A total of 520 non-duplicate S. serovar Infantis isolates from human infections (n = 84), poultry meat (n = 418) and other sources (n = 18) were collected during 2010, 2013 and 2015 at the National Centre for Enteropathogenic Bacteria and Listeria (NENT), Switzerland. The isolates had been forwarded by hospitals, diagnostic laboratories or surveillance programs of retail markets and food or feed producing facilities for final species-level identification according to the White-Kauffmann-Le Minor scheme (Grimont and Weill, 2008).

Antimicrobial susceptibility testing was performed using the disk-diffusion method and the antibiotics ampicillin (AM), amoxicillin-clavulanic acid (AMC), cefotaxime (CTX), nalidixic acid (Na), ciprofloxacin (CIP), gentamicin (GM), kanamycin (K), streptomycin (S), sulfamethoxazole (Su), trimethoprim (TMP), tetracycline (T), and chloramphenicol (C) (Becton Dickinson, Heidelberg, Germany). Results were interpreted according to Clinical and Laboratory Standards Institute (CLSI) performance standards (CLSI, 2016). For sulfamethoxazole, for which breakpoints are not listed separately from trimethoprim, an inhibition zone of ≤10 mm was interpreted as resistant. Isolates displaying resistance to three or more classes of antimicrobials (counting β-lactams as one class) were defined as multidrug-resistant (MDR). Synergistic effects between AMC and CTX were regarded as an indication of the presence of an ESBL producer (Kaur et al., 2013).

Putative ESBL producers were grown on Brilliance^TMESBL agar (Oxoid, Hampshire, United Kingdom). The presence of bla_ESBLs was confirmed by PCR by screening for bla_TEM, bla_SHV, and bla_CTX-M alleles belonging to CTX-M groups 1, 2, 8, 9, and 25 as described previously (Woodford et al., 2006; Geser et al., 2012; Zurfluh et al., 2015). Synthesis of primers and DNA custom sequencing was carried out by Microsynth (Balgach, Switzerland) and nucleotide sequences were analyzed with CLC Main Workbench 6.6.1. For database searches the BLASTN program of NCBI¹ was used.

Pulsed-field gel electrophoresis was performed according to the PulseNet protocol of the Centers for Disease Control and Prevention (CDC)², using the restriction enzyme XbaI (Roche, Mannheim, Germany) and Salmonella serovar Braenderup strain H9812 (ATCC BAA 664) as the reference strain. Restricted DNA was separated in a 1% agarose gel (BioRad, Cressier, Switzerland) at 12°C for 20 h at 6 V/cm under linear ramp with switch times from 2 to 64 s and 120° included angle using a CHEF-DR III system (BIO-RAD, Munich, Germany). Gels were stained with ethidium bromide and visualized under UV light with Gel Doc (BIO-RAD, Munich, Germany). GelCompar II software (Applied Maths NV, Sint-Martens-Latem, Belgium) was used for analysis. Pairwise similarities between the XbaI PFGE patterns were calculated by the DICE similarity coefficient. Clustering was based on the unweighted pair-group method with averages, setting tolerance at 1% and optimization at 0.5%.

Whole genome sequencing was performed with a representative subset of 32 isolates selected with regard to their PFGE pattern, source, year of isolation or presence of bla_CTX-M-65. DNA extraction was performed with the Wizard^® Genomic DNA Purification Kit according to the manufacturers protocol (Promega AG, Dübendorf, Switzerland). Sequencing was performed on a MiSeq sequencer (Illumina, San Diego, CA, United States), utilizing a 600 cycle Nextera XT library kit. Trimmed Fastq data sets were de novo assembled using the recommended workflow on CLC Genomics Work bench version 8.0 (CLC bio, Aarhus, Denmark). Genomic contigs were annotated using the RAST annotation server (Aziz et al., 2008). The assembled sequences were uploaded to the http://www.genomicepidemiology.org/ server. Sequences of the seven housekeeping genes (aroC, dnaN, hemD, hisD, purE, sucA and thrA) were analyzed to identify multilocus sequence types (MLST)³, antibiotic resistance genes⁴, and plasmid replicon types⁵, using each website’s algorithms and databases. Routine processing of genome datasets was carried out by in-house perl scripts (available upon request). A local customized database of Salmonella chromosomal and plasmid genomes from NCBI was created and used for annotation, plasmid analysis, and homology detection with BLAST suite (Altschul et al., 1990). Phylogenetic analysis was conducted using a multiloci analysis based on a published core gene dataset (Leekitcharoenphon et al., 2012). Alleles in 2780 core gene loci across all the genomes were identified with Salmonella enterica subsp. enterica serovar Typhimurium LT2 genome as the reference. Of these loci, 1,500 core gene loci were randomly chosen and alleles were concatenated. The data matrix of alleles was subject to multiple alignment and phylogenetic analysis using tools available in MEGA7 suite (Kumar et al., 2016). The phylogenetic tree was built using the Neighbor-Joining algorithm in the MEGA 7 suite with 9,562 positions across 50 genomes for the final analysis.

For genome comparison, whole genome draft sequences of Salmonella enterica serovar Infantis strains from public databases were retrieved, including four recent strains of the prevalent Hungarian clone B of S. serovar Infantis from Hungary (Wilk et al., 2016, 2017), five bla_ESBL harboring, and eight non-bla_ESBL harboring strains from Italy, Israel and the United States. (Aviv et al., 2014; Franco et al., 2015), and one fully susceptible strain from the United Kingdom (Olasz et al., 2015). An overview of the strains and their GenBank accession numbers is given in Table 1. For plasmid comparison, annotations were obtained from RAST (Aziz et al., 2008), and compared with pCSAM042077 (Tate et al., 2017) on the SEED server for additional verification (Overbeek et al., 2013).

The distribution of resistance phenotypes among the S. serovar Infantis isolates is summarized in the Table 2 and shown in detail in technical Supplementary Tables S1–S3.

Multidrug resistant was detected in 389 (74.8%) of the isolates. Thereof, the majority (379/520, 72.9% of all isolates) showed a combined resistance pattern to nalidixic acid, sulfamethoxazole and tetracycline, with (NaSSuT) or without streptomycin (NaSuT). The NaSSut pattern was detected in 189 (36.3% of all isolates) and the NaSuT pattern in 190 (36.5% of all isolates). Resistance to ampicillin and ciprofloxacin was verified in 20 (3.8%) and 24 (4.6%) of the isolates, respectively, and showed a rising prevalence between 2010 and 2015 for both antimicrobials, as illustrated in Figure 1. A total of 496 (95.4% of all isolates) tested within the intermediate range for ciprofloxacin, according to CLSI breakpoints. An ESBL phenotype and growth on Brilliance^TMESBL agar was recorded for three strains (0.6% of all isolates), whereof two originated from humans (isolates 21-13 and 125-15) and one (isolate 144-13) from food (Table 3 and Supplementary Tables S2, S3).

The genomic characteristics of the 32 genomes are listed in Supplementary Tables S1–S3. The average genome size, %G+C, and number of coding DNA sequences (CDS) of the isolates was 4.88 mbp (range:4.56–4.98 mbp), 52.2% (range: 52.1–52.3%), and 4812 CDS (range: 4443–4912 CDS), respectively.

All three isolates displaying an ESBL phenotype showed the presence of the bla_CTX-M-65 gene by PCR. WGS analysis using ResFinder confirmed the presence of the bla_CTX-M-65 genes. All bla_CTX-M-65-positive isolates carried other resistance genes aph4-la, aadA1 and aac3-IVa (aminoglycoside resistance), fosA (fosfomycin resistance), floR (phenicol resistance), sul1 (sulphonamide resistance), tetA (tetracycline resistance), and dfrA14 (trimethoprim resistance) (Table 3 and Supplementary Tables S2, S3), while strains 21-13 and 144-13 also carried the aph3′-lc gene (aminoglycoside resistance). WGS of 23 MDR, non-ESBL producing isolates revealed the presence of aadA1, sul1, and tetA resistance genes throughout. Food isolate 115-10 additionally carried the non-ESBL bla_TEM-116 gene. The environmental isolate 70-15 from Israel additionally harbored floR and dfrA14 and MDR strain 119-15 isolated from animal feed also carried aph3′-lc and dfrA14 (Table 3 and Supplementary Table S3).

Whole genome sequencing did not detect any resistance genes in six non-MDR isolates (Table 3 and Supplementary Tables S1–S3), except for human isolate 153-13 which showed intermediate resistance to streptomycin and harbored aadA1 and sul1 (Table 3 and Supplementary Table S2).

By mapping WGS assemblies to the 320 kb plasmid pCFSAN42077 from bla_CTX-M-65 carrying food isolate CFSAN42077 (Tate et al., 2017), contigs or regions similar to this plasmid in isolates 21-13, 144-13 and 125-15 were detected (data not shown). Specific contigs bearing the bla_CTX-M-65 gene and surrounding loci were identified (Figure 2). The ∼5 kb cassette consisted of bla_CTX-M-65 followed by an IS5/IS1182 family transposase and a gene encoding a TonB-dependent siderophore receptor. The 5′-end is flanked by an IS6 family transposase upstream of the fipA gene, which is truncated in pCFSAN42077. By comparison, fipA is absent in p14026835, a bla_CTX-M-65 harboring plasmid from human S. Infantis isolate 14026835 from Italy (ERR1014119), as shown in Figure 2. As predicted, no replicons were detected in the genomes of these three bla_CTX-M-65 harboring isolates (Franco et al., 2015).

For other isolates in this study, WGS analysis showed the presence of incompatibility group IncFII (p96A) plasmid (accession #JQ418521), in the fully susceptible human isolate 99-15 (N15-1280), Incl1 (accession #AP005147) in fully susceptible food isolate 100-13 (N13-1368), and pESA2 and IncFII (pCTU2) (accession nos. #CP000784 and #FN543095, respectively) in fully susceptible human isolate 126-15 (N15-1729) (data not shown).

The accession numbers of the 32 sequenced isolates are listed in Table 3 and Supplementary Tables S1–S3. This Whole Genome Shotgun project has been deposited at DDBJ/ENA/GenBank under the accession NAOX00000000. The version described in this paper is version NAOX00000000.1.