Antibiotic, particularly penicillin, was greatly appreciated at the time of World War II. Over many long years, the overuse of antibiotics has led to the rapid emergence of resistance in bacteria1). In particular, antibiotic-resistant Salmonella spp. can cause severe life-threatening illness and become serious health problems2). The prevalence of antibiotic resistance can arise due to the activation in efflux mechanism, alteration of cell membrane permeability, and enzymatic inactivation of antibiotics3,4). Hence, new therapeutic methods are required to control the antibiotic resistance. Recently, bacteriophages (phages) have considered as an alternative therapeutic option over current antibiotics5).
Virulent lytic phages can be used for the control and detection of bacterial pathogens due to the high specificity of phage-bacterium interactions6,7). Phage specificity predominantly depends on the phage-host binding efficacy, which is directly associated with the lytic activity of phages8,9). The bacterial surface receptors and phage tail fiber affect the host ranges of phages. Phage-binding receptors on the surface of bacteria include outer membrane proteins, flagella, and lipopolysaccharides10). There is still a challenging question whether phages are applicable to control antibiotic-resistant bacteria. However, relatively few studies have characterized the phage behavior in antibiotic-resistant pathogens. Therefore, the aim of this study was to evaluate the possibility of using phages to control antibiotic-resistant S. Typhimurium.
Materials and Methods
Bacterial strains and culture conditions
Strains of Salmonella enterica subsp. enterica serovar Typhimurium ATCC 19585 (STWT), S. Typhimurium KCCM 40253 (STKCCM) and S. Typhimurium CCARM 8009 (STCCARM) were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA), Korean Culture Center of Microorganism (KCCM, Seoul, Korea), and Culture Collection of Antibiotic Resistant Microbes (CCARM, Seoul, Korea), respectively. Ciprofloxacininduced S. Typhimurium ATCC 19585 strains (STCIP) was obtained by a serial passage method11) from 0 to 2 μg/mL of ciprofloxacin in trypticase soy broth (TSB; Difco, Becton, Dickinson and Co., Sparks, MD, USA). All S. Typhimurium strains were cultured in TSB with aerobic condition at 37°C for 20 h and harvested by a centrifugation at 3,000 × g for 20 min at 4°C. The collected cells were washed twice with phosphate-buffered saline (PBS, pH 7.2).
Bacteriophage propagation
Salmonella phages, P-22, P-22 B1, PBST-10, PBST-13, PBST-32, and PBST-35, were purchased from ATCC and Bacteriophage Bank (PB) at Hankuk University of Foreign Studies (Yongin, Gyeonggi, Korea). All phages were propagated in TSB with the host strain STKCCM at 37°C for 20 h. The proliferated phages were harvested by centrifugation at 5,000 × g for 10 min then filtered using 0.2 μm pore-size disposable syringe filters. The phage stock solution was determined using the agar overlay assay and stored at 4°C. In brief, the serially diluted phages (1:10) were gently mixed to avoid air bubble with the host strain (108 CFU/ml) in TSB12). The mixtures were poured onto the pre-heated TSA plates and incubated at 37°C for 20 h to count the phage plaques which were estimated as plaque-forming unit (PFU).
Determination of phage host range
The host ranges of phages were analyzed using a phage spot test with minor modifications13). In brief, the bacterial strains used in this study were STWT, STKCCM, STCCARM, STCIP, Klebsiella pneumoniae ATCC 23357, K. pneumoniae KCCM 11257, K. pneumoniae CCARM 10237, ciprofloxacin-induced K. pneumoniae ATCC 23357, Staphylococcus aureus ATCC 15564, S. aureus KCCM 13236, S. aureus CCARM 3080, ciprofloxacin-induced S. aureus ATCC 15564, and oxacillininduced S. aureus ATCC 15564. Soft-agar was overlaid with 200 μL of each strain on pre-solidified trypticase soy agar (TSA; Difco, Becton, Dickinson and Co.) plate and then 2 μL of phages (108 PFU/mL) were spotted and incubated at 37°C for 20 h. Proper hosts were determined by the formation of clear zone, showing the lytic capability of phages.
Efficacy of phage adsorption
The adsorption assay was determined to evaluate the ability of phage to bind the host cell surface14). Each bacterial strain (105 CFU/mL) was infected with phage (P- 22, P-22 B1, PBST-10, PBST-13, PBST-32, or PBST-35,) at multiplicity of infection (MOI) of 1 and allowed to adsorb at 37°C for 20 min. After incubation, bacterial cells were centrifuged at 13,000 × g for 5 min and the titers of adsorbed phages were determined using a soft-agar overlay assay.
One-step growth curve analysis
The one-step growth curves were used to estimate the phage burst size against antibiotic-susceptible and antibioticresistant host strains. The phages infected to bacterial cells (104 CFU/mL) at MOI of 1 for 20 min were centrifuged at 13,000 × g for 5 min to remove unabsorbed free phages. The collected phage-absorbed cell pellets were resuspended in TSB and incubated at 37°C for 30 min. The phage titers were estimated at every 5 min for 30 min by using the agaroverlay assay. The phage burst size was estimated by the ratio of the total number of phages released to the initial number of infected bacterial cells15).
Determination of lytic activity
To evaluate the lytic activity of phages13), the bacterial cells (104 CFU/mL) were infected with phages at MOI of 1 and incubated at 37°C for 12 h. The bacterial cells were centrifuged at 3,000 × g for 20 min. The collected bacterial cells were serially diluted (1:10) with phosphate buffer saline (PBS, pH7.2) and plated on TSA plate using an Autoplate® Spiral Plating System (Spiral Biotech Inc., USA). After incubation at 37°C for 24 h, the viable cells were enumerated using a QCount® Colony Counter (Spiral Biotech Inc., USA).
Statistical analysis
All experiments were conducted with three replicates. Data were analyzed within the Statistical Analysis System software 9.4 (SAS, Institute Inc., Cary, NC, USA) using the general linear model (GLM) and least significant difference (LSD) procedures. The significant mean differences between treatments were determined at P<0.05.
Results and Discussion
This study describes the interactions between phage and antibiotic-resistant S. Typhimurium in association with adsorption rate, one-step growth curve, and lytic activity. It is worth understanding the phage-host interactions in order to develop effective control and detection systems for antibiotic-resistant pathogens.
Host range and adsorption rate of Salmonella phages
The host ranges of Salmonella phages (PBST-10, PBST- 13, PBST-32, PBST-35, P-22, and P-22 B1) were determined by using phage spot assay against STWT, STKCCM, STCCARM, STCIP, KPWT, KPKCCM, KPCCARM, KPCIP, SAWT, SAKCCM, SACCARM, SACIP, and SAOXA. The lytic spectrum of phages varied against S. Typhimurium strains (data not shown). All phages have a narrow host range, which only lysed all Salmonella strains excepting STCCARM. The result suggests that the phage-binding receptors were altered in multidrug-resistant STCCARM. In addition, no intraspecies and interspecies infections were observed for Salmonella phages against K. pneumoniae and S. aureus strains. In general, phages can infect bacterial pathogens at close range16). The adsorption of phages (PBST-10, PBST-13, PBST-32, PBST- 35, P-22, and P-22 B1) to the surfaces of host cells (STWT, STKCCM, STCCARM, and STCIP) were determined after 20 min of infection (Fig. 1). The phage adsorption to host cells is the initial step of phage infection process8). The highest adsorption rates of P-22 (85%), PBST-35 (80%), PBST-10 (75%), and PBST-32 (71%) were observed for STWT, while PBST-13 and P-22 B1 showed the lowest adsorption rates, respectively, 47% and 54%. The adsorption rates of PBST- 10, P-22, PBST-35, PBST-32, and PBST-13 were 95%, 80%, 77%, 75%, and 73%, respectively, to STKCCM. In case of STCIP, the highest adsorption rate was more than 90% when infected with PBST-32 and PBST-35, followed by PBST-10 (80%), whereas the least adsorption rates of PBST-13, P-22, and P-22 B1 were 62%, 61%, and 59%, respectively, after 20 min of infection. The increased adsorption rates of PBST- 32, PBST-35, and PBST-10 to STCIP (Fig. 1) are in a good agreement with previous report that ciprofloxacin enhanced the phage adsorption of Φ13 to S. aureus17). However, STCCARM showed the least adsorption rates when infected with all phages (<40%). STCCARM, a clinically isolate, was highly resistance to β-lactam antibiotics (ampicillin and penicillin G) and aminoglycoside antibiotics (streptomycin and kanamycin)3). Therefore, the alteration in the surface receptors of STCCARM resulted in the decrease in phage adsorption18). The phage adsorption is directly associated with bacterial surface receptors such as lipopolysaccharide, pili, flagella, and membrane porin19). Our data suggests that the alteration of bacterial membrane proteins and phagebinding receptors on the host cell surface were responsible for the reduction in the phage adsorption and binding specificity to STCCARM9).
One-step growth curve and lytic activity of Salmonella phages
The one-step growth curves of PBST-10, PBST-13, PBST- 32, PBST-35, P-22 and P-22 B1 were used to estimate the phage latent times and burst sizes against STWT, STKCCM, STCCARM, and STCIP (Fig. 2). The burst sizes varied from 24 to 1,156 PFU. The burst sizes of PBST-10, PBST-13, PBST- 32, PBST-35, P-22 and P-22 B1 were 63, 63, 350, 267, 347, and 43 PFU against STWT, 158, 1154, 400, 37, 271, and 530 PFU for STKCCM, 167, 335, 1067, 24, 350, and 500 PFU for STCCARM, and 500, 1000, 100, 66, 1000, and 309 PFU for STCIP, respectively, at MOI of 1 (Fig. 2). PBST-13, PBST- 32, and P-22 showed the highest burst sizes against STKCCM (1153 PFU), and STCCARM (1067 PFU), and STCIP (1000 PFU), respectively (Fig. 2B, 2C, and 2E). The short latent time with high burst sizes is the best criterion for a potential phage therapeutic application20). The lytic activity of PBST- 10, PBST-13, PBST-32, PBST-35, P-22, and P-22 B1 was evaluated against STWT, STKCCM, STCCARM, and STCIP, at the MOI of 1 (Fig. 3). The phage lytic patterns varied depending on bacteria stains. The highest lytic activity was observed for PBST-35 against STWT, followed by PBST-10 and PBST- 13 (Fig. 3A). The P-22 showed the highest lytic activity against STKCCM, while other phages such as PBST-10, PBST- 13, PBST-32, PBST-35, and P-22 B1 (23 ± 0.01%) showed the least lytic activities (p<0.05) (Fig. 3B). No significance in the lytic activity of P-22 and P-22 B1 was observed against STCCARM when compared to the control (P>0.05) (Fig. 3C). The growth of STCIP was well inhibited by all phages at the early of incubation period, excepting P-22 B1 (Fig. 3D). The phage lysis pattern and resistance range depend on the host type and phage species21). The low lytic efficacy of most phages against STCCARM may be attributed to the low binding affinity due to the alteration in phagebinding receptors22). Phage-host interaction is highly associated with the presence of phage-binding proteins and receptors23). Therefore, the phage resistance may occur as a result of host cell surface receptor modification that leads to reduce the lytic efficacy. The phage resistance of the host bacteria occurs due to the inhibition of phage adsorption and entry24,25). Further understanding the phage resistance mechanisms is essential to improve lytic activity of phages.
In conclusion, the most significant finding was that the phage behaviors of PBST-10, PBST-13, PBST-32, PBST-35, P-22, and P-22 B1 were varied depending on the degree of antibiotic resistance in S. Typhimurium, STWT, STKCCM, STCIP, and STCCARM. The highest adsorption rates were observed for all phages to STWT, STKCCM, and STCIP, except STCCARM. The phages burst sizes and lytic activities depended on the alteration of phage-binding surface receptors. This study provides useful information for designing phage control system.