Phylogenetic structure of *Salmonella Enteritidis* provides context for a foodborne outbreak in Peru

Salmonella spp. are a major cause of morbidity and mortality, particularly among children in developing countries, with an estimated 93.8 million global non-typhoidal salmonellosis cases annually, most related to foodborne transmission. Person-to-person transmission is uncommon; thus, animal-source foods are considered the principal exposure route. Salmonella enterica subsp. enterica serovar Enteritidis (S. Enteritidis) commonly colonizes the reproductive tract of chickens and other birds, often asymptomatically, and uses eggs as a main vehicle of infection via internal and external contamination. Additional factors such as cross-contamination of foods and asymptomatic infected food handlers contribute to human infections. Since the 1980s, S. Enteritidis has replaced S. Typhimurium as the primary global cause of salmonellosis, with high isolation rates reported in Latin America. Whole genome sequencing (WGS) is widely used for foodborne pathogen surveillance, enabling real-time outbreak detection and phylogenetic analysis to study emergence, spread, and transmission. This study explored the phylogenetic structure of Peruvian S. Enteritidis strains collected over nearly two decades (2000–2018) and examined their relationship to a 2018 outbreak in Callao, Peru.

The paper references global and regional data showing high incidence of Salmonella infections and the predominance of S. Enteritidis in many areas since the 1980s. Prior studies have outlined mechanisms of egg contamination by S. Enteritidis and factors affecting its survival in foods like mayonnaise. WGS-based phylogenetics has been successfully applied to investigate S. Enteritidis emergence and transmission in Europe, Latin America, and elsewhere, including identification of lineages correlating with distribution networks and differing clinical presentations. Reports also describe the role of prophage regions such as SE20 in pathogenicity and the global expansion of S. Enteritidis, and the high clonality of this serovar detected via MLST and SNP analyses.

Sampling and population: A total of 180 Salmonella spp. strains received by the National Reference Laboratory of Enteropathogens of the Instituto Nacional de Salud (INS) from different parts of Peru (2000–2018) were included. Nine strains were isolated during investigation of a 2018 outbreak in Callao: five from patients with acute diarrheal disease after dining at a restaurant, and three from the suspected source (mayonnaise prepared at the restaurant). Ethics: Use of human biological material was approved by the INS Institutional Research Ethics Committee (Register No 20516–2018) with complete anonymization of patient information. Culturing and biochemical confirmation: Strains were recovered on selective media. Each was inoculated in Trypticase Soy Broth (37 °C, 6–8 h), then plated on Salmonella Shigella Agar (37 °C, 18–24 h). Salmonella genus confirmation used conventional biochemical tests. Serotyping employed O and H agglutination antisera following the Kauffmann-White scheme. Library preparation and sequencing: All 180 strains serotyped as S. Enteritidis were selected. DNA was extracted with DNeasy Blood & Tissue kit (Qiagen), quantified by Qubit 3.0. Libraries were prepared using Nextera XT (Illumina) and sequenced on Illumina MiSeq. Genomic data processing: Read quality was assessed with FastQC v0.11.5; adapters and low-quality bases trimmed with Trimmomatic v0.38. De novo assemblies were generated using A5-miseq v20160825. Genus identification and contaminant contig detection were performed using Kraken v1.0 with Minikraken. Inclusion criteria required <300 contigs and ≥92% contigs belonging to Salmonella spp. Sequence types were determined with MLST v2.10. Phylogenetics: S. Gallinarum strain 287/91 (GenBank AM933173.1) served as outgroup. Genome alignments were produced with Parsnp v1.2; locally colinear blocks (LCBs) and SNPs were extracted using HarvestTools v1.2. Maximum likelihood phylogeny was inferred with RAxML v8.0 under the GTR+G model with 1000 bootstrap replicates. Recombinant regions were identified and accounted for using ClonalFrameML v1.1. Population structure was analyzed with hierBAPS (R implementation), including singleton SNPs, maximum depth=2, and up to 20 populations. Visualization was done via Microreact (project: https://microreact.org/project/9G7UFQLRAwiS4S1J13M69C20). Virulence factors and accessory genome: BLAST searches against VFDB identified pefB, shdA, spvB, ssa and regions of the S. Enteritidis ST64B-like prophage SE20. Coding sequences were predicted using Prodigal v2.6.3. Homologous gene identification used BLAST with >90% identity and >60% alignment coverage relative to reference genes. Annotation code: http://github.com/OrsonMM/Blast-score-ratio-for-genomics. Genomic data deposited in GenBank (BioProject PRJNA552561).

- All 180 strains, including outbreak-related isolates, were identified as S. Enteritidis by conventional microbiology and serotyping (1,9,12: g,m:—). - WGS metrics: average 3.3 million reads, 77 contigs, 52.0% GC, mean genome size 4.8 Mb. - MLST: all sequenced strains were sequence type ST11. - Core genome alignment: 65.7% core; 6208 SNPs within >3.2 Mb LCBs. Low recombination: R/θ=0.07; r/m=0.26. - Phylogeny: Two well-supported clades identified. Clade A contained 143 strains (including outbreak strains); Clade B contained 37 strains. hierBAPS split the dataset into four populations (all with 100% bootstrap). Clade A corresponded to population 1; Clade B comprised populations 2–4. - Population composition and timelines: • Population 1 (green): clinical, environmental, and animal sources (2001–2013); notably, 48.6% of this population were strains from 2001–2002; later years were rare. • Population 2 (orange): two environmental strains from 2017. • Population 3 (blue): predominantly clinical (94.3%) plus environmental strains (2000–2017). • Population 4 (red): clinical, food, and environmental strains (2008–2018); included the 2018 outbreak strains. - Virulence and accessory genes: • pefB present in 98.9% of strains; spvB in 84.4%; ssa in 50.6%. • shdA absent in 91.7% overall but present in the 2018 outbreak strains. • Prophage SE20 regions detected in 80.6% of strains, present across all four populations. - Outbreak linkage: Clinical and food (mayonnaise) isolates from the 2018 Callao outbreak clustered together within population 4, implicating mayonnaise as the infection vehicle and highlighting the shdA-associated virulence of the outbreak strain.

The Peruvian S. Enteritidis isolates over nearly two decades displayed high clonality, with all strains belonging to ST11 and exhibiting a low recombination rate, consistent with prior reports for this serovar. Maximum likelihood phylogeny revealed two distinct clades, suggesting two circulating lineages in Peru. Population structure analysis identified four discrete populations without segregation by source (clinical, food, environmental, animal), indicating intermingled transmission across sources and regions. Within population 4, clinical outbreak isolates were phylogenetically linked to mayonnaise-derived isolates, incriminating mayonnaise as the vehicle of transmission during the 2018 Callao outbreak. The presence of the shdA gene—an intestinal colonization adhesin associated with prolonged fecal shedding and expanded host range—in outbreak strains supports increased virulence and persistence. Population 1 was enriched for older isolates (2001–2002), suggesting a lineage that circulated more prominently early in the study period and persists at low frequency. Population 2 contained only recent environmental isolates, potentially reflecting higher variation outside animal hosts. Prophage SE20 regions were common across populations, aligning with previous evidence that SE20 contributes to pathogenicity and may have facilitated the global emergence of S. Enteritidis. The lack of a clear geographic distribution pattern for lineages across Peru suggests widespread, overlapping transmission networks similar to observations in other South American countries.

Peruvian S. Enteritidis isolates from 2000 to 2018 segregate into two well-defined clades, indicating two distinct circulating lineages. Four populations with particular characteristics were identified, reflecting discrete groups of clinical strains transmitting across multiple geographic areas without clear spatial structuring. The 2018 Callao outbreak was associated with a highly virulent strain carrying the shdA gene and linked to contaminated mayonnaise as the vehicle. Future work should include pre-2000 isolates and broader geographic sampling to clarify temporal dynamics, assess the historical prevalence of prophage SE20, and refine understanding of lineage spread and virulence determinants.

- Lack of isolates prior to 2000 and incomplete coverage of all 25 regions of Peru limited temporal and geographic representativeness. - Confirmation of the hypothesized association between emergence of S. Enteritidis and prophage SE20 requires analysis of older (pre-1995) isolates to verify the majority absence of SE20 regions. - Source attribution precision is constrained by limited environmental and food sampling within certain populations (e.g., population 2).