The Journal of Clinical Investigation|March 2001|Volume 107|Number 5Es - PDF document

The Journal of Clinical Investigation|March 2001|Volume 107|Number 5Escherichia coli, a venerable workhorse for biochemicaland genetic studies and for the large-scale production ofrecombinant proteins, is one of the most intensivelystudied of all organisms. The natural habitat of in humans, this species is the most common facultativeharmless symbionts, there are many pathogenic humans. In addition, from an evolutionary perspective,are so closely related phylo-genetically that they are included in the group of organ-(1, 2). Pathogenic differ from those that predominate in the enteric floraof healthy individuals in that they are more likely toexpress virulence factors Ñ molecules directly involvedin pathogenesis but ancillary to normal metabolic func-tions. Expression of these virulence factors disrupts thenormal host physiology and elicits disease. In additionways unavailable to commensal strains, and thus tospread and persist in the bacterial community.It is a mistake to think of species. Most genes, even those encoding conservedmetabolic functions, are polymorphic, with multiplealleles found among different isolates (1). The compo-The fully sequenced genome of the laboratory K-12strain, whose derivatives have served an indispensablerole in the laboratories of countless scientists, shows evi-ed that the K-12 lineage has experienced more than 200lateral transfer events since it diverged from about 100 million years ago and that 18% of its con-temporary genes were obtained horizontally from otherspecies (4). Such fluid gain and loss of genetic materialare also seen in the recent comparison of the genomicsequence of a pathogenic O157:H7 with the K-12genome. Approximately 4.1 million base pairs of Òback-boneÓ sequences are conserved between the genomes,but these stretches are punctuated by hundreds ofsequences present in one strain but not in the other. Thepathogenic strain contains 1.34 million base pairs of lin-eage-specific DNA that includes 1,387 new genes; someof these have been implicated in virulence, but manyhave no known function (5).pathotypes were acquired from numerous sources,of other bacteria. Pathogenicity islands, relatively large�(10 kb) genetic elements that encode virulence factorsand are found specifically in the genomes of patho-genic strains, frequently have base compositions thatdiffer drastically from that of the content of the rest ofgenome, indicating that they were acquiredfrom another species. Here, we explore some of theknown virulence factors that contribute to the hetero-strains, and we review what is knownPathogenic forms of animal diseases are remarkably diverse. Certain path-ogenic strains cause enteric diseases ranging in symp-toms from cholera-like diarrhea to severe dysentery;other may colonize the urinary tract, resulting incystitis or pyelonephritis, or may cause other extrain-testinal infections, such as septicemia and meningitis.In discussing the diversity of pathogenic forms of thisversatile species, we distinguish between an isolateÕspathotype, a classification of have a similar mode of pathogenesis and cause clini-cally similar forms of disease, and the ognized pathotypes of (Table 1) but many morethe same pathogenic clone represent a monophyleticbranch of an evolutionary tree and typically carrymany of the same mobile genetic elements, includingthose that determine virulence.Early evidence for the clonal nature of pathogenic was seen in the repeated recovery of identicalserotypes and biotypes from separate outbreaks of dis-support from the study of protein polymorphisms, Pathogenesis and evolution of virulence in enteropathogenic and enterohemorrhagic and Thomas S. WhittamDivision of Infectious Diseases, University of Maryland, Baltimore, Baltimore, Maryland, USANational Food Safety & Toxicology Center, Michigan State University, East Lansing, Michigan, USAAddress correspondence to: Michael S. Donnenberg, Division of Infectious Diseases, University of Maryland, Baltimore, 10 S. PinBaltimore, Maryland 21201, USA. Phone: (410) 706-7560; Fax: (410) 706-8700; E-mail: mdonnenb@umaryland.edu. and James M. Musser, The Journal of Clinical Investigation|March 2001|Volume 107|Number 5first with patterns of the major outer proteins andenzyme electrophoresis (1). Recent sequence compar-isons have shown that a phylogenetic approach basedon the clone concept, however, is complicated byrecombination events, which, like mutations, con-tribute to the divergence of bacterial genomes innature (reviewed in refs. 6, 7).The diversity of pathotypes and their genetic related-includes strains of the pathotypes associated withenteric disease and strains representing the major phy-logenetic groups (groups A, B1, B2, and D) of the Reference (ECOR) collection, a set of natural isolateschosen to represent genetic variation in the strains of the most common clones of five serogroups(O26, O111, O55, O128, and O157) associated withinfectious diarrheal disease; these widespread clones arereferred to as the DEC (diarrheagenic addition, there are representatives of the commonclones of enteroinvasive (8). The genetic distancebetween clones based on alleles detected by enzyme elec-sequence divergence in housekeeping genes (Figure 1b).The sequence data indicate that the deepest branches inthe dendrogram reflect about 8% divergence at synony-mous sites. It should be emphasized that because ofpast recombination, the dendrogram cannot be a truephylogeny but can only serve as a framework for inves-tigating the evolution of the various clones.Pathotypes of groups, although some pathotypes are found in multi-ple lineages (Figure 1). In particular, there are two clus-ated with infantile diarrhea and two clusters ofclusters are highly divergent, whereas both EPEC 2 andEHEC 2 are more closely related to one another and fallinto the B1 group of ECOR. The finding that inde-cause clinically similar disease indicates that certainpathotypes have evolved multiple times in differentclonal groups (7). EPEC and EHEC groups are phylo-genetically distinct from the enteroinvasive (EIEC), bacteria that cause dysentery and are mostclosely related to strains of the ECOR group A. Thedifferent from those recovered in extraintestinal infec-found near the bottom of the dendrogram in the B2Below, we focus on the virulence factors and patho-genic mechanisms of two major pathotypes, EPEC andEHEC, for which data exist both on the genetic basis ofdisease and on the phylogenetic history of the strains.These examples are put forward to demonstrate howgenetic polymorphisms among foundly influence disease. The reader is referred else-where for reviews of diarrheagenic (11) and extrain-EPEC was the first group of strains recognized aspathogens, an insight that followed from serologicalstudies comparing strains cultured from devastatingoutbreaks of neonatal diarrhea with other strains iso-lated from healthy infants. Although such outbreaksare now rare in developed countries, EPEC strains con-from developing countries worldwide (11). In recentyears, the pathogenesis of EPEC infection has provedto be amenable to genetic dissection, and severalthemes have emerged.. In the early 1980s, investigators in the labora-tory of James Kaper reported that a particular patternof adherence to tissue culture cells by EPEC strains wasfactor (EAF) plasmid (11). Rather than covering tissueculture cells uniformly, EPEC strains form denselypacked three-dimensional clusters on the surface of thecell, a pattern known as localized adherence. This pat-tern of adherence is so characteristic of and specific toEPEC strains that it can be used as the basis for diag-nosis (11). The ability to perform localized adherencecan be transferred to nonpathogenic laboratory strains by transformation with the EAF plasmid. Con-versely, EPEC strains lose this ability and demonstrateThe principal factor responsible for the localizedadherence phenotype is a surface appendage known asthe bundle-forming pilus (BFP), a member of the typeIV fimbria family that is encoded on the EAF plasmidity of BFP to reversibly aggregate into ropelike bundles.If any of the genes required for the formation of BFPare inactivated by mutation, the bacteria fail to formaggregates and do not display localized adherence (14).ly polymorphic protein encoded by operon. Another protein, BfpF,which is predicted to be a cytoplasmic nucleotide-bind-ing protein, plays a special role in aggregation. Whenis mutated, the bacteria continue to make pili thataggregate and allow the bacteria to do the same (15);however, the pili fail to form higher-order bundles andestingly, despite the fact that they remain capable offurther steps in pathogenesis, mutants are signif- and James M. Musser, Thus, it appears that not only the BFP structure, butalso intact BFP function, is required for full virulence.A chromosomal pathogenicity island encoding a type IIIsecretion system and the ability to alter the host cytoskeletonthe formation of intestinal lesions caused by the abil-ity of bacterial cells to attach intimately to the hostcell membrane, destroy microvilli, and induce theformation of cuplike pedestals composed ofcytoskeletal proteins upon which the bacteria sit(Figure 2b). This ability, known as attaching andeffacing activity, has been observed in vitro and induodenal and rectal biopsies from infants with EPECinfection (11, 13). A 35-kb genetic element known asthe locus of enterocyte effacement (LEE) is necessaryfor this effect and, when cloned from EPEC strainE2348/69 into a nonpathogenic ficient to confer attaching and effacing activity (18).because it contains virulence loci, it is not found innonpathogenic strains, it is inserted into theat specific sites (tRNA genes), andfinally because its distinctive G+C content (38%) indi-cates its origin in another species. The LEE is insert-ed near different tRNA loci in different EPEC strainswhich encode a type III secretion system and variousproteins secreted via this system, including anadhesin and its cognate receptor, a regulator, and sev-eral proteins of unknown function. Type III secretionsystems are found in bacteria from several Gram-neg-ative genera that have close relationships witheukaryotic hosts (19). These systems can transportbacterial proteins across the inner and outer mem-brane and can deliver effector proteins to the surfaceThe proteins secreted via type III systems can be divid-ed into two classes: the effector proteins, which aretranslocated to the host cell, and the components of thetranslocation apparatus, which are required to deliverthe effector proteins into the host cell. The best-charac-terized EPEC effector protein is called Tir, for translo-cated intimin receptor. Tir is encoded by the LEE and isit is inserted in the plasma membrane (18, 20). Muta-tions in components of the type III secretion system orin the genes encoding two of the secreted proteins, EspAand EspB, prevent the translocation of Tir. Thus, EspAand EspB can be classified as part of the translocationapparatus. Tir has two membrane-spanning domainsand is oriented so that both the amino- and the carboxy-termini protrude into the host cell cytoplasm (21). Onceinserted into the host cell membrane, Tir serves as areceptor for intimin, an outer membrane proteinrequired for virulence. Intimin is the product of the gene, located just downstream of EPEC have evolved an adherence mechanism in whichthe bacteria synthesize both the adhesin (intimin) andits receptor (Tir); the latter is inserted directly into thehost cell by the LEE secretion apparatus.Luo et al. (22) recently determined the three-dimen-bound to the extracellular domain of Tir. They identifieda series of immunoglobulin-like domains (Dgive intimin a rigid, roughly cylindrical shape and a dis-tal carboxy-terminal domain (Dincomplete C-lectin structure. In the cell membrane, Tirforms a dimer with each molecule consisting of a pair ofhelices separated by a hairpin turn. Theentire structure is a four-helix bundle with the hairpinloops protruding from either side (Figure 2c). Intimindimer binds to two intimin molecules. Tir forms con-tacts with intimin along one side of the C-lectin domain. The Journal of Clinical Investigation|March 2001|Volume 107|Number 5 Table 1 Clinical and epidemiological features and virulence factors of various PathotypeClinical featuresEpidemiological featuresVirulence factorsEnteropathogenicWatery diarrhea and vomitingInfants in developing countriesBundle-forming pilus, attaching and effacingEnterohemorrhagicWatery diarrhea, hemorrhagic colitis, Food-borne, water-borne outbreaks Shiga toxins, attaching and effacinghemolytic-uremic syndromein developed countriesEnterotoxigenicWatery diarrheaChildhood diarrhea in developing countries, Pili, heat-labile and travelerÕs diarrheaheat-stable enterotoxinsEnteroaggregativeDiarrhea with mucousChildhood diarrheaPili, cytotoxinsEnteroinvasiveDysentery, watery diarrheaFood-borne outbreaksCellular invasion, intracellular motilityDiffuse-adheringPoorly characterizedOlder children??UropathogenicCystitis, pyelonephritisSexually active womenType I and P fimbriae, hemolysin, pathogenicity islandsMeningitis-associatedAcute meningitisNeonatesK1 capsule, S fimbriae, and James M. Musser, To achieve this configuration, both intimin and Tirappear to be oriented roughly parallel to both the bacte-rial and the eukaryotic cell membranes. This orientationaccounts for the close contact (10 nm) between the bac-While Tir is clearly an effector protein, the roles ofthree other proteins, EspA, EspB, and EspD, which areencoded in an operon in the LEE and are secreted byEPEC via the type III system, are still being defined.EspA appears to be purely a component of the translo-cation apparatus. EspA molecules form a surfaceappendage that can be seen by electron microscopydence that EspA molecules penetrate the host cell cyto-plasm or membranes. EspD has several putative trans-membrane domains and has been observed in the hostcell membrane (23). Because it is required for thetranslocation of EspB, EspD is also a part of thetranslocation apparatus. Interestingly, when mutated, EspA filaments are much shorter than nor-mal, suggesting a role for EspD in formation or stabi-The function of the EspB protein is more enigmatic.While EspB is required for the translocation of Tir,indicating that it is a component of the translocationThe protein has a hydrophobic stretch that could act asa transmembrane domain, and EspB molecules havethese observations, some investigators have suggestedthat EspB forms part of a pore that enables the passageof Tir into the host cell (18). However, when host cellsare transfected with a vector that enables them tolose stress fibers, suggesting that EspB also acts as aneffector protein and affects cytoskeletal regulation (24).What triggers the molecular events in the host cellsthat lead to the attaching and effacing activity? A recentstudy shows that the Arp2/3 complex, which nucleatespedestals of attaching and effacing lesions (25). Mem-bers of the Wiskott-Aldrich syndrome protein (WASP)family, which activate the Arp2/3 complex, are alsolocalized within the pedestals, and dominant-negativeforms of WASP prevent attaching and effacing activity.Thus it has been proposed that EPEC activates WASP toRecent work has shed light on the role in pathogene-sis of another secreted protein, EspF. An mutantstrain exhibits normal attaching and effacing activity(26) but fails to provoke a decrease in transepithelialelectrical resistance Ñ a phenotype, found in wild-typeEPEC strains, that may be related to loss of intestinalbarrier function and diarrhea in vivo (in this issue, ref.mutant fails to induce apop-tosis in host cells, another feature of the EPECÐhost The Journal of Clinical Investigation|March 2001|Volume 107|Number 5 and James M. Musser, Figure 1) The dendrogram is based on analysis of polymorphism at 36 proteinloci studied by multilocus enzyme electrophoresis. Isolates mentionedrepeatedly in the text are shown in red. The number of differencesbetween strains is converted to a genetic distance assuming that eachdifference results from at least one amino acidÐaltering mutation at theDNA level. The diagram can be interpreted as a hypothetical phylogenyof strains that can be tested by gathering independent data. Mainbranches representing pathotypes are labeled. The A, B1, B2, and Dgroups are the clusters from the ECOR set. The triangles mark positionsat which major acquisition of virulence factors are postulated to haveoccurred. () Nucleotide substitutions for seven housekeeping genesplotted against genetic distance. Nucleotide differences were analyzedseparately for synonymous sites (), positions in codons where pointmutations do not predict amino acid replacements, and nonsynony-mous sites (), where point mutations result in amino acid changes.The points are averages of the comparison of pairs of strains (marked. UTI, urinary tract infection. or of cells has no effect, but synthesis of EspF in trans-fected cells results in rapid cell death. Interestingly,EspF contains proline-rich repeats that may serve asSrc-homology 3 binding domains, allowing it to inter-act with as-yet unidentified host proteins. Thesedomains could mediate the effects of EspF on intestin-A large toxin that inhibits lymphocyte activation. Severalyears ago, a factor was described that is produced bylymphocyte activation. This heat-labile factor blockslymphocyte proliferation and the production of IFN-IL-2, IL-4, and IL-5. Although lymphocytes exposed tothe factor are nonresponsive, there is no evidence thatencoding this factor, lymphostatin, was cloned andmutated, the resulting strain could no longer inhibitlymphocyte function (29). A relatively short stretch ofthe sequence from this very large protein is homolo-cytotoxins, which covalently inactivate members of theRho family of small mammalian GTPases. SequencesThe mechanism by which lymphostatin blocks lym-phocyte activation and the role, if any, of lymphostatinin disease have not been established.Two divergent groups of EPECAs seen in Figure 1, two distinct phylogenetic groupshave been identified that have a concentration ofplay the serotypes that were first implicated in out-The first group, EPEC 1, includes some of the origi-nally identified adherent strains, most notably, strainE2348/69 (serotype O127:H6), the widely used modelorganism of human EPEC infection. This group com-prises widespread clones with EPEC serotypes O55:H6,teria of these clones usually carry both the LEE and theEAF plasmid, and they display typical localized adher-ence. EPEC 2 consists of other classical EPEC serotypes,Some of these clones are common and very widespread.For example, DEC 12 (serotype O111:H2) has histori-recovered from out-frequently recovered O111 clone associated with diar-includes strain B171, an intensively studied O111 strainoriginally recovered from a diarrhea outbreak (17, 32).The divergence between EPEC 1 and EPEC 2 is seennot only in their allelic differences in housekeepingsites at which the LEE pathogenicity island is insertedinto the bacterial genome (18). In both EPEC groups,sequenced in its entirety lacks genes for transmissionstructural subunit of BFP, shows considerablesequence variability. The eight known alleles can beseparated into two groups ( The Journal of Clinical Investigation|March 2001|Volume 107|Number 5 Figure 2Pathogenesis of EPEC infection. () Electron micrograph of a culture ofEPEC bacteria grown under conditions that lead to the production of typeIV fimbria known as bundle-forming pili (BFP). BFP are required for bac-terial aggregation and localized adherence to epithelial cells. () Electronmicrograph of an EPEC bacterium engaged in attaching and effacing activ-ity with a host intestinal epithelial cell. Note the loss of microvilli and theformation of a cuplike pedestal to which the bacterium is intimatelyattached. () A model of EPEC pathogenesis. A bacterial aggregate, con-nected by bundles of BFP fibers, is shown near an intestinal epithelial cell(panel 1). As infection proceeds, the bacteria detach from the pilus fibers,disaggregate, and become connected to the host cell through a surfaceappendage that contains EspA (panel 2). It is believed that Tir, EspB, andEspF travel through this appendage to the host cell. EspF is not requiredfor attaching and effacing activity but plays a role in disruption of intes-tinal barrier function and host cell death. EspB and Tir are required forattaching and effacing activity (panel 3). The bacterial outer membraneprotein intimin, composed of three immunoglobulin-like extracellular, light blue) and a receptor-binding lectin-like domain, dark blue), binds to Tir in the host cell membrane (panel 4). Tir formsa four-helix bundle composed of two molecules each containing twoantiparallel helices connected by a hairpin loop. One intimin moleculebinds to each loop of the dimer. Wiskott-Aldrich syndrome protein(WASP) is recruited to the pedestal where it activates the Arp2/3 complexto nucleate and polymerize actin. and James M. Musser, bundlin alleles are distributed in both EPEC groups, itappears that the plasmids have recently spread hori-zontally (33). Comparison of the sequences of bundlin also indicates an excess of nonsynonymousend of the gene. This finding sug-gests the influence of positive selection for amino acidreplacements and enhanced polymorphism in bundlin,In summary, the interactions between EPEC and thehost are complex (Figure 2c). A plasmid-encoded typeIV BFP is essential for full virulence, but exactly how itfacilitates infection is not clear. The LEE pathogenici-ty island encodes a type III secretion system, an outermembrane adhesin, and its cognate receptor necessaryfor attaching and effacing activity. An additional pro-and a loss of intestinal barrier function. A large toxinwith lymphocyte inhibitory activity may aid the bacte-ria in forestalling an immune response. Finally, thecombination of virulence factors that define EPEC hasemerged at least twice in the evolutionary radiation ofgroups differ genetically and in virulence or epidemio-logical properties has not been fully explored.Enterohemorrhagic EHEC was first recognized as a cause of infectious diar-rheal disease as a result of several outbreaks of severeSince then, EHEC strains, particularly serotypeO157:H7, have been implicated worldwide in outbreaksof food- and water-borne disease in developed countries.The nomenclature for this group of organisms is con-strains known as Shiga toxinÐproducing which are defined by their ability to produce Shiga tox-ins (Stx). (For historical reasons, these same toxins arealternatively referred to as verotoxins and the organismsthat produce them as VTEC.) EHEC are a subset ofSTEC that carry the LEE and exhibit attaching andeffacing activity. EHEC strains of serotype O157:H7have caused both the largest number of outbreaks andepidemics that have involved the greatest numbers ofpatients. Strains with this serotype have also caused themajority of sporadic STEC infections (34Ð36). AlthoughEHEC O157:H7 strains contain large plasmids similarto those of EPEC, they lack the genes required for syn-thesis of BFP. Instead, EHEC plasmids carry a homo-gene encoding lymphostatin, genesencoding a type II secretion system, catalase-peroxidasekatP), a secreted serine protease (operon (37, 38). It is not clear what, if any, proteins aresecreted by the type II system. Indeed, the role of theEHEC plasmid in pathogenesis has not been confirmedin animal models of infection (11).The most serious complication of EHEC infection isnated capillary thrombosis and ischemic necrosis (34).The kidneys are the end organs most severely affected,but ischemic necrosis of the intestines, central nervoussystem (stroke), and indeed any organ may occur.Approximately 15% of those with HUS either die or areleft with chronic renal failure (35). Because of the dan-ger of HUS, EHEC strains and mutants cannot be test-ed in volunteers. EHEC strains of serotype O157:H7can tolerate acidic environments and have a very lowtestinal tract of cattle and may contaminate groundbeef during processing. A variety of other foods includ-ing milk, juices, lettuce, and sprouts have been involvedin outbreaks. The infection is acquired by ingestion ofcontaminated food or water or by person-to-personspread through close contact.The recently completed genomic sequence of EHECO157:H7 suggests that there exist many potential vir-ulence genes, including fimbrial operons, otheradhesins, toxins, secretion systems, and iron acquisi-tion systems that have yet to be explored (5). In thisreview we will concentrate on two potential virulencesystems that have been relatively well studied.. The Shiga toxins are the mostimportant factor that differentiates EHEC from EPEC(11). These toxins are encoded by bacteriophages relatedStx1 and Stx2, the two most prevalent forms of the toxinfound in EHEC strains pathogenic for humans, areencoded by closely related bacteriophages. Each toxin iscomposed of a single A subunit noncovalently associat-ed with a pentamer composed of identical B subunits.The B subunits bind specifically to globotrioacylceramide and related glycolipids on host cells. The A sub-unit is taken up by endocytosis and transported to theendoplasmic reticulum. The toxin target is the 28SrRNA, which is depurinated by the toxin at a specificadenine residue, causing protein synthesis to cease andinfected cells to die by apoptosis (39). Receptors for Stxare found on endothelial cells. Renal microvascularendothelial cells appear to be particularly sensitive to thetoxin. It is presumed that Stx enter the systemic circula-tion after translocation across the intestinal epithelium(11) and damage endothelial cells, which leads to activa-tion of coagulation cascades, formation of microthrom-bi, intravascular hemolysis, and ischemia.. The composition of the EHEC LEEfrom a serotype O157:H7 strain is very similar to thatof a distantly related EPEC O127:H6 strain (18). Thegene order of the elements from the two strains is iden-tical and the predicted amino acid sequences of mostof the proteins that compose the type III secretion sys-tems are nearly identical. Three differences between the The Journal of Clinical Investigation|March 2001|Volume 107|Number 5 and James M. Musser, LEE elements of EPEC and EHEC stand out. Theence in the former of the remnants of a lysogenic bac-EHEC LEE subsequent to acquisition of the LEE by aLEE show much greater sequence divergence in thesequences encoding intimin and the secreted proteinsthan in those for the secretion system. It has been sug-gested that this divergence could be the result of selec-tive pressure exerted by the immune system of the hostcould be involved in cell death and loss of intestinalbarrier function. Interestingly, the predicted EHECEspF protein has four proline-rich motifs, rather thanthree, as does the EPEC protein. Finally, unlike the LEEfrom EPEC strain E2348/69, the divergent LEE fromO157:H7 EHEC does not confer attaching and effacingactivity upon nonpathogenic strains of Evolution of EHEC groups. Like EPEC, EHEC strains fallinto two divergent clonal groups. EHEC 1 includes theO157:H7 clone complex and the closely related O55:H7O55:H7 clone (DEC 5) have the intimin but most lack the EAF plasmid encoding BFP,and they do not typically display localized adherence(41). Bacteria of this clone invariably carry the but otherwise they display a diverse array of virulencepropensity to acquire new virulence factors.O157:H7 is unusual in that these organisms do not fer-(GUD) activity, in contrast to most commensal (42). However, one sorbitol-positive (Sor), nonmotileStx2 has been implicated in an outbreak of HUS in Ger-many. Because the restriction digests of these SorO157:HÐ strains differed from typical O157:H7 inpulsed field gel electrophoresis, Feng and coworkersthe clonal relationships among a variety of Stx-pro-ducing O157 strains. Their analysis revealed that thesestrains comprise a cluster of five closely related elec-tropherotypes that differ from one another by only oneor two enzyme alleles. The SorGermany belong to the most divergent clone of thecomplex and appear to represent a new clone with sim-ilar virulence properties to those of O157:H7.Stepwise evolution of E. coli O157:H7.From the geno-typic and phenotypic data, Feng and colleagues for-mulated an evolutionary model that posits a series ofmodel is based on the assumption that during diver-exceeds that of gain of function for metabolic genes,transfer of genes, and that the sequence of eventsinvoking the fewest total steps is the most likely model.The evolutionary steps are outlined in Figure 3a,which begins at the left with the ancestral or primitivestates and progresses to the right to the contemporaryor derived states. The model begins with an EPEC-likeancestor that is assumed to resemble most present-day). From this EPEC-likeancestor, the immediate ancestor with the O55 somat-ic and the H7 flagellar antigens evolved. This ancestral The Journal of Clinical Investigation|March 2001|Volume 107|Number 5 Figure 3Cladograms of major evolutionary steps in the divergence of EPEC and EHEC clones. The two cladograms are based on the presence ) loci. The diagrams are models of a branching order for the ancestry of the chromosomal backgrounds or clonal frames inferredfrom multilocus analysis. Branch lengths are arbitrary and not set to an evolutionary scale. Points of acquisition of principal virulence factors thatdefine EPEC and EHEC are marked on the branches. Gains and losses of genes or phenotypes are marked below branches. The circles designateancestral nodes referred to in the text. The EPEC (EAF) plasmid has two arrows to denote the possibility that it may have been acquired multipletimes, a hypothesis to account for the ) alleles occurring in both EPEC groups. HPI, high pathogenicity island. and James M. Musser, cell, labeled A1, represents the most recent commonO157:H7 and its relatives. A1 is assumed to have inher-ited its LEE (which is found near the teria of this lineage) from an early EPEC-like ancestorcarrying the to A2, was the acquisition of , presumably by trans-duction by a toxin-converting bacteriophage, resultingin Stx2-producing O55:H7 strains. The next stageinvolved two changes, the acquisition of the EHECO157. From here, the model proposes that two distinctlines evolved. In the lower path, the bacterial lineagelost motility, but it retained the Stx2 and theprimitive phenotypes, to give rise to theeage lost GUD activity and the ability to ferment sor-bitol, and acquired the gene (presumably by phageconversion) to give rise to the phenotype of the com-mon O157:H7 clone that has spread globally. Recentgenes and motility, in nature or during iso-lation and culture, would account for the variantsThe stepwise model of Feng et al. (42) makes specificpredictions about the history of descent and the orderof acquisition of virulence factors in the emergence ofthe EHEC pathotype. The model predicts that bothO157:H7 and the German O157:HÐ were derived froman EPEC-like O55:H7 ancestor that carried the LEEand acquired the ported by the similarities between these strains in sequence (43) and by the presence of identical muta-tions in the gene for O157:HÐ clone, however, represents an early-divergingmember of the EHEC clone complex, which retainedGUD activity. The hypothesis of early divergence of thisnonmotile clone is also supported by the observationthat there are multiple mutations in fliCresult of the long-term silencing of flagellin expression.was acquired once,before the somatic antigen transition to O157 and priorto the acquisition of the EHEC plasmid and . Recentevidence, however, indicates that different O157:H7strains harbor diverse Stx2-encoding phages (44). Therelative significance of mutation and of recombinationor gene conversion in explaining the diversity of toxin-converting phages remains to be elucidated.. An unexpected finding of theevolutionary analysis was that the O157:H7 cluster isonly distantly related to a second group of Stx-produc-ing strains (primarily serotypes O26:H11 andO111:H8), which were originally classified as EHECalong with O157:H7. Bacteria of these two EHECgroups have in common a large plasmid (pO157) thatencodes a variety of putative virulence factors (37).Much less is known about the virulence properties,epidemiology, and evolution of the EHEC 2 group,Stx-producing strains. Although they often haveserotypes O111:H8, O111:HÐ, O26:H11, or O26:HÐ,members of EHEC 2 include diverse O:H serotypes,and many of these strains are nonmotile or nonty-pable with standard antisera. Because members ofthis group have the same principal virulence factorsrecovered from patients with hemorrhagic colitis andHUS, they have been classified together withO157:H7 as EHEC. However, evolutionary geneticanalysis indicates that this group is sufficiently diver-EHEC 2 includes several widespread clones, includ-ing, for example, a common nonmotile O111 clonethat occurs in both North and South America (29).Members of this clone have and produce both Stx1and enterohemolysin (31). Interestingly, the EHEC 2pathogens, such as RDEC-1, an O15:NM isolate fromorganism for human EPEC infection.A stepwise evolutionary model can be hypothesizedogenic lineage is thought to begin with the acquisitionbecause this is a conserved characteristic found in bothintimin gene, which is found amongthe diverse serotypes in these groups. From the ances-tral EPEC-like strain (A1), one lineage led to the EPEC2 group of strains characterized by the localized adher-ence phenotype encoded on the EAF plasmid, and theother lineage (A2) led to the EHEC 2 group of strains.The subsequent stages in the evolution of the EHEC2 group are not yet clear but apparently involved mul-tiple gains and losses of Shiga-toxin genes and patho-genicity islands. In Figure 3b, we have assembled theinformation into a sequence of events that is highlyspeculative and requires further study. We posit that A2was an ancestral O26:H11 strain that eventuallyacquired an phage and an EHEC plasmid to giveshift to O111 to produce the EHEC O111 clone. Datafrom multilocus sequencing and multilocus enzymeelectrophoresis show that these two EHEC clones areclosely related genetically, indicating that these eventsoccurred recently in evolution.Other important genetic changes have also occurred.Karch and colleagues (45) have recently shown that the The Journal of Clinical Investigation|March 2001|Volume 107|Number 5 and James M. Musser, gous to sequences from pathogenic Yersiniathis so-called high pathogenicity island (HPI) is notfound in the closely related O111 strains. From an evo-lutionary perspective, this observation suggests thatthe HPI was either very recently acquired in the O26lineage or recently lost in the O111 lineage. Furtherdivergence of the O26:H11 and O111:H8 EHEC clonesalso appears to have involved recombination within theintimin gene. EHEC O111 strains carry a mosaic) with the sequence of the con-served trans-membrane domain resembling the gene and the sequences encoding the variable external(C.L. Tarr and T.S. Whittam,unpublished results). The nature of the recombinationevent and its influence on the intimin-Tir interactionhas yet to be illuminated.The dynamic nature of clonal evolution in the EHEC2 group is perhaps best seen in a recent finding thatdetected with pulsed field gel electrophoresis compar-over the past several years to high frequency (46). Pre-sumably this new type has been recently derived fromare also common in the bovine reservoir, it is possiblethat these organisms will emerge as important food-borne pathogens in North America.serves as a prime example of the role of polymor-pathotypes cause distinct diseases.Genetic variation, both acquired through the horizon-tal spread of virulence factors and present in certainsible for these diverse clinical entities. Studies of twopathotypes, EPEC and EHEC, have been particularlyrevealing, and the molecular and cellular basis ofpathogenesis for both of these pathotypes is emerging.In addition, studies of clonal relationships have illu-minated the evolution of these pathogens. One of theimportant themes that has emerged from studies ofpolymorphisms within virulence factor genes is thepresence of increased rates of nonsynonymous substi-tution (amino acidÐaltering mutations) in surface-exposed and secreted proteins, implying the influenceof diversifying selection on polymorphism. This effectis seen in the divergence of the LEE-borne genes ofEPEC and EHEC: the genes for Tir, intimin, and sever-al of the EspÕs have levels of nonsynonymous changefive to ten times greater than seen in housekeepinggenes. Bundlin is also highly polymorphic and hasexperienced an accelerated rate of nonsynonymousincreased diversity helps the individual organism toescape the immune response within a host or favorsspread of a variant in a population against the effectsof herd immunity. Evidence for recombination withinvirulence factor genes also illustrates the potential forreintroduction of mobile genetic elements containingvirulence factors into established pathogens to increasediversity. may thus be viewed as a rapidly evolv-ants that can foil host protective mechanisms andThis work was supported by Public Health Serviceawards AI-32074, AI-37606, and DK-49720 (to M.S.Donnenberg) and AI-43291 (to T.S. Whittam) fromNIH. The authors are grateful to Rick Blank for sup-plying the electron micrograph shown in Figure 2a. Anearlier, more extensively referenced version of thisschool.umaryland.edu/infeMSD/som.html.1.Whittam, T.S. 1996. Genetic variation and evolutionary processes in nat-Escherichia coliEscherichia coli and Salmonella: cellu-lar and molecular biology.F.C. Neidhardt, editor. ASM Press. Washington,DC, USA. 2708Ð2720.2.Pupo, G.M., Karaolis, D.K.R., Lan, R.T., and Reeves, P.R. 1997. Evolu-tionary relationships among pathogenic and nonpathogenic Escherichiasequence studies. 3.Blattner, F.R., et al. 1997. The complete genome sequence of EscherichiaK-12. 4.Lawrence, J.G., and Ochman, H. 1998. Molecular archaeology of theEscherichia coliProc. Natl. Acad. Sci. USA.:9413Ð9417.5.Perna, N.T., et al. 2001. Genome sequence of enterohaemorrhagicEscherichia coliO157:H7. Nature.6.Milkman, R. 1997. Recombination and population structure inEscherichia coli7.Reid, S.D., Herbelin, C.J., Bumbaugh, A.C., Selander, R.K., and Whittam,T.S. 2000. Parallel evolution of virulence in pathogenic Escherichia coliNature.:64Ð67.8.Martinez, M.B., Whittam, T.S., McGraw, E.A., Rodrigues, J., and Trabul-si, L.R. 1999. Clonal relationship among invasive and non-invasivestrains of enteroinvasive Escherichia coli:145Ð151.9.Picard, B., et al. 1999. The link between phylogeny and virulence inEscherichia coli10.Bingen, E., et al. 1998. Phylogenetic analysis of Escherichia colicausing neonatal meningitis suggests horizontal gene transfer from a11.Nataro, J.P., and Kaper, J.B. 1998. Diarrheagenic Escherichia coliMicrobiol. Rev.:142Ð201.12.Donnenberg, M.S., and Welch, R.A. 1996. Virulence determinants ofEscherichia coliUrinary tract infections: molecular patho-H.L.T. Mobley and J.W. Warren, editors.ASM Press. Washington, DC, USA. 135Ð174.13.Donnenberg, M.S. 1999. Interactions between enteropathogenicEscherichia coli14.Anantha, R.P., Stone, K.D., and Donnenberg, M.S. 2000. Effects of mutations on biogenesis of functional enteropathogenic Escherichia coli15.Anantha, R.P., Stone, K.D., and Donnenberg, M.S. 1998. The role ofBfpF, a member of the PilT family of putative nucleotide-binding pro-teins, in type IV pilus biogenesis and in interactions between enteropath- The Journal of Clinical Investigation|March 2001|Volume 107|Number 5 and James M. Musser, Escherichia coli:122Ð131.16.Knutton, S., Shaw, R.K., Anantha, R.P., Donnenberg, M.S., and Zorgani,A.A. 1999. The type IV bundle-forming pilus of enteropathogenicEscherichia coli17.Bieber, D., et al. 1998. Type IV pili, transient bacterial aggregates, andEscherichia coli18.Frankel, G., et al. 1998. Enteropathogenic and enterohaemorrhagicEscherichia coli: more subversive elements. :911Ð921.19.Lee, C.A. 1997. Type III secretion systems: machines to deliver bacterialproteins into eukaryotic cells? Trends Microbiol.20.DeVinney, R., Knoechel, D.G., and Finlay, B.B. 1999. Enteropathogen-Escherichia coliCurr. Opin. Microbiol.21.Kenny, B. 1999. Phosphorylation of tyrosine 474 of the enteropatho-Escherichia coli(EPEC) Tir receptor molecule is essential for actinnucleating activity and is preceded by additional host modifications.:1229Ð1241.22.Luo, Y., et al. 2000. Crystal structure of enteropathogenic Escherichia coliintimin-receptor complex. Nature.:1073Ð1077.23.Wachter, C., Beinke, C., Mattes, M., and Schmidt, M.A. 1999. Insertionof EspD into epithelial target cell membranes by infecting enteropath-Escherichia coli:1695Ð1707.24.Taylor, K.A., Luther, P.W., and Donnenberg, M.S. 1999. Expression ofthe EspB protein of enteropathogenic Escherichia coliaffects stress fibers and cellular morphology. 25.Kalman, D., et al. 1999. Enteropathogenic acts through WASPand Arp2/3 complex to form actin pedestals. Nat. Cell Biol.:389Ð391.26.McNamara, B.P., and Donnenberg, M.S. 1998. A novel proline-rich pro-tein, EspF, is secreted from enteropathogenicEscherichia coliIII export pathway. 27.McNamara, B.P., et al. 2001. Translocated EspF protein fromEscherichia colidisrupts host intestinal barrier func-J. Clin. Invest.28.Crane, J.K., McNamara, B.P., and Donnenberg, M.S. 2001. Role of EspFEscherichia coliMicrobiology.29.Klapproth, J.-M., et al. 2000. A large toxin from pathogenic Escherichiastrains that inhibits lymphocyte activation. 30.Whittam, T.S., and McGraw, E.A. 1996. Clonal analysis of EPECRevista de Microbiologia.31.Campos, L.C., Whittam, T.S., Gomes, T.A.T., Andrade, J.R.C., and Tra-Escherichia coliserogroup O111 includes several clonesof diarrheagenic strains with different virulence properties. 32.Tobe, T., et al. 1999. Complete DNA sequence and structural analysisEscherichia coli33.Blank, T.E., Zhong, H., Bell, A.L., Whittam, T.S., and Donnenberg, M.S.) genes from diverseEscherichia coli34.Boyce, T.G., Swerdlow, D.L., and Griffin, P.M. 1995. Current concepts:Escherichia coli35.Tarr, P.I. 1995. Escherichia colidemiological aspects of human infection. 36.Griffin, P.M., and Tauxe, R.V. 1991. The epidemiology of infectionsEscherichia coliO157:H7, other enterohemorrhagic Epidemiol. Rev.37.Burland, V., et al. 1998. The complete DNA sequence and analysis of theEscherichia coliO157:H7. Nucleic Acids Res.38.Makino, K., et al. 1998. Complete nucleotide sequences of 93-kb andEscherichia coliderived from Sakai outbreak. DNA Res.39.Yoshida, T., et al. 1999. Primary cultures of human endothelial cells aresusceptible to low doses of Shiga toxins and undergo apoptosis. 40.Elliott, S.J., Yu, J., and Kaper, J.B. 1999. The cloned locus of enterocyteeffacement from enterohemorrhagic Escherichia colito confer the attaching and effacing phenotype upon K-12. 41.Pelayo, J.S., et al. 1999. Virulence properties of atypical EPEC strains. 42.Feng, P., Lampel, K.A., Karch, H., and Whittam, T.S. 1998. Genotypicand phenotypic changes in the emergence of Escherichia coliO157:H7. 43.McGraw, E.A., Li, J., Selander, R.K., and Whittam, T.S. 1999. Molecularevolution and mosaic structure of Escherichia coliMol. Biol. Evol.44.Wagner, P.L., Acheson, D.W., and Waldor, M.K. 1999. Isogenic lysogensof diverse shiga toxin 2-encoding bacteriophages produce markedly dif-ferent amounts of Shiga toxin. 45.Karch, H., et al. 1999. A genomic island, termed high-pathogenicityisland, is present in certain non-O157 Shiga toxin-producing Escherichia:5994Ð6001.46.Zhang, W.L., et al. 2000. Molecular characteristics and epidemiologicalsignificance of Shiga toxin-producing Escherichia coli The Journal of Clinical Investigation|March 2001|Volume 107|Number 5 and James M. Musser,