Complete Genome Sequences of Two Escherichia coli O145:H28 Outbreak Strains of Food Origin.

Escherichia coli O145:H28 strain RM12581 was isolated from bagged romaine lettuce during a 2010 U.S. lettuce-associated outbreak. E. coli O145:H28 strain RM12761 was isolated from ice cream during a 2007 ice cream-associated outbreak in Belgium. Here we report the complete genome sequences and annotation of both strains.

GENOME ANNOUNCEMENT

Shiga toxin-producing Escherichia coli O145 was identified as a cause of hemolytic uremic syndrome (HUS) at least two decades ago (1–3), and now is recognized as one of the six non-O157 serotypes that are most frequently associated with human disease in the United States (4, 5). Food-borne outbreaks of O145 infection have been reported worldwide (6–9). E. coli O145:H28 strain RM12581 was isolated from bagged romaine lettuce during a multistate outbreak of O145 infections associated with romaine lettuce consumption in April and May 2010 in the United States (9). Strain RM12761 was isolated from ice cream during an outbreak of O145 and O26 infections among ice cream consumers in Belgium in 2007 (7, 10). We reported previously the complete genome sequence of a clinical isolate linked to each of the two outbreaks (11).

Whole-genome sequencing (WGS) and 8-kb insert paired-end (PE) 454 sequencing libraries were prepared and sequenced on an FLX Genome Sequencer (Roche) using Titanium chemistry. Illumina library preparation and sequencing (101 bp PE) were run at Ambry Genetics (Aliso Viejo, CA) on an HiSeq2000 sequencer. PacBio libraries for circular consensus sequence (CCS) reads were prepared and PacBio SMRT sequencing was run on a PacBio RS instrument using C2 chemistry. Draft genomes were assembled using Newbler (v2.3) for 454 reads and VELVET (v1.0.13) for Illumina PE reads as described previously (11). A total of 13 scaffolds composed of 158 contigs and 2 scaffolds composed of 6 contigs were generated for RM12581 and RM12761, respectively, after gap closure and misassembly correction using GapResolution and dupFinisher (12, 13). PacBio CCS reads were then aligned to the contigs using Geneious (v5.1) software (14), and the remaining gaps were manually closed in silico. Annotation was carried out using RAST (http://rast.nmpdr.org) (15).

The E. coli RM12581 genome is composed of a 5,585,611-bp chromosome and two large plasmids, pO145-12581 (81,120 bp) and pRM12581 (64,562 bp), and contains 5,769 coding DNA sequences (CDSs), 22 rRNAs, and 104 tRNAs. The E. coli RM12761 genome is composed of a 5,402,281-bp chromosome and two large plasmids, pO145-12761 (98,067 bp) and pRM12761 (58,666 bp), and contains 5,493 CDSs, 22 rRNAs, and 98 tRNAs. Single nucleotide polymorphisms (SNPs) between the strains of food and clinical origins linked to the same outbreak (RM12581 and RM13514; RM12761 and RM13516) were identified using breseq (v0.23) (16, 17). A total of 18 SNPs were identified in non-repeat regions for the 2007 Belgian ice cream-associated outbreak strains, of which 11 are in intergenic regions and seven are in coding regions (4 synonymous SNPs, 2 nonsense mutations, and 1 nonsynonymous mutation). The two CDSs with a nonsense mutation encode an endonuclease (ECRM13516_0820) and a hypothetical protein (ECRM13516_2566). The nonsynonymous mutation was detected in the gene encoding a valyl-tRNA synthetase (ECRM13516_5090). In contrast, no SNPs were identified in non-repeat regions between the two strains linked to the 2010 U.S. lettuce-associated outbreak. Nearly identical methylomes were detected between the two E. coli O145 strains of food and clinical origins that were linked to the same outbreak (11).

Nucleotide sequence accession numbers.

The E. coli O145:H28 strain RM12581 and RM12761 genome sequences were deposited at DDBJ/EMBL/GenBank under the accession no. CP007136 to CP007138 and CP007133 to CP007135, respectively.