Short-term evolution of Shiga toxin-producing Escherichia coli O157:H7 between two food-borne outbreaks

Authors: COWLEY, LAUREN - Public Health England (PHE); DALLMAN, TIMOTHY - Public Health England (PHE); FITZGERALD, STEPHEN - Roslin Institute; IRVINE, NEIL - Public Health Agency; ROONEY, PAUL - Royal Victoria Hospital; MCATEER, SEAN - Roslin Institute; DAY, MARTIN - Public Health England (PHE); PERRY, NEIL - Public Health England (PHE); Bono, James - Jim; JENKINS, CLAIRE - Public Health England (PHE); GALLY, DAVID - Public Health England (PHE)

Submitted to: Microbial Genomics
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/8/2016
Publication Date: 9/20/2016
Publication URL: http://handle.nal.usda.gov/10113/5678127
Citation: Cowley, L.A., Dallman, T.J., Fitzgerald, S., Irvine, N., Rooney, P.J., McAteer, S., Day, M., Perry, N.T., Bono, J.L., Jenkins, C., Gally, D. 2016. Short-term evolution of Shiga toxin-producing Escherichia coli O157 between two food-borne outbreaks. Microbial Genomics. doi:10.1099/mgen.0.000084.

Interpretive Summary: Shiga toxin-containing Escherichia coli O157:H7 (STEC O157) is a bacterium that can cause severe diarrhea, bloody diarrhea, and kidney failure in humans. In this study, we investigated two strains of STEC O157 that caused outbreaks, eight weeks apart, at the same restaurant. While traditional genetic fingerprinting methods had indicated that the strains were different, and potentially not closely related to each other, whole genome DNA sequencing showed that their core genomes were virtually identical, indicating that they were very closely related and probably diverged from a common ancestor strain in the last year. Different DNA sequencing technologies enabled us to show that while the core genomes were virtually identical, strains from the first and second outbreaks had several significant differences in their DNA. Thus, our findings indicated that while the two outbreak strains were considered virtually identical at a core genome level, their ‘accessory’ genomes had changed significantly, these changes, may have resulted in a strain more efficient at making people sick. The results of this study demonstrate the utility of using whole genome DNA sequencing as a component of outbreak investigations.

Technical Abstract: Background: Shiga toxin-producing Escherichia coli (STEC) O157 is a public health threat and outbreaks occur worldwide. STEC O157 has a mosaic genome with extensive prophage integration, including bacteriophage-encoded Shiga toxins. Here, we investigate genomic differences in a strain of STEC O157 that caused two outbreaks, eight weeks apart, at the same restaurant. Four infected individuals were identified in the first outbreak while the second was linked to over 140 confirmed cases. Results: Traditional phage typing identified that the first outbreak was caused by PT8 isolates and the large second outbreak by PT54. Short read genome sequencing defined two closely-related sub-clusters (PT8 & PT54), but only separated by 3 SNPs in their core genomes. Isolates did not cluster with local strains but with those associated with foreign travel to the Middle East/North Africa. The investigation suggested its introduction via an imported food and that an infected/colonized food handler may have contributed to the spread. Combined long read sequencing approaches and optical mapping of an isolate from each outbreak revealed that the accessory genome had undergone significant microevolution between the two outbreaks. This included acquisition of a 240 Kbp IncHI2 plasmid responsible for the phage type switch from PT8 to PT54. There was also prophage gain, loss and recombination including a prophage-rich 200kbp inverted repeat in the PT8 isolate. Phenotyping and fitness assays demonstrated significant metabolic differences between the isolates with the PT54 isolate having a general fitness advantage in rich medium including an increased capacity to use specific amino acids and dipeptides as a nitrogen source. Conclusions: Short-term microevolution between two closely related outbreaks has been demonstrated by WGS approaches. The second outbreak was considerably larger, the organism was detected in a food handler, and there were multiple secondary cases indicative of effective human-to-human transmission. While the strain was imported, a food handler responsible for garnishing food in the restaurant during the large second outbreak was infected with the PT54 strain. We propose that plasmid acquisition and prophage changes have further adapted the strain for human infection and transmission with increased fitness of the PT54 strain. While a dynamic accessory genome can confound traditional typing methods, our study shows the added insight provided by combining WGS approaches for outbreak investigations.