Mini ReviewWhat defines extraintestinal pathogenic Escherichia coli?
Introduction
Already shortly after the initial description of “Colibacterium commune” by Theodor Escherich in 1886 (Escherich, 1886), Alphonse Lesage postulated that this species comprises harmless strains as well as variants with different pathogenic potential (Lesage, 1897). Since then, non-pathogenic E. coli have been distinguished from intestinal or extraintestinal E. coli pathotypes based on the type of infection. Different diarrheagenic E. coli, also designated intestinal pathogenic E. coli (IPEC), variants could be identified, i.e. enterotoxigenic E. coli (ETEC), enteropathogenic E. coli (EPEC), enterohaemorrhagic E. coli (EHEC), enteroinvasive E. coli (EIEC), and enteroaggregative E. coli (EAEC). Characteristic virulence factors were defined for these individual pathotypes which play a decisive role in pathogenesis. These markers allow to allocate E. coli isolates to diarrheagenic pathotypes and to discriminate them from non-pathogenic E. coli and from ExPEC (Kaper et al., 2004). E. coli strains isolated from infections outside of the intestinal tract, e.g., uropathogenic E. coli (UPEC), neonatal meningitis-associated E. coli (NMEC) and sepsis-causing E. coli (SEPEC) have been grouped as ExPEC by (Johnson and Russo, 2002, Johnson and Russo, 2005, Russo and Johnson, 2000, Smith et al., 2007). In addition, strains which cause systemic infection in poultry, avian pathogenic E. coli (APEC), resemble certain human-pathogenic ExPEC variants (Bélanger et al., 2011, Clermont et al., 2011, Ewers et al., 2007, Johnson et al., 2010, Moulin-Schouleur et al., 2007). Two new animal pathogenic subgroups have been proposed: mammary pathogenic E. coli (MPEC) (Shpigel et al., 2008) and endometrial pathogenic E. coli (EnPEC) (Sheldon et al., 2010). ExPEC are part of the intestinal microflora of a fraction of the healthy population and normally asymptomatically colonize the gut. Once they get access to niches outside of the gut, they are, however, able to efficiently colonize these niches and cause disease in man, i.e. urinary tract infection (UTI), septicemia or meningitis in newborns, as well as UTI or systemic disease in many animals. Human and animal pathogenic strains share common genetic backgrounds (Clermont et al., 2011, Johnson et al., 2008b). Although several important ExPEC virulence factors and their role during pathogenesis have been described (Kaper et al., 2004, Mokady et al., 2005, Smith et al., 2007), many ExPEC cannot be unambiguously distinguished from commensal E. coli based on a set of discriminatory virulence factors as ExPEC can use multiple virulence factors in a mix-and-match fashion. Nevertheless, ExPEC classification has been proposed on the basis of isolation site and the detection of two virulence-associated genes (VAG) typical of the specific pathotype (Johnson and Russo, 2005), respectively two VAGs for non-host samples like food samples (Johnson et al., 2005).
This high level of genetic heterogeneity within the species E. coli mirrors the genomic plasticity of this generally clonal group of organisms which results from frequent acquisition and loss of genomic information as well as high recombination rates within the flexible genome (Brzuszkiewicz et al., 2009, Schubert et al., 2009, Tenaillon et al., 2010, Touchon et al., 2009). It has also been shown that the genetic background plays a role in the acquisition and expression of foreign DNA (Escobar-Paramo et al., 2004). Our current knowledge of the marked genomic heterogeneity among different members of this species contrasts the rather traditional view of E. coli pathotypes with a standard set of characteristic virulence genes when it comes to diagnostics, strain classification and treatment of clinical isolates.
A reliable differentiation between commensal strains and the different E. coli pathotypes is a prerequisite for risk assessment, epidemiological and ecological studies as well as for population genetics. Nevertheless, an approach to unambiguously discriminate ExPEC subtypes (UPEC, NMEC and SEPEC) or between commensals and ExPEC has not yet been established. The search for a typing system that enables a fast and precise determination of the pathogen, its evolutionary history, fitness and pathogenic potential is still ongoing. One classical bacterial typing effort is the analysis of the somatic, capsular and flagellar (O:K:H) serogroups. As these antigens are non-randomly associated with E. coli strains, serotyping allows discrimination of important groups of isolates and was for a long time the bacterial characteristic used for the definition of IPEC pathotypes as the majority of pathogenic clones belong to a limited number of well-recognized O:H serotypes (Karch et al., 2005, Stenutz et al., 2006). To a certain extent, this holds also true for ExPEC, where highly virulent variants with phenotypes epidemiologically linked with disease can be allocated to a limited number of O serogroups and virulence gene sets (Table 1) (Smith et al., 2007). Multilocus enzyme electrophoresis (MLEE) has been introduced as another phylogrouping effort. MLEE analyses demonstrated that among E. coli species with their high genetic diversity only a few distinctive genotypes exist and led to the establishment of the major phylogroups A, B1, B2, D, and E. MLEE was also helpful to assess whether E. coli isolates may represent ExPEC which mainly belong to phylogroup B2 and to a lesser extend to group D, while group A and B1 strains are often intestinal pathogens or commensals (Boyd and Hartl, 1998, Selander and Levin, 1980). DNA sequence-based molecular typing methods such as multilocus VNTR analysis (MLVA) provide a fast and comparable system for outbreak detection and surveillance. MLVA has already been used for epidemiological analysis of IPEC, especially that of EHEC (Jenke et al., 2010, Lindstedt, 2005). There is also a generic MLVA scheme available for E. coli, but this has not yet been broadly used (Lindstedt et al., 2007). The currently most well-accepted system for strain analysis, representing the gold standard for ease of use and global comparability, is multilocus sequence typing (MLST) which has been introduced by Maiden et al. (1998). There are many different E. coli MLST schemes available (Escobar-Paramo et al., 2004, Mokady et al., 2005, Reid et al., 2000, Wirth et al., 2006), which, nevertheless, result in a similar delineation of the phylogenetic diversity of E. coli isolates. Today, although the use of nucleotide sequences derived from multiple chromosomal segments does not unambiguously reveal the phylogenetic structure of E. coli populations (Leopold et al., 2011), the allocation to the E. coli phylogroups is often based on MLST data (Jaureguy et al., 2008, Tenaillon et al., 2010, Walk et al., 2009, Wirth et al., 2006) using Bayesian modeling approaches like “Structure” (Falush et al., 2003a, Falush et al., 2003b, Hubisz et al., 2009, Pritchard et al., 2000). Molecular epidemiology based on MLST demonstrated that the species E. coli is composed of a variety of individual strains with a different phylogenetic background. Furthermore, MLST data provide good arguments that phylogroup B2 which includes the majority of ExPEC isolates represents the evolutionary oldest lineage within the species (Tenaillon et al., 2010).
To extend the assessment of phylogeny and disease association of ExPEC strains, the distribution of IPEC and ExPEC within the different clonal complexes included in the E. coli MLST database (http://mlst.ucc.i.e./mlst/dbs/Ecoli) was evaluated. The E. coli MLST database (as of 12th January 2011) contained 3454 entries distributed into 1859 sequence types (ST). An updated assignment to clonal complexes (CCs) according to eBurst (http://eburst.mlst.net/v3/instructions/3.asp) identified 114 CCs that comprised 1197 STs including 2610 database entries. The analysis of the overall distribution of ExPEC relative to IPEC and commensal, non-pathogenic E. coli required the assignment of the individual database entries to these general pathogroups. After this, from the 3454 database entries 1117 (32%) represented ExPEC, 922 (27%) IPEC, 100 (3%) Shigella spp., and 535 (16%) non-pathogenic isolates, respectively. 780 (23%) database entries could not be assigned to any of these groups. The assignment to STs or ST complexes showed that only a few of them predominantly (>90%) include isolates of one pathotype and that these mainly represent IPEC (ST 1, 11, 25, 270, 272, 280, 582, 731, 1200, 1283), Shigella spp. (ST145, 148, 245), or non-pathogenic E. coli (ST 184, 1163). A few STs (ST 567, 1219, 1386, 2012) so far only include ExPEC isolates, but as the number of database entries for these STs are very small, it is unclear whether these STs indeed are “ExPEC-specific”. Instead, the vast majority of STs and CCs including ExPEC are also composed of non-pathogenic strains or other pathotypes. The composition of the STs and CCs is visualized in (Fig. 1). In the center of this minimum spanning tree (MST) CC 10 is located, which represents the largest group of closely related STs present in the database. This clonal complex includes many IPEC, non-pathogenic and not defined isolates, but only a smaller proportion of ExPEC strains. Similarly, the right branch of the MST is dominated by CCs mainly comprising IPEC, non-pathogenic as well as not defined E. coli or Shigella spp. isolates. One exception is CC 69 which is located at the very end of the right branch and mainly comprises ExPEC isolates. The left branch of the MST encompasses CCs which primarily include ExPEC and non-pathogenic strains (Fig. 1). In this “ExPEC area”, nine of the 15 biggest ExPEC CCs are located (CC 95, 73, 131, 127, 141, 17, 14, 12, 144), some of which have already been discussed before to represent successful ExPEC clones (Croxall et al., 2011, Peirano and Pitout, 2010, Rogers et al., 2011, Tartof et al., 2005, Wirth et al., 2006, Zdziarski et al., 2008). Obviously, the genetic background or specific traits represented by these CCs favor extraintestinal virulence relative to other CCs within the E. coli population. The “ExPEC branch” of the MST includes nearly all B2 strains deposited in the database, thus corroborating that phylogroup B2 is correlated with the ExPEC pathotype. These CCs are not only large in sample size, which may be a result of sampling bias, but they also belong to the CCs with the highest number of STs (Table 2), suggesting that the founder is a successful variant which is broadly disseminated. Although only a limited number of CCs seems to be correlated with the ability to cause extraintestinal infection, there is no possibility to distinguish between commensals and ExPEC, or even between different ExPEC subtypes (UPEC, NMEC or APEC) based on their allocation to individual CCs.
To further subtype individual members of an evolutionary lineage regarding their fitness or different pathogenic potentials, additional information is still needed as limited diversity in their housekeeping genes hampers a higher resolution of phylogenetic analysis of ExPEC by MLST. Thus, conventional molecular epidemiological approaches, including MLST or MLVA, have to be combined with, e.g., virulence gene detection, serogrouping or other phenotyping results to enable the identification or characterization of specialized subtypes within clonal lineages (Bidet et al., 2007, Brandal et al., 2007, Homeier et al., 2010, Mordhorst et al., 2009). MLST is a valuable tool to describe microbial populations. Nevertheless, the attempt to discriminate ExPEC from other E. coli variants based on MLST demonstrates that molecular epidemiological approaches still have to be improved. Next generation sequencing has already been demonstrated to increase resolution of the population structure and evolutionary aspects of bacterial species like Escherichia coli (Baker et al., 2010, Lukjancenko et al., 2010, Mellmann et al., 2011, Rasko et al., 2011). Large scale whole-genome sequencing has also already been applied to E. coli isolates to allow in-depth phylogenetic analyses and strain typing (Touchon et al., 2009).
Although certain clonal lineages with a high ExPEC virulence potential can be deduced from MLST analysis, the inability to clearly distinguish between ExPEC and non-pathogenic E. coli corroborates that ExPEC strains are facultative pathogens (Dobrindt et al., 2010, Tenaillon et al., 2010). An extraintestinal infection is a multifactorial process which usually involves a set of virulence factors which can also be considered to be fitness factors broadly distributed among commensals. This hypothesis is supported by the early emergence of the B2 phylogroup within the species (Le Gall et al., 2007). Furthermore, ExPEC virulence factors probably evolved because of their importance for a commensal lifestyle as they can promote intestinal colonization and survival within the normal gut environment. ExPEC adhesins, siderophore systems, and toxins can be correlated with successful gut colonization in humans (Diard et al., 2010, Johnson et al., 2008a, Nowrouzian et al., 2006, Wold et al., 1992). Additionally, lipopolysaccharide (LPS) and other ExPEC virulence factors can protect bacteria against predation in the environment (Alsam et al., 2006, Diard et al., 2007). Consequently, extraintestinal virulence has been interpreted as a coincidental by-product of commensalism (Le Gall et al., 2007, Tenaillon et al., 2010). Presumably, ExPEC virulence factors have not been evolved in order to increase the ability to cause extraintestinal infections. There is no strong selective pressure to become a perfect ExPEC as the vast majority of intestinal E. coli variants carrying “ExPEC virulence genes” never cause disease, but live as commensals. In contrast, the finding that strains of phylogroup B2 or with high ExPEC virulence gene content represent members of the intestinal E. coli flora only in a subgroup of the human population does not support the idea that “ExPEC virulence factors” confer a fitness advantage in the intestine and promote commensalism.
Although our recent geno- and phenotypic as well as epidemiological knowledge allows us to define certain clonal lineages with high ExPEC virulence potential, an unambiguous discrimination between ExPEC and commensal E. coli could not yet be achieved. Furthermore, the observation that ExPEC infections in the elderly or in immuno-compromised patients are frequently caused by E. coli variants without or with only a very low ExPEC virulence gene content suggests that the ability to cause such infections does not always rely on a specific set of virulence determinants or certain genetic backgrounds.
Consequently, for ExPEC strain typing, it has to be considered that different groups of ExPEC isolates, i.e. those with varying sets of archetypal ExPEC virulence factors as well as others comprising “unconventional” ExPEC or commensal-like strains, exist. Future research should aim at an increased resolution of the phylogenetic structure within candidate ExPEC clonal lineages. In addition, an advanced understanding of the interaction between ExPEC and their hosts in intestinal and extraintestinal niches as well as of host susceptibility factors are required for an improved ExPEC strain typing and risk assessment.
Section snippets
Acknowledgements
This work was supported by a grant of the German Federal Ministry of Education and Research (BMBF, grant no. 0315219B). U. Dobrindt was also supported by the German Research Foundation (DO 789/4-1). The excellent technical assistance of B. Plaschke (Würzburg) and O. Mantel (Münster) is appreciated.
References (64)
- et al.
Navigating the future of bacterial molecular epidemiology
Curr. Opin. Microbiol.
(2010) - et al.
Octaplex PCR and fluorescence-based capillary electrophoresis for identification of human diarrheagenic Escherichia coli and Shigella spp
J. Microbiol. Methods
(2007) - et al.
Animal and human pathogenic Escherichia coli strains share common genetic backgrounds
Infect. Genet. Evol.
(2011) - et al.
Caenorhabditis elegans as a simple model to study phenotypic and genetic virulence determinants of extraintestinal pathogenic Escherichia coli
Microbes Infect.
(2007) - et al.
Decreasing the effects of horizontal gene transfer on bacterial phylogeny: the Escherichia coli case study
Mol. Phylogenet. Evol.
(2004) - et al.
Avian pathogenic, uropathogenic, and newborn meningitis-causing Escherichia coli: how closely related are they?
Int. J. Med. Microbiol.
(2007) - et al.
Extraintestinal pathogenic Escherichia coli: “the other bad E. coli”
J. Lab. Clin. Med.
(2002) - et al.
Molecular epidemiology of extraintestinal pathogenic (uropathogenic) Escherichia coli
Int. J. Med. Microbiol.
(2005) - et al.
Enterohaemorrhagic Escherichia coli in human medicine
Int. J. Med. Microbiol.
(2005) - et al.
Study of polymorphic variable-number of tandem repeats loci in the ECOR collection and in a set of pathogenic Escherichia coli and Shigella isolates for use in a genotyping assay
J. Microbiol. Methods
(2007)