Identification of pathogenic genes in Campylobacter jejuni isolated from broiler carcasses and broiler slaughterhouses

Campylobacter jejuni is one of the most common causes of foodborne diseases worldwide. There are few reports on Campylobacter strains isolated from Latin-American countries. Here, 140 C. jejuni strains isolated from cloacal and transport boxes swabs, water from chiller tanks, and broiler carcasses of five poultry companies in Southern Brazil were identified using phenotypic and genotypic methods. Polymerase chain reaction (PCR) was used to analyze eight C. jejuni virulence markers: flaA, cadF, and invasion-associated (iam) genes, cdtABC operon (associated with the cytolethal distending toxin), and plasmidial virB11 and wlaN genes were present in 78.5%, 77.8%, 0%, 74.2%, 22.1%, and 10.7% of samples, respectively. There were 25 different virulence profiles: 1 (cdtA, cdtB, cdtC, flaA, and cadF), 2 (cdtA, cdtB, cdtC, flaA, cadF, and virB11), and 3 (cdtA, cdtB, cdtC, flaA, cadF, and wlaN) were the most common (> 60% of strains). We provide insight into factors related to the occurrence of this pathogen and their epidemiology.

www.nature.com/scientificreports/ distending toxin (CDT) holotoxin consists of three subunits encoded by the genes cdtA, cdtB, and cdtC, which are genetically arranged as an operon 18,19 . The expression of all three cdt genes is required for maximum toxin activity 20 . The cdtA and cdtC gene products are two subunits responsible for toxin binding to the cell membrane and for the delivery of the cdtB gene product (CdtB), which is the enzymatically active subunit 14 .
Campylobacter exhibits bipolar flagella composed of a hook-basal body complex, which is composed of several different proteins, and the extracellular filament structural components, which are composed of a major flagellin protein, FlaA (coded by flaA), and a minor flagellin protein, FlaB (coded by flaB) 14 . The flaA gene seems to be highly conserved among Campylobacter isolates, and its transcription level is usually higher than that of flaB 11,21 . Thus, flagella are crucial for the attachment to intestinal epithelial cells and are involved not only in motility and chemotaxis, but also in the secretion of virulence proteins, autoagglutination, microcolony formation, and avoidance of the innate immune response 22 . The cadF (Campylobacter adhesion to fibronectin F) gene encodes an adhesion and fibronectin-binding protein involved in the process of invasion, influencing the microfilament organization in host cells 14,23 . Another virulence gene linked to Campylobacter invasiveness is the invasion-associated marker (iam) gene 24 . It has been suggested that the pVir plasmid, which encodes for proteins of a type IV secretion system, and the mutation of the plasmid virB11 gene results in a reduced adherence and invasion potential in vitro, as well as in less severe symptoms in vivo 25 . The wlaN gene is presumably involved in the expression of ganglioside mimics in Guillain-Barré Syndrome and may encode β-1,3-galactosyltransferases with identical enzymatic activities 26,27 .
Considering that Campylobacter may be transferred from animals to humans, and that poultry is its main reservoir 2, 8 , it is important to know the genetic profile of circulating strains in broiler populations and whether all Campylobacter isolates obtained from different companies and sources of isolation have the same virulence potential. In addition, many aspects concerning epidemiology, public health significance, and risk assessment studies are recent and still poorly known in Brazil, and there are no Brazilian legal standards established regarding Campylobacter presence or counts in food products. The purpose of this study was to determine whether eight genes previously identified as coding pathogenic markers in Campylobacter were present in the 140 samples obtained from several broiler sources and collected from five poultry companies using polymerase chain reaction (PCR).

Results
Out of a total of 140 screened samples, all of them were confirmed as having Campylobacter jejuni species. This method is based on the detection of 16S rRNA and mapA genes. The occurrence of the virulence factor genes flaA, cadF, iam, virB11, wlaN, and the gene cluster cdtABC was analyzed, and the virulence profiles that were observed are described in Table 1. Distribution of C. jejuni strains has shown 25 different virulence profiles. Profiles 1 (cdtA, cdtB, cdtC, flaA, and cadF), 2 (cdtA, cdtB, cdtC, flaA, cadF, and virB11), and 3 (cdtA, cdtB, cdtC, flaA, cadF, and wlaN) were the most common, representing 30% (42/140), 20% (28/140), and 10.7% (15/140) of the strains, respectively. Together, these three profiles represented more than 60% of the strains. The distribution of profiles according to the poultry company is described in Table 1. For companies A, D, and E, the most common profile was 1, corresponding to 36%, 37.5%, and 80% of the strains, respectively. For company B, profile 2 was the most common (33%), and for company C, profiles 3 (26.4%) and 1 (24.5%) were the most common.
The results showed that only one strain did not present any virulence markers. A total of 61.4% (86/140) of C. jejuni strains possessed five or six virulence determinants, 24.3% (34/140) presented three or four virulenceassociated genes and the presence of one or two genes was detected in 13.6% (19/140) of isolates. No strain had more than six virulence markers.
The PCR detection results of eight virulence-associated genes are described in Tables 2 and 3, according to the source and company from which they were isolated. The comparison among isolation sources has shown no significant differences, regardless the virulence gene (Table 2). Significant differences in gene frequency among the isolation companies were observed for virB11 between companies B and C (p = 0.0013); in addition, cdtB were significantly less common in company B (p = 0.002) ( Table 3).

Discussion
The purpose of this study was to determine the presence of C. jejuni virulence markers as indicators of the potential role of these strains in the pathogenesis of human diseases. An understanding of the epidemiology of all foodborne zoonotic agents is essential for the implementation of control strategies and interventions 28 . Phylogenetic studies have demonstrated a large heterogeneity among isolates of the same species, with the absence of clones, indicating a high level of diversity of circulating genotypes 29 . Previous studies have demonstrated the increased pathogenic potential of Campylobacter jejuni isolates from Brazilian poultry during the last decade 30 . In the present study, the identification and frequency of this pathogen throughout the slaughter line revealed potential virulence-associated markers, emphasizing the zoonotic risk of avian-derived Campylobacter strains.
Similar to previous reports 31 , the flaA gene presented a high frequency (78.6%) in this study. The bacterial flagellum and the virulence-associated injectisome are complex and structurally related to nanomachines that bacteria use for locomotion or the translocation of virulence factors into eukaryotic host cells 32 . It has been suggested that Campylobacter can survive under strong acid-shock conditions, and this is linked with the increased transcription of a subset of flagellar biosynthetic genes and stress responses, as well as the downregulation of www.nature.com/scientificreports/ genes involved in cell division and metabolism 33 . The flagellum not only has a distinct function in bacterial motility and cell binding, but also acts as a type III secretion system (T3SS) 9,34 . Flagellin is a potent activator of a broad range of host cell types involved in innate and adaptive immunity 35 . Rizal et al. 17 examined the presence of the flaA factor in C. jejuni and C. coli derived from chicken and human isolates and obtained a 100% prevalence.
Other authors have examined the presence of factors in C. jejuni and C. coli derived from humans, poultry meat, broiler, and bovine feces and obtained similar results 24 . Datta et al. 27 also determined this factor in a group of 111 C. jejuni strains isolated from different sources and found that all of them were positive. Previous reports in Brazil have indicated that flaA is present in approximately 80% of Campylobacter spp. isolates from human and poultry samples 30,36 .     Plucking (n = 9) 100 (9/9) a 100 (9/9) a 100 (9/9) a 100 (9/9) a 100 (9/9) a 100 (9/9) a 0 33. www.nature.com/scientificreports/ Another virulence gene examined in this study was cadF, one of the markers determining the adherence of Campylobacter. This gene was present in 77.9% of the isolates analyzed, similarly to previously reported results [37][38][39] . Previous reports in Japan, Australia, India, Poland, and Brazil have indicated that cadF is present in almost all C. jejuni isolates from poultry samples 17,24,27,30,35 . This gene encodes an adhesion factor used by C. jejuni to attach and, eventually, invade mammalian cells by binding to fibronectin, a component of the extracellular matrix 34,35 . Investigations reported that ΔcadF mutant strains were not able to colonize the gut of chickens 40 . Monteville et al. 23 verified that the adhesion and transmigration of ΔcadF mutant strain to human INT-407 cells was reduced by 50%, when compared to a wild-type strain. Thus, the cadF gene, which appears to be essential for chicken gut colonization, may presumably have a similar role in the pathogenesis of human infection 41 . This gene is also an important mediator of material and information transfers between cells and their environment, and between compartments within cells. These surface-exposed proteins are conserved in C. jejuni strains and are also highly immunogenic in chicks 42 .
A bacterial toxin, CDT, may potentially play a role in disease development. This cytotoxin arrests eukaryotic cells to the G2 phase of the cell cycle, preventing them from entering mitosis, and leading to cell death. The toxin was named according to the morphological changes associated with its cytoplasmic distension action and its function is well documented 8,43 . In the present study, 74.3% of the C. jejuni strains carried the cdt complex. Datta et al. 27 detected a 100% frequency of these genes in their studied samples, which included chicken feces, cattle feces, and human clinical samples. Rozynek et al. 41 found a 100% prevalence in isolates from broiler carcasses. More than 76% of samples isolated from different sources were positive for cdtA, cdtB, and cdtC 24 . Previous studies have demonstrated extremely high rates of occurrence of this complex in C. jejuni isolated from several sources, including humans and poultry 35,44 . In Brazil, there are few studies on CDT in C. jejuni. Carvalho et al. 45 stated that only 36.4% of C. jejuni samples from Brazilian broiler carcasses contained the CDT complex. In contrast, Melo et al. 30 showed that almost all Campylobacter strains presented the cdt complex.
A significant percentage of C. jejuni carries the virB11 gene localized on the pVir plasmid. The pVir plasmid encodes several genes that are homologous to a type IV secretion system and contributes to the ability of Campylobacter jejuni subsp. jejuni 81-176 to invade INT-407 cells in vitro, a marker that was correlated with virulence in a diarrheal disease ferret model 25,46 . Moreover, other studies showed both an association 47 and a lack of association 48 between the pVir plasmid and bloody diarrhea in C. jejuni enteritis. We found this gene in 22.1% of C. jejuni isolates. The occurrence of virB11 varies among studies. In 2008, Wieczorek and Osek 24 detected this gene in 18.5% of strains. Ten years later, the same authors found this gene in less than 5% of strains 35 . A small percentage of this marker in poultry samples has also been reported by other authors 27,41 . The reasons underlying this significant variation are still unknown.
The iam gene is a genetic marker that is preferentially associated with adherence and invasion of HEp-2 cells in vitro 49 . There is a good correlation between the clinical presentation of diarrhea and the isolation of Campylobacter strains that adhere to and invade HEp-2 cells. However, the involvement and function of the iam marker in the process of campylobacteriosis have not been precisely explained 17,50 . In this study, the iam sequence was not detected in any of the investigated isolates. Rozynek et al. 41 detected this gene in 53.8% of C. jejuni samples isolated from chickens. The presence of this molecular marker is not restricted to C. jejuni, but is also present in C. coli and C. lari. Wieczorek and Osek 24 showed the presence of this gene in 66% of the investigated strains. It was found that the marker was predominant in C. coli (82.9% of the isolates were positive), whereas only 46.7% of C. jejuni strains possessed this gene. The variable occurrence depends not only on the origin, but also on the isolated Campylobacter species. The fact that in some invasive strains the gene was not identified using PCR supports the existence of important polymorphisms and high heterogeneity in the iam locus or suggests that other genetic markers of invasion may occur in different loci 45 . The genetic characterization of C. jejuni through whole-genome sequencing (WGS) would probably result in the detection of this marker. Previous studies have detected iam in 99% of Campylobacter strains, showing that WGS has a high discriminatory power for this genus 51 .
The wlaN gene is involved in variations of the lipooligosaccharide structure, facilitating the avoidance of C. jejuni in the host immune system. Our results showed that this marker was present in 10.7% of the tested C. jejuni strains. According to the investigation by Datta et al. 27 , this gene was present in 4.7% of isolates from broiler feces and in 23.8% of poultry meat isolates. The presence of the gene can vary depending on the geographic area and on the high rate of polymorphisms within C. jejuni strains. Koga et al. 50 reported that the genetic polymorphism of C. jejuni determines the reactivity of autoantibodies and the clinical presentation of the Guillain-Barré syndrome, possibly through the modification of the host molecule mimic. Importantly, these positive isolates may have a higher pathogenic potential, and therefore, a higher capacity to induce autoimmune disease in their www.nature.com/scientificreports/

hosts. A previous study showed a correlation between an increased capacity of cell invasion (in vitro and in vivo)
, and the presence of the wlaN gene, since 80% of the strains that lack this gene presented a lower or a total lack of invasiveness 52 . Table 4 summarizes the occurrence of virulence markers among species (human, poultry, swine, and cattle). The relative frequencies of virulence genes observed in C. jejuni strains in the present study were similar or slightly lower than those observed in other species. These results indicate that strains isolated from poultry sources are potential carriers of virulence genes to humans.
It is important to note that the genetic characterization of C. jejuni, which is based on the molecular detection of genes, corresponds to a qualitative analysis. Thus, it is not possible to establish a linear relationship between the presence of the genes and in vivo virulence. However, molecular tools for virulence marker detection are essential to determine the genetic virulence profile of the strains circulating in poultry establishments and, thus, to determine specific control measures to mitigate the risks of dissemination. It is noteworthy that this study did not perform a temporal analysis of the data, and it is not known if the relative differences are stable over time or are restricted to the time of sample collection.
Brazil has an important position in the global trade market of chicken meat. However, unlike the European Union and United States, there is no specific legislation in the country for the analysis of Campylobacter. Moreover, campylobacteriosis is still underreported, and epidemiological and molecular studies on the diversity and virulence of C. jejuni are still few in the country 29,53 . Our results reinforce the urgent need for specific legislation adoption to prevent the spread of potentially pathogenic strains through the consumption of poultry products.
In conclusion, the present study revealed that eight potential pathogenicity genes of C. jejuni circulate in slaughter lines, among other poultry sources in Southern Brazil. The frequency of C. jejuni virulence markers, their relationship with clinical severity in humans, and the expression of virulence factors should be further investigated. The results provide more insight into factors related to the occurrence of this pathogen and in the understanding of their epidemiology.

Methods
Bacterial strains and growth conditions. This research was carried out in federally inspected slaughterhouses from five companies (A, B, C, D, and E) in Southern Brazil. A total of 20 visits were made to the processing plants. All selected establishments were representative of slaughterhouses located in the region; broilers were processed for production, with an average of 250,000 birds slaughtered per day, each having a standard live weight of 2.0 kg (4.4 lb). The processing line was operated under standard commercial conditions. In total, 980 samples were collected from swabs (cloacal and transportation boxes), water from the chiller tank, and carcasses throughout the slaughter line, as previously described by Borges et al. 53 . The samples were stored in isothermal ice boxes and taken to the laboratory for identification.
Buffered peptone water (BPW) (400 mL, 0.1%) was previously added to each carcass and manually shaken in a vigorous way for 1 min. An aliquot (1 mL) of each sample (swabs, water, and carcasses) was homogenized in Bolton broth (9 mL) supplemented with antimicrobials (cefoperazone 20 mg/L, vancomycin 20 mg/L, trimethoprim 20 mg/L, cycloheximide 50 mg/L; SR0183, Oxoid) and incubated under microaerophilic conditions using a gas tank with a mixture (10% CO 2 , 5% O 2 , and 85% N 2 ) for 48 h at a temperature of 41.5 °C. After incubation, the suspension (100 µL) was spread onto an acetate mem deoxycholate agar (brane having 0.65-µm pores and placed on the surface of a modified charcoal cefoperazone mCCDA) plate with selective supplements (cefoperazone, 32 mg/L; amphotericin B, 10 mg/L; SR0155, Oxoid) for 30 min. The direct application of membrane filters to the surface of the plating media has been used to exclude unwanted microflora for the selective isolation of Campylobacter 54 . Due to its reduced pores, this membrane promotes the filtration of the suspension and provides a reduction of the bacterial contamination. The plate was incubated under microaerophilic conditions at 41.5 °C for 48 h. Campylobacter spp. suspected colonies (grayish, often with a metallic sheen, flat, and moist, with a tendency to spread) 31 were examined for cell morphology by using phase-contrast microscopy. Subsequently, , extracted template DNA (2 µL) and each of the primers (0.5 µL each, 10 pmol/L). Sterile Milli-Q water was added in sufficient quantity to achieve a volume of 25 µL. All amplification reactions were performed in a thermal cycler. The cycles were performed as described in Table 2. For the visualization of the PCR products, 10 µL aliquots were subjected to electrophoresis in a 2% agarose gel in Tris-acetylated EDTA (TAE) buffer. DNA bands were stained with ethidium bromide for 2 h at 100 V, viewed under an ultraviolet transilluminator, and photographed. The size of the PCR amplicons was compared to that of the 100 bp DNA ladder.
Statistical analysis. The obtained data were subjected to statistical analysis using the PASW Statistics software (IBM, Hong Kong). Descriptive statistics (frequency distribution) was used to determine the presence of the virulence genes according to the source and the poultry company from which the sample was isolated. Chisquare (χ 2 ) and Fisher's exact tests were used to compare the frequencies of virulence-associated genes among each other and among sources and companies from which samples were isolated. Statistical significance was defined at a p < 0.05, and Bonferroni correction was applied to adjust confidence intervals for multiple hypothesis testing. Adjusted p-values are indicated at each Table. Data availability The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.  CTA TTT TAT TTT TGA GTG CTT GTG  GCT TTA TTT GCC ATT TGT TTT ATT A  589  C. jejuni species identification 55   flaA  flaAF  flaAR   GGA TTT CGT ATT AAC ACA AAT GGT GC  CTG TAG TAA TCT TA AAC ATT TTG   94 °C/5