Molecular Characterization and Antimicrobial Susceptibility of C. jejuni Isolates from Italian Wild Bird Populations.

Poultry is considered a major reservoir of human campylobacteriosis. It also been reported that not only poultry, but also wild birds, are capable of carrying C. jejuni, thus demonstrating to be a risk of spreading the bacteria in the environment. To gain insight into the population structure and investigate the antimicrobial resistance genotypes and phenotypes, we analyzed a collection of 135 C. jejuni from 15 species of wild birds in Italy. MLST revealed the presence of 41 sequence types (STs) and 13 clonal complexes (CCs). ST-179 complex and the generalist ST-45 complex were the most prevalent. Core genome MLST revealed that C. jejuni from ST-45 complex clustered according to the bird species, unlike the ST-179 complex which featured 3 different species in the same cluster. Overall we found a moderate prevalence of resistance to tetracycline (12.5%), ciprofloxacin and nalidixic acid (10%). The novel ST isolated from one pigeon showed resistance to all the antibiotics tested. The ST-179 complex (33.3%) was identified with significantly higher nalidixic acid resistance relative to other tested STs. Nine AMR genes (tet(O), cmeA, cmeB, cmeC, cmeR, aad, blaOXA-61, blaOXA-184 and erm(B)) and 23S rRNA and gyrA-associated point mutations were also described, indicating a concordance level between genotypic and phenotypic resistance of 23.3%, 23.4% and of 37.5% for streptomycin, tetracycline and quinolones/fluoroquinolones, respectively. We recommend that particular attention should be given to wild birds as key sentinel animals for the ecosystem contamination surveillance.


Introduction
Campylobacter disease continues to be the leading cause of bacterial gastroenteritis worldwide with a significant economic impact [1,2]. In human patients, apart from acute gastroenteritis, campylobacteriosis may lead to more severe, occasionally long-term sequelae, such as Guillain-Barre' syndrome, reactive arthritis, and irritable bowel syndrome [3,4]. It is well known that most livestock species, carry C.jejuni asymptomatically [5], making control at the farm level difficult. However, some clones may cause gastroenteritis and abortion in bovine or sheep [6]. It is estimated that 50-80% of strains causing the disease in humans derived from the chicken reservoir, 20-30% from the cattle reservoir, and the remaining part originated from other reservoirs, including environment and wild animals [7]. Numerous findings suggest that wild birds, given their ability to fly long distances and their ubiquity, play an important role in the epidemiology and evolution of Campylobacter spp. [8,9].
One strain from a grey heron (Ardea cinerea) was not assigned to a specific Campylobacter spp. by specific PCR. The sequence of the16s rDNA gene resulted in potential identification of Campylobacter volucris. The genome assembly from the heron isolate confirmed the C. volucris species by Genome-to-Genome Distance Calculator (GGDC), an in-silico DNA-DNA hybridization method (DDH), using the DDH model "formula 2" as recommended for draft genomes.  Figure 1). Five novel STs, not previously represented in the PubMLST database (https://pubMLST. org/campylobacter), were identified (ST-10211, ST-10212, ST-10213, ST-10214, ST-10215 and ST-10216) ( Figure 1). Eighteen STs resulted with not assigned clonal complex (Figure 1). ST-179 complex and ST-45 complex were the most frequently CCs found in the 53.3% and in the 16.3% of the isolates, respectively, confirming their association to the wild bird's reservoir ( Figure 1). They showed a wide geographical distribution and were found present throughout the study period. ST-45 was isolated from 4 different wild bird species (pigeons, magpies, starling and greenfinch) ( Table 1). However, a very few numbers of STs were shared among different species. In particular, ST-2655 and ST-4447 were isolated in pigeons and crow; ST-2538 was isolated in one jay and in one thrush; ST-2116 in one pigeon and in one white wagtail, while ST-220 was isolated in many pigeons and in one owl (Table 1). Differently, a great number of STs revealed to be species-specific (Table 1). For example 9 STs were found only in crows; 14 different STs only in pigeons and 2 STs showed a host specificity toward mallards (Table 1, Figure 1).
Pathogens 2020, 9, 304 5 of 15 one owl (Table 1). Differently, a great number of STs revealed to be species-specific (Table 1). For example 9 STs were found only in crows; 14 different STs only in pigeons and 2 STs showed a host specificity toward mallards (Table 1, Figure 1). The wild bird's geographic distribution in the North Italian regions is shown in Figure 2. The wild bird's geographic distribution in the North Italian regions is shown in Figure 2. The comparison of cgMLST allelic profiles using a cluster distance threshold of 13 different alleles from ST-45 complex and ST-179 complex revealed the presence of five and 1 clusters respectively (Figures 3 and 4). In detail, for the ST-45 complex we observed two distinct clusters with 6 and 3 pigeons; 1 cluster with 2 greenfinches, 1 cluster with 3 magpie and 1 cluster composed of two buzzards ( Figure 3). All the wild birds featured in the same clusters were captured in the same area. The comparison of cgMLST allelic profiles using a cluster distance threshold of 13 different alleles from ST-45 complex and ST-179 complex revealed the presence of five and 1 clusters respectively (Figures 3 and 4). In detail, for the ST-45 complex we observed two distinct clusters with 6 and 3 pigeons; 1 cluster with 2 greenfinches, 1 cluster with 3 magpie and 1 cluster composed of two buzzards ( Figure 3). All the wild birds featured in the same clusters were captured in the same area. The comparison of cgMLST allelic profiles using a cluster distance threshold of 13 different alleles from ST-45 complex and ST-179 complex revealed the presence of five and 1 clusters respectively (Figures 3 and 4). In detail, for the ST-45 complex we observed two distinct clusters with 6 and 3 pigeons; 1 cluster with 2 greenfinches, 1 cluster with 3 magpie and 1 cluster composed of two buzzards ( Figure 3). All the wild birds featured in the same clusters were captured in the same area.   The connecting lines between STs depict the number of allelic differences between them. Genotypes complexes with a distance of up to thirteen alleles are colored in grey.

Antimicrobial Resistance Phenotypes
Altogether, 20% of the C. jejuni strains from wild birds were resistant to at least one of the six antibiotics tested. The MIC test revealed that 12.5% of the isolates were resistant to tetracycline, 10% of the strains showed resistance to nalidixic acid and ciprofloxacin, while 6.7% and 4.2% were resistant to streptomycin and erythromycin, respectively ( Table 2). Few strains resulted resistant to gentamicin (2.5%) ( Table 2). The 33.3% of the strains resistant to nalidixic acid, and the 25% of strains resistant to ciprofloxacin were assigned ST-179 complex. The 26.7% of C. jejuni resistant to tetracycline and the 16.7% of those resistant to ciprofloxacin and nalidixic acid were assigned to ST-353 complex. Finally, the 20% of C. jejuni resistant to tetracycline was assigned to ST-952 complex and the 13.3% to ST-45 complex. We identified seven C. jejuni multidrug resistance pattern (MDR), indicated in Fig.1. One C. jejuni strain isolated from a pigeon, showed resistance to all tested antibiotics ( Figure 1). The most common MDR pattern (cipNatet) was observed for C. jejuni isolated from 2 pigeons, 1 crow and 1 white wagtail, while the other 2 (ciperygen) and (Naerystm) were observed in 2 pigeons (Figure 1). The connecting lines between STs depict the number of allelic differences between them. Genotypes complexes with a distance of up to thirteen alleles are colored in grey.

Antimicrobial Resistance Phenotypes
Altogether, 20% of the C. jejuni strains from wild birds were resistant to at least one of the six antibiotics tested. The MIC test revealed that 12.5% of the isolates were resistant to tetracycline, 10% of the strains showed resistance to nalidixic acid and ciprofloxacin, while 6.7% and 4.2% were resistant to streptomycin and erythromycin, respectively ( Table 2). Few strains resulted resistant to gentamicin (2.5%) ( Table 2). The 33.3% of the strains resistant to nalidixic acid, and the 25% of strains resistant to ciprofloxacin were assigned ST-179 complex. The 26.7% of C. jejuni resistant to tetracycline and the 16.7% of those resistant to ciprofloxacin and nalidixic acid were assigned to ST-353 complex. Finally, the 20% of C. jejuni resistant to tetracycline was assigned to ST-952 complex and the 13.3% to ST-45 complex. We identified seven C. jejuni multidrug resistance pattern (MDR), indicated in Figure 1. One C. jejuni strain isolated from a pigeon, showed resistance to all tested antibiotics (Figure 1). The most common MDR pattern (cipNatet) was observed for C. jejuni isolated from 2 pigeons, 1 crow and 1 white wagtail, while the other 2 (ciperygen) and (Naerystm) were observed in 2 pigeons (Figure 1).

Detection of Resistance Genes, Mutations and Levels of Concordance Among the Two Type of Resistances
The following genes of antimicrobial resistance were investigated: tet(O), cmeA, cmeB, cmeC, cmeR, aad, oxa184, oxa61 and erm(B). The tet (O) gene, encoding tetracycline resistance ribosomal protection protein (TetO), was detected in 53.4% of wild bird strains. The multidrug efflux pump (CmeABC) and its regulator (CmeR), conferring resistance to a wide range of antimicrobials, were found in all the isolates; while aad gene, known to confer streptomycin resistance, was present in the 28.6% of wild bird isolates. The two major B-lactamase genes were differently observed, with a prevalence of oxa-184 (86.7%) versus oxa-61 (10%). Differently, erm(B) gene and point mutations on 23s rRNA regarded to be associated with erythromycin resistance, were not identified in any isolate.
C. jejuni strains showing resistance to quinolones and fluoroquinolones, were tested for point mutations in gyrA gene. Two types of mutations linked to ciprofloxacin and nalidixic acid resistance were detected in the analyzed wild bird populations on gyrA gene.
In the detail, 25% possessed the C257T point mutation (2 pigeons and 1 crow), resulting in a T86I substitution in the gyrA gene, while 8.3% presented the A256G point mutation (1 pigeon) producing a T86V substitution. Weak correlations were found for phenotypic and genotypic resistances, reaching levels of concordance of 23.3%, 23.4% and of 30% for streptomycin, tetracycline and quinolones/fluoroquinolones, respectively ( Table 2).

Discussion
Reducing the occurrence of campylobacteriosis is a food safety issue of high priority in European Union; however, continuing invasion on wildlife habitat by humans through a massive urbanization and intensive agriculture increases the opportunities for contact between wild birds, domestic animals, and people [9]. Wild birds, being natural reservoir of Campylobacter, thus represent a serious risk for public health, either directly through the consumption of wild bird game meat, or indirectly, by disseminating Campylobacter into the environment and domestic livestock [14]. The aim of this study was to assess the genomic characterization of circulating wild bird species in Italy and indagate their capability to disseminate in the environment important antimicrobials factors, in order to evaluate their influence on public health. Wild birds are usually considered to have a marginal role in human illness. However, specific STs and generalist lineages of C. jejuni found in hospitalized patients with gastroenteritis have been reported in several species of wild bird [13,[15][16][17]21], indicating their potentiality as reservoirs of human infection. ST-45 and ST-267, isolated in humans, have been found among both blackbird and chicken isolates from Sweden, magpies, greenfinches, starling, pigeons, owl and blackbird in Italy [26,33,34]. In a previous study [26] other STs (ST-48, 2116 and 1044) from Italian human patients, were shared with starling, magpies and white wagtail circulating in Italy. Differently, the high occurrence (73.2%) of species-specific STs, would confirm the existence of particular C. jejuni lineages for specific wild bird species. ST-1224 that we isolated in one magpie, has only been found in wild birds from USA and in environmental water in Canada, confirming on one side a species-specificity, on the other side the bacteria dissemination in environmental waters through wild birds acting as vectors [35,36]. Moreover, ST-2353 isolated in a seagull, has been found only in a gull in New Zealand in 2008 so far, as well as in environmental water in the same area (www.pubMLST.org), reinforcing the idea that gulls are not a major source of C. jejuni in human infections. In our study, two major CCs were reported in 53.3% (ST-179 complex) and in 16.3% (ST-45 complex) of wild bird population; thus, in order to analyze the genomic diversity within the most detected clonal complexes, a comparative genomics using cgMLST was performed. The minimum spanning tree based on the cgMLST of ST-179 complex clearly indicates that although the isolates clustered in three groups defined by different STs, they form a cohesive single population with only 1 or 2 allelic differences. Furthermore, in addition to 70 pigeons, we found a crow and an owl within the cluster, confirming the idea that these STs are commonly shared between different animal host species. In detail, the crow belonged to the same town of the pigeons, while the owl was collected from a sampling site 400 km away. Otherwise, the ST-45 complex, comprised seven different species, nonetheless the isolates were more closely related to those of the same species, independently from the sampling site or time of isolation. In our study, the sampling size and the variety of species were defined by the use of passive surveillance without active case finding, therefore, it is not possible to make a general conclusion on the whole wild bird population. Nonetheless, the results aided several other studies and seem to corroborate the idea that wild birds not only can disseminate C. jejuni in different regions, but they can harbor different lineages of C. jejuni. Our results are in line with a recent study demonstrating the existence of certain sub lineages of ST-45 forming genetically isolated clades containing C. jejuni strains with extremely similar genomes regardless of time and location of sampling [27]. In addition to the highly mobile nature of many wild bird species to disseminate bacteria in the environment [32], the World Health Organization in 2017 (World Health Organization, 2017) showed that many of these microorganisms exhibited resistance to antimicrobials considered of highest importance to human medicine [37,38]. Interestingly, C.jejuni strains isolated in this study showed a moderate percentage resistance to tetracycline, to ciprofloxacin and nalidixic acid, suggesting that the antimicrobial resistance could be acquired by horizontal gene transfer [39] or, to the usage antibiotics as a therapeutic agent in dairy cattle and poultry farms [40]. Lower resistance rates have been observed for streptomycin, erythromycin and gentamicin. The prevalence observed in our study is partially in line with data obtained by Aksomaitiene at al. [41] for tetracycline and erythromycin in Lithuania. A low number of streptomycin-resistant C.jejuni from wild birds has been previously described [42]. In contrast, a worrying resistance rate for quinolones, fluoroquinolones and tetracyclines has been detected in the same C. jejuni strains. Krawiek et al, shows higher level of resistance, respect to our data, for all antibiotics with the exception of nalidixic acid, slightly lower than that reported from us probably related to different environmental features [24].
We also studied our wild birds in the context of MDR, defined as resistance to at least three different antimicrobials [43]. Antimicrobial resistance is a complex problem and the wildlife role could be underestimated. In our dataset we found only 5.2% of wild birds resistant to three different antibiotic classes. However, one C.jejuni from a bird usually living strictly with human people (pigeon) showed a worrying MDR to all the antimicrobials tested in the study, reinforcing the idea that wild birds represent a real risk in spreading resistant bacteria. Wild bird studies depend on opportunistic sample collection; therefore, some birds are more likely to be sampled than others, as happen in this study with the pigeons that are the majority in this set of wild birds analyzed.
C.jejuni isolates resistant to cip and Na were screened for the presence of the mutations in the quinolone resistance-determining regions (QRDR) of the gyrA gene. We found T86I amino acid substitution as the most common (25% of pigeons and crow isolates), followed by only a pigeon with T86V aminoacid substitution. The rest of wild birds resulted without mutations, confirming other mechanisms of resistance such as pump efflux system, as reported previously [44]. We found the genes involved in this multidrug CmeABC efflux system, together to his repressor cmeR, present in all wild bird isolates. However, our wild strains miss the known genetic mechanisms for the ery resistance, implying that resistance to ery could be the result of another unknown system in our strains.
These findings are in line with the study of Yao et al. that revealed the existence of a "super" CmeABC variant used by these bacteria to enhance multidrug resistance, desensitizing Campylobacter to many antimicrobials and conferring a high-level resistance to fluoroquinolone in association with the C257T mutation in gyrA [45].
We screened for other anti-resistant factors such as blaoxa-184 and blaoxa-61 gene, described to confer beta-lactam resistance in C.jejuni strains by Alfredson et al 2005 [46]. We found the 86.7% and the 10% of the genes respectively presents, in line with other studies [45,47]. This is worth to investigate further because this antimicrobial class usually is not used for the Campylobacter infections. Finally, in this study we found a low correlation between phenotypic resistance to tetracycline, streptomycin and quinolones/fluoroquinolones and the existence of resistance genes or nucleotide polymorphisms.
In conclusion, our study demonstrated that wild birds are mostly colonized by host-adapted STs, possessing a minor set of STs shared with further hosts, including poultry, livestock and humans. Although, we found a low and moderate prevalence of antibiotic resistance in our data set, it is extremely important to monitor the situation also in wild birds. Understanding of the prevalence of these potentially dangerous bacteria in wild birds is necessary to guide public health policies about the control of foodborne illness in humans. Similarly, the occurrence of resistant bacteria in wildlife has been detected in many animal species across different geographical areas, highlighting the importance and complexity of wild animals, in the transmission of resistant bacteria, even if they are not usually exposed directly to antimicrobials. Thus, continued surveillance of multi-resistant bacteria in wild animals is warranted.

Bacterial Strains and Species Identification
A total of 135 wild bird isolates from Piedmont and Veneto regions were included in the study. In particular, 67/135 were previously analyzed for antimicrobial susceptibility [44] and added to the others sixtyeight. Strains were isolated during five years (2015-2019) via passive surveillance monitoring by the Istituti Zooprofilattici Sperimenatali (IIZZSS) network and/or sent to the National Laboratory Reference of Campylobacter (NRL, http://www.izs.it/IZS/Eccellenza/Centri_nazionali/LNR-Campylobacter), for species identification or molecular characterization. The strains were cultured on Columbia blood agar plates in microaerobic atmosphere at 42 • C for 48h and stored at −80 • C in Microbank ™ until further analysis. After an initial phenotypic characterization, suspected colonies confirmed as thermotolerant Campylobacter and identified to species level using a multiplex and a simplex PCR, as described previously

Sequence Analysis and Identification of Antibiotic Resistance Genes
Total genomic DNA was used to prepare sequencing libraries using Nextera XT Library Preparation Kit (Illumina, Inc., San Diego, CA, USA). The libraries were then sequenced using Illumina NextSeq 500 sequencer. Sequence reads (150-bp, pair-end) were demultiplexed and the adapters were removed. Subsequently the reads were trimmed with Trimmomatic tool (version 0.36) and de novo assembled using SPAdes version 3.11.1 with the 'careful' option selected [48]. The sequence reads generated in this study were deposited in NCBI Sequence Read Archive (SRA) in Bioproject PRJNA623711. Additionally, 15 previously published sequences used in the analysis can be found in NCBI Bioproject PRJNA510785 [49].
C.jejuni genome assemblies, were genotyped by using MLST and cgMLST. The assemblies were also investigated for the genomic AMR traits. Gene-by-gene analysis was performed using the SeqSphere+ v4.1.1 software (RidomGmbH, Münster, Germany). The MLST profiles were assigned using a C.jejuni/coli task template MLST 7 loci, schema available at https://pubmlst.org/campylobacter/ accessible through in Ridom SeqSphere+ software. The cgMLST profiles were assigned using C.jejuni task template with 949 target core genomes available in the Ridom SeqSphere+ database (https: //www.ridom.de/seqsphere/u/Task_Template_Sphere.html). A Neighbor joining tree was generated for MLST profiles, while a Minimum spanning tree (MST) was generated by pairwise comparison of cgMLST target genes using default settings parameters of Ridom. Missing alleles were ignored in the pairwise comparisons.
AMR genes were identified in silico using PointFinder v. 3.1.0 and ABRicate v. 0.8 (https://github. com/tseemann/abricate/) by querying the publicly available Comprehensive Antibiotic Resistance Database (CARD) [50,51]. Prokka v1.13 [52] was used to annotate the assemblies and gyrA sequences were extracted applying the query_pan_genome function in Roary v3.12.0 [53]. gyrA genes were aligned using Uniprot UGENE v1.18.0 [54], from which the gene variants were identified. Only mutations in the quinolone resistance-determining region (QRDR) of gyrA were assessed to be the determinants of resistance, as only these loci have been linked with phenotypic resistance to quinolones. In particular we analyzed the amino acid changes at position 86.

Antimicrobial Susceptibility
A total of 68/135 wild bird strains were subjected to the antimicrobial test and added to the previously analyzed [44]. Susceptibility to antimicrobials was evaluated with the microdilution method using the Sensititre automated system (TREK Diagnostic Systems, Venice, Italy). Colonies were sub cultured on Columbia agar for 24 hours and then seeded in Mueller Hinton Broth supplemented with blood (Oxoid, Basingstoke, UK). Subsequently, they were dispensed into Eucamp2 microtiter plates (TREK Diagnostic Systems, Venice, Italy), containing known scalar concentrations of the following antibiotics, indicating the distribution % of Minimum Inhibitory Concentration (MIC): ciprofloxacin (0.12-16 µg/mL), erythromycin (1-128 µg/mL), gentamicin (0.12-16 µg/mL), nalidixic acid (1-64 µg/mL), streptomycin (0.25-16 µg/mL), and tetracycline (0.5-64 µg/mL). After inoculation, the plates were incubated at 42 • C in microaerobic atmosphere for 24 h and then screened. Campylobacter jejuni strain NCTC 11351 was used as control. The strains were classified as resistant (R), and susceptible (S) to the examined antimicrobials based on MIC breakpoints, by using Swin v3.3 Software (Thermo Fisher Scientific) in accordance with the epidemiological cutoff values (ECOFFs) as defined by EUCAST (European Committee on antimicrobial breakpoints) (www.eucast.org) to interpret their antimicrobial susceptibilities. MIC breakpoints of resistance were ≥0.5 µg/mL for ciprofloxacin, ≥4 µg/mL for streptomycin, ≥4 µg/mL for erythromycin, ≥2 µg/mL for gentamicin, ≥16 µg/mL for nalidixic acid and ≥1 µg/mL for tetracycline. Details of the wild bird isolates are summarized in supplementary Table S1.