Molecular Typing of Campylobacter jejuni and Campylobacter coli Isolated from Various Retail Meats by MLST and PFGE

Campylobacter species are one of the leading causes of foodborne disease in the United States. Campylobacter jejuni and Campylobacter coli are the two main species of concern to human health and cause approximately 95% of human infections. Molecular typing methods, such as pulsed-field gel electrophoresis (PFGE) and multilocus sequence typing (MLST) are often used to source track foodborne bacterial pathogens. The aim of the present study was to compare PFGE and MLST in typing strains of C. jejuni and C. coli that were isolated from different Oklahoma retail meat sources. A total of 47 Campylobacter isolates (28 C. jejuni and 19 C. coli) isolated from various retail meat samples (beef, beef livers, pork, chicken, turkey, chicken livers, and chicken gizzards) were subjected to pulsed-field gel electrophoresis (PFGE) and multilocus sequence typing (MLST). PFGE was able to group the 47 Campylobacter isolates into two major clusters (one for C. jejuni and one for C. coli) but failed to differentiate the isolates according to their source. MLST revealed 21 different sequence types (STs) that belonged to eight different clonal complexes. Twelve of the screened Campylobacter isolates (8 C. jejuni and 4 C. coli) did not show any defined STs. All the defined STs of C. coli isolates belonged to ST-828 complex. The majority of C. jejuni isolates belonged to ST-353, ST-607, ST-52, ST-61, and ST-21 complexes. It is worthy to mention that, while the majority of Campylobacter isolates in this study showed STs that are commonly associated with human infections along with other sources, most of the STs from chicken livers were solely reported in human cases. In conclusion, retail meat Campylobacter isolates tested in this study particularly those from chicken livers showed relatedness to STs commonly associated with humans. Molecular typing, particularly MLST, proved to be a helpful tool in suggesting this relatedness to Campylobacter human isolates.

have previously been determined [14,20] and are available on the MLST website [21]. The MLST website also carries information on several different sequence types and is used to share this information with scientists worldwide. Alternate schemes that do not use all the same genes are also available [22,23].
The objective of this study was to determine the genetic relatedness among 47 strains of C. jejuni and C. coli, isolated from different retail meat sources and to determine if one typing method is superior to the other one in determining such relatedness.

Bacterial Isolates
Forty-seven C. jejuni and C. coli, previously isolated from retail meat samples, were used in this study for MLST and PFGE [24,25]. The isolates were selected to represent both species (C. jejuni and C. coli), several meat brands, as well as different retail meat sources, such as chicken (breast and thighs), turkey (breast, thighs, neck pieces, and ground), beef livers, pork (tongue), chicken livers, and chicken gizzards ( Table 1). All isolates were kept frozen at −80 °C in Brucella broth (Becton Dickinson, Sparks, MD, USA) with 20% glycerol.

Pulsed-Field Gel Electrophoresis Typing
The isolates were typed by PFGE following the PulseNet protocol for Campylobacter [26]. Briefly, isolates were grown on MH agar with 5% laked-horse blood and then diluted to the required concentration and agarose-embedded plugs were made and washed. They were then digested with SmaI restriction enzyme (Promega, Madison, WI, USA). The digested plugs were run in Seakem agarose gel (Lonza, Allendale, NJ, USA) with 0.5× Tris-Borate EDTA (TBE) buffer (Amresco, Solon, OH, USA) to separate the bands on the CHEF Mapper PFGE system (Bio-Rad) by running for 16 h at 14 °C switching directions every 6.76 s and ending with 35.38 s (25). Salmonella enterica serovar Braenderup digested with XbaI (Promega, Madison, WI, USA) was used as the molecular reference marker. Gels were stained with ethidium bromide and viewed and recorded under UV transillumination (UVP, Upland, CA, USA). Gel images were analyzed using BioNumerics software (Applied Maths, Austin, TX, USA). The banding patterns were clustered using Dice coefficients using unweighted pair group method, with arithmetic mean (UPGMA), and a 3% band tolerance.
The selected isolates were tested for the presence of the seven different housekeeping genes used in the MLST scheme for C. jejuni by PCR reactions. The primers used were available at the MLST website [29,30] and are shown in Table 2 Once the cycles were complete, reactions were held at 4 °C until gel electrophoresis. Ten microliters of PCR product was subjected to horizontal electrophoresis in a 1% agarose gel in 1× Tris-acetate-EDTA (TAE) buffer. A 1 kb plus ladder (Bioneer, Alameda, CA, USA) was used as the molecular marker. Gels were viewed and recorded by ultraviolet transillumination, using a UV imager (UVP). Sterile water was used as the negative control.
The PCR products were purified using ExoSAP-IT enzyme (Affymetrix, Santa Clara, CA, USA). The sequencing PCR reaction was prepared according to a modified ABI 3130xl manufacturer′s sequencing protocol (Applied Biosystems, Foster City, CA, USA). Briefly, sequencing reactions were prepared to a 15 µL volume containing 3.5 µL purified PCR product, 1.5 μL primer, 0.5 μL sequencing buffer, 2 μL betaine, 0.5 μL BigDye, and 2 μL RNase-free water. The cycling conditions were set up according to the ABI capillary sequencer instructions (Applied Biosystems, Foster City, CA, USA). The sequenced products were then read using the ABI 3130xl (Applied Biosystems) and analyzed using BioNumerics software (Applied Maths, Austin, TX, USA), which has a function to determine STs and clonal complexes by directly submitting the sequences to the MLST website. Table 2. Polymerase chain reaction (PCR) and sequencing primer sets for the multilocus sequence typing (MLST) scheme for C. jejuni and C. coli.

Pulsed-Field Gel Electrophoresis
The results of the PFGE showed that the isolates studied were separated into two major groups according to their species (C. jejuni and C. coli) (Figure 1). The isolates were also able to be separated by their ST-complexes within the species groups. PFGE was able to group the 47 isolates into 2 major clusters (one for C. jejuni and one for C. coli) but wasn't able to differentiate the isolates by meat source within the species (Figure 1). By PFGE separation, the isolates were found to also cluster into their ST-complexes, such as for ST-61, ST-21, ST-52 for C. jejuni, and ST-828 for C. coli (Figure 1). Among the C. jejuni isolates, ST-61 was related to beef liver isolates and ST-21 and ST-52 related to chicken sources. ST-607 isolates did not all cluster closely together but they all belonged to poultry sources and to the same ST (1212). ST-353 isolates also did not cluster together, but they all belonged to poultry sources (Figure 1). When using PFGE for genotyping, genetic variation among Campylobacter strains becomes a concern [12]. Some strains are not typable, using either of the commonly used restriction enzymes SmaI or KpnI, which creates questions about the usefulness of PFGE with Campylobacter species [11,13]. In our study PFGE was able to group the 47 Campylobacter isolates into two major clusters (one for C. jejuni and one for C. coli) but failed to differentiate the isolates according to their source.

Isolate
Source Species ST ST Complex

Multilocus Sequence Typing
MLST was able to separate the isolates into 21 different STs, which belonged to eight different clonal complexes (Figure 1). Twelve of the isolates (eight C. jejuni and four C. coli) were not assigned STs (Figure 1). All the defined STs of C. coli isolates belonged to ST-828 complex. The majority of the C. jejuni isolates grouped into ST-353, ST-607, ST-52, and ST-61 clonal complexes (Figure 1). The most common clonal complex observed in this study was the ST-828 complex, which consists mostly of C. coli isolates. Other clonal complexes identified were ST-21, ST-61, ST-52, ST-206, ST-353, ST-443, and ST-607. Cluster analysis by Minimum Spanning Tree created using the BioNumerics software also shows that the isolates clustered into two distinct groups according to their species, which are C. jejuni and C. coli (Figure 2). Campylobacter jejuni appears to be more diverse than Campylobacter coli in regards to their STs and clonal complexes distributions (Figure 2).  The isolates were separated according to the species of Campylobacter they belong to and, within those groups, there is a separation into ST-complexes. The unidentified isolates were also added and show affiliation with one group or the other. The most common clonal complex was ST-828 consisting of C. coli isolates, which has been previously observed [11,17,27,31,32]. ST-828 was also reported as the most common complex among C. coli isolates by other studies [11,27,[31][32][33][34]. This was also the case with the C. coli isolates in this study for all of our defined isolates. Most human infections are caused by complexes ST-828 and ST-1150 [17]. In fact, the ST-828 complex is commonly associated with human isolates, as well as chicken meat or offal, according to data from the MLST website [17]. ST-21 has also been reported to be the most commonly detected complex among the isolates that were published on the MLST website with the most isolates from human origins [17], and the next most common belong to chicken isolates [17,35,36]. ST-21 has also been found in bovine and ovine isolates [37,38]. Chicken is also the source of ST-52, ST-61, ST-206, ST-353, and ST-828 in other studies [5]. In our study, all of these ST complexes belonged to various poultry isolates. ST-353 and ST-21 were also previously reported among C. jejuni isolates [5,34,37,38]. Colles et al., 2003 [39], reported in their study that the ST-61 complex was associated with sheep isolates. Data collected from the MLST website in 2012 by Colles and Maiden [17], found that ST-61 was most commonly associated with human isolates and the next most common source was beef offal or meat. ST-206 was also most associated with human isolates in that study [17].
The fact that all the C. coli isolates belong to the same complex could be due to the more conservative nature of the C. coli genome and the C. jejuni genome being more variable [27,40]. Colles et al. [39], and Manning et al. [22], found that there was no association of the STs with the host, inferring that this could be due to the lack of diversity among C. coli or possibly due to their sample not being very diverse. Miller et al. [31], reported that, in their MLST studies, there was some association among C. coli and their STs to specific hosts suggesting that source tracking would be possible with C. coli.
Most of the STs identified in this study were found to be associated with human and other sources of infection. Most of the STs found in the chicken livers were STs associated with human infection only according to the MLST website [29]. Adding to the importance of chicken livers as a public health risk is the recent discovery by Strachan et al. [41], that molecular source attribution by MLST demonstrated that Campylobacter strains from chicken livers were most similar to those found commonly in humans, which provides further evidence that chicken liver is a probable source of human infection.

Conclusions
In conclusion, retail meat Campylobacter isolates tested in this study particularly those from chicken livers showed relatedness to STs commonly associated with humans. Molecular typing particularly MLST proved to be a helpful tool in suggesting this relatedness to Campylobacter human isolates and can be regarded as superior to PFGE in this regard. The genetic variation among C. jejuni strains appeared higher than that among C. coli strains using MLST in our study.