Associations between Bacteriophage Resistance and Antibiotic Resistance Phenotypes in Laboratory and Clinical strains of Salmonella Typhimurium

Bacteriophages have received great attention as alternative over antibiotics due to the host specificity. Therefore, this study was designed to evaluate the associations between bacteriophage-insensitive (BI) and antibiotic-resistant mutants of Salmonella Typhimurium strains. Bacteriophage-sensitive Salmonella Typhimurium ATCC 19585 (BSSTWT), ciprofloxacin-induced S. Typhimurium ATCC 19585 (BSSTCIP), S. Typhimurium KCCM 40253 (BSSTLAB), and clinically isolated multidrug-resistant S. Typhimurium CCARM 8009 (BSSTMDR) were used to induce the bacteriophage-insensitive mutants (BISTWT, BISTCIP, BISTLAB, and BISTMDR) against bacteriophage P22.

4 with bacteriophage adsorption, antibiotic susceptibility, and gene expression.

In vitro Stepwise Selection Assay
To induce antibiotic-resistant S. Typhimurium ATCC 19585 was exposed with serially increasing ciprofloxacin concentrations according to the serial passage procedure [16].
Salmonella Typhimurium ATCC 19585 was repeatedly cultured in TSB and trypticase soy agar (TSA) containing ciprofloxacin concentrations from 0.0078 to 1 μg/mL. The ciprofloxacin-induced resistant S. Typhimurium ATCC 19585 was stable for more than ten passages in antibiotic-free TSB at 37°C for 20 h prior to use.

Bacteriophage Propagation
Salmonella bacteriophages, P22, P22-B1, PBST-10, PBST-13, PBST-32, and PBST-35, were obtained from ATCC and Bacteriophage Bank at Hankuk University of Foreign Studies (Yongin, Gyeonggi, Korea). All bacteriophages were propagated at 37°C for 20 h in TSB 5 containing S. Typhimurium KCCM 40253. The propagated bacteriophages were collected by centrifuging at 6000 × g for 10 min, filtered through by a 0.2-μm filter to remove bacterial lysates, and further purified using polyethylene glycol (PEG) precipitation assay [17]. The titers of bacteriophage were determined by using a soft-agar overlay method [18]. In brief, the selected bacteriophages were serially (1:10) diluted with PBS and gently mixed with the host cells (10 7 CFU/mL) in TSB containing 0.5% agar. The mixture was poured onto the pre-warmed base agar and solidified at room temperature, and then incubated at 37°C for 20 h to enumerate the bacteriophages expressed as a plaque-forming unit (PFU).

Lytic Activity of Bacteriophage
Salmonella bacteriophages (P22, P22-B1, PBST-10, PBST-13, PBST-32, and PBST-35) were used to evaluate the lytic activity against ST W T , ST CIP , ST LAB , and ST MDR . The selected strains (10 8 CFU/mL each) were mixed with bacteriophage (10 10 PFU/mL each) and incubated at 37 o C for 10 min. The incubated cultures were centrifuged at 6,000 × g for 5 min, serially diluted (1:10) with PBS, and plated on TSA using an Autoplate ® Spiral Plating System (Spiral Biotech Inc.). The plates were incubated at 37°C for 24-48 h. The lytic activity was expressed as log N/N 0 ; N and N 0 denote the counts of bacterial cells treated with and without bacteriophages, respectively.

Fluctuation Assay
The fluctuation assay was used to determine mutant distribution whether ST W T , ST CIP , ST LAB , and ST MDR mutants were spontaneous or inducible in the presence of bacteriophages [19]. In brief, ST W T , ST CIP , ST LAB , or ST MDR cells (10 3 CFU/mL) were 6 distributed into 10 test tubes (0.2 mL; Group A) and one test tube (2 mL; Group B). The tubes were incubated at 37 o C for 3 h. Group A (1 replicate/tube) and Group B (10 replicates/tube) were plated on TSA with P22 and incubated for 37 o C for 24-48 h to enumerate viable cells.

Lysogenic Induction Assay
Lysogenic cells were induced by mitomycin C [18]. In brief, the cultured BSST W T , BSST CIP , BSST LAB , BSST MDR , BIST W T , BIST CIP , BIST LAB , and BIST MDR were treated with mitomycin C (0.5 μg/mL) at 37 o C for 2 h. After incubation, the mixtures were centrifuged at 7000 × g for 5 min and filtered through a 0.2-μm filter. The collected supernatants were spot tested to confirm the lytic bacteriophages against BSST LAB .

Bacteriophage Adsorption Assay
Bacteriophage adsorption rates were estimated to evaluate the bacteriophage-binding

Quantitative RT-PCR Assay
Total RNA was extracted from BSST W T , BSST CIP , BSST LAB , BSST MDR , BIST W T , BIST CIP , BIST LAB , and BIST MDR according to the protocol of RNeasy Protect Bacteria Mini kit protocol (Qiagen, Hilden, Germany). The pre-cultured cells were mixed with 1 mL of RNA protect Bacteria Reagent to stabilize RNA, and the mixture were centrifuged at 5,000 × g for 10 min. The collected cells were lysed with a lysozyme-containing buffer TE (10 mM Tris·Cl, 1 mM EDTA, pH 8.0). The lysate cells were mixed with 95% ethanol to extract RNA through an RNeasy mini column. According to the QuantiTech reverse transcription procedure (Qiagen), cDNA was synthesized. Briefly, the RNA extracts were rinsed with a Wipe buffer to remove genomic DNA and mixed with a master mixture containing reverse transcriptase, RT buffer, and RT primer mix. The mixture was incubated at 42°C for 15 min followed by 95°C for 3 min. For amplification, the PCR mixture (20 μl) containing 10 μl of 2× QuantiTect SYBR Green PCR Master, 2 μl of each primer, and 2 μl of cDNA, and 4 μl of RNase-free water was denatured at 95°C for 30 sec, followed by 45 cycles of 95°C for 5 sec, 55°C for 20 sec, and 72°C for 15 sec using an QuantStudio™ 3 Real-Time PCR System (Applied Biosystems™, USA). The synthesized oligonucleotide primers used in this study are listed in Table 1. The relative gene expression levels were determined using the comparative method [21].

Statistical Analysis
All analyses were performed in duplicate on three replicates. Data were analyzed using Statistical Analysis System (SAS). The general linear model (GLM) and least significant difference (LSD) procedures were used to determine significant mean differences among treatments at P < 0.05.

Variability in Inherited and Induced Mutation in Salmonella Typhimurium
BSST MDR and BIST MDR at PBST-13, and PBST-35 had similarities of 99% and 95%, respectively. No similarities were observed between BSST MDR and BIST MDR at PBST-10 and PBST-32.

Mutants
The lysogen induction assay were performed to evaluate whether the bacteriophage resistance was owing to BIST W T , BIST CIP , BIST LAB , and BIST MDR lysogenic cells (Figure S4).
BIST W T and BIST LAB treated with mitomycin C exhibited lytic growth, indicating that the P22-resistant BIST W T and BIST LAB strains have inducible prophages. However, BIST CIP and BIST MDR treated with mitomycin C did not show phage plaque, suggesting that the P22 was not lysogenized in the BIST CIP and BIST MDR ( Figure S4).

Mutants
The antibiotic susceptibilities of BSST W T , BSST CIP , BSST LAB , and BSST MDR were evaluated using disk diffusion assay and compared with those of BIST W T , BIST CIP , BIST LAB , and BSST MDR , respectively (Figure 3). No noticeable changes in antibiotic susceptibilities were observed between bacteriophage-sensitive and bacteriophage-insensitive strains ( Figure 3A). The susceptibilities of BIST CIP , BIST LAB , and BIST MDR were significantly increased to ciprofloxacin (Figure 3B), ampicillin (Figure 3C), and tetracycline ( Figure   3D), respectively, compared to BSST CIP , BSST LAB , and BSST MDR .

Gene Expression in Bacteriophage-insensitive Salmonella Typhimurium Mutants
The relative gene expression of btuB, fhuA, fliK, fljB, ompC, ompF, rfaL, seiA, stn, and tolC were observed in BIST W T , BIST CIP , BIST LAB , and BIST MDR compared to BSST W T , BSST CIP , BSST LAB , and BSST MDR ( Figure 4 ). Most genes were slightly overexpressed in BIST W T and BIST LAB , while the relative expression levels of most genes were significantly decreased in BIST CIP and BIST MDR . The sieA was highly overexpressed in BIST W T by 11-fold and BIST LAB by 18-fold, while the relative expression levels of stn was decreased in the BIST CIP by 7fold and BIST MDR by 4-fold.

Discussion
The application of bacteriophages for combatting antibiotic-resistant bacteria is often impeded by bacteriophage resistance [22]. However, bacteria under selection pressure can lead to a trade-off between bacteriophage-resistance and antibiotic resistance [23].
Therefore, this study describes the association between bacteriophage-resistance and antibiotic resistance in S. Typhimurium with different levels of antibiotic resistance, which needs to design effective bacteriophage-based therapy to control antibiotic-resistant bacteria.
P22 was able to most effectively lyse ST W T , ST CIP , and ST LAB ( Figure 1). The efficacy of lytic activity of bacteriophages depend on the specific recognition between bacterial cell surface receptors and receptor-binding proteins of bacteriophages [8]. As shown in Figure   2, the large variation from group A and small variation from group B imply that mutation was induced before bacteriophage infection [19]. The mutation occurred spontaneously before exposure to selection pressure. The alteration in bacteriophage-binding receptors on the host cells can lead bacteriophage-insensitive mutants [24]. The resistance of 13 BIST LAB to P22 was also due to the lysogenic conversion, which can lead superinfection exclusion [25]. This was confirmed by the lysogenic induction assay, showing that prophages were induced from the BIST W T and BIST LAB after mitomycin C treatment ( Figure   S4).
Numerous Salmonella-specific bacteriophages that use LPS as a receptor can modify LPS to protect from superinfection when host cells are lysogenied [26,27]. Recently, two copies of lipopolysaccharide modification acyltransferase and GtrA are found on the genome of lysogenic P22-like bacteriophage [28]. Although the modification of LPS protects the lysogeny from superinfection by LPS targeting bacteriophage, the lysogeny is still susceptible to bacteriophages that target other receptors such as flagella [28] ( Figure S3). The multiple resistance of BIST CIP might be attributed to the mutation in genes encoding bacteriophage-binding receptors [24]. The loss of bacteriophage-binding receptors is directly associated with the decrease in lytic ability of bacteriophages [8].
Typhimurium strains were well corresponded to the lytic activities with the exception of BSST W T and BIST W T against P22-B1, BSST W T and BIST W T against PBST-13, BSST LAB and BIST LAB against P22-1, and BSST CIP and BIST CIP against PBST-12 ( Figure S3). The results 14 might be due to the alteration of bacteriophage-binding receptors, resulting in the change in binding affinity. The bacteriophage-insensitivity and antibiotic resistance contributed to the alteration of bacteriophage-binding receptors on the host cells, which may ultimately result in the decrease in bacteriophage adsorption rate and lytic activity [29]. However, a high adsorption rate does not always linked to a lytic activity [30]. The bacteriophage adsorption rate to the host cells having various receptors are not much affected by their alteration [31]. The antibiotic susceptibility patterns of bacteriophage-insensitive mutants are more likely to increase in the BIST CIP , BIST LAB , and BIST MDR ( Figure 3 ). The bacteriophage resistant mutants produce an evolutionary trade-off in antibiotic-resistant bacteria, which can change the phage binding receptors and efflux pump system, resulting in increased susceptibility to several classes of antibiotic [23].
The decrease in relative expression of btuB, fhuA, fliK, fljB, ompC, ompF, rfaL, and tolC was well corresponded to the low adsorption rates of bacteriophages to BIST CIP and BSST MDR ( Figure 4 ). Bacteriophage tail proteins bind to the host surface proteins, polysaccharides, and lipopolysaccharides, responsible for host specificity and range [32].
BtuB (vitamin B12 transporter), FhuA (Ferrichrome outer membrane transporter), FliK (flagella hook), OmpC (outer membrane protein), OmpF (outer membrane protein), RfaL (O-antigen ligase), and TolC (innate efflux pump) can serve as surface receptors for bacteriophages [28,29,32]. The OmpC, OmpF, and TolC contribute to multidrugresistance in bacteria. This confirms the antibiotic susceptibilities were increased against BIST CIP , BIST LAB , and BIST MDR when compared to the BSST CIP , BSST LAB , and BSST MDR (Figure 3 ). Bacteria can adapt to the selective pressure imposed by bacteriophages, leading to antibiotic resistance. This suggests that the bacterial resistance to bacteriophages is related to the antibiotic resistance [33]. The overexpression of sieA in BIST W T and BIST LAB ( Figure 4) implies that these mutants might be due to the lysogenic conversion, preventing the entry of superinfecting bacteriophage DNA into the host [34].
The product of sieA is an inner membrane protein of P22-lysogenic Salmonella Typhimurium [14]. The suppression of stn in the BIST CIP and BIST MDR is in a good agreement with the previous observations that the bacteriophages-insensitive mutants exhibited the reduced virulence in antibiotic-resistant bacteria [35].

Conclusions
The most significant findings in this study were that (i) P22-induced BIST CIP mutant had multiple resistance to P22, P22-B1, PBST-32, and PBST-35, (ii) the superinfection exclusion occurred at P22-induced BIST W T and BIST LAB , (iii) the adsorption rates were varied between BS and BI S. Typhimurium strains, (iv) antibiotic susceptibilities were increased in the BIST CIP , BIST LAB , and BIST MDR , and (v) the virulence was reduced in the BIST CIP and BIST MDR . The results pointed out that the bacteriophage-binding receptors were altered in the BI mutant S. Typhimurium strains, which did not induce cross-resistance to antibiotics.
The results provide useful information for designing effective treatments in bacteriophage alone or combination of bacteriophages and antibiotics that reduce the risk of antibiotic resistance in bacteria. However, further study is needed to understand the association between bacteriophage insensitivity and antibiotic resistance.