The World Health Organization (WHO) has recently published a list of bacteria for which new antibiotic therapies are urgently needed, with Campylobacter classified as high priority (WHO, 2017). This is of concern as campylobacteriosis is the most commonly notified bacterial foodborne infection in the European Union (European Food Safety Authority [EFSA]/European Centre for Disease Prevention and Control [ECDC], 2016a). The disease is principally caused by Campylobacter jejuni and Campylobacter coli following the ingestion of contaminated food and drink, with consumption of poultry meat recognized as a major risk factor (Wilson et al., 2008; Whiley et al., 2013). Infection can be associated with extra intestinal pathologies and sequelae such as reactive arthritis or Guillain-Barré syndrome (Nachamkin et al., 1998), but it is usually self-limiting. Treatment of severe infection occasionally requires antimicrobial therapy, often with erythromycin (European Food Safety Authority [EFSA]/European Centre for Disease Prevention and Control [ECDC], 2016b) and to a lesser extent with gentamicin, the later used occasionally when infection becomes systemic (Lehtopolku et al., 2009). Although fluoroquinolones were commonly used in the past, the rising of resistance among Campylobacter isolates makes these antibiotics ineffective (Kassem et al., 2016).

Erythromycin inhibits protein synthesis by binding to the ribosome and blocking the exit of the nascent peptide chain (Fyfe et al., 2016). Erythromycin resistance in bacterial isolates from animals and humans is associated with the presence of erm genes (Weisblum, 1995). The most widely distributed erm gene class is erm(B), which encodes an rRNA methylase which produces cross-resistance to macrolides, lincosamides and streptogramins B (MLS_B phenotype) (Leclercq, 2002). The erm(B) encoded enzyme acts on the 23S rRNA gene by methylating an adenine residue that hinders antibiotic binding-to the ribosome (Weisblum, 1995). The erm(B) gene is present in a variety of Gram-positive bacteria, including enterococci, streptococci, and staphylococci (Jensen et al., 1999). The potential for interspecies horizontal gene transfer (HGT) (Conwell et al., 2017) has facilitated the emergence of resistance in multiple species including Gram-negative bacteria of the genera Bacteroides, Shigella, Escherichia, Klebsiella, and recently Campylobacter (Shoemaker et al., 2001; Soge et al., 2006; Phuc Nguyen et al., 2009; Qin et al., 2014). Resistance in Campylobacter has been associated with ribosomal mutations, efflux pumps and the erm(B) gene that has been identified in isolates from China and Spain (Qin et al., 2014; Florez-Cuadrado et al., 2016; Fyfe et al., 2016). Recent work has reported high-level erythromycin resistance (MIC ≥ 1024 mg/L) in a C. coli isolate carrying erm(B)- in a genomic island along with other determinants conferring resistance to aminoglycosides, tetracycline and streptothricin (Florez-Cuadrado et al., 2016). Thus, eight types of erm(B)-carrying genomic islands have been differentiated (Qin et al., 2014; Wang et al., 2014; Florez-Cuadrado et al., 2016), all of which share aminoglycoside resistance genes in addition to other determinants, likely leading to co-selection after genetic mobilization (Chapman, 2003).

Improved understanding of the distribution of resistance genes within bacterial species in different host niches, and the mobility of these genes between populations, could be important for identifying source and sink populations. In the case of Campylobacter, the erm(B) gene has been identified in C. coli isolates from chicken, ducks, swine, and humans and from C. jejuni isolated from chicken (Qin et al., 2014; Wang et al., 2014; Deng et al., 2015) but other host species may be relevant. Turkeys are among the top 10 farmed animals in Europe and the United States with an estimated 323 million birds reared anually (FAOSTAT, 2017). While studies have shown that turkeys are an important host species harboring large numbers of C. jejuni and C. coli, the resistance status of these strains is not well-characterized. This has lead to the inclusion of this animal species in international surveillance programs to evaluate the levels of antibiotic resistance. In this study we carried out combined molecular microbiology and whole genome sequencing approaches to evaluate the presence of erm(B) and it’s genetic background in Campylobacter isolates obtained from turkeys sampled in Spain. The comparison with erm(B) sequences from other host species might allow further description of the microevolutionary events associated with the acquisition of this antibiotic resistance genes in Campylobacter.

Definitive characterization of erythromycin resistance in bacterial pathogens is an important objective defined by the EFSA and the European Centre for Disease Prevention and Control (ECDC) (European Food Safety Authority [EFSA]/European Centre for Disease Prevention and Control [ECDC], 2017). To date, the erm(B) gene has been identified in Campylobacter isolates from swine, chickens, ducks and humans from China and in one broiler sample from Spain (Wang et al., 2014; Florez-Cuadrado et al., 2016). Monitoring antimicrobial resistance in Campylobacter isolated from turkeys has been mandatory since 2014 in European countries where the production of turkey meat exceeds 10,000 tons per year (2016). The occurrence of erythromycin resistant C. coli isolates from turkeys in Germany, Romania and Spain had risen to 43.3% in 2014 (2016), compared to previous surveys where lower frequencies were detected for poultry (14.5%) and swine (20.7%) (European Food Safety Authority [EFSA]/European Centre for Disease Prevention and Control [ECDC], 2015, 2016b). Since Campylobacter infections are related to consumption of food from animal origin, these levels of resistance could potentially produce therapeutic failure of antibiotic treatment for campylobacteriosis in humans.

Increased antimicrobial resistance among Campylobacter populations is a consequence of the widespread acquisition of antimicrobial resistance and clonal expansion of resistant lineages (Wimalarathna et al., 2013). Campylobacter can acquire DNA, including antimicrobial resistance genes, from relatively distantly related lineages through HGT, involving the replacement of homologous sequences, or the acquisition of mobile genetic elements (MGEs). There is evidence that plasmid acquisition mediates Campylobacter resistance to tetracycline, chloramphenicol and aminoglycosides (Courvalin et al., 1978; Taylor et al., 1987; Wang and Taylor, 1990) but in some cases, these resistances might be conferred by polymorphism of chromosomal sequences. This is also the case in Campylobacter for resistance to fluoroquinolones and macrolides, mediated by mutations in gyrA or 23S rDNA sequences, respectively (Engberg et al., 2001). Mutations that confer antimicrobial resistance can occur independently in multiple lineages but can also spread by natural transformation followed by homologous recombination, leading to the dissemination of antimicrobial resistance among bacteria that share an ecological niche (Meric et al., 2015).

The increasing number of studies identifying erm(B) genes in Campylobacter suggest that horizontal transfer is related to macrolide resistance in Campylobacter (Qin et al., 2014; Wang et al., 2014; Deng et al., 2015; Florez-Cuadrado et al., 2016; European Food Safety Authority [EFSA]/European Centre for Disease Prevention and Control [ECDC], 2017). High levels of resistance among C. coli isolates have been previously reported in isolates from livestock (Thakur and Gebreyes, 2005), potentially reflecting a selective advantage in these niches and clonal expansion of resistance lineages. This study shows the location of erm(B) gene in a genomic island along with other antimicrobial resistance genes including tet(O), encoding resistance to tetracycline and aph(3′)-IIIa, aacA-aphD, aad9 and aadE, conferring resistance to different aminoglycosides (Qin et al., 2014; Wang et al., 2014; Florez-Cuadrado et al., 2016), a clustering that implies the possibility of co-selection as an evolutionary mechanism (Chapman, 2003). In addition, linked erm(B) and aph(2”)-IIIa genes detected in the genome of C. coli ZTA14/01426 isolate, could be transferred to virulent strains limiting seriously the effectiveness of the two main choices for treatment of severe campylobacteriosis, erythromycin and gentamicin (Lehtopolku et al., 2009).

Although the presence of the erm(B) gene was evaluated on a total of 85 Campylobacter isolates, the scope of the genomic comparison was limited to only two C. coli because they were positive for the erythromycin resistance gene. Despite the limited number of genomes sequenced, the erm(B)-carrying genomic islands identified in Campylobacter isolated from Spain show genetic differences in comparison with ones from China. Thus, the genomic islands identified in Spain do not correspond to any of the six types of genomic islands identified in C. coli from China (Qin et al., 2014; Wang et al., 2014; Florez-Cuadrado et al., 2016). All of the erm(B)-carrying genomic islands posses aminoglycoside resistance genes but the gentamicin resistance gene aph(2”)-IIIa is present only in Spanish isolates. Resistance to gentamicin was represented in the Chinese erm(B)-carrying sequences with the presence of the aacA-aphD gene and was identified in C. coli isolates from humans, poultry, and swine (Wang et al., 2014).

Identical antibiotic resistance gene nucleotide sequences have been detected in both Gram-positive and Gram-negative bacteria. This is consistent with HGT facilitating the spread of resistance genes between distantly related species (Trieu-Cuot et al., 1985; Courvalin, 1994). Gram-positive bacteria including Enterococcus, Streptococcus, and Staphylococcus are widely known to harbor various resistance genes (Zilhao et al., 1988; Pinto-Alphandary et al., 1990) and recently, tetracycline and aminoglycoside resistance genes from Gram-positive bacteria have been identified in Gram-negative bacteria including Escherichia coli or Klebsiella pneumoniae (Arthur et al., 1987; Brisson-Noel et al., 1988) and Campylobacter (Lambert et al., 1985; Zilhao et al., 1988; Pinto-Alphandary et al., 1990). Four erm(B) alleles were identified in Campylobacter in this study (Supplementary Table S3) and comparison of these alleles with those from other bacteria provides useful information about possible routes of transmission (Figure 2). First, 100% nucleotide identity with erm(B) genes from Enterococcus, Streptococcus, Peptoclostridium, Anaerostipes, Arcanobacterium, Eggerthella, Lactobacillus, Lactococcus, Macrococcus, and Selenomonas suggests horizontal acquisition from Gram-positive bacteria, as previously described (Qin et al., 2014; Wang et al., 2014; Florez-Cuadrado et al., 2016). Second, the erm(B) alleles identified in this study have been previously identified mainly in human and porcine bacterial pathogens of Asian and European origin (Figure 2). The pig and human pathogen S. suis is considered a reservoir of antibiotic resistance genes (Palmieri et al., 2011) and transferable genetic elements carrying the erm(B), tet(O) and aminoglycoside resistance genes have previously been reported in S. suis isolates (Palmieri et al., 2012). The use of combined antibiotic therapies on swine farms potentially selects for bacterial isolates carrying resistance to different antibiotics (Zhu et al., 2013).

Allele 2 of the erm(B) gene has been identified in more bacterial genera than the other alleles in this study (Figure 2). All bacterial genera where erm(B) allele 2 has been identified belong to the Firmicutes, with two exceptions that belong to the Actinobacteria. These bacterial genera are present in the gastrointestinal tract and species like Enterococcus, Streptococcus, and Staphylococcus share the niche with Campylobacter (Zilhao et al., 1988). The transfer of antibiotic resistance genes in the human colon between bacteria of different genera has been reported (Huddleston, 2014) and it is possible that the antibiotics supplied in both animals and humans could facilitate the selection of strains carrying these multiresistant genomic islands and favor their dispersion. Identifying erm(B) genes and MDR genomic islands in Campylobacter isolates from turkeys, adds another component to the already extensive network of bacterial genera and hosts involved in the possible dispersion of critical antibiotic resistance genes and MGEs. This study highlights the need to sequence a greater number of Campylobacter genomes with the objective of evaluating the impact of genomic islands on the dispersion of antimicrobial resistance genes in this genus.

Contributions to the conception: DF-C, AQ, MP, LD, and SS; design of the work: DF-C, MU-R, GM, BP, JS-L, and GO; analysis or interpretation of data: DF-C, MU-R, GM, and BP; drafting the work or revising it critically: DF-C, MU-R, GM, AQ, MP, BP, JS-L, GO, LD, and SS; final approval of the version published: DF-C, MU-R, GM, AQ, MP, BP, JS-L, GO, LD, and SS.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.