FITNESS AND PATHOGENICITY OF OUTBREAK CAUSING Salmonella UPON SHORT-TERM EXPOSURE TO GROUNDWATER WITH RESIDUAL ANTIBIOTICS

The changes in survival and pathogenicity of three Salmonella enterica subsp. enterica serotypes upon short term exposure to groundwater with residual antibiotics have been studied in relationship to overall microbial fitness. A wild type flagellated Salmonella enterica ser. Typhimurium outbreak strain, a mutant Salmonella enterica ser. Typhimurium strain, and a wild type avian disease-causing Salmonella enterica ser. Pullorum strain were exposed to a range of ionic strength (3-30 mM) groundwater with residual antibiotics for 6-24 hours. Exposed organisms’ pathogenicity was tested in vitro exposure to a human epithelial cell line (HEp2). Resistance profiles against 10 common antibiotics were also tested and compared to unexposed controls. Results show minor antibiotic resistance changes for S. enterica ser. Typhimurium strains in response to some antibiotic classes mediated with active efflux pumps. This trend was not observed for S. enterica ser. Pullorum, suggesting that resistance found in groundwater exposed organisms might be strain-dependent. In vitro epithelial cell invasion assays showed bacterial invasion of HEp2 cells initially decreases with time and increases after 24 hours. It is concluded that S. enterica serotypes reaching groundwater environments in the presence of residual antibiotics may exhibit increased levels of pathogenicity, strain-dependent resistance to antibiotics, and sustained levels of viability.


Introduction
Salmonella is a robust zoonotic pathogen threatening both human and animal health significantly.Nontyphoidal Salmonella infections are increasing worldwide with at least nine to ten incidents reported each year since 2009 only in the U.S. (Vugia et al. 2004).A recent major outbreak caused by S. enterica ser.Typhimurium has affected 714 individuals in 46 states (CDC 2009).One of the reservoirs for pathogenic bacteria including Salmonella following excretion from infected individuals is groundwater environments (Schmoll et al. 2006).Pathogens released from various point and non-point sources of human and/or animal waste will likely reach groundwater and create health related issues in individuals who are directly or indirectly exposed to the groundwater (Crane & Moore 1984).Considering the facts that antibiotics are frequently detected in various groundwater systems (Barber et al. 2009, Barnes et al. 2008, Vulliet & Cren-Olivé 2011) and 26% of the total freshwater demand is supplied through groundwater (Hutson 2004), it is essential to understand the effects of groundwater environments with residual antibiotics on the fitness of waterborne pathogens as it pertains to several physiological 2 B.Z. Haznedaroğlu characteristics such as pathogenicity, survivability and gained or lost resistance to certain antibiotics.
Another unknown issue with the ion-rich nature of groundwater systems is how bacterial efflux pumps are affected with respect to antimicrobial resistance, especially against aminoglyglycosides (Bradford 2008), tetracyclines (Chopra & Roberts 2001) and quinolones (Poole 2000).It has been previously documented that working mechanisms of these bacterial efflux pumps are directly and/or indirectly linked to the ionic components in the external environment.For example, the insertion of aminoglycosides into bacterial LPS (Bradford 2008) follows a proton gradient (energy) dependent transport across the cytoplasmic membrane.Additionally, a tetracycline efflux pump system is induced under the presence of Mg 2+ in the external environment (Roberts 1996, Orth et al. 2000), which results in the excretion of the antibiotic in an energy-dependent manner with proton exchange for a tetracycline-cation complex (Yamaguchi et al. 1993).The majority of efflux pumps work in this manner, i.e., compound-ion antiporter, and are reviewed in detail elsewhere (Kumar & Schweizer 2005).Similarly, efflux pump systems in Salmonella also assist the bacteria in resistance to antibiotics such as ciprofloxacin (Giraud et al. 2000), β-lactams (Nikaido et al. 1998), as well as creating simultaneous resistance to multiple drugs (Abouzeed et al. 2008, Piddock et al. 2000, Quinn 2006).Resistance to β-lactams in Salmonella is mediated by the AcrAB efflux system and AcrB is also required for virulence of S. enterica ser.Typhimurium (Lacroix 1996), suggesting that drug transporters may contribute not only to antibiotic resistance but also to virulence (Kunihiko Nishino 2006).Despite these findings, comprehensive studies on changes of antibiotic resistance in waterborne Salmonella spp.exposed to groundwater with residual antibiotics have not been performed so far.
This study aims to investigate whether the fitness of Salmonella is affected when exposed to ion-rich groundwater conditions with residual antibiotics in the external environment.The fitness was evaluated with respect to decreased antibiotic resistance, viability and pathogenicity against human cell lines.For these purposes, an outbreak-causing wild type S. enterica ser.Typhimurium, a non-motile mutant phenotype of S. enterica ser.Typhimurium and S. enterica ser.Pullorum, an important avian outbreak-causing strain, were selected as model organisms.These particular strains were chosen due to their different genotypic and phenotypic variation and their low-level susceptibility to the tested antibiotics as control groups to monitor any suspected changes in their profiles.

Bacterial strain selection and preparation
All Salmonella strains used in the study were obtained from the Salmonella Genetic Stock Centre (SGSC) of University of Calgary, Alberta, Canada.S. enterica ser.Typhimurium strain ST5383 is a wild type strain and was originally isolated from an outbreak infecting more than 1700 people (Bezanson et al. 1985).The strain designated as SGSC1512 is a non-motile mutant of S. enterica ser.Typhimurium TM2 (Yamaguchi et al. 1986).The nonmotile, non-flagellated avian pathogenic S. pullorum wild type strain SA1685 (CDC number 2863-65) was originally isolated from an infected turkey.
S. enterica serotypes used during the course of the study was pre-cultured in Luria-Bertani (LB) broth (Fisher Scientific, Fair Lawn, NJ) at 37 °C overnight, and shaken continuously at 120 rpm.Overnight grown cultures were inoculated into fresh LB broth in 1:100 dilutions and harvested at the mid-exponential phase determined by the growth curve analysis for each bacterium.A refrigerated bench-top centrifuge (5804R; Eppendorf, Hamburg, Germany) equipped with a fixed angle rotor (F-34-6-38; Eppendorf) was used to pellet the cells with an applied 3,700×g force for 15 min at 4 °C.Growth medium was decanted and the pellet was resuspended in 10 mM potassium chloride (KCl) solution.The process was repeated twice in order to ensure complete removal of the growth medium.Electrolyte solutions used in cell preparation and other experiments were prepared with deionized water (DIW) (Millipore, Billerica, MA) and reagent-grade KCl (Fisher Scientific).The concentration of the cell stock solution was determined by using a cell counting hemocytometer (Bürker-Turk, Germany) under a light microscope (Fisher Scientific).

Cell viability analysis
At the end of exposure to GW, viability of both the control and the test groups was determined by using the Live/Dead BacLight® kit (L-7012; Molecular Probes, Eugene, OR) according to the manufacturer's directions.Direct counting of the stained live and dead cells was done using an inverted microscope (IX70; Olympus, Japan) operated under red/green fluorescence filter set (Chroma Technology Corp., Brattleboro, VT).
Invasion of Salmonella into epithelial cells was quantified with slight modifications of the protocols described in Burnham et al. (2007).Briefly, a monolayer of HEp-2 cells was grown until confluence in 24-well plates (Corning, Corning, NY).Salmonella, at a concentration of 10 5 cells per mL, was inoculated into cell-culture plates and incubated for 2 hours at 37 °C to allow internalization.Following the incubation, each well of the plates was washed three times with phosphate buffered saline (PBS) to remove unbound bacteria.Bacteria that were bound to epithelial cells but had not been internalized were killed by applying fresh growth medium containing 5 µg penicillin per mL and 100 µg gentamicin per mL, and incubated for 2 hours at 37 °C.Following the incubation, cells were washed with PBS, treated with trypsin-EDTA complex (ATCC), and lysed with 0.1 % Triton-X100 (Fisher Scientific).The lysates were spread onto LB agar plates and incubated for 18 hours at 37 °C.The number of colony forming units (CFUs) was counted to quantify the number of Salmonella that had successfully invaded the monolayer of the epithelial cells.

Statistical analysis of data
Changes in the diameters of antibiotic inhibition zones (for determining antibiotic susceptibility profiles), viability, and the number of CFUs of Salmonella infecting HEp2 cells were statistically compared using the unpaired Student's t-test using Minitab ® Version 14 (State College, PA).Differences between the control and the test groups were considered to be significant at 95% confidence interval (P < 0.05).

Light microscopy imaging of epithelial cells
Representative images of both uninfected and Salmonella-invaded HEp2 cells were obtained by using a light microscope (Micromaster, Fisher Scientific) (Fig. 1).The Giemsa stain (Fisher Scientific) which differentially stains human and bacterial cells in purple and pink, respectively was used.Chambered cover glass slides (Nunc Lab-Tek, Fisher Scientific) were used to grow HEp2 cells until desired confluence and infected with Salmonella as described above.After the chambered cover glasses were manually removed, the slides were placed in Giemsa stain (2.5%) for 45-60 minutes.Slides were rinsed by dipping in a buffer solution [0.59% (w/v) Na2HPO4, 0.36% (w/v) NaH2PO4H2O] three-four times and left to air dry.Stained slides were observed under light microscope and their images were taken at 400× magnification.

Antibiotic susceptibility analysis
The control group of the non-motile mutant SGSC1512 was susceptible to all antibiotics except cephalothin and nalidixic acid, where intermediate level resistance was observed (Table 1).SGSC1512 exposed to GW for 6, 12, 18 and 24 hours did not change in resistance to antibiotics except for cephalothin to which the isolates became susceptible after exposure periods of 12, 18, and 24 hours at all IS (with the exception of 30 mM IS exposure at 12 hours).For 6 hours of exposure, SGSC1512 showed decreased susceptibility to cephalothin only at 3.33 mM IS, whereas at 10 and 30 mM IS the isolates remained with intermediate resistance (Table 1).
The control group of the wild type S. enterica ser.Typhimurium strain ST5383 showed full susceptibility to all antibiotics except amoxicillin-clavulanic acid and nalidixic acid, for which only intermediate resistance was observed (Table 2).Interestingly, exposure to GW increased the resistance of ST5383 against nalidixic acid for all conditions except for 12 hours at 30 mM IS. ST5383's intermediate resistance to amoxicillinclavulanic acid decreased to the susceptible level when it was exposed to GW for 24 hours at all IS and after 18 hours at 10 mM IS (Table 2).
The control group of the wild type and non-motile S. enterica ser.Pullorum strain SA1685 was susceptible to all tested antibiotics.This strain did not show any changes in its antibiotic resistance profile and remained susceptible to all antibiotics tested for all experimental conditions as shown in Table 3.

Epithelial cell invasion assay
The number of CFUs for control groups invading HEp2 cells ranged approximately from 300 to 420 (Fig. 1).After exposure to GW for 6 and 12 hours at all IS tested, the number of all Salmonella strains invading the epithelial cells lines decreased significantly.The same trend was also observed after 18 hours of exposure to GW for all strains except ST5383 at 10 mM and 33 mM and SGSC1512 at 30 mM IS. Longer exposure periods (i.e.24 hours) to GW was associated with an increase in the number of bacteria invading the HEp2 cells at all IS, with the exception of 3.33 mM for SGSC1512 and ST5383.The invasiveness of SA1685 exposed to all IS for 24 hours increased significantly compared to the control group (Fig. 1).
Representative images of SGSC1512 strain invading HEp2 cells are presented in Fig. 2.

Viability of Salmonella under stress
The percentages of viable cells of both the control and test groups of all Salmonella isolates exposed to 3.33, 10, and 30 mM GW with residual antibiotics for 6, 12, 18, and 24 hours were given in Fig. 3.As can be seen in Fig. 3, viability values of the control groups were determined to be over 90%.Throughout the spectrum of exposure conditions, residual antibiotics did not cause major viability loss; the lowest viability values observed were approximately 85%, 83% and 86% for SGSC1512, ST5383 and SA1685, respectively.

Discussion
In this study, the changes in pathogenicity and viability of outbreak causing S. enterica serotypes exposed to groundwater with residual antibiotics were investigated to determine whether this groundwater matrix might impose fitness changes in terms of pathogenicity (invasiveness to target cell lines) and survivability.In addition, different IS conditions were tested to see if any correlation would exist between bacterial efflux pumps (as antibiotic resistance mechanisms) and the external GW environment with varying IS.The results showed increased virulence to be more evident at longer durations (i.e.24 hours), sustained viability, and strain-dependent decreased antibiotic resistance among S. enterica serotypes upon exposure to GW with residual antibiotics.However, no conclusive evidence has been found to link the contribution of efflux pumps to antibiotic resistance of the bacteria under varying IS.
The first organism tested was S. enterica ser.Typhimurium strain SGSC1512 with dysfunctional flagella making it a non-motile strain.During exposure to GW, SGSC1512's resistance to cephalothin decreased from an intermediate resistance to a susceptible level for the exposure period of 12 hours and further (with the exception of 30 mM IS exposure at 12 hours).Although cephalothin is a member of the β-lactam class antibiotics, the same trend was not observed for the other β-lactam antibiotics.The influence of efflux pumps (that are likely to be impacted with varying IS of the external environment) on resistance to β-lactams class antibiotics was previously documented in the literature (Lacroix, 1996;Nikaido et al., 1998) but the results of the present study did not confirm this influence.Both the control and the test groups of SGSC1512 showed intermediate levels of susceptibility when tested against nalidixic acid.Nalidixic acid is a quinolone class antibiotic which is normally pumped out of the Gram-negative bacteria cytoplasm via their active efflux pumps.This is in agreement with previous studies showing the effects of efflux pumps mediating reduced quinolone susceptibility in Salmonella (Cebrián et al. 2005, Chu et al. 2005) However, we did not observe an increased resistance in the case of ciprofloxacin, another quinolone antibiotic, most probably due to the slightly different physical and chemical structures of the two drugs, with possible variations in the influx and efflux trends.
The second organism tested was a wild type, outbreakcausing S. enterica ser.Typhimurium strain (ST5383).Exposure to GW resulted in increased resistance of the control group from susceptible level to an intermediate resistance for all tested conditions, which was apparent by the statistically significant increased zone diameters at 3.33 mM for 12 hours and at 10 mM for 18 hours against nalidixic acid.As in the case of resistance profile of SGSC1512, no change was observed against the other quinolone ciprofloxacin for the tested conditions.At longer exposure conditions (i.e.24 hours for all IS) the intermediate level of resistance of the control group of ST5383 against amoxicillin-clavulanic acid decreased to susceptible levels.The same decrease was also observed at 10 mM GW after 18 hours of exposure.As mentioned earlier, S. enterica serotypes shows resistance to βlactams by the AcrAB efflux system.However, there was no change in the resistance profiles of ST5383 against cefoxitin, ceftriaxone, and cephalothin, all in β-lactam group.This suggest that there is either no direct correlation between the efflux pumps assisting the ST5383 strain to sustain resistance against amoxicillinclavulanic acid or the fitness of the strain was weakened following the 24 hours GW exposure.Another possibility is the secretion and release of β-lactamases by pathogenic bacteria is a more direct way of mediating the β-lactams.Long-term GW exposure might have negatively impacted the synthesis of β-lactamase in ST5383, which is observed as increased susceptibility.
The last organism tested was S. enterica ser.Pullorum strain SA1685, a non-flagellated serovar responsible for avian diseases in the poultry industry.The control group of SA1685 was observed to be susceptible to all antibiotics tested.The exposure of this strain to GW did not cause a change in the antibiotic resistance profile.Since SA1685 is a genotypically and phenotypically different serotype than SGSC1512 and ST5383, this trend might be due to the morphological features and overall fitness of the bacterium and its response to GW stress conditions.Epithelial cell invasion assay results showed that exposure to GW for 6 and 12 hours resulted in decreased invasiveness of S. enterica serotypes into HEp2 cell lines for all tested IS conditions (Fig. 1).In addition, the number of SA1685 that can invade epithelial cells also decreased at all IS for 18 hours of exposure.The number of bacteria that can invade HEp2 cells decreased for 3.33 mM and 10 mM GW, and 3.33 mM for 18 hours of exposure in the cases of SGSC1512 and ST5383, respectively.However, at 24 hours of exposure to GW, increased number of ST5383 and SGSC1512 that can invade the epithelial cells were observed at 10 and 30 mM IS.In the case of SA1685, the number of invading CFU increased for all tested IS for 24 hours of exposure.The decreases in the number of invading bacteria observed up to 18 hours of exposure might be related to the decreased fitness induced by the residual antibiotics in the GW.After 24 hours of exposure, the bacteria might have metabolically become accustomed to the GW matrix, which in return increased the number of invading bacteria.Another possibility is that prolonged exposure to GW with divalent cations such as Ca 2+ in this case may have contributed to the increased number of CFUs infecting HEp2 cells, which is in parallel to previous studies that showed increased invasion of human cell lines with increased Ca 2+ concentrations (Niesel & Peterson 1987, Peterson 1988).
In conclusion, although the occurrence of residual antibiotics in groundwater systems is common, there is very limited knowledge on the effects of these antibiotics on the fitness levels of pathogenic bacteria with respect to their pathogenicity and survivability.In this initial study, outbreak-causing S. enterica serotypes with different morphologies were exposed to groundwater conditions with residual antibiotics to test for any changes in invasiveness against human cell lines.Their viabilities were also monitored.In addition, as an ion-rich environment, groundwater may affect the active efflux pumps of bacteria, which in turn change their resistance profiles against certain antibiotic classes.The results showed that relatively short-term exposure to residual antibiotics at µg per liter levels will not decrease viabilities, but may still induce increased invasiveness of S. enterica serotypes into human epithelial cell lines, which was more evident at 24 hours of exposure in the present conditions.Although there was no clear evidence to link the efflux pump mediated antibiotic resistance to different concentrations of ions tested in this study, subtle changes in resistance profiles were observed in certain quinolone and β-lactam class antibiotics.Further analyses might elaborate these findings to investigate the fate of S. enterica serotypes in groundwater settings.

Fig. 1 .Fig. 2 .
Fig. 1.Changes in the number of Salmonella CFUs infecting HEp2 cells upon exposure to GW with residual antibiotics at different IS (in mM) for four different time periods.Data represent average values of three biological replicates and the error bars reflect standard deviations.

Fig. 3 .
Fig. 3. Changes in the percent viability of Salmonella isolates upon exposure to GW with residual antibiotics at different IS (in mM) for four different time periods.Data represent average values of three biological replicates and the error bars reflect standard deviations.

Table 1 .
Antibiotic susceptibility profile of SGSC1512 exposed to different IS (in mM) GW with residual antibiotics for four different time periods.The letters denote the degree of resistance as S; Susceptible, I; Intermediate Susceptible and R; Resistant.
b Numbers denote the concentration (in µg) of the given antibiotic impregnated on 6mm disks.

Table 2 .
Antibiotic susceptibility profile of ST5383 exposed to different IS (in mM) groundwater (GW) with residual antibiotics compared for four different time periods.The letters denote the degree of resistance as S; Susceptible, I; Intermediate Susceptible and R; Resistant.

Table 3 .
Antibiotic susceptibility profile of SA1685 exposed to different IS (in mM) groundwater (GW) with residual antibiotics compared for four different time periods.The letters denote the degree of resistance as S; Susceptible, I; Intermediate Susceptible and R; Resistant