Foodborne Origin and Local and Global Spread of Staphylococcus saprophyticus Causing Human Urinary Tract Infections

Abstract

Staphylococcus saprophyticus is a primary cause of community-acquired urinary tract infections (UTIs) in young women. S. saprophyticus colonizes humans and animals but basic features of its molecular epidemiology are undetermined. We conducted a phylogenomic analysis of 321 S. saprophyticus isolates collected from human UTIs worldwide during 1997-2017 and 232 isolates from human UTIs and the pig-processing chain in a confined region during 2016-2017. We found epidemiologic and genomic evidence that the meat-production chain is a major source of S. saprophyticus causing human UTIs; human microbiota is another possible origin. Pathogenic S. saprophyticus belonged to 2 lineages with distinctive genetic features that are globally and locally disseminated. Pangenome-wide approaches identified a strong association between pathogenicity and antimicrobial resistance, phages, platelet binding proteins, and an increased recombination rate. Our study provides insight into the origin, transmission, and population structure of pathogenic S. saprophyticus and identifies putative new virulence factors.

Keywords: MRSA and other staphylococci; One Health; UTIs; WGS; antimicrobial resistance; bacteria; clone; food; food safety; livestock; origin; population structure; urinary tract infections; virulence; whole-genome sequencing; zoonoses.

Figure 1

Maximum-likelihood tree of Staphylococcus saprophyticus isolates recovered from human infections and colonization globally, 1997–2017. The tree was constructed by using 9,134 SNPs without recombination. Among analyzed isolates, 321 were recovered from UTIs, 12 from blood, and 4 from colonization. Each node represents a strain; nodes with identical color belong to the same lineage. The assembled contigs were mapped to the reference genome S. saprophyticus ATCC 15305 (GenBank accession no. AP008934.1; black star). Polymorphic sites resulting from recombination events in the single-nucleotide polymorphism (SNP) alignments were filtered out by using Gubbins version 2.3.4 (12). Maximum likelihood tree was reconstructed by using RAxML version 8.2.4 (https://github.com/stamatak/standard-RAxML). We performed generalized time-reversible nucleotide substitution model with gamma correction with 100 bootstraps random resampling for support. We visualized the tree by using Interactive Tree of Life (iTOL; https://itol.embl.de). Black triangles represent isolates fully sequenced by using the long-read nanopore technologies and used as reference to estimate r/m in the respective lineage. Cream color represents clusters G1, G2, G3, G4, and S1, which had dissemination and transmission in same country and in different countries. The outer ring represents isolates’ country of origin; blocks with identical color represent isolates from the same country. Of note, cluster G4 contains a pair of isolates collected in 2016 that had only 10 SNPs difference; one is a blood isolate from Barcelona, Spain (KS266) and the other is a UTI isolate recovered in Lisbon, Portugal (KS135). Scale bar indicates number of substitutions per site. UTI, urinary tract infection; r/m, recombination to mutation ratio.

Figure 3

Maximum-likelihood tree depicting the genetic relatedness of Staphylococcus saprophyticus isolates belonging to clonal lineage G recovered from human infections or slaughterhouse contamination, Portugal, 2016–2017. Each node represents a strain. The tree was visualized by using Interactive Tree of Life (iTOL; https://itol.embl.de). Clusters highlighted in cream in the innermost ring represent admixture of isolates in clusters G5–G11, which were recovered from different sources and are closely related by single-nucleotide polymorphism. Ring 1 represents genetic relatedness of isolates recovered from different sites in the slaughterhouses and those recovered from infection in the community. Ring 2 depicts the isolates recovered from different hospitals. Scale bar indicates nucleotide substitutions per site.

Figure 4

Growth rate of Staphylococcus saprophyticus clonal lineages in tryptic soy broth (TSB) and in different concentrations of female sex hormones. All assays were performed in triplicate and each experiment was repeated 3 times. A) Growth rate of S. saprophyticus strains in different concentrations of progesterone. First panel represents growth rate in TSB at 37°C; isolates belonging to lineage S grew significantly faster (p = 0.0007) than isolates in lineage G in TSB without hormones. However, no statistically significant difference in the growth rate of either lineage was noted in physiologic (2.0–200 ng/mL) and higher concentrations of progesterone. B) Growth rate of S. saprophyticus strains in TSB (first panel) and different concentrations of estradiol. Lineage S isolates grew faster in physiologic concentrations (350 pg/mL–350 ng/mL) and higher of estradiol, suggesting that this lineage is better adapted to the hormone-rich environment of the urine and the vagina than lineage G. Error bars indicate 95% CIs; horizontal lines indicate medians. C) Growth rate mean values of S. saprophyticus strains in progesterone and estradiol.

Figure 2

Phylogenomic analysis and distribution of Staphylococcus saprophyticus isolates collected from human infections and a slaughterhouse, Portugal, 2016–2017. A) Maximum-likelihood tree of 232 isolates from human infections or slaughterhouse contamination. The tree was constructed by using 14,110 single-nucleotide polymorphisms (SNPs) without recombination. Each node represents a strain; nodes with identical color belong to the same lineage. The assembled contigs were mapped to the reference genome S. saprophyticus ATCC 15305 (GenBank accession no. AP008934.1; black star). SNPs generated from each genome were concatenated to single alignment corresponding to position of the reference genome. Polymorphic sites resulting from recombination events in the SNP alignments were filtered out by using out by using Gubbins version 2.3.4 (12). Tree was reconstructed by using RAxML version 8.2.4 (https://github.com/stamatak/standard-RAxML). The generalized time-reversible nucleotide substitution model with gamma correction was performed with 100 bootstrap random re-samplings for support. The tree was visualized by using Interactive Tree of Life (iTOL; https://itol.embl.de). The clusters highlighted in cream represent admixture of isolates recovered from different sources that are closely related by SNPs in clusters G5–G11 and S2–S4. The inner ring (ring 1) represents genetic relatedness of isolates recovered from different sites inside the slaughterhouses and those recovered from infection in the community. The center ring (ring 2) identifies the isolates recovered from different hospitals. The outer rings (rings 3–8) represent the distribution of 6 genes that convey antimicrobial resistance. Scale bar indicates nucleotide substitutions per site. B) Source-based distribution of S. saprophyticus isolates in the lineage G and lineage S. Lineage G consisted isolates from infections, colonization, and contamination. Almost all (97%) isolates in lineage S are from human colonization and infection.

Figure 5

Genetic determinants that contribute to the distinction of clonal lineages and lifestyle of Staphylococcus saprophyticus. The graph displays determinants that contribute (A) and mediate (B) adaptation of S. saprophyticus to either infection or contamination. We used the genome-wide association study (GWAS) method to identify genetic factors by using 2 association comparisons: lineage G versus lineage S and human infection versus surface contamination. We used the pairwise comparison and included a core-SNP phylogenetic tree without recombination to remove the lineage effect in the analysis. Hits with Benjamini Hochberg corrected p<0.05 and odds ratio >1 were considered statistically significant. We grouped the identified genes into biologic functions based on gene annotation predicted by Prokka (https://vicbioinformatics.com/software.prokka.shtml). Some genetic factors that were associated with infections and contamination also were associated with the lineages despite subjecting the GWAS to lineage correction.

Figure 6

A proposed model for the dissemination and transmission of Staphylococcus saprophyticus in the community. The arrows represent the dissemination and transmission of S. saprophyticus isolates that belonged to lineage G (green) and lineage S (yellow). Lineage G S. saprophyticus strains are of animal origin and enter the slaughterhouse through production animals, such as pigs, persist on the equipment, and contaminate the meat in processing chain. Lineage G strains could enter the community through contaminated meat and workers colonized in the slaughterhouse. Lineage S strains most likely are of humans and primate origin and probably are disseminated by person-to-person contact within the community.