Genomic characterization of a clonal emergent Salmonella Minnesota lineage in Brazil reveals the presence of a novel megaplasmid of resistance and virulence

ABSTRACT

Salmonella Minnesota has emerged in Brazil as the predominant serovar in poultry and poultry products, along with Salmonella Heidelberg. To understand the emergence of Salmonella Minnesota over the last few years in Brazil, we performed a comparative analysis between 69 selected S. Minnesota genomes from Pathogen Detection database and 65 clonal emergent genomes isolated from Brazil. We demonstrate the presence of multidrug resistance genes against tetracycline [tet(A)], sulfonamide (sul2), and AmpC beta-lactamase (bla_CMY-2) in emergent genomes, along with the carriage of a megaplasmid of resistance and virulence (~210 kb), designated pESM (plasmid for emergent Salmonella Minnesota). pESM is an IncC/A2 plasmid predicted to increase S. Minnesota environmental tolerance to mercury (mer operon) and provide resistance to tetracycline and ampicillin due to the presence of tet(A) and bla_CMY-2, respectively. Moreover, pESM carries the yersiniabactin siderophore (high-pathogenicity island of Yersinia) related to the iron uptake. The temporal inference demonstrated that the most recent common ancestor dated from ~1978 and that the clonal emergent genomes carrying the pESM belong to a completely different lineage of S. Minnesota. Our results indicate that the presence of pESM likely contributes to the emergence of S. Minnesota and is precisely related to the successful spread of this particular clonal lineage in Brazil.

IMPORTANCE

Salmonella Minnesota has emerged in Brazil as one of the leading serovars related to human and animal infection, presenting high virulence and antibiotic resistance to drugs classified as the highest priority for clinical treatment in humans. This study performed whole-genome sequencing, temporal analysis, and phylogenetics to understand the genetic insights related to the emergence of Salmonella Minnesota in Brazil. Long-read sequencing has led to the identification and characterization of a unique megaplasmid carrying virulence, antibiotic resistance, and heavy-metal tolerance genes, which may play a central role in S. Minnesota’s successful emergence in Brazil and possibly worldwide. The potentially high transmissibility of this plasmid between clones and serovars represents a risk to public health since its acquisition may increase Salmonella’s fitness, virulence, resistance, and persistence. Understanding the genetic aspects related to the emergence of serovars can help devise measures to mitigate the spread of hazardous multidrug-resistant strains.

INTRODUCTION

Nontyphoidal Salmonella enterica (NTS) is one of the major pathogens associated with human gastroenteritis and diarrheal diseases worldwide, reaching about 75 million cases per year (1). Many of these infections are mostly self-limiting, but some of them develop bacteremia, leading to death in many cases, especially in the elderly, children, and immunosuppressed people (2). NTS can colonize the intestines of livestock asymptomatically, potentially contaminating the food production chain and promoting a risk of infection to humans through ingesting contaminated food (3). The virulence genes encoded by these bacteria can determine whether they cause bacteremia, in addition to increasing their survival and pathogenicity (4–7). When associated with antimicrobial-resistant genes, they may increase the bacteria’s ability to persist in the environment, as this makes them difficult to control (8). Many of Salmonella’s virulence and resistance genes are encoded on mobile elements such as plasmids, which can be acquired through horizontal gene transfer, a crucial vehicle of genetic information transmission between bacteria (9). These elements play a vital role in the bacteria’s evolution, intrinsically affecting fitness and Salmonella persistence in the environment (9).

Over 2,600 Salmonella serovars are known (10), although only a few are extensively described due to their high impact on public health. For example, some nations, including Brazil (11), have implemented control strategies only for specific serovars such as Salmonella Typhimurium and Enteritidis due to their high incidence and virulence during the 2000s (12). However, the control strategies for S. Enteritidis over the last 20 years may have led to the emergence and spread of serovars that were not previously common in Brazil since when the prevalence of one serovar decreases, other serovars emerge and assume their ecological niche (13). In this scenario, Salmonella Minnesota became one of the most reported serovars in Brazil recently, along with Salmonella Heidelberg, being isolated mainly from poultry and poultry production environments (14–17).

Studies involving Salmonella Minnesota in Brazil have reported multidrug resistance to AmpC beta-lactamase (bla_CMY-2), tetracycline [tet(A)], and sulfonamide (sul2), besides the presence of IncC replicon plasmids (14, 15, 17) and the yersiniabactin siderophore, related to persistence and biofilm formation (16). The emergency of S. Minnesota in Brazil dates from 2010 (12), and the high prevalence remains stable today. Although these characteristics represent a possible pattern related to this serovar’s emergence, the molecular and evolutionary processes have not yet been fully elucidated in Brazil.

In this context, genomic epidemiology studies are efficient tools that have become important for the phylogenetic and molecular evaluation of pathogens without sharing biological samples (18–20). Accurate genome sequencing of microorganisms is crucial to understanding bacterial evolution, speciation, and the impact of genomic diversity in pathogen disease (21). Long-read sequencing has been successfully applied in bacterial genomes with the advantage that the entire genome can be assembled in a single contig (22). While sort-read sequencing produces reads of 600 bp, long-read sequencing can generate reads of over 10 kb (23). This advantage can be used to close plasmids and megaplasmids entirely, given the opportunity to discover and characterize new plasmids.

Here, we aimed to bring insights into the genomic characteristics associated with the emergence of S. Minnesota in Brazil over the last few years. Hence, we compared 65 emergent S. Minnesota sequenced genomes from Brazil and 69 S. Minnesota genomes from the National Center for Biotechnology Information’s (NCBI’s) Pathogen Detection database isolated worldwide between 2000 and 2024. Resistance and persistence markers were determined, along with other genetic features, to understand genotypic factors associated with S. Minnesota’s emergence in Brazil and temporal analysis to presume when this lineage’s emergence began. Long-read sequencing revealed a novel megaplasmid in the emergent lineage of S. Minnesota, carrying virulence and antibiotic and heavy-metal-resistant genes. These findings significantly advance our understanding of S. Minnesota’s emergence in Brazil.

RESULTS

In silico typing, MLST, and cgMLST characterization

A total of 65 emergent Salmonella Minnesota genomes isolated in 2020 and 2021 were analyzed. The isolates were obtained from the Laboratory of Enterobacteria at Oswaldo Cruz Institute (LABENT-IOC) strain collection, which is responsible for receiving and analyzing Enterobacteria isolates from all over Brazil (https://ioc.fiocruz.br/en/labent). DNA from these isolates was extracted and sequenced using the Illumina NextSeq 1000 platform with paired-end 300 bp reads. Using in silico typing resources (see Materials and Methods), we confirmed the serovar and determined the multi-locus sequence typing (MLST) and core genome multi-locus sequence typing (cgMLST). To explore the evolutionary aspect of this clonal lineage from Brazil, we also included 69 genomes isolated from other countries in different years (see Materials and Methods). These genomes were obtained from the Pathogen Detection database and mainly isolated from poultry and poultry products (https://github.com/anamariadossantos/Suplementary_material.git, Table S1). In total, we analyzed 134 genomes.

All 65 emergent genomes from this study were confirmed as Salmonella Minnesota. Sixty-eight out of the 69 genomes obtained from the Pathogen Detection database were also confirmed as S. Minnesota, and one reported as S. Minnesota was indeed identified as Salmonella Schwarzengrund (GCA_010617425.1). Most emergent genomes, 95.4% (62/65), belonged to the ST548, and 4.6% (3/65) were undetected. About the selected genomes from the NCBI database, 92.7% (64/69) belonged to the ST548, 2.9% (2/69) were ST3088, 1.4% (1/69) were ST285, and 1.4% (1/69) were not detected. The S. Schwarzengrund genome belonged to the ST96.

The cgMLST results classification demonstrated a large variability of individual cgSTs, although 13 cgSTs were assigned to more than 1 genome. Most of these cases of duplicity happened to the emergent genomes from this study (https://github.com/anamariadossantos/Suplementary_material.git, Table S1). Interestingly, the cgST 3286369549 was assigned to four genomes: two emergent genomes from Brazil isolated in 2020, one from Saudi Arabia from 2020, and one from Germany isolated in 2019. The cgST 3708552441 was assigned to a genome from the United Kingdom isolated in 2017 and a genome from Saudi Arabia isolated in 2020. The cgST 1978129241 was assigned to genomes from three countries (Brazil, United Kingdom, and Saudi Arabia) and isolated from different years (2020, 2016, and 2021, respectively). cgST 11564871784 was assigned to two genomes from Brazil, isolated in 2016 (from the Pathogen Detection database) and 2020 (from this study). cgST 2688871353 was assigned to genomes from the USA isolated from different years, 2004 and 2007, respectively.

Plasmid replicons and AMR genotypes

A total of 21 incompatibility groups were identified among the 134 analyzed genomes (Fig. 1). IncA/C2 was present in 60.5% (81/134) of the analyzed genomes, followed by Col440I in 48.5% (65/134). Among the 65 Brazilian emergent genomes, 95.4% (62/65) of IncA/C2 and 69.2% (45/65) of Col440I. Furthermore, 69% (55/81) of the genomes that presented IncA/C2 also presented Col440I.

Fig 1.

Phylogeny of Salmonella Minnesota based on core genome single-nucleotide polymorphisms (SNPs) using Salmonella Minnesota ATCC 49284 as reference. SNP tree of all 134 genomes analyzed, comprising this study’s 65 sequenced emerging genomes. Heatmaps represent the presence or absence of yersiniabactin siderophore (green), antibiotics resistance genes (red), and plasmid replicons (purple). The tree was annotated using iTOL.

Twenty-four antimicrobial resistance genes were identified among the 134 analyzed genomes. The most prevalent were tet(A) (56.7%, 76/134), followed by sul2 (55.9%, 75/134), bla_CMY-2 (53.7%, 72/134), aph(3')-Ia (44.8%, 60/134), ant(3")-Ia (43.3%, 58/134), and qnrB19 (41%, 55/134). Among the genomes that displayed IncC/A2 incompatibility group, 91.4% (74/81) presented tet(A), 88.9% (72/81) presented sul2, 86.4% (70/81) presented bla_CMY-2, 71.6% (58/81) presented aph(3')-Ia, 67.9% (55/81) presented ant(3")-Ia, and 64.2% (52/81) presented qnrB19.

The merR gene, associated with mercury resistance regulation, was detected in 63.4% (85/134) of the genomes (https://github.com/anamariadossantos/Suplementary_material.git, Table S2), with 82 from clade I in the phylogenetic analysis and only 3 genomes from Brazil, Ireland, and the Netherlands in clade II presented it (Fig. 1).

Seventeen insertion sequences (IS) were identified among the 134 analyzed genomes (https://github.com/anamariadossantos/Suplementary_material.git, Table S3). The emergent genomes presented a similar pattern of IS, with ISNCY, ISAS1, IS21, IS256, IS3, IS4, and IS200/IS605 being the most prevalent. A new IS, with 851 bp length, was identified in 12.3% (8/65) of emergent genomes, which was also present in 13.0% (9/69) of the genomes selected from the Pathogen Detection database (https://github.com/anamariadossantos/Suplementary_material.git, Table S3).

Clonal emergent lineage of Salmonella Minnesota carries megaplasmid of resistance and virulence

The gene distribution in the clonal lineages of Salmonella Minnesota was almost homogeneous, mainly due to the high prevalence of yersiniabactin siderophore, mer operon, and AMR genes as bla_CMY-2, sul2, and tet(A) linked to the presence of an IncA/C2 replicon. To fully disclose the genomic content conferring antibiotic resistance and the fitness determinants of the clonal isolates, we performed long-read sequencing in one strain (CFSAN130365). This isolate carried the resistance genes bla_CMY-2 and tet(A), along with a yersiniabactin siderophore, mer operon, toxin/antitoxin system ccdA/B, repB, and a diverse of insertion sequences (ISNCY, IS3, IS21, IS256, ISAS1, and IS200/IS605) according to the previous short-read genomic analysis (see Materials and Methods). These characteristics were also identified in the other clonal isolates of Salmonella Minnesota (https://github.com/anamariadossantos/Suplementary_material.git, Tables S2 and S3). The long-read sequencing assembly resulted in three contigs total, including the chromosome (4,765,113 kb—circular) and two circular plasmids (210,709 and 2,989 kb, respectively).

The complete genome sequence of the strain CFSAN130365 confirmed that the megaplasmid (210,709 kb) (Fig. 2) belonged to the IncA/C2 incompatibility group, identified in almost all of the emergent genomes in this study. We performed BLAST search within NCBI and confirmed that it is a previously uncharacterized plasmid containing genetic elements involved with plasmid transfer, stress, virulence, and antibiotic resistance. Due to its exclusivity, we named this megaplasmid pESM (plasmid for emergent Salmonella Minnesota) following the same naming pattern used for the Salmonella Infantis megaplasmid (pESI) characterized by Aviv et al. (24). The BLAST screening also revealed high similarity (85.2%) with pSH-359.42, a Salmonella Heidelberg plasmid with 214,320 kb (accession NZ_CP080427.1), disposable in the NCBI database. To understand more about the genomic aspects of pESM, we performed a genetic environment study with pSH-359.42 and the pESI (plasmid for emergent Salmonella Infantis, accession NZ_CP047882.1) since it presents similar gene content and has been responsible for the emergent of Salmonella Infantis in Europe and the US in recent years (25–28) (Fig. 3). pESM harbors the genes trbA/B, involved with the plasmid conjugal transfer (29), nikAB genes required for relaxometry formation (30), parAB responsible for partitioning (31), umuCD needed for the SOS response for DNA damage (32), and the ccdAB related to the toxin/antitoxin system (33, 34). The main difference between the pSH-359.42, pESI, and pESM megaplasmid is the absence of tra genes responsible for conjugal transfer in pESM (Fig. 3). The tra genes are intimately related to type IV secretion machines (35), which are also absent in pESM but present in pSH-359.42 and pESI. The most remarkable similarity between these three megaplasmids is the block containing the yersiniabactin siderophore operon and a large hypothetical protein, which is a autotransporter-associated beta strand repeat-containing protein that can be used as a identifier for the three megaplasmids (Fig. 3). These analogous proteins may indicate that these megaplasmids have underwent similar selective pressures or have the same origin.

Fig 2.

Plasmid for emerging Salmonella Minnesota representation and genetic environment against megaplasmid pSH-342.59 of Salmonella Heidelberg. Megaplasmid with 210,709 bp carrying yersiniabactin siderophore (orange), mer operon (pink), and AMR genes (red). Arrows in navy blue indicate the insertion sequences along the plasmid. %GC is indicated by a bar chart over the plasmid where green %GC and purple %AT, automatically created by dnaplotter software.

Fig 3.

Genetic environment among pSH-359.42, pSEM, and pESI. Comparison of the megaplasmids from Salmonella Heidelberg (pSH-359.42,), Salmonella Minnesota (pESM), and Salmonella Infantis (pESI) comprising the function core genes, yersiniabactin siderophore, mer operon, resistance genes, and insertion sequences. Regions over 67% are indicated in gray shadow.

We also carried out the characterization of the second plasmid (2,989 bp) and confirmed through PlasmidFinder analysis (see Materials and Methods) that it is a Col440I plasmid carrying the qnrB19 gene (Fig. 4). The qnrB19 gene was next to the pspF gene, as previously identified in other studies (36–38). BLAST search matched 100% identity to pFA27 plasmid from Escherichia coli (accession KX452394.1) and pFDA1182898 from Salmonella enterica (accession CP158214.1). We performed a genetic environment study proving their high similarity (Fig. 4).

Fig 4.

QnrB19-plasmid representation and genetic environment comparison. QnrB19-plasmid part of Col4401 incompatibility group with 2,989 bp carrying the qnrB19 gene (red) predicted to confer quinolone resistance. Regions over 99% identity are indicated in gray shadow.

AMR, yersiniabactin siderophore, and mer operon are integrated into pESM

Genomic analysis revealed higher GC content in some parts of pESM, mainly the ones containing mobile genetic elements, such as transposons and insertion sequences (Fig. 3). Between these elements, we confirmed the presence of yersiniabactin siderophore, mer operon, and the AMR genes bla_CMY-2 and tet(A).

The yersiniabactin siderophore was identified in 92.3% (60/65) of this study’s emergent genomes and in 29% (20/69) of selected genomes from the Pathogen Detection database (Fig. 1).

The mer operon was between the recD gene and a Tn3/TnAs2 transposon. The bla_CMY-2 was next to an ISEcp1 insertion sequence, as previously described in other Salmonella bla_CMY-2-positive strains (17). The tet(A) gene related to tetracycline resistance was between an IS91 and a Tn3/TnAs2 (Fig. 3). The other difference observed between this pESM and pSH-359.42 is the absence of sul2 next to the tet(A) gene in pESM.

To gain a broader perspective on the presence of pESM in other genomes from this study, we performed a BLAST search against our complete pESM sequence. pESM displayed 244 genes. Because short-read sequencing is unable to elucidate the complete pESM sequence, we consider the entire presence of pESM in short-read sequenced genomes presenting over 90% similarity of our complete pESM genome along with plasmid core function genes repB (replication), parB (partitioning), nikAB (relaxase), and trb (conjugal transfer) (39). Genomes between 60% and 89% of the pESM genes were considered to have a partial pESM. Seventy genomes out of the 134 analyzed presented over 90%, 14 genomes between 60% and 89%, 6 genomes between 10% and 50%, and 43 genomes presented less than 6% of the pESM genes (https://github.com/anamariadossantos/Suplementary_material.git, Table S4).

Phylogenetics and temporal inference of clonal emergent Salmonella Minnesota

The generated single-nucleotide polymorphisms (SNPs) tree consisting of 3,160 SNPs showed two well-defined clades grouped according to the genome’s lineages. The SNP distance within these clades ranges from 0 to 193 (CFSAN130308). Sixty-four out of 65 (98.5%) emergent genomes from Brazil were grouped in clade I, and 84.6% (55/65) presented the complete pESM. The other genomes grouped in clade I were from the United Kingdom, Saudi Arabia, Chile, Australia, Germany, and the United Arab Emirates, isolated between 2015 and 2021. Only the genome from the United Arab Emirates presented a partial pESM, while the others presented a complete pESM.

The remaining genomes from the Pathogen Detection database are grouped in clade II. None of these genomes displayed a complete pESM. The genomes from the Netherlands (2010) and Ireland (2008) were the only ones in clade II to display a complete yersiniabactin siderophore along the merR gene and an IncC replicon, which is the primary pattern identified in the clonal lineage in clade I. However, these genomes presented ~65% of the pESM genes, which could represent an incomplete pESM or a pESM-like plasmid since they present the plasmid core function genes.

The time-scaled analysis conducted to ascertain the possible emergence date of the clonal lineage of S. Minnesota suggested that the most recent common ancestor (MRCA) was around 1978 (1978.55) (Fig. 5). This date matches the first split that separates the two clades between emerging clonal lineages and the other Salmonella Minnesota genomes selected from the Pathogen Detection database. The MRCA estimate demonstrates a variation of almost 40 years, with some clusters of the emerging clonal lineage overlapping with those of the other genomes. Support material for this analysis can be seen in the root-to-tip charts and parameters in the models’ convergence in Fig. S1 and S2. The effective size parameters were acceptable: µ = 107.66, σ = 358.25, and α = 116.46.

Fig 5.

Time-scaled inference of emergent and pre-emergent Salmonella Minnesota. (A) Root-to-tip distance against the sampling date with R² = 0.33. (B) Time-scaled tree of emergent and pre-emergent Salmonella Minnesota genomes based on the SNP tree from the selected genomes.

DISCUSSION

Salmonella Minnesota became one of Brazil’s most isolated serovars in poultry production, along with Salmonella Heidelberg. A few years ago, this position was occupied by Salmonella Enteritidis and Typhimurium, which were endemic in Brazil (12). Alikhan et al. (17) suggested that the sudden change of serovar is linked to the introduction of the mandatory vaccine against S. Enteritidis in Brazilian poultry production in 2003, which may have led to the rise of other serovars not commonly found, such as S. Minnesota and S. Heidelberg (17). In fact, long-term changes in the epidemiology of an existing population can lead to an increase in the incidence of other pathological agents (40), which would explain the emergence of S. Minnesota from a vaccine theory perspective.

However, much of this emergence can be explained by the possible selection of resistant and persistent serovars due to selective pressure caused by external factors, such as the increased use of antibiotics in animal production (41, 42). Likewise, plasmids carrying key factors for the survival of these bacteria can also influence their emergence (24). S. Minnesota presenting the complex tet(A)/sul2/bla_CMY-2 associated with IncA/C2 plasmids are becoming expansively common in Brazil, mainly in poultry and poultry products (12, 15, 17), indicating that, besides the possible selective pressures in this field, the increased transmissibility of these plasmids and its components can be responsible for the emergency success of S. Minnesota.

Here, we used whole-genome sequencing to evaluate genomic adaptative changes in a clonal emergent lineage of S. Minnesota isolated from Brazil in 2020 and 2021. To understand the evolutionary aspects of this clonal lineage, we compared Salmonella Minnesota isolated from other countries in different years (2000 to 2024). Furthermore, the long-read analysis demonstrated the presence of a unique and previously uncharacterized megaplasmid designated plasmid to emergence Salmonella Minnesota, carrying mobile elements encoding AMR genes and fitness determinants in most of the emergent clonal lineage analyzed in this study.

The pESM is a megaplasmid (210,709 bp) harboring resistance genes against AmpC beta-lactamase and tetracycline, besides mer operon (mercury resistance) and yersiniabactin siderophore (iron uptake). The presence of these specific components within mobile elements in the clonal population of S. Minnesota follows a pattern observed in other megaplasmids previously characterized in Salmonella, such as pESI (Salmonella Infantis) (24, 25, 27, 43) and pSH-359.42 (Salmonella Heidelberg) (accession NZ_CP080427.1). The varied composition of G + C within the mobile elements (Fig. 2) indicates that pESM possesses a mosaic structure, a characteristic of successful megaplasmids (44) defined by the insertion of several components with different functions, allowing the bacteria to survive a challenging environment (44). We suggest that pESM contributes significantly to the emergency success of Salmonella Minnesota in Brazil, considering its MDR genotype and the presence of specific genes that enhance bacteria persistence, virulence, and fitness. The circulation of this plasmid in the poultry environment may pose a relevant health risk, given its potential to transfer to other Salmonella or bacteria species. The presence of a toxin/antitoxin system such as ccdAB is an example of this plasmid stability and integrity (45). This genetic element is related to bacterial adaptability to stress conditions and the maintenance of plasmids and is also responsible for the ecological success of other Salmonella serovars, such as Salmonella Infantis pESI positive (46).

Yersiniabactin siderophore in Salmonella is primarily associated with poultry but is not restricted to Salmonella Minnesota. This system is part of the so-called Yersinia high-pathogenicity island (HPI) and is related to iron uptake and biofilm formation in Escherichia coli (47) and Klebsiella (48). We first identified the HPI in S. Heidelberg isolated from poultry in Brazil (49), and ever since, HPI has been widely reported in studies with recent MDR S. Minnesota also isolated from poultry environment (16). HPI in Salmonella is predicted to increase environmental persistence and to influence the expression of virulence genes not associated with the island itself (50), which explains the persistence of emerging S. Minnesota in poultry environments. Our late studies showed the first description of pESI in S. Infantis isolated from Brazil (51) and a large-scale genomic analysis in which we described the poultry environment as a central point for pESI dissemination (42). We also noticed that the pESI in Brazil does not seem to spread as quickly as in other countries, given its low prevalence in genomes isolated from Brazil (42). The main reason may be that the adaptation of Salmonella Infantis undergoes the predominance of other emerging serovars, such as Minnesota and Heidelberg in Brazil (12, 13), or the possible dominance of pESM, which could be preventing the perpetuation of other megaplasmids in poultry environment. Given the prevalence of pESM identified in the emerging genomes in this study, we suggest that once S. Minnesota acquires the pESM, it could likely quickly spread in the local population, given its fitness components (35). The spread of AMR genes in combination with yersiniabactin siderophore and mer operon has been carried together not only by pESM (IncA/C2) but also by IncHI2-St2 megaplasmids in Salmonella Schwarzengrund and Newport isolated from Brazil (52). This pattern indicates that different selective pressures are probably being applied simultaneously in Brazil’s poultry production, leading to the selection and acquisition of megaplasmids by Salmonella, as we observed in this study.

Despite current Brazilian legislation on nontyphoidal Salmonella prohibiting the use of antimicrobial drugs as growth promoters in the production chain (53), antibiotics continue to be widely employed prophylactically in traditional medicine practices and animal production (54), further exacerbating the selection pressure for MDR strains. The tet(A) gene was the most prevalent among our S. Minnesota genomes in this study, which seems to be compatible with the Brazilian reality since tetracycline is still the most widely used antibiotic for treating several diseases in Brazil’s veterinary and animal production (54). Also, the high concentration of tetracycline leads to increased conjugation of IncC plasmids (55), which may explain the high presence and likely increased transmissibility of pESM among the genomes for the emerging clonal lineage of S. Minnesota. The sul2 and bla_CMY-2 were the second most prevalent among our S. Minnesota genomes, following the pattern already observed in several genomic epidemiology studies in Brazil (15, 17). These findings highlight the need to continue monitoring acquired resistance to AmpC beta-lactams and sulfonamides to prevent their spread in the animal and food production chain.

The SNP phylogenetic tree demonstrated two well-defined clades that seem to be separated by the pESM presence or absence. The phylogenetic distribution and temporal inference show that a foreign strain was likely introduced in Brazil and underwent clonal expansion, given the current nature of the emerging clonal genomes represented in clade I (Fig. 1). The early genome to present complete pESM was from Australia in 2015 and shares the same monophyletic group as a genome from Brazil isolated in 2018. The Australian genome isolation source was clinical, so this introduction could have happened through humans with travel history, considering that Brazil and Australia do not share a long-term history of exporting animal products (56). On the other hand, given the phylogenetic proximity between the Brazilian genomes and Saudi Arabia and the United Arab Emirates genomes, it is possible that pESM-positive S. Minnesota was imported from Brazil and underwent clonal expansion in these countries. Since Saudi Arabia and the United Arab Emirates are Brazil’s largest chicken meat importers (56), and they displayed the same cgMLST and year of isolation, this theory becomes even more accurate. pESM-positive S. Minnesota was also identified in genomes from the United Kingdom, Germany, and Chile, which share the same cgMLST. It demonstrates the capacity of possible pESM dissemination among the production chain regardless of national borders, affecting animal production and representing a risk to human health through food contamination.

Three genomes from a different lineage of the clonal emerging genomes presented a partial pESM. These genomes were isolated from the Netherlands (2010), Ireland (2008), and Brazil (2010). The mosaic structure of pESM allows new genes to be inserted according to the selective pressure while maintaining its backbone, which already happens with pESI (24, 27). Thus, the partial potential presence of pESM in these genomes may represent an incomplete megaplasmid, which has not yet undergone all selective pressures necessary for acquiring other genes, or an IncC plasmid with similar characteristics to the pESM. IncC plasmids carrying AMR genes such as tet(A), sul1, and bla_CMY-2 next to the IS91 and ISEcp1, respectively, have already been reported in S. Heidelberg in meat imported from Brazil to the Netherlands (57). Also, Alikhan et al. (17) demonstrated the co-carriage of these IncC plasmids in S. Heidelberg and S. Minnesota in Brazil. Since these genomes from the Netherlands and Ireland belong to a different lineage, it is unlikely that Brazil’s clonal expansion of the emerging S. Minnesota pESM positive has originated from them.

The majority of emerging genomes exhibited the presence of qnrB19, which was associated with the Col4401 incompatibility group (Fig. 3). The occurrence of Col4401 and the IncC replicon has been documented previously and appears to be a recurring characteristic in emerging Salmonella species (58), serving as the predominant plasmid-mediated quinolone resistance (PMQR) gene observed in Brazil (59). We supported this hypothesis by showing that only recent Salmonella Isangi isolates from Brazil (2020–2021) exhibited the qnrB19 gene (60). Considering the emergence and persistence of S. Minnesota, it is increasingly probable that these plasmids are being exchanged within the environment from a reservoir of other Enterobacteria, similar to what has been observed with other members of Enterobacteriaceae (61, 62). The increasing PMQR gene dissemination underscores the urgency of monitoring and surveillance efforts concerning MDR Salmonella Minnesota and MDR plasmid replicons.

Our findings highlight the presence of a singular megaplasmid designated pESM in the emerging clonal lineage of Salmonella Minnesota in Brazil. The pESM components, as the toxin/antitoxin system, are the key to maintaining this megaplasmid in the S. Minnesota population once it is acquired. Furthermore, the extensive use of antibiotics in animal production, added to heavy metals contamination such as mercury compounds in agriculture, may favor the selection of pESM-positive S. Minnesota in poultry environments, further representing a risk to human public health. Given the current nature of the clonal lineage presented in this study, it could quickly replace the local population of S. Minnesota in a short period, meaning that critical imperative for ongoing surveillance to monitor and prevent MDR S. Minnesota is necessary. Our results also underline the effectiveness of combining genomic data from different countries and sources to gather information from a global perspective, also assisting with a One Health approach. Overall, updates on the presence of new plasmids and their components are essential to help surveillance zoonotic diseases and outline control measures worldwide.

MATERIALS AND METHODS

Bacterial isolates, short-read whole-genome sequencing, assembly, and quality filtering

The 65 bacterial strains were obtained from the Brazilian Reference LABENT-IOC-Salmonella collection. DNA extraction was performed using the DNeasy Blood & Tissue Kit (QIAGEN, Germany) following the manufacturer’s instructions. Libraries were generated from genomic DNA with the Illumina DNA prep and then sequenced on the NextSeq 1000 platform with the P2 300-bp paired-end reagent cartridge (Illumina, San Diego, CA). Genome assembly was performed using the Unicycler software v0.5.0 (63). The quality of draft genomes was evaluated through QUAST v5.2.0.0 (64), and sequence annotations were obtained with Prokka software v1.14.6 (65).

Acquisition and selection of additional Salmonella Minnesota genomes

We obtained the genomes from the Pathogen Detection database at the National Center for Biotechnology Information. Initially, utilizing the browser filtering tools available as of 10 June 2024, we selected genomes from isolates collected from various sources among poultry, humans, and the environment. Subsequently, we excluded genomes within the same SNP cluster to mitigate clonality and retained those from diverse countries and isolation year. Thus, our selection criteria prioritized the SNP cluster, followed by country, isolation, and quality metrics such as N50 and contig number. Following this rigorous screening process, we selected 69 genomes isolated between 2000 and 2024 as the complementary genomic data set (https://github.com/anamariadossantos/Suplementary_material.git, Table S1).

In silico short-read genome characterization

Salmonella in silico typing resource v1.1.1 (66) was used to confirm the serovar of Salmonella genomes in the data set. StarAMR software v0.9.1 (https://github.com/phac-nml/staramr) was used to perform MLST and cgMLST v2.19.0 (67). We identified plasmids with the Plasmidfinder database (68). We used ABRicate software v1.0.1 (https://github.com/tseemann/abricate) to identify virulence genes against the Virulence Factor Database (http://www.mgc.ac.cn/VFs/), and used the Resfinder database (69) to identify resistance genes. Minimum nucleotide identity and coverage thresholds of 90% and 60%, respectively, were employed for all analyses. To detect and evaluate metal resistance genes, we used AMRfinderplus software v3.11.14 (70). BLAST Ring Image Generator was used to compare plasmid sites to the pSH-359.42 (accession NZ_CP080427.1), a Salmonella Heidelberg IncC plasmid.

Phylogenetic analysis and time-scaled inference

The detection and prediction of whole-genome variants were executed utilizing rapid haploid variant calling and core genome alignment—Snippy v4.6.0 (https://github.com/tseemann/snippy). The ATCC 49284 (accession CP019184.1) strain is employed to identify and eliminate recombinant regions within the core genome alignment to enhance the accuracy of phylogenetic reconstructions. Additionally, Phastaf v0.1.0 (https://github.com/tseemann/phastaf) and Barrnap v0.9 (https://github.com/tseemann/barrnap) were utilized for the identification and masking of phage regions and ribosomal RNA, respectively, in the reference genome.

The final high-quality core genome SNPs were extracted from the alignment file using SNP-sites v2.5.1 (71). Subsequently, Iqtree v2.0.3 was utilized to generate a maximum likelihood phylogeny from this core genome alignment, employing a GTR + F + I + G4 substitution model with 1,000 bootstrap replications to support the tree nodes. The resulting tree was visualized using iTOL (72).

BactDating v1.0.1 software in Rstudio was employed for time-scaled inference on the phylogeny. The final labeled tree from Gubbins, with recombination correction, served as input for inferring sequence variation in the core genome. The Markov chain Monte Carlo was run with 10⁶ interactions, ensuring convergence and mixing chains with µ = 107.66, σ = 358.25, and α = 116.46, as calculated using the R coda package, all exceeding 100 as recommended by the authors (73).

Long-read whole-genome sequencing and complete genome assembly

A single genome from our selection underwent long-read sequencing (CFSAN130365) with PacBio Revio. The isolate was selected for sequencing based on the AMR diversity and the presence of key factors that indicated the presence of a plasmid (IncC plasmid replicon, ccdAB, RepB, trb, and nikAB), besides other fitness components previously described in megaplasmids, such as the yersiniabactin siderophore and the mer operon. The high molecular weight extraction of DNA was proceeded using the commercial kit MagAttract (QIAGEN) following the manufacturing instructions. The initial quantification was proceeded with Nanodrop and Qubit (DNA-kit high sensitivity), and the Pippin Pulse was used to analyze the DNA integrity. The library was prepared with the SMRTell Prep Kit 3.0. The long-read assembly was made using Unicycler v0.5.0 (63). Genomes were annotated with Prokka v1.14.6 (65).

Plasmids characterization and genetic environment

The complete assembled plasmids after the long-read sequencing underwent characterization using the Plasmidfinder database within the ABRicate software. We confirmed the incompatibility groups and proceeded with the plasmid’s annotation using Prokka v1.14.6 (65). Plasmids visualization was made using Dnaplotter from Artemis software (74). We used BLAST against the NCBI database (75) to identify previously characterized plasmids against our plasmid’s complete genomes. The genetic environment was accessed using Easyfig software (76).

pESM screening among the emergent clonal lineage of Salmonella Minnesota

We used pESM’s Prokka annotation to identify the total number of genes and create a database within the ABRicate software following the GitHub instructions. This database was used to perform a massive screening in all Salmonella Minnesota genomes analyzed in this study and predict the pESM percentage in these genomes. We also evaluated the presence of critical genes, such as repB, ccdAB, nikAB, and trb, to confirm plasmid presence along with this massive screening.

DATA AVAILABILITY

Whole-genome sequence data are available in GenBank under the BioProject number PRJNA1083980. The long-read sequence data can be accessed in the same BioProject, under the accessions CP168003, CP168004 (pESM), and CP168005 (QnrB19-plasmid). The subsequent information about the sequenced genomes is in the supplemental material.