MEAT SCIENCE Meat Science 71 (2005) 62–71 www.elsevier.com/locate/meatsci Review Use of comparative genomics as a tool to assess the clinical and public health significance of emerging Shiga toxin-producing Escherichia coli serotypes Mohamed A. Karmali * Laboratory for Foodborne Zoonoses, Public Health Agency of Canada, 110 Stone Road West, Guelph, Ont., Canada N1G 3W4 Abstract Shiga toxin (Stx)-producing Escherichia coli (STEC) cause sporadic or epidemic food- or water-borne illness whose clinical spec- trum includes diarrhea, hemorrhagic colitis, and the potentially fatal hemolytic uremic syndrome (HUS). Over 200 STEC serotypes have now been implicated in human disease. Serotype O157:H7 is associated most outbreaks and most cases of HUS. Other sero- types are also associated with outbreaks and HUS but less commonly than serotype O157:H7, and some cause HUS but are typically non-epidemic. Many STEC serotypes have been associated with diarrhea, but not with outbreaks or HUS, while others, isolated from cattle, have never been linked to human disease. The only proven virulence strategies for STEC are Stx production and, in some strains, a characteristic attaching and effacing cytopathology on enterocyte that is mediated by factors encoded on a patho- genicity island (PAI) known as the locus of enterocyte effacement (LEE). But Stx subtypes and LEE cannot fully explain the appar- ent differences in virulence between STEC subgroups. However, publication of the genome sequences of two E. coli O157:H7 strains has revealed new candidate PAIs and has stimulated the use of novel approaches for assessing differences in virulence potential between groups of strains. This paper highlights the clinico-pathological features and pathogenesis of STEC infection, new infor- mation arising from E. coli O157:H7 genome sequences, and progress in the use of of comparative genomics for assessing potential differences in virulence and public health significance between STEC subgroups. Crown Copyright Ó 2005 Published by Elsevier Ltd. All rights reserved. Keywords: Shiga toxin-producing Escherichia coli; Non-O157 STEC; Comparative genomics; Pathogenicity island; O-island 122; Public health Contents 1. Introduction . 63 2. Clinicopathological features of STEC infection. 63 3. Pathogenesis of STEC infection and virulence factors . 63 3.1. The attaching and effacing (AE) lesion and the locus of enterocyte effacement (LEE) . 64 3.2. Shiga toxins and the pathogenesis of HUS . 64 3.3. Plasmid-mediated virulence factors . 65 3.4. Other candidate virulence factors . 65 4. The genome of E. coli O157:H7 .............................................................. 65 4.1. Pathogenicity islands (PAI) . 66 5. Comparative genomic approaches to assess the clinical and public health significance of STEC serotypes . 66 * Tel.: +1 519 822 3300. E-mail address: [email protected]. 0309-1740/$ - see front matter. Crown Copyright Ó 2005 Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.meatsci.2005.03.001 M.A. Karmali / Meat Science 71 (2005) 62–71 63 5.1. EDL 933 genomic O-island 122 (OI-122). 67 5.2. Incomplete OI-122. 68 6. Conclusions. 68 References . 68 1. Introduction 2001; World Health Organization, 1999). Others (e.g., O113:H21 and O91:H21) rarely cause outbreaks (Paton, Shiga toxin (Stx)-producing Escherichia coli (STEC) Woodrow, Doyle, Lanser, & Paton, 1999), but are asso- (Calderwood, Acheson, Goldberg, Boyko, & Don- ciated with sporadic episodes of HUS (Johnson et al., ohue-Rolfe, 1990), also referred to as Verocytotoxin- 1996; Nataro & Kaper, 1998). A large number of STEC producing E. coli (VTEC) (Konowalchuk, Speirs, & serotypes have been isolated from patients with diar- Stavric, 1977), are causes of a potentially fatal food- rhea, but have not been associated with outbreaks or or water-borne illness whose clinical spectrum includes HUS (Johnson et al., 1996; Nataro & Kaper, 1998; non-specific diarrhea, hemorrhagic colitis, and the World Health Organization, 1999), and yet others, iso- hemolytic uremic syndrome (HUS) (Karmali, 1989; lated from cattle, have never been linked to human dis- Karmali et al., 1985; Karmali, Petric, Steele, & Lim, ease (World Health Organization, 1999). Thus STEC 1983). The serious public health concern about STEC serotypes appear to differ in pathogenic potential, but infection is due to the risks of massive outbreaks (Carter the scientific basis for this is not known. Consequently et al., 1987; CCDR, 2000; CDC, 1993; Michino et al., the assessment of the clinical and public health risks is 1998; Riley et al., 1983) and of HUS, the leading cause greatly compromised when non-O157 STEC are found of acute renal failure in children (Karmali, 1989). Up in humans, foods, animals, and the environment. Publi- to 40% of the patients with HUS develop long-term re- cation of the genome sequences of two E. coli O157:H7 nal dysfunction and about 3–5% of patients die during strains (Hayashi et al., 2001; Perna et al., 2001)has the acute phase of the disease (Loirat et al., 1984; Siegler opened up new approaches for assessing differences in et al., 1991; Trompeter et al., 1983). There is no specific virulence potential between groups of strains. The objec- treatment for HUS, and vaccines to prevent the disease tive of this paper is to highlight the clinico-pathological are not yet available. features and pathogenesis of STEC infection, new infor- Ruminants, especially cattle, are the main reservoir of mation arising from whole genome sequencing of E. coli STEC, which are transmitted to humans primarily via O157:H7 strains, and progress in the use of of compar- contaminated foods and water (Griffin, 1995; Griffin & ative genomics for assessing potential differences in vir- Tauxe, 1991; Karch, Bielaszewska, Bitzan, & Schmidt, ulence between STEC subgroups. 1999; Karmali, 1989; Nataro & Kaper, 1998). Although serotype O157:H7 has been implicated in most out- breaks and in most cases of HUS globally (Griffin, 2. Clinicopathological features of STEC infection 1998; Karch et al., 1999; Karmali, 1989; Nataro & Ka- per, 1998), there is growing concern about the risk to hu- The characteristic features of STEC O157:H7 infec- man health associated with non-O157 STEC serotypes tion include a short period of abdominal cramps and (Johnson et al., 1996; Tarr & Neill, 1996), over 200 of non-bloody diarrhea, without significant fever, which which have now been associated with human illness may be followed, in some cases, by hemorrhagic colitis (World Health Organization, 1999). First reported in and the hemolytic uremic syndrome (HUS) (Karmali, association with HUS in Canada (Karmali et al., 1985; 1989). Karmali et al., 1983), non-O157 serotypes have since Defined by the triad of features: acute renal failure, been more commonly implicated in HUS than serotype thrombocytopenia, and microangiopathic hemolytic O157:H7 in Latin America (Lopez, Contrini, & de Rosa, anemia, HUS develops in about one-tenth to one quar- 1998) and Australia (Elliott et al., 2001a), and their fre- ter of the cases of STEC infection (Griffin, 1995; Griffin quency may be rising in Europe (Gerber, Karch, Aller- & Tauxe, 1991; Karch et al., 1999; Karmali, 1989; Nat- berger, Verweyen, & Zimmerhackl, 2002; Tozzi et al., aro & Kaper, 1998). 2003). Up to 20% of HUS cases in North America may be associated with non-O157 STEC (Banatvala et al., 1996). Some non-O157 STEC serotypes (e.g., sero- 3. Pathogenesis of STEC infection and virulence factors types O26:H11, O103:H2, O111:NM, O121:H19, and O145:NM) are associated with outbreaks and HUS The two main virulence strategies of STEC are Stx- but less commonly than serotype O157:H7 (Griffin & production and the ability to colonize the bowel (Nataro Tauxe, 1991; Johnson et al., 1996; McCarthy et al., & Kaper, 1998). E. coli O157:H7 and certain other 64 M.A. Karmali / Meat Science 71 (2005) 62–71 STEC serotypes (e.g., O26:H11, O103:H2, O111:NM, adhesin, intimin (Jerse et al., 1990; Nataro & Kaper, O121:H19, and O145:NM) colonize the mucosal epithe- 1998), and genes that encode the translocated intimin lial cells of the large bowel with a characteristic ‘‘attach- receptor known as Tir (Kenny et al., 1997) or EspE ing and effacing’’ (AE) cytopathology which is encoded (Deibel, Kramer, Chakraborty, & Ebel, 1998), and the for by genes on the locus of enterocyte effacement (LEE) Tir chaperone, CesT (Abe et al., 1999; Elliott et al., pathogenicity island (PAI) (Jerse, Yu, Tall, & Kaper, 1999). The LEE 4 operon encodes the structural needle 1990; McDaniel & Kaper, 1997). Serotypes associated protein, EscF (Wilson, Shaw, Daniell, Knutton, & Fran- with the AE cytopathology are also referred to as kel, 2001), the translocator proteins EspA, EspB, and enterohemorrhagic E. coli (EHEC) (Levine et al., EspD (Nataro & Kaper, 1998), and the effector protein 1987). While Stxs and LEE-encoded factors are the only EspF (McNamara & Donnenberg, 1998). Other LEE- proven virulence components, STEC strains produce encoded effector proteins include EspG (Elliott et al., many other putative virulence factors whose precise role 2001b), EspH (Tu, Nisan, Yona, Hanski, & Rosenshine, in pathogenesis is not known. E. coli O157:H7 has a low 2003), and Map (Kenny & Jepson, 2000). The TTSS also infectious dose (100–500 organisms) (Griffin, 1998), secretes a number of other effector molecules that are en- which is thought to reflect its resistance to gastric acid coded outside the LEE on other pathogenicity islands. (Lin et al., 1996; Waterman & Small, 1996). After inges- The precise role of these so-called non-LEE-encoded tion, the organism passes through the stomach and effectors (NLEs) (Deng et al., 2004; Gruenheid et al., small bowel and is thought to colonize enterocytes of 2004) has yet to be fully established. The first gene the large bowel mucosa by a characteristic attaching (Ler), in the LEE 1 operon, activates transcription of and effacing (AE) cytopathology (Nataro & Kaper, the LEE 2, 3, and 4 operons (Elliott et al., 2000; Mellies, 1998).