Comparative genomic analysis and characteristics of NCCP15740, the major type of enterotoxigenic Escherichia coli in Korea

Escherichia coli is a rod-shaped, gram-negative, facultative anaerobic and non sporulating bacterium belonging to the family Enterobacteriaceae. E. coli inhabits the intestines of all humans and animals. Most E. coli are harmless, but some induce various diseases; thus, the species is considered an opportunistic pathogen. E. coli strains that cause diarrhea can be categorized into six groups according to virulence elements in the genome: enterotoxigenic (ETEC), enteropathogenic, nonsporulating, enteroaggregative, enteroinvasive, and diffusely adherent [1]. ETEC is a major cause of traveler’s diarrhea and is responsible for 700,000 diarrhea-related deaths per year in young children of less than 5 years of age in developing countries [2, 3]. Among the major virulence factors, two enterotoxins, i.e., a heat-labile toxin (LT) and a heat-stable toxin (ST), induce watery diarrhea in ETEC. The LT toxin is encoded by the eltAB gene. ST toxins are classified into two types, STh and STp; human-derived STh is encoded by estA, and porcine-derived STp is encoded by st1 [4]. In addition to serotyping, ETEC strains are classified by the combination of the O antigen of the lipopolysaccharide, H antigen of the flagellin, and K antigens. Although there are over 100 different O antigens and 34 H antigens associated with ETEC [5, 6], O6, O8, O25, O78, O128, and O153 and H7, H12, H16, H21, H45, and H49 are the most common, respectively [7]. In addition to enterotoxins, ETEC strains possess adhesive pili called colonization factors (CFs), which mediate adherence to the small intestinal wall. Over 30 CFs have been described in human ETEC strains to date. The most prevalent CFs are CFA/I and CS1–CS6, and strains typically carry two or three CFs, such as CS1 + CS3, CS2 + CS3, and CS5 + CS6.

In a previous study [8], 258 isolates from patients with diarrhea in Korea and 33 isolates from travelers visiting other Asian countries were analyzed, and two major sequence types were identified by multilocus sequence typing (MLST). In particular, ST171 (n = 62) was identified as the most prevalent ETEC type in Korea, but ST949 (n = 5) was the most frequent among inflow isolates. Although ST171 was a major MLST type of ETEC in Korea, the genomic characteristics, including enterotoxin genes, CF genes, and virulence factors, had not yet been investigated. In the present study, we selected one ST171 strain identified in this previous work, i.e., NCCP15740, isolated in 2010 from a patient with diarrhea, with serotype O6:H16, and performed whole-genome sequencing. We compared the genome of NCCP15740 with other whole-genome sequences of ETEC strains reported as O6:H16 isolates over a similar time period.

ETEC is responsible for 700,000 diarrhea-related deaths per year in young children of less than 5 years of age and is a main cause of traveler’s diarrhea [3]. However, the type and relative proportions of ETEC enterotoxins differ depending on the geographical source.

The enterotoxin types of Korean isolates from 2003 to 2011 were reported to be similar to those of isolates from Asia and the Middle East, but different from those of isolates from South America [8]. In this study, we investigated the characteristics of the major type of ETEC in Korea at the genomic level by sequencing an ST171 isolate, NCCP15740, and performing a comparative analysis with the genome sequences of other O6 strains. According to the whole-genome phylogeny, NCCP15740 belonged to one of the two groups of strains that were isolated in 2011, but belonged to the group that included the majority of O6 strains in the MLST-based phylogeny (Fig. 2). There are many genomic changes that determine the branch of a strain in a phylogenetic tree, including SNPs, insertions, deletions, prophages, and other insertion sequence elements. However, MLST genes are housekeeping genes and are more conserved than other genomic loci. Therefore, whole-genome-based phylogeny is more sensitive than MLST-based phylogeny, although it is more difficult to group strains with whole-genome-based phylogeny. Accordingly, it is necessary to select whole-genome- or MLST-based phylogeny according to the needs of the study design and aim. An MLST-base phylogeny is suitable for clustering strains according to their MLST type, whereas whole-genome phylogeny provides a better representation of the differences between strains than MLST-based phylogeny.

We investigated the toxin types of NCCP15740 and reference strains. Strains that only express LT are generally less pathogenic [26]. The NCCP15740 genome had genes encoding both STh and LT enterotoxin types. The presence of both STh and LT was common in O6 strains (21 out of 24).

The genomes of ETEC strains harbor genes encoding more than one type of CF [27]. According to a previous study [8], the CS3/CS21 genes are the most prevalent CF genes in Korean isolates, and CS3-CS21-CS1/PCF071 (15/64) and CS2-CS3-CS21 (13/64) are the most frequent CF genes in ST171. In contrast, NCCP15740 had CS2/CS3 genes that were only observed in two ST171 isolates. However, 20 out of 24 O6 strains carried CS2/CS3/CS21 genes, the major CF genes in ST171, even though they had different MLST types.

Based on the virulence factor investigation, NCCP15740 carried fewer virulence factors (59.4%) than the average number of virulence factors (62.0%) in the strains used in this study. In particular, flgD and flgE were not present in NCCP15740, but were detected in all of the reference strains. With respect to toxins, enterotoxin-related genes (entA, entB, entC, and entD) [28] were present in all of the O6 strains, including NCCP15740, whereas alpha-hemolysin-related genes (hlyA, hlyB, hlyC, and hlyD) [29] were only present in E. coli ATCC 25922. Alpha-hemolysin is a major virulence factor in ETEC, Shiga toxin-producing E. coli, and enteropathogenic strains and is thought to be acquired by horizontal gene transfer via conjugative plasmids [30].

Interestingly, none of the nonclassical virulence genes were detected in the NCCP15740 genome, and only three nonclassical virulence genes, i.e., etpA, irp2, and tia, were detected in other O6 strains. The O6 reference strains were isolated from patients in the USA, but the nonclassical virulence gene profiles were quite different from those of South American isolates. The eatA, irp2, and fyuA genes were the most prevalent in Colombian and Chilean ETEC strains [25], but none of the genes were detected in O6 strains. In addition, the tia and leoA genes were less frequent in Bolivia [31], Chile [32], Guatemala, and Mexico [33], although 11 out of 24 O6 strains had the tia gene.

In summary, NCCP15740, representing the major type of ETEC in Korea, appeared to belong to the O6 serotype and ST4. Unlike other ST4 strains, NCCP15740 did not carry the CS21 gene. Moreover, the strain harbored fewer classical virulence factors than the O6 reference strains and did not contain any nonclassical virulence factors. These results provided important insights into the development of ETEC vaccine candidates. However, because the results were obtained from in silico analyses, experimental confirmation of the results is required.