Genomic investigation of a Shiga toxin-producing Escherichia coli O26:H11 outbreak

In a rural nursery, small children started getting diarrhea, and public health officials found shiga toxin-producing Escherichia coli (STEC) in the stool samples. Some children in the nursery developed hemolytic uremic syndrome (HUS), while others experienced diarrhea or were asymptomatic. This prompted a study to identify the source of the outbreak. The investigation focused on the children's contact with cows, as three bovine samples were suspected to be the origin of the STEC.

Were the cows the origin of the outbreak? In this article, we see how whole genome sequencing can be used to characterize outbreaks and discover possible origins.

‍

This is a Solu open-access article. You can view the genomes on Solu Platform

Background

Shiga toxin-producing Escherichia coli (STEC) is one of the six recognized pathotypes of diarrheagenic E. coli. It’s a foodborne pathogen responsible for causing infections in humans, including mild and severe diarrhea, hemorrhagic colitis, and sometimes hemolytic uremic syndrome (HUS) as a complication.

In this case study we reproduce the whole genome sequencing analyses of Moran-Gilad et al., (2017) using the Solu Platform. The paper studied an outbreak of 11 shiga-toxin producing E. coli isolates in a rural area. The outbreak spread in children in a nursery, some of which developed HUS, some diarrhea, and some were asymptomatic.

This study also included 3 bovine samples (cow/calf with STEC) that were suspected to be the origin because the children had been in contact with the animals.

Link to original article

Analysis

We downloaded genomes from ENA using the accessions from the original study. The 11 isolates in question were available as genomes, (sequenced with Illumina Miseq), and de novo assembled with the Qiagen GLC workbench. We uploaded the files to Solu Platform and reproduced the analyses from the original paper.

Species and typing

Solu identified all samples as E.coli, as expected. Six different PubMLST types were found: ST7217, ST11, ST2836, ST677, ST515, and ST659, with most samples belonging to the ST7217 type. The paper had reported the ST7217 isolates as a "new sequence type," and the isolates with ST11, ST659, ST515, and ST677 matched Solu's results. However, we couldn't replicate the paper's ST6 result, with Warwick identifying it as ST2836 and Pasteur as ST481. This discrepancy may be due to database and scheme updates. The Solu platform currently does not support in silico serotyping for E.coli, keep an eye on our changelog for updates regarding this feature.

Assembly quality

Most of the assemblies in this case study have more than 200 contigs, indicating low contiguity. Unfortunately, we were unable to access the reads to attempt de novo assembly. However, the genome size (genome fraction) appeared plausible for this species and strain, and the coverage used in the original publication was decent, measuring 49 or higher.

Antimicrobial resistance

In the analyzed samples, we observed the following antimicrobial resistance (AMR) highlights:

• All 11 samples presented genes coding for antimicrobial efflux systems including acr, emr and mdt genes.
• Antimicrobial resistance was identical for isolates within the same ST-group, and slightly different across different STs.
• Two isolates had the blaTEM-1 beta-lactamase gene.
• The mutation pmrB_Y358N, which confers resistance to colistin, was found in isolates from STs 677,  2836, 11, 7217, and 659.
• Resistance to metals and biocides were identified in all isolates.
• The original article did not report genotypic or phenotypic resistance.

Virulence

The Solu platform automatically analyzes the virulence factors associated with shiga toxin. By exporting the samples, we can observe the presence of virulence genes (pictured below), where the value shown is the gene’s coverage multiplied by identity.

The results show good agreement with the original article, with minor variations in two samples. In the MRWA01 sample, Solu identified an additional indication of stx2d. In the MRVY01 sample, Solu detected the presence of stx1a, stx2c, and stx2a subunits, which were not reported in the original article.

Phylogenetic analysis of the outbreak

The Solu platform automatically performs SKA distance for phylogenetics and creates a phylogenetic tree. By exporting the tree in Newick format and uploading it to iTOL, and midpoint rooting, we obtain this tree that helps visualize the likely connections. The SNP distances are helpful for determining how close the genetic relationships between isolates are.

Samples MRVX and MRVY can be said to be unrelated to this outbreak as they had SNP distances of >29,000 and >5,800, respectively, from the outbreak samples, which aligns with the conclusion of the original paper.

The bovine samples (MRWA01, MRWB01, MRVZ01) had SNP distances of >15,000 from the human isolates, indicating that they are not related to the outbreak, which is consistent with the original paper's findings.

Six samples were found to be close to each other with a SNP difference of within 81, and genotypic analysis confirmed that they likely belong to the same outbreak. This conclusion is supported by epidemiological data.

Conclusions

Genome sequencing can offer practical help in investigating an outbreak and its causes. In this case study, it was found that the cows were not the source of the outbreak. Fast action is necessary to prevent further spread of these organisms, as highlighted in the paper's main finding regarding the importance of the public health response.

References

MORAN-GILAD J, ROKNEY A, DANINO D, et al. Real-time genomic investigation underlying the public health response to a Shiga toxin-producing Escherichia coli O26:H11 outbreak in a nursery. Epidemiology and Infection. 2017;145(14):2998-3006. doi:10.1017/S0950268817001923