Public health implications of Yersinia enterocolitica investigation: an ecological modeling and molecular epidemiology study

BACKGROUND

Yersinia enterocolitica has been sporadically recovered from animals, foods, and human clinical samples in various regions of Ningxia, China. However, the ecological and molecular characteristics of Y. enterocolitica, as well as public health concerns about infection in the Ningxia Hui Autonomous Region, remain unclear. This study aims to analyze the ecological and molecular epidemiological characteristics of Y. enterocolitis in order to inform the public health intervention strategies for the contains of related diseases.

METHODS

A total of 270 samples were collected for isolation [animals (n = 208), food (n = 49), and patients (n = 13)], then suspect colonies were isolated and identified by the API20E biochemical identification system, serological tests, biotyping tests, and 16S rRNA-PCR. Then, we used an ecological epidemiological approach combined with machine learning algorithms (general linear model, random forest model, and eXtreme Gradient Boosting) to explore the associations between ecological factors and the pathogenicity of Y. enterocolitis. Furthermore, average nucleotide identity (ANI) estimation, single nucleotide polymorphism (SNP), and core gene multilocus sequence typing (cgMLST) were applied to characterize the molecular profile of isolates based on whole genome sequencing. The statistical test used single-factor analysis, Chi-square tests, t-tests/ANOVA-tests, Wilcoxon rank-sum tests, and Kruskal-Wallis tests.

RESULTS

A total of 270 isolates of Yersinia were identified from poultry and livestock (n = 191), food (n = 49), diarrhoea patients (n = 13), rats (n = 15), and hamsters (n = 2). The detection rates of samples from different hosts were statistically different (χ² = 22.636, P < 0.001). According to the relatedness clustering results, 270 isolates were divided into 12 species, and Y. enterocolitica (n = 187) is a predominated species. Pathogenic isolates made up 52.4% (98/187), while non-pathogenic isolates made up 47.6% (89/187). Temperature and precipitation were strongly associated with the pathogenicity of the isolates (P < 0.001). The random forest (RF) prediction model showed the best performance. The prediction result shows a high risk of pathogenicity Y. enterocolitica was located in the northern, northwestern, and southern of the Ningxia Hui Autonomous Region. The Y. enterocolitica isolates were classified into 54 sequence types (STs) and 125 cgMLST types (CTs), with 4/O:3 being the dominant bioserotype in Ningxia. The dominant STs and dominant CTs of pathogenic isolates in Ningxia were ST429 and HC100_2571, respectively.

CONCLUSIONS

The data indicated geographical variations in the distribution of STs and CTs of Y. enterocolitica isolates in Ningxia. Our work offered the first evidence that the pathogenicity of isolates was directly related to fluctuations in temperature and precipitation of the environment. CgMLST typing strategies showed that the isolates were transmitted to the population via pigs and food. Therefore, strengthening health surveillance on pig farms in high-risk areas and focusing on testing food of pig origin are optional strategies to prevent disease outbreaks.

Prevalence, antimicrobial resistance, and genotype diversity of Salmonella isolates recovered from retail meat in Hebei Province, China

This study investigated the prevalence of Salmonella in 210 retail meat samples (105 raw chicken and 105 raw pork) collected from supermarkets and wet markets in 13 areas of Hebei Province, China, from June to October 2018. Whole-genome sequencing was performed on all 125 Salmonella isolates to investigate their genetic relationship. Core genome multilocus sequence typing of 77 representative isolates was used to further elucidate the genetic relatedness among the Salmonella isolated from retail meat. The mean detection rate of Salmonella in all samples was 59.5% (125/210). The prevalence of Salmonella was 53.3% (56/105) in chicken and 65.7% (69/105) in pork. Chicken and pork samples collected in July had the highest detection rate of Salmonella among the sampling months. The isolates were assigned to 19 serotypes, with S. Derby, S. London, and S. Thompson being the most frequent serotypes. Resistance to tetracycline (primarily used for the treatment of bacterial infections) was observed in 89.6% of the isolates, and 84.0% were resistant to doxycycline (also a tetracycline antibiotic) or gemifloxacin (commonly used for clinical treatment of human acute bronchitis). More than 80% of the isolates were multidrug resistant. A total of 21 sequence types were identified. Sequence type 40 (ST-40), the predominant genotype among all isolates, was found only in pork; the sequence types of chicken isolates were more diverse. A total of 58 different antibiotic resistance genes (ARGs) were detected in the 125 isolates. Most types of ARGs were associated with aminoglycoside and β-lactam resistance. Nevertheless, the tetracycline resistance gene tet(A) was the most frequently occurring ARG in all isolates at 78.4%. Multiple isolates of ST-26 contained 20 ARGs. All isolates of ST-40 were divided into two clusters, with at least 160 allelic differences between them. The findings highlight the need to continually monitor ARGs in foodborne Salmonella with particular emphasis on ST-40 and ST-26; the monitoring should include as many retail meat types as possible in the study area.

Identification of Closely Related Listeria monocytogenes Isolates with No Apparent Evidence for a Common Source or Location: A Retrospective Whole Genome Sequencing Analysis

ABSTRACT

Public health and regulatory agencies worldwide sequence all Listeria monocytogenes isolates obtained as part of routine surveillance and outbreak investigations. Many of these entities submit the sequences to the National Center for Biotechnology Information Pathogen Detection (NCBI PD) database, which groups the L. monocytogenes isolates into single nucleotide polymorphism (SNP) clusters based on a pairwise SNP difference threshold of 50 SNPs. Our goal was to assess whether isolates with metadata that suggest different sources or locations could show evidence for close genetic relatedness indicating a recent common ancestor and a possible unknown common source. We compared the whole genome sequencing (WGS) data of 249 L. monocytogenes isolates sequenced here, which have detailed metadata, with WGS data of nonclinical isolates on NCBI PD. The 249 L. monocytogenes isolates originated from natural environments (n = 91) as well as from smoked fish (n = 62), dairy (n = 56), and deli meat (n = 40) operations in the United States. Using a combination of subtyping by core genome multilocus sequence typing and high-quality SNP, we observed five SNP clusters in which study isolates and SNP cluster isolates seemed to be closely related and either (i) shared the same geolocation but showed different source types (one SNP cluster); (ii) shared the same source type but showed different geolocations (two SNP clusters); or (iii) shared neither source type nor geolocation (two SNP clusters). For one of the two clusters under (iii), there was, however, no strong bootstrap support for a common ancestor shared between the study isolates and SNP cluster isolates, indicating the value of in-depth evolutionary analyses when WGS data are used for traceback and epidemiological investigations. Overall, our results demonstrate that some L. monocytogenes subtypes may be associated with specific locations or commodities; these associations can help in investigations involving multi-ingredient foods such as sandwiches. However, at least some L. monocytogenes subtypes can be widespread geographically and can be associated with different sources, which may present a challenge to traceback investigations involving these subtypes.

HIGHLIGHTS

[Molecular epidemiology of Listeria monocytogenes isolated from ready-to-eat food in 2017 in China]

Objective: To analyze the molecular characteristics of Listeria monocytogenes strains from ready-to eat food in China. Methods: A total of 239 Listeria monocytogenes strains isolated from ready-to-eat food in 2017, all strains underwent whole-genome sequencing (WGS) , and comparisons uncovered population structure derived from lineages, clonal complex, serogroups, antimicrobial susceptibility and virulence, which were inferred in silico from the WGS data. Core genome multilocus sequence typing was used to subtype isolates. Results: All strains were categorized into three different lineages, lineage Ⅱ was the predominant types in food, and IIa was the main serogroups. CC8, CC101 and CC87 were the first three prevalent CCs among 23 detected CCs, accounting for 49.4%. Only 4.6% (11 isolates) of tested strains harbored antibiotic resistance genes, which were mostly trimethoprim genes (7 isolates, 2.9%). All strains were positive for LIPI-1, and only a part of strains harbored LIPI-3 and LIPI-4, accounting for 13.8% (33 isolates) and 14.2% (34 isolates), respectively. ST619 carried both LIPI-3 and LIPI-4. 51.5% (123 isolates) of strains carried SSI-1, and all CC121 strains harbored SSI-2. Different lineages, serogroups and CCs can be separated obviously through cgMLST analysis, and 24 sublineages were highly concordant with CCs. Conclusion: Ⅱa was the main serogroups in ready-to-eat food isolates in China; CC8, CC101 and CC87 were the prevalent CCs, and CC87 isolates was hypervirulent isolates, cgMLST method can be adopted for prospective foodborne disease surveillance and outbreaks detection.

Core genome MLST and resistome analysis of Klebsiella pneumoniae using a clinically amenable workflow

Whole genome sequencing (WGS) is replacing traditional microbiological typing methods for investigation of outbreaks in clinical settings. Here, we used a clinical microbiology laboratory core genome multilocus sequence typing (cgMLST) workflow to analyze 40 isolates of K. pneumoniae which are part of the Antimicrobial Resistance Leadership Group (ARLG) isolate collection, alongside 10 Mayo Clinic K. pneumoniae isolates, comparing results to those of pulsed-field gel electrophoresis (PFGE). Additionally, we used the WGS data to predict phenotypic antimicrobial susceptibility (AST). Thirty-one of 40 ARLG K. pneumoniae isolates belonged to the same PFGE type, all of which, alongside 3 isolates of different PFGE types, formed a large cluster by cgMLST. PFGE and cgMLST were completely concordant for the 10 Mayo Clinic K. pneumoniae isolates. For AST prediction, the overall agreement between phenotypic AST and genotypic prediction was 95.6%.

Whole genome sequencing analyses of Listeria monocytogenes that persisted in a milkshake machine for a year and caused illnesses in Washington State

Background

In 2015, in addition to a United States multistate outbreak linked to contaminated ice cream, another outbreak linked to ice cream was reported in the Pacific Northwest of the United States. It was a hospital-acquired outbreak linked to milkshakes, made from contaminated ice cream mixes and milkshake maker, served to patients. Here we performed multiple analyses on isolates associated with this outbreak: pulsed-field gel electrophoresis (PFGE), whole genome single nucleotide polymorphism (SNP) analysis, species-specific core genome multilocus sequence typing (cgMLST), lineage-specific cgMLST and whole genome-specific MLST (wgsMLST)/outbreak-specific cgMLST. We also analyzed the prophages and virulence genes.

Results

The outbreak isolates belonged to sequence type 1038, clonal complex 101, genetic lineage II. There were no pre-mature stop codons in inlA. Isolates contained Listeria Pathogenicity Island 1 and multiple internalins. PFGE and multiple whole genome sequencing (WGS) analyses all clustered together food, environmental and clinical isolates when compared to outgroup from the same clonal complex, which supported the finding that L. monocytogenes likely persisted in the soft serve ice cream/milkshake maker from November 2014 to November 2015 and caused 3 illnesses, and that the outbreak strain was transmitted between two ice cream production facilities. The whole genome SNP analysis, one of the two species-specific cgMLST, the lineage II-specific cgMLST and the wgsMLST/outbreak-specific cgMLST showed that L. monocytogenes cells persistent in the milkshake maker for a year formed a unique clade inside the outbreak cluster. This clustering was consistent with the cleaning practice after the outbreak was initially recognized in late 2014 and early 2015. Putative prophages were conserved among prophage-containing isolates. The loss of a putative prophage in two isolates resulted in the loss of the AscI restriction site in the prophage, which contributed to their AscI-PFGE banding pattern differences from other isolates.

Conclusions

The high resolution of WGS analyses allowed the differentiation of epidemiologically unrelated isolates, as well as the elucidation of the microevolution and persistence of isolates within the scope of one outbreak. We applied a wgsMLST scheme which is essentially the outbreak-specific cgMLST. This scheme can be combined with lineage-specific cgMLST and species-specific cgMLST to maximize the resolution of WGS.