 7.0.2 (CLC Bio Aarhus, Denmark) and compared with the reported sequences from GenBank (http://blast.ncbi.nlm.nih.gov/Blast.cgi).
PCR-based replicon typing was used with genomic DNA of the transconjugants using the methods described by Johnson et al. [15]. Briefly, pellets from 1 mL of overnight culture were resuspended with 200 L of nanopure water and placed in a heating block at 100 ∘ C for 10 min. The lysed suspension was cooled to room temperature and centrifuged briefly to pellet debris. Supernatant was transferred to a new vial and stored at −80 ∘ C until ready for testing against 18 different sets of primers [15] that were grouped into three multiplex primer panels [15]. The following PCR conditions were used: 5 min at 94 ∘ C, 30 cycles of 30 s at 94 ∘ C, 30 s at 60 ∘ C, and 90 s at 72 ∘ C, and a final extension of 5 min at 72 ∘ C. The amplified PCR products were visualized using 1.5% Tris-acetate-EDTA agarose gel containing 0.2 g/mL ethidium bromide alongside a 1 kb ladder (Gene ruler 1 Kb, Life Technologies).

Discussion
Plasmid-mediated horizontal transfer of multidrug resistance traits plays a key role in the dissemination of antimicrobial resistance around the world [27]. Our study shows that E. coli isolated from Tanzanian water sources harbor multiple plasmids belonging to major plasmid replicon types such as IncF, A/C, l1, and N. Most of these plasmids were associated with transfer of antibiotic resistance traits via conjugation with rates that varied between 10 −1 and 10 −7 using filter-mating assays. Moreover, these plasmids harbor multiple antibiotic resistance genes that are associated with plasmid replicon types such as IncF, A/C, N, l1, H12, and B/O. Studies show that plasmid-mediated horizontal gene transfer occurs within and between E. coli and Pseudomonas isolated from sewage and lake water [13]. Given the presence of resistant E. coli in biologically contaminated water from Tanzania, it is likely that their presence contributes to the long-term persistence of resistance traits in people and animals who share these water resources.
Among the 11 replicon types found, IncF group plasmids were detected more frequently than other tested groups. This is in accordance with previous studies where IncF plasmids were found predominantly in E. coli from clinical samples (rectal samples, gastric aspirate samples, and vaginal sample) [28,29] and in E. coli from people (feces and UTI patients) and poultry (fecal swab) [15]. IncFIB was the most frequently detected (16%) replicon type, similar to what has been reported for E. coli isolates collected from fecal samples of healthy people and cattle in Nigeria [30]. The overall proportion of IncF-positive E. coli was 49%, which is lower than that observed in Germany (71%) [31] but higher than that observed in E. coli isolates from fecal samples of healthy people and food animals in Switzerland (45%) [27]. IncF type plasmids have a "narrow" host range although they are well adapted to E. coli and are frequently associated with the presence of tet(A), TEM-1 , and CTX-M [31,32]. In this study, plasmid type IncF was associated with tet(A), TEM-1 , and CTX-M-15 . The CTX-M-15 -lactamases are disseminated worldwide and are usually located in the conjugative plasmids [33]. Detection of this trait in E. coli isolated from water sources is a public health concern because CTX-M-15 -lactamases are commonly associated with urinary tract infections [30].   The total percentage sums to 106 because some isolates were positive for more than one replicon type.
Another study from a hospital in Tanzania found CTX-M- 15 in Klebsiella pneumoniae that can be associated with neonatal sepsis [34]. In this study, detection of CTX-M-15 in E. coli from water suggests a possible contamination of human, livestock, and wild animal excreta and thus consumption of this untreated water is clearly a potential risk for transmission back to people [35]. The IncI1 plasmid type was the second most prevalent replicon type (13%) and these plasmids also harbor multiple resistance genes. E. coli can reportedly maintain IncI1 plasmids without antibiotic selection pressure and with little or no apparent fitness cost to the host bacterium [36]. Importantly, it is also a conjugative plasmid commonly detected in E. coli recovered from humans and animals with the conjugation efficiencies ranging between 10 −2 and 10 −7 [15,16,28,34]. The IncI1 plasmid carrying CTX-M-15 and TEM-1 has been associated with the recent 2011 outbreak of E. coli O104 in Germany [37]. In addition, bacteria carrying lncl1 plasmids were responsible for community and hospital acquired infections [38,39].
IncA/C plasmids are typically larger (∼150 kb) than others with lower conjugation efficiencies [24,40,41]. About 10% of E. coli isolates from our water samples harbored IncA/C plasmids and these were associated with resistance to ampicillin, streptomycin, sulfonamide, tetracycline, and trimethoprim. Importantly, IncA/C plasmids can harbor a large number of antimicrobial resistance genes and the broad-host spectrum coupled with an ability to spread via conjugation transfer within bacteria communities means that they can transfer an arsenal of resistance traits to pathogens of people and animals [40,42,43]. Isolation of E. coli with IncA/C from environmental water samples that are consumed by humans and animals on daily basis is a major public health concern. The replicon typing methods have some pitfalls including the obvious inability to detect unknown replicons [15]. For example, 14 isolates in the current study transferred their resistance phenotypes to the recipients' cells but no plasmid replicon was detected. Detection of these public health important conjugative plasmids and antibiotic resistant traits in Tanzanian water suggests the possible pollution of these water sources from human, livestock, and wild animal wastes and also shows the potential of these water sources in the maintenance and transmission of these resistance traits between environment, animals, and people. Therefore, appropriate intervention strategies should be identified and implemented to reduce the water pollution.