Molecular characterization of antimicrobial resistant non-typhoidal Salmonella from poultry industries in Korea

Background Antimicrobial resistant Salmonella strains are a direct threat to human health when this resistance interferes with treatment and an indirect threat when resistance can be transferred to other human pathogens. The objective of the present study was to characterize antimicrobial resistant non-typhoidal Salmonella (NTS) isolates recovered from poultry industries, including a description of genetic diversity and virulence profiles. Results In total of 93 Salmonella isolates shown antimicrobial resistance to one or more drugs, all isolates exhibited common resistance to streptomycin, nalidixic acid and cephalothin but no ciprofloxacin resistance. Among 26 virulence gene profiling, 12 virulence genes, invA, orgA, prgH, sopB, tolC, sipB, gatC, msgA, pagC, spiA, sifA, and sitC were found in all antimicrobial-resistant NTS isolates. In comparing the data from ERIC-PCR clusters, virulence profiles and resistance profiles, some Salmonella isolates grouped into the same cluster were found to exhibit similar virulence and resistance patterns. Conclusions Virulence profiling combined with ERIC-PCR offered a rapid approach to characterize antimicrobial-resistant NTS.

and iron uptake [8]. For example, genes such as invA, orgA, prgH, sipB and spaN in SPI-1 encode a type 3 secretion system 1 (T3SS-1) which allows Salmonella to invade phagocytic and non-phagocytic cells. Genes such as spiA in SPI-2 encode a type 3 secretion system 2 (T3SS-2), which allows the survival and replication of Salmonella in host cells [27].
Other chromosomal genes such as lpfA and pefA encode fimbriae-associated proteins that are important for adherence [11]. Moreover, plasmidal genes such as spvB contribute to colonization of deeper tissues, among other functions [2]. The virulence potential of Salmonella determines the differences in pathogenicity among Salmonella serotypes.
The increase in antibiotic-resistant food-borne pathogens is also a major public health problem. A high rate of antimicrobial-resistance in Salmonella strains has been reported in Korea [5,24]. Antimicrobial resistance and virulence of Salmonella strains play an important role in systemic infections with these pathogens [14].
Pulsed-field gel electrophoresis (PFGE) is the gold standard subtyping method used to assess relatedness among Salmonella strains from different sources [28]; however it is time consuming and labor intensive [16]. DNA-based fingerprinting techniques such as enterobacterial repetitive intergenic consensus (ERIC), repetitive extragenic palindromic (REP), and BOX repeat-based (BOXAIR) PCR methods are relatively easy to perform, rapid, and sensitive in discriminating between closely related strains [12,16]. In recent years, the ERIC-PCR fingerprinting method has been used to confirm epidemiological relationships between various isolates, and this method has shown high discriminatory power [1,32]. The objective of the present study was to characterize antimicrobial-resistant NTS isolates recovered from poultry industries, including a description of genetic diversity and virulence profiles.

Salmonella isolates
Ninty-three Salmonella isolates showing antimicrobial resistance to one or more drugs were tested in this study. All isolates were recovered from chilled chicken carcasses (n = 25) and chillers (n = 23) in chicken slaughter houses; chilled carcasses (n = 15) and chillers (n = 8) in duck slaughter houses; egg belts (n = 2), feeders (n = 1), feces (n = 5), dust (n = 3) and egg shells (n = 8) from layer farms; and raw shell eggs (n = 2) and egg contents (n = 1) from retail markets between 2008 and 2014 ( Table 1). Salmonella isolates were serotyped according to the Kauffmann-White scheme following slide agglutination testing with Salmonella-specific O and H antisera (Difco, Detroit, MI).

Analysis of virulence genes
The DNA for all analyses was extracted by the boiling method [7], and 5 μl of DNA template (approximate 60 ng) was used in each PCR reaction. Primers details are presented in Table 2. To establish the reproducibility of the experiments, PCR reactions were performed twice. The amplified PCR products were visualized by electrophoresis on 1.5% agarose gels stained with ethidium bromide (0.5 μg/ml).

ERIC-PCR fingerprints analysis
Genotyping of isolates was performed by ERIC-PCR using a pair of primers (F: 5'-ATG TAA GCT CCT GGG GAT TCA C-3'; R: 5'-AAG TAA GTG ACT GGG GTG AGC G-3') [36]. The PCR reaction was performed using a lyophilized PCR master mix according to the manufacturer's instructions. (AccuPower PCR PreMix, Bioneer, Korea). A thermocycler (Bio-Rad, Singapore) was programmed as follows: initial denaturation at 95°C for 7 min, followed by 30 cycles of denaturation at 90°C for 30 s, annealing at 52°C for 1 min and extension at 65°C for 8 min, and a final extention step at 65°C for 16 min [39]. A negative control consisting of the same reaction mixture without a DNA template was included in each reaction. ERIC-PCR reactions were repeated at least twice for each isolate to determine the reproducibility of banding patterns. Data were analyzed using the software package BioNumerics 7.5 (Applied Maths, Keistraat, Belgium). A similarity dendrogram was constructed by the UPGMA method with a 1% tolerance limit and 1% optimization, using the DICE similarity coefficient. Clusters were identified based on an 80% similarity cut-off [10]. The discrimination index for ERIC-PCR was calculated using Simpon's diversity index [19].

Results
The results of ERIC-PCR, virulence gene profiling, and antimicrobial susceptibility test are summarized in Fig. 1.
The ERIC-PCR analysis showed that 60 Salmonella isolates grouped into 18 clusters (E1 to E18) at 80% genetic similarity, whereas the remaining 33 isolates remained unclustered. The discrimination index of ERIC-PCR typing in this study was 0.974. Except for five clusters (E1, E3, E11, E13 and E14), isolates within each of the other 13 clusters belonged to the same serotype. Irrespective of their sources and places of isolation, all S. Bareilly isolates grouped into only one cluster (E12), whereas isolates of other serotypes could be found in different clusters, despite isolation from the same source and location. For example, 18 of 23 S. Enteritidis isolates grouped into eight clusters (E3, E4, E5, E7, E8, E10, E11 and E13) and the remaining five isolates of this serotype were not grouped. Two of 8 S. Montevideo isolates clustered into E17, and the remaining six isolates were not found in any cluster. Some of these Salmonella isolates representing the same serotype, sources, and location clustered separately. R-CGTTCAGGTTTTGTCGCTGG Fig. 1 The ERIC-PCR analysis of the non-typhoidal Salmonella from Korean poultry industry is displayed using dendrograms generated by Bionumerics software. The vertical line shows the delineation level of 80%. The red color indicates the presence of the gene while the pink color indicates the absence of the gene. The dark blue color indicates resistance to the corresponding antibiotic while the light blue color indicates susceptibility. The 17 chicken slaughter houses, nine duck slaughter houses, eight commercial layer farms and three retail markets were designated C1~C17, D1~D9, L1~L8, R1~R3, respectively. AM ampicillin, C choloramphenicol, CAZ ceftazidime, CIP ciprofloxacin, CF cephalothin, CTX cefotaxime, GM gentamicin, K kanamycin, NA nalidixic acid, S streptomycin, STX trimethoprim/sulfamethoxazole, TE tetracycline

Discussion
Antimicrobial-resistant Salmonella strains are a direct threat to human health when this resistance interferes with treatment and an indirect threat when resistance can be transferred to other human pathogens [13]. In Korea, S. Enteritidis, S. Montevideo, S. Typhimurium is most common serotypes in poultry slaughter houses [4,38], and the most frequently observed Salmonella serovars in the layer farms were S. Bareilly [20]. Cephems, quinolones and aminoglycosides have been widely used in the poultry industries in Korea [18,24]. Therefore, Salmonella isolates tested in study showed consistent resistance to S, NA, and CF. Our finding showed that 29% among 93 antimicrobial resistance isolates were MDR. All MDR isolates recovered from chicken or duck slaughter houses. However, there was no MDR isolates recovered from layer farms or eggs of retail markets. Generally, Salmonella isolated from the broilers demonstrated greater MDR compared to those isolated from layer and eggs [21].
The ability of antimicrobial resistant NTS strains to cause invasive disease can be attributed to various virulence genes, and virulotyping rapidly allows the discrimination of isolates with diverse pathogenic potential [17]. In this study the genes-invA, orgA, prgH, sopB, tolC, sipB, gatC, msgA, pagC, spiA, sifA, and sitC, located in SPI-1, SPI-2, SPI-5, SPI-11, or others, were found in all antimicrobial-resistant NTS isolates. In addition, spaN, located in SPI-1, and iroN, located in an effector were highly conserved in the isolates. Similar findings were reported in a study of samples from both sick and healthy birds [34], clinical samples, and environmental samples from poultry houses [31], and free-living birds [26].
Recently some researchers have reported that typhoidassociated virulence genes (cdtB, tcfA, and hlyE) in NTS serotypes of human and poultry origin are increasing [11,26,34]. NTS strains containing cdtB, tcfA, and hlyE genes were found in this study, and, to our knowledge, this is the first report to detect the presence of cdtB in the S. Hogton, S. Give, S. Newbrunswick, S. Thomson, S. Kortrijk, S. Coquilhativille, and S. Binza serotypes.
In this study, lpfC, one of three fimbrial genes, was more prevalent than the others, pefA and tcfA, which is consistent with previous studies [8,15]. However, no S. Bareilly isolates tested in this study harbored the lpfC gene; whereas tcfA was presented in all S. Bareilly isolates, and pefA was highly conserved. Gong et al. [15] showed that the presence or absence of specific fimbrial genes in certain Salmonella serovars might have diagnostic value, as fimbrial genotypes can be used to determine certain Salmonella serotypes. Our results of fimbrial gene profiles are consistent with this report.
Molecular typing of Salmonella is vital to determining potential sources of infection and implementing effective epidemiological surveillance and control [9]. In this study, 60 out of 93 antimicrobial-resistant NTS isolates grouped into 18 ERIC-PCR clusters, and 33 isolates remained unclustered. This variability might be due to a difference in the sources of samples or in serotypes. In this study, the discrimination index of ERIC-PCR was 0.974. Based on a recommendation by Hunter et al. [19], a D value > 0.9 is desirable for good differentiation; our ability to discriminate between isolates was high. Based on these criteria, ERIC typing is useful for Salmonella typing, and our report showed that ERIC-PCR differentiated Salmonella strains indistinguishable to levels of heterogeneity of various serotypes.
The distribution of profiles among serotypes demonstrated that different serotypes showed similar fingerprinting patterns. These results are consistent with findings of Ranjbar et al. [33] who found that every Salmonella isolate had a unique fingerprinting but the serotypes were not grouped together in major branches.
In this study, the correlations among ERIC-PCR clusters, virulence profiles and resistance profiles were analyzed. We found that virulence genes and resistance profiles correlated with ERIC-PCR subtypes. Some isolates showed the same or similar virulotype or resistance pattern, irrespective of serotypes. The simultaneous presence of a resident virulence plasmid and resistance gene in the same bacterial cell has been reported in Salmonella [3]. Therefore, assessing the prevalence of virulence genes as well as the antibiotic resistance status in Salmonella serotypes would be useful to better understanding Salmonella pathogenicity.

Conclusion
In conclusion, this study provides a molecular characterization of antimicrobial-resistant NTS from poultry industries in Korea. Virulence profiling combined with ERIC-PCR may offers a rapid approach to characterize antimicrobial-resistant NTS isolates. Therefore, determination of their definitive correlations will require future studies with isolates from various source of animal.