A total of seven NTS isolates, one per farm, was recovered from the 22 sampled farms. Serotyping and WGS analysis determined that two isolates (G10A and G11A) belonged to serotype S. Schwarzengrund. The remaining five isolates (G3A, G12A, G13A, G15A, and G17A) were typed as S. Infantis. All S. Infantis isolates exhibited MDR phenotypes (Table 1). All the isolates were susceptible to colistin (CL), azithromycin (AZM), tigecycline (TGC), and meropenem (MEM), (data not shown). One S. Infantis isolate (G17A) could not be recuperated for WGS analysis and characterization.

Genomic analysis revealed that the S. Schwarzengrund isolates lacked resistance genes against most of the antibiotics tested in our panel and displayed the corresponding drug-sensitive phenotypes (Table 2). The qnrB19 gene was present in both S. Schwarzengrund isolates; however, only G10A exhibited reduced susceptibility to ciprofloxacin (CIP) (Table 2). All S. Infantis isolates possessed mutation D87Y in the gyrA gene conferring decreased susceptibility to CIP. S. Infantis isolates possessing tetracycline (tetA), trimethoprim (dfrA14), and sulfonamide (sul1) resistance genes displayed the corresponding resistance phenotypes. Similarly, multiple aminoglycoside resistance genes, aph(4)-la, aadA1, aac(3)-IVa, and aph(3′)-Ia, were detected in the S. Infantis genomes and corresponded to gentamicin (GEN) resistant phenotypes. Three S. Infantis isolates (G12A, G13A, and G15A) possessed the ESBL-producing bla_CTX-M-65 gene and contigs mapped closely to the 316,160-bp IncFIB-like plasmid pCVM44454 (>298,000 identical bp), including the resistance region (Figure 1). Plasmid replicons belonging to type Col440II were detected in the genomes of both S. Schwarzengrund isolates.

Seven-gene (7-gene) MLST analysis revealed that the S. Schwarzengrund isolates (G10A and G11A) clustered in ST96, while all S. Infantis isolates clustered as ST32 (Table 2). We then used SNVPhyl to perform SNP-based phylogenetic analysis on the four S. Infantis isolates from Galapagos and ten previously reported S. Infantis isolates from the United States that were shown to carry bla_CTX-M-65. The results indicated that the S. Infantis isolates from Galapagos were highly related to each other (0 pairwise SNP differences) and were also closely related to the S. Infantis isolates from the United States (1–3 pairwise SNP differences; Supplementary Table S1). Five of the United States isolates were highly related (0 pairwise SNP differences), while the remainder were as distantly related to each other, and to the other United States isolates (2–5 pairwise SNP differences), as they were from the Galapagos S. Infantis isolates.

A ParSNP analysis focusing on the S. Infantis isolates from the Galapagos and the United States showed a more distinct branching of the Galapagos isolates. The United States isolates split into two groups (veterinary/retail and human), of which the veterinary/retail clustered together, while the remaining human isolates exhibited varying degrees of uniqueness but were phylogenetically closer to the Galapagos isolates (Figure 2).

Highly discriminatory prophage sequence typing further distinguished among the Galapagos isolates while preserving major branching patterns previously reported by Tate et al. for related S. Infantis isolates (Figure 3; Supplementary Table S2) [25]. The distinguishing features of each isolate, regardless of location, were demonstrable by the numbers and types of prophages present. Each of the S. schwarzengurd isolates possessed four prophages, but they all differed either in identity (1 prophage) or sizes of the prophage genomes (three phages). On the other hand, the four S. Infantis isolates had 5–7 prophages with distinct features, whereas the United States isolates reported by Tate et al. had 8–10 prophages [25].

There is some evidence that the global prevalence of AMR NTS isolates is partially driven by the overuse of antibiotics in human and veterinary medicine, as well as the movement of humans, animals and food commodities between different regions [30,31,32,33]. In this study, we identified MDR clones of S. Infantis in the Galapagos Island of Santa Cruz that are genetically similar to strains isolated from the United States and Ecuador. Here, we speculate that the importation of live poultry from Ecuador for commercial farming is a potential route for the entry of MDR S. Infantis into the Galapagos Islands. Furthermore, our study highlights the role of a large IncFIB-like plasmid in the global dissemination of ESBL-producing and MDR strains of S. Infantis, through poultry production.

Despite the apparently low recovery of Salmonella in the sampled poultry farm environments, both recovered serotypes (S. Schwarzengrund and S. Infantis) are known human pathogens. Interestingly, neither serotype was detected in previous surveillance studies of wildlife in several Galapagos Islands, including Santa Cruz [27,28,29]. As previously highlighted, poultry production in the Galapagos Islands receives all inputs, including day-old chicks, from Ecuador—a country in which S. Infantis has been reported to have a prevalence rate of ~42% in broiler chicken farms [16]. In fact, S. Infantis was the most abundant serotype contaminating chicken carcasses destined for retail in Ecuador [17,34]. In contrast, S. Schwarzengrund does not appear to be highly prevalent in Ecuador, but has been isolated from poultry and multiple food sources in Brazil [18,35,36] and Argentina [37]. To our knowledge, this is the first report on the occurrence of NTS in poultry from the Galapagos Islands.

WGS revealed that the S. Infantis isolates from this study were closely related and likely share a common ancestor. All isolates clustered into ST32, a globally disseminated clone that is highly prevalent in poultry and has been associated with diarrheal disease in affected humans [38]. Likewise, ST96 isolates of S. Schwarzengrund, also detected in this study, have been isolated from poultry-related products and the associated environments in Latin America and other countries [39,40,41]. Interestingly, SNP analyses showed a close relationship between S. Infantis isolates from the Galapagos and previously reported isolates originating from food sources in the US [25]. Additional resolution of the accessory genome using PST further revealed distinguishing features in each S. Infantis isolates. The high resolution provided by PST can be exploited to track the Galapagos isolates back to their potential place of origin in Ecuador and may prove to be an important laboratory support for future epidemiological investigations of the serovar S. Infantis (Table S2). We observed that the genome of Salmonella phage SJ46 was only present in the S. Infantis isolates from the United States, whereas the Salmonella phage g431c was present in both groups of S. Infantis isolates but not in the S. Schwarzengrund isolates. Three of the Galapagos S. Infantis had unique prophage sequences (Bacillus phage phi STI in G13A, Clostridium phage phi CTC2A in G3A and Escherichia phage pro 483 in G15A). Similarly, phage Entero P4 was unique to G12A among the Galapagos isolates. However, the majority of the United States isolates also had this phage.

The qnrB19 gene was the only AMR determinant detected in the genomes of the S. Schwarzengrund isolates and encodes plasmid-mediated quinolone resistance (PMQR), conferring reduced susceptibility to quinolones/fluoroquinolones by protecting the bacterial gyrase from interactions with the antibiotic [42]. The Col440II-like plasmid was detected in the S. Schwarzengrund isolates; however, a recent study highlighted that a small pPAB19-4-like plasmid plays an important role in the dissemination of qnrB19 throughout Chile [43]. Not surprisingly, NAL susceptibility was detected in both qnrB19-positive S. Schwarzengrund isolates from this study, and one displayed an intermediate CIP phenotype. The ability of qnrB19 to confer reduced quinolone/fluoroquinolone susceptibility in NTS of human and animal origin is well established [44,45,46]. Full resistance is usually exhibited by isolates that bear multiple quinolone resistance genes, including mutations within the quinolone-resistance-determining regions (QRDR) of the gyrA, gyrB, parC and parE genes [47]. Nonetheless, NAL and CIP-resistant S. Schwarzengrund isolates that possess qnrB19, but lack QRDR mutations, have been isolated from chicken by-products in Brazil [48].

The S. Infantis isolates in this study were all resistant to NAL but displayed intermediate MIC to CIP. Although these isolates lacked qnrB19, they possessed the D87Y mutation in the chromosomally encoded GyrA enzyme. Quinolones/fluoroquinolones target the bacterial DNA gyrase and topoisomerase enzymes, and mutations within the QRDR of gyrA can reduce their binding affinity for the antibiotics, rendering them ineffective [49]. Single-amino acid gyrA mutants display reduced quinolone susceptibility in multiple NTS serotypes, and full resistance has been observed in double-mutants [50,51] and strains that possess an additional PMQR such as qnrB [47]. All isolates from this study possessed the D87Y mutation commonly observed in quinolone-resistant strains of NTS [49]. The contribution of QRDR mutations to the rise of reduced quinolone/fluoroquinolone susceptibility and resistance in Latin America is well documented. For example, a surveillance study highlighted that the prevalence of quinolone-resistant NTS isolates, half of which possessed the D87N mutation in gyrA, was twice as high in Latin America (~14%) compared to North America (6.1%) [52].

MIC testing demonstrated that all S. Infantis isolates from the Galapagos Islands are multi-drug resistant. This feature has been reported in other studies in Ecuador, where most S. Infantis isolates presented multi-resistant phenotypes [16,17]. Moreover, resistance to third-generation cephalosporines mediated by the bla_CTX-M-65 gene has been reported in Latin American countries, the United States and Europe [22,53,54]. Recently, this phenotype was reportedly associated with travel to South America [24], but more research is needed to ascertain whether S. Infantis strains carrying the bla_CTX-M-65 gene are now endemic in countries where they are being increasingly detected.

ESBL-producing bacteria are characterized by resistance to AMP, extended-spectrum third-generation cephalosporins and monobactams [55]. Three S. Infantis isolates from this study possessed the bla_CTX-M-65 gene that encodes a CTX-M β-lactamase. These enzymes constitute the most globally disseminated ESBL subgroup found in Gram-negative bacteria [56,57,58,59] and are often present in bacteria that contaminate live poultry and poultry by-products [60]. We detected the bla_CTX-M-65 gene in a resistance region that was genetically similar to that in the IncFIB-like plasmid identified in ESBL-producing strains of S. Infantis (Figure 1) isolated from chicken by-products and patients in the US and Italy [25].

Escherichia coli and Salmonella isolates that contain bla_CTX-M genes have been reported in Latin American countries including Brazil and Argentina [61]. Furthermore, bla_CTX-M-65-containing and ESBL-producing strains of S. Infantis have been isolated from poultry in Ecuador [16,62]. The use of antibiotics as prophylactics during the first week of life of chicks is a common practice in poultry production on Ecuador’s mainland; however, this is rarely practiced in the poultry industry of the Galapagos. It is noteworthy that cephalosporin antibiotics such as ceftiofur are often injected into fertile eggs at hatcheries to prevent E.coli-induced omphalitis in day-old chicks [63]. Moreover, a Canadian study revealed a strong correlation between this practice and the rise of ceftiofur-resistant strains of Salmonella Heidelberg [64]. In addition, Dierikx et al. demonstrated that the prevalence of ESBL-producing E. coli present in the poultry environment increased significantly after the use of β-lactam antibiotics [65]. The CTX-M family of ESBL enzymes are highly prevalent in E. coli [59], and the presence of E. coli strains that possess bla_CTX-M-65 in Ecuadorian broilers chickens [66] highlights a potential source for the horizontal transfer of the gene to poultry-associated S. Infantis isolates in Ecuador. ESBL-producing strains of S. Infantis not only occupy specific niches on the Latin American mainland but can also disseminate to other localities. This was demonstrated when MDR clones that contained bla_CTX-M-65 were isolated from foods and patients in the United States and shown to be closely related to an outbreak strain in Peru and Ecuador [24,25].

Taken together, the data from this study suggest that MDR strains of S. Infantis could potentially be transmitted from Ecuador to the Galapagos Islands through the movement of poultry-related inputs including day-old chicks, feed, personnel and other supplies between both places. This observation has significant implications from a public health standpoint, owing to the potential risk of transmission to humans and wildlife, and the potential difficulty in treating MDR infections. Indeed, suggestions have been made to construct hatcheries on the Galapagos Islands and to replace imported day-old chicks with locally available chickens [26]. This would reduce the reliance on externally sourced inputs, thus limiting potential incoming MDR NTS. However, further research is needed to pinpoint the source of MDR NTS in the Galapagos, and some of the tools described in this communication appear adequate to shed light on this need and thus provide optimism that measures can be developed and instituted to stem MDR NTS propagation in the Galapagos and elsewhere.

Twenty-two broiler chicken farms stocked with the Cobb 500 breed, located on the island of Santa Cruz, which represented 54% of the broiler chicken farms in the Galapagos (n = 41), were sampled from February 2016 to April 2017. All farms were privately owned and reared between 6000 to 10,000 broiler chickens. Chickens were fed with compound feed prepared in Ecuador’s mainland. No growth promotors nor antibiotics were used during the rearing period, and chickens were between 35 to 42 days of age at the time of sampling. Farms were sampled once, resulting in the collection of 22 samples during the study period. Sampling was performed by walking twice inside and along the entire length of one barn per farm while wearing sterilized disposable overshoes. The used overshoes were then aseptically removed and stored in sterile bags on ice, then sent to a laboratory in Quito (Ecuador) within 12 h. Salmonella was isolated from the samples using the ISO 6579-1:2007 protocol [67]. All isolates were serotyped according to the Kauffmann–White scheme [68].

The AMR phenotype of all confirmed NTS isolates was determined by evaluating MIC values with broth microdilution EUVSEC plates (Thermo Scientific, West Palm Beach, USA) according to the manufacturer instructions. The following antibiotics were evaluated: sulfamethoxazole (SMX), trimethoprim (TMP), gentamicin (GEN), ciprofloxacin (CIP), nalidixic acid (NAL), ampicillin (AMP), cefotaxime (CTX), ceftazidime (TAZ), tetracycline (TET), chloramphenicol (CHL), colistin (CL), azithromycin (AZM), tigecycline (TGC) and meropenem (MEM). E. coli ATCC 25922 was used as the quality control strain.

Epidemiological cutoff values (ECOFF) derived from EUCAST were used to determine the presence and level of phenotypic resistance in the Salmonella isolates [69]. For those antibiotics for which ECOFF values are not published (SMX, CL and AZM), clinical breakpoint values from the Clinical and Laboratory Standards Institute (CLSI) or previously recommended criteria were used [70]. All intermediate phenotypes obtained with breakpoint values from CLSI were considered as resistant since they are expected to harbor genetic determinants of antimicrobial resistance. Multi-resistant isolates were defined as those having resistance to three or more antibiotics.

Genomic DNA was extracted and purified using Qiagen DNeasy Blood & Tissue Kit (Qiagen Sciences Inc., Germantown, Frederick, MD, USA). DNA concentrations were measured using the Qubit^® Fluorimeter for quantification of double-stranded DNA and the Qubit dsDNA BR Assay kit (Invitrogen/Thermo Fisher Scientific, Waltham, MA, USA). Additional quality assessments were made using the NanoDrop 2000 UV-Vis (Thermo Fisher Scientific, Waltham, MA, USA) for determination of A_260/280 values. Whole-genome sequencing (WGS) was performed using the MiSeq platform (Illumina, San Diego, CA, USA) according to the harmonized FDA GenomeTrakr/CDC PulseNet protocol (https://www.cdc.gov/pulsenet/pathogens/protocols.html; last accessed on 8 January 2018).

Raw reads were submitted to GenomeTrakr and assembled de novo using their internal analysis pipeline. Genome assemblies were submitted to EnteroBase (https://enterobase.warwick.ac.uk/species/index/senterica; last accessed on 15 January 2020) in order to determine the 7-MLST profile of each isolate based on the Acthman scheme and confirm serotype designation using the SISTR algorithm [71]; https://lfz.corefacility.ca/sistr-app/; last accessed on 16 January 2020. Antimicrobial resistance genes were identified using the ResFinder database (https://bitbucket.org/genomicepidemiology/resfinder_db; last accessed on 17 January 2020, 90% ID and 60% gene coverage cutoffs). For mutational resistance, gyrA and parC sequences were extracted from genome assemblies using a custom Perl script and aligned to identify mutations. For plasmid analysis, the assembled contigs from each sequenced isolate were annotated using Prokka (V1.14.0) and Galileo AMR (https://galileoamr.arcbio.com/mara/; last accessed on 30 January 2020) and mapped to the reference plasmid pCVM44454 (GenBank Accession CP016413), isolated from clinical S. Infantis from the US [25]. Single-nucleotide variations among the S. Infantis Galapagos isolates were examined using two separate tools, namely SNVPhyl [72] and ParSNP [73] and compared to another ten isolates of S. Infantis obtained from the US as previously reported [25]. A fully assembled S. Infantis genome (4,710,675 bp, LN649235) was used as a reference for the SNVPhyl analysis, while the auto-recruit option was used for the choice of the reference genome for ParSNP analysis. The tools provided different degrees of stringency for evaluating relatedness among the different isolates. To further evaluate the relationship among the isolates, we used a highly discriminatory phage sequence typing (PST) tool [74], capable of exploring changes in the mobile accessory genome of Salmonella isolates, which usually contains prophages [75]. We identified the presence of prophage sequences in the genome of each strain using PHASTER (https://phaster.ca/; last accessed on 3 February 2020) and clustered related sequences CD-HIT-EST (http://weizhongli-lab.org/cd-hit/; last accessed on 3 February 2020) with sequence identity and length parameter cut off set at 99%, and the relationships were displayed as a phylogenetic tree by means of QIIME software (https://qiime2.org/; last accessed on 4 February 2020).