HAL Id: hal-00902571https://hal.archives-ouvertes.fr/hal-00902571

Submitted on 1 Jan 1999

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

Avian pathogenic Escherichia coli (APEC)Maryvonne Dho-Moulin, John Morris Fairbrother

To cite this version:Maryvonne Dho-Moulin, John Morris Fairbrother. Avian pathogenic Escherichia coli (APEC). Vet-erinary Research, BioMed Central, 1999, 30 (2-3), pp.299-316. �hal-00902571�

Review article

Avian pathogenic Escherichia coli (APEC)

Maryvonne Dho-Moulina John Morris Fairbrotherb

a Pathologie aviaire et parasitologie, Inra, centre de Tours, 37380 Nouzilly, Francen Groupe de recherche sur les maladies du porc (GREMIP), faculté de médecine vétérinaire,

université de Montréal, CP5000, Saint-Hyacinthe, Quebec J2S 7A6, Canada

(Received 4 January 1999; accepted 26 January 1999)

Abstraet- Avian pathogenic Escherichia cnli (APEC) cause aerosacculitis, polyserositis, septicemiaand other mainly extraintestinal diseases in chickens, turkeys and other avian species. APEC arefound in the intestinal microflora of healthy birds and most of the diseases associated with them aresecondary to environmental and host predisposing factors. APEC isolates commonly belong to cer-tain scrogroups, Ol, 02 and 078, and to a restricted number of clones. Several experimental mod-els have been developed, permitting a more reliable evaluation of the pathogenicity of E. coli for chick-ens and turkeys. Hence, virulence factors identified on APEC are adhesins such as the Fl and Pfimbriae, and curli, the aerobactin iron sequestering system. Kl capsule, temperature-sensitivehemagglutinin (Tsh), resistance to the bactericidal effects of serum and cytotoxic effects. Experimentalinfection studies have shown that the air-exchange regions of the lung and the airsacs are importantsites of entry of E. coli into the bloodstream of birds during the initial stages of infection and that resis-tance to phagocytosis may be an important mechanism in the development of the disease. They havealso demonstrated that F fimhriae arc expressed in the respiratory tract, whereas P fimbriae areexpressed in the internal organs of infected chickens. The role of these fimbrial adhesins in thedevelopment of disease is not yet, however, fully understood. The more recent use of geneticapproaches for the identification of new virulence factors will greatly enhance our knowledge ofAPEC pathogenic mechanisms. Diagnosis of APEC infections is based on the clinical picture, lesionsand isolation of E. coli. This may be strengthened by serotyping and identification of virulence fac-tors using immunological or molecular methods such as DNA probes and PCR. Approaches for theprevention and control of APEC infections include the control of environmental contamination andenvironmental parameters such as humidity and ventilation. Antibiothcrapy is widely used, althoughAPEC are frequently resistant to a wide range of antibiotics. Vaccines containing killed or attenuatedvirulent bacteria protect against infection with the homologous strain hut are less efficient against het-ei-ologous strains. Hence, vaccination for colibacillosis is not widely practised because of the largevariety of serogroups involved in field outbreaks. O Inra/Elscvier, Paris.

avian / Escherichia coli / virulence / fimbriae / capsule / aerobactin

* Correspondence and reprintsTel.: (33) 2 47 42 77 63; fax: (33) 2 47 42 77 74; e-mail: dhoinouli (-d tours.int-a. li-

Résumé -Escherichia coli pathogènes aviaires (APEC). Les Escherichia coli pathogènes aviaires(ou APEC) sont responsables d’aérosacculite, de lésions fibrineuses des séreuses, de septicémie etd’autres pathologies extra-intestinales chez le poulet, la dinde et d’autres espèces de volailles. Les APECsont présents dans la flore intestinale des oiseaux sains et la plupart des pathologies qui leur sont asso-ciées sont secondaires à l’action de facteurs prédisposants. Les APEC appartiennent fréquemment àtrois sérogroupes : O1, 02 et 078, ainsi qu’à un nombre limité de clones. Plusieurs modèles d’infec-tion expérimentale permettent une évaluation fiable de la virulence pour le poulet et la dinde. Les fac-teurs de virulence identifiés chez les APEC comprennent des adhésines (fimbriae FI, P et curli), lesystème aérobactine d’acquisition du fer, la capsule KI, l’hémagglutinine thermosensible Tsh, larésistance au pouvoir bactéricide du sérum, ainsi qu’un pouvoir cytotoxique. Des infections expéri-mentales chez le poulet ont mis en évidence que les zones d’échanges gazeux, sacs aériens et poumons,constituaient des sites de pénétration des APEC dans la circulation sanguine. La résistance à la pha-gocytose serait un autre mécanisme important dans le développement de l’infection. Il a été démon-tré également que les fimbriae FI ne sont exprimés que dans le tractus respiratoire, alors que lesfimbriae P sont exprimés dans les organes internes des poulets infectés. Le rôle de ces adhésinesdans la pathogénie n’est pas complètement élucidé. L’identification de nouveaux facteurs de virulencepar les approches génétiques en cours devrait permettre d’accroître les connaissances sur les APEC.Le diagnostic de colibacillose repose habituellement sur le tableau clinique, l’observation de lésionscaractéristiques et l’isolement de E.coli. Il peut être confirmé par la détermination du sérotype et ladétection de facteurs de virulence par des méthodes immunologiques ou moléculaires (PCR, sondesADN). Dans l’élevage, le contrôle de la contamination de l’environnement ainsi que des paramètrestels que la température et l’humidité constituent des méthodes de prévention de la colibacilloseaviaire. Les traitements antibiotiques sont largement utilisés bien que les APEC soient fréquem-ment résistants à plusieurs antibiotiques. Des vaccins à base de bactéries tuées ou à virulence atténuéepeuvent protéger contre la souche homologue d’Ec-oli, mais sont peu efficaces contre des souches hété-rologues. De fait, la vaccination contre la colibacillose est peu utilisée en raison de la grande diver-sité des sérotypes d’E. coli qui peuvent être impliqués. © Inra/Elsevier, Paris.

Escherichia coli / aviaire / virulence / fimbriae / capsule / aérobactine

1. INTRODUCTION

Although Escherichia coli is present inthe normal microflora of the intestinal tractand other host mucosal surfaces and in thebird’s environment, only certain of thesestrains possessing specific virulence

attributes, designated as avian pathogenicE. coli (APEC), are able to cause disease.APEC are mostly associated with extra-intestinal infections, principally of the res-piratory tract or systemic infections, andresult in a variety of diseases which areresponsible for severe economic losses 146J.Most of the diseases associated with APECare secondary to environmental and hostpredisposing factors. Therefore, losses dueto these diseases may be greatly reduced bycontrolling these factors. In the last severalyears, the increasing use of more sophisti-cated molecular approaches for the study ofbacterial pathogenesis have led to someexciting insights into the virulence deter-minants of APEC and the mechanisms bywhich these bacteria are able to developinfection and cause disease, although thepathogenic mechanisms of APEC have notyet been fully elucidated. In this chapter,we will highlight the recent advances in thisarea, and approaches currently being under-taken to further our knowledge.

2. DISEASE SYNDROMESASSOCIATED WITH APEC

APEC are mostly associated with infec-tion of extraiiitestinal tissues in chickens,turkeys, ducks and other avian species withthe exception of a possihle relationship withthe development of enteritis [44,46], Yolksac infections are most frequently observedtowards the end of the egg incubating period,usually following fecal contamination ofthe egg surface. They often result in embry-onic mortality or death of the young birds forup to 3 weeks following hatching. Reten-tion of the infected yolk sac and omphalitisare often observed. The most important dis-

ease syndrome associated with APEC beginsas a respiratory tract infection and may bereferred to as aerosacculitis or the airsac dis-ease. If unchecked, this infection may evolveinto a bacteriemia and a generalized infec-tion which manifests as a polyserositis. Therespiratory tract complex is most oftenobserved in birds of 4 to 9 weeks of age and

may result in extensive economic losseswith up to 20 °l° mortality as well as reducedgrowth and feed efficiency and an increasedcondemnation rate at the abattoirs. APECinfection of the respiratory tract is secondaryto initial infection with one or more of the

respiratory tract agents: Newcastle diseasevirus (NDV), infectious bronchitis virus(IBV) and Mycoplasma gallisepficul11, orwith the NDV or IBV vaccine viruses [74,75, 98]. Susceptibility of birds to APECinfection is increased by deciliation of theepithelial cells of the upper respiratory tractfollowing exposure to ammonia and dust inthe immediate environment of the birds.

Respiratory tract infection with APECresults in depression, fever and death. Res-piratory lesions include aerosacculitis witha serous to fibrinous exudate, an initial infil-tration with heterophils and a subsequentpredominance of mononuclear phagocytes.In more generalized infections, lesions ofpericarditis and perihepatitis are alsoobserved. In adult birds, an acute form ofsepticemia due to APEC may occur. Lesionsmay be absent or include pericarditis, peri-tonitis and bile-staining and necrotic foci inthe liver. In laying birds, APEC may infectthe oviduct via the left abdominal airsac

leading to salpingitis and loss of egg layingability. Alternatively, APEC may sporadi-cally invade the peritoneal cavity via theoviduct leading to peritonitis and death. Inbroilers, broilcr breeders and layers, APECmay cause a specific syndrome called theswollen head syndrome, following an ini-tial infection with turkey rhinotracheitisvirus [64] or possibly other viruses such asIBV or NDV. Lesions observed in this syn-drome include cellulitis and oedema of thefacial skin and periorbital tissues. In broilers,

APEC are also associated with a cellulitisor necrotic dermatitis of the lower abdomenand thighs [33, 41, 71]. This disease hasbeen reported more and more frequently inrecent years and although it does not causeclinical signs or mortalities, the associatedsubcutaneous fibrinous lesions result inextensive economical losses due to con-demnation of carcasses [34].APEC are probably not a cause of intesti-

nal diseases such as enteritis in poultry,although enterotoxigenic E. coli have occa-sionally been associated with outbreaks ofdiarrhea in poultry [4, 59].

3. EPIDEMIOLOGY OF APECINFECTIONS

E. coli are normal inhabitants of the lower

digestive tract of many avian species, and10!-10! colony forming units (cfu) per gramare usually present in the intestinal contentsof birds. E. coli also colonize the upper res-

piratory tract (pharynx and trachea), and canbe isolated from skin and feathers, depend-ing on the level of environmental contami-nation [53]. Pathogenic as well as non-pathogenic E. coli isolates can be recoveredat these sites from healthy birds.

The contamination of birds with E. colioccurs in the first hours following hatching,and E. coli strains rapidly multiply in theintestine. Many different strains can beacquired during the life of a bird. Verticalcontamination results from the transmissionof E. coli from breeders, via contaminatedshells during hatching, or in ovo, as a resultof salpingitis [46!. ] .

Horizontal contamination with E. coli

usually occurs through contact with otherbirds, or through feces, contaminated waterand feed. Birds are frequently contaminatedby inhalation of particles present in dustwhich can contain as many as 10!’ cfu ofbacteria per gram. Carlson and Whenham

[ 17 have demonstrated that the risk of col-ibacillosis increases with the level of envi-ronmental contamination.

4. CHARACTERIZATION OF APEC

4.1. Serogroups

APEC isolates commonly belong to cer-tain 0 serogroups, particularly 1, 2, 8, 15,18, 35, 78, 88, 109 and 115 [46]. As firstdemonstrated by Sojka and Carnaghan [99],three of these, 0 I, 02 and 078 are the mostfrequently recovered from colibacillosis inthe different countries worldwide, and theyrepresent 15-6l °Io of the total number ofisolates, depending on the study [ 13, 20, 29,40]. However, many pathogenic isolates donot belong to these identified pathogenicserogroups, and they are commonly desig-nated as ‘untypable’. This results in diffi-culties in identifying APEC strains in vet-erinary laboratories, because at presentdiagnosis mainly relies on serogrouping.

4.2. Biochemical properties

Several biochemical characteristics havebeen associated with APEC, such as the fer-mentation of dulcitol or of salicin, but, infact, this represents more a correlation withthe serogroup of the strains rather than withtheir virulence. An example is the positivecorrelation between the fermentation ofadonitol by 035 E. coli strains that wereresponsible for numerous cases of coli-bacillosis in Delaware (USA) between 1981 1and 1983 [20].

4.3. Clonal relationships

The molecular characterization of APEC

strains, which permits the evaluation ofgenetic relatedness, has been widely docu-mented. Classification of isolates is basedon electrophoretic types (ET), as defined byelectrophoretic detection of allelic variantsof enzyme-coding genes. Certain APECstrains isolated from different countries andat different times are genetically related andbelong to the same clone [114!. Further-

more, it has been demonstrated that someclones are specific to APEC: in a compari-son of 45 E. coli isolates from poultry, Whiteet al. [115] showed that 83 % of pathogenicisolates belong to only five clones, whereaseach of ten non-pathogenic strains belongto different clones. Similar results have beenobtained by other methods such as ribotyp-ing (Coulange et al., unpublished results).It is noteworthy that some APEC strainsbelong to the same clones as do pathogenicE. coli isolated from extra-intestinal infec-tions in humans [ 1 14].

4.4. Experimental testing of APECin animals

Several experimental models have beendeveloped, allowing the evaluation of thepathogenicity of E. coli for chickens orturkeys. Pathogenic E. coli isolates are ableto kill embryos or 1-day-old chicks follow-ing subcutaneous inoculation (25, 52]. Bothof these models give rapid results and permitthe measurement of the virulence of the iso-lates according to their 50 % lethal dose.However, they bypass the natural route ofinfection by avoiding the respiratory tract.Birds can also be inoculated intravenously[97].

Other experimental models which repro-duce natural disease in birds at susceptibleages that correspond to those of field dis-ease (2-4 weeks old) have also been used.Bacteria can be aerosolized [43], inoculatedin the naso-pharynx [98 j, or directly inocu-lated into the trachea [15, 42!, following apreliminary challenge with a triggering agentsuch as a virus (infectious bronchitis virus,or Newcastle virus), mycoplasma, or anincrease in ammonia which impairs the nat-ural defences of the respiratory tract. Typi-cal lesions of colibacillosis are thus repro-duced and several criteria, such as weightgain, presence of fibrin on airsacs, lesions ofpericarditis and perihepatitis, and contami-nation of internal organs and blood can be

recorded. This permits confirmation of thepathogenicity of the tested strain.

The direct inoculation of E. coli into theairsacs is a ’high performance’ model as itdoes not require the preliminary action oftriggering agents, and as it results in typi-cal lesions of colibacillosis with greaterhomogeneity in bird responses, as comparedto inoculations in the upper respiratory tract[87,89j. ].

In all these models, the use of specificpathogen-free (SPF) birds is a necessarycondition to avoid cross contamination withother pathogens. In some cases, axenicchickens have been used. This approach hasallowed the visualization of the inoculated E.coli strain in the contaminated tissues with-out the interference of commensal E. colistrains [30].

5. VIRULENCE FACTORS OF APEC

Several potential virulence factors havebeen identified on APEC, mainly from apositive correlation between phenotypiccharacteristics and virulence for chickens.The study of the involvement of these fac-tors in virulence using experimental mod-els of infection is just beginning. Thisinvolvement includes the adherence abilityof bacteria to the respiratory tract, mediatedby fimbriae, the resistance of bacteria to theimmunological defences, the multiplicationof bacteria in the host physiological liquidsthrough the expression of iron siderophores,and the ability to produce cytopathic effects.More recently, genomic methods have beenused, providing very interesting additionalhypotheses.

5.1. Fimbriae

5.1.1. f7y!fM<’Wae (type 1)

Evidence that the ability of E. coli toadhere to the epithelium of the respiratorytract of chickens could be a virulence factor

was first provided by the observation thata virulent and fimbriated strain was less eas-

ily cleared from the trachea of turkeys thanan avirulent and less fimbriated strain [8].These results were strengthened by thedemonstration that virulent E. coli strainswere better colonizers of the chicken tra-chea than avirulent strains, and that theseadhesive properties were mediated by typeI fimbriae [24, 26 Adhesion of type I fim-

briae to chicken epithelial cells of the phar-ynx and trachea was demonstrated both invivo and in vitro [25, 761. Gyimah and Pan-igrahy [49! blocked the specific adherenceof APEC strains to chicken tracheal sectionswith specific anti-type I fimbriae serum andby D-mannose which is the cellular receptorof the adhesin of type 1 fimbriae.

Type fimbriae consist of a major pro-tein, FimA, associated with ancillary pro-teins, FimF, FimG and the adhesin FimH.They are encoded by the fim gene cluster,which is located at 98 mn on the chromo-some of E. coli, and comprises nine genes,seven of which are present in a single operonwhose expression is controlled by an invert-ible element containing the promoter [85].

The presence of type I fimbriae is morefrequent on pathogenic than on non-pathogenic avian E coli strains, even thoughthese fimbriae are common among E. coli.Dozois et al. [291 demonstrated the pres-ence of the,fiwD gene in 74 % of the I 12 2APEC isolates but only in 55 % of the E.coli isolated from healthy birds. Wooley etal. [ 1 16] found that 100 % of the APECstrains produced type 1 fimbriae as com-parcd with 57.5 % of the commensal strains.

Several variants of typc 1 fimbriae havebeen described on APEC and they seem tobe related to the scrogroup of the strains

[27. 102]. They differ from classical typeIA fimbriae with respect to the molecular

weight of the major fimbrial subunit and itsimmunological reactivity. More recently,four variable regions have been identifiedin thelin>A gene of an APEC isolate, amongwhich at least two regions could he specific

for isolates of the 02 serogroup [68J. TheFimH adhesin is located at the tip of thefimbriae, or both at the tip and along thefimbriae, depending on the APEC strain[ 18 The significance of these different loca-tions is unknown.

In vivo, type I fimbriae are expressedmainly in the trachea and in the lungs andairsacs, but their expression has never beenobserved in other organs nor in the blood

)30. 9i This could result from the phasevariation of type I fimbriae, depending onthe in vivo environmental conditions.

The role of type I fimbriae in infection isunclear. By using a mutant deleted for theentire,fim operon, Marc et al. !69] showedthat the expression of type I fimbriae is not

necessary for the colonization of the tracheaand airsacs, but that these fimbriae could

play a role in the colonization of the lungs.

Type I fimbriae could also play a role inthe interaction of APEC strains with theimmune system, although this role is con-troversial. Orndorff et al. [85 suggestedthat type 1 fimbriae could protect E. colifrom phagocytosis. Other studies demon-strated that the resistance to the bacterici-dal effects of serum was positively corre-lated with the presence of type I fimbriae[29, I 16]. It was recently demonstrated thattype 1 fimbriae could be mastocyte activa-tors via the FimH adhesin, and that this acti-vation would result in phagocytosis of bac-teria and recruitment of neutrophils at thesite of infection [65-67J. Furthennore, Pour-bakhsh et al. [901 showed that highly viru-lent APEC strains were resistant to the bac-tericidal effect of macrophages when theydid not express type fimbriae, but weresensitive when they did express these fim-briae.

Further studies are needed to clarify therole of type I fimbriae in the virulence ofAPEC as a favorable or unfavorable factor.This topic is also presently controversial forother extraintestirial pathogenic E. coli, suchas those responsible for urinary tract infec-tions.

5.1.2. P fimbriae

P fimbrial adhesins were initially foundon E. coli associated with upper urinary tractinfections in humans. They mediate bacterialadherence to uroepithelial cells and are animportant virulence determinant in thedevelopment of pyelonephritis [60!. P fim-briae are encoded by the pap gene cluster,which is chromosomally located and con-sists of 11 genes whose structure and func-tion have been extensively studied [5 1 ]. Pfimbriae consist of a major fimbrial subunit,PapA, and a terminally located fimbrialadhesin, PapG. At least 11 serotypes of Pfimbriae, F7-F16, have been recognizedbased on antigenic differences in the majorfimbrial subunits [51]. Receptor specificityof P fimbriae is conferred by the adhesinPapG, for which three variants, classes I, 11and III have been identified [101]. Thesevariants recognize different isoreceptors ofthe globoseries of glycolipids, which containthe disaccharide gal-gal and may also bedistinguished by their mannose-resistanthemagglutination (MRHA) of different ery-throcytes.

Certain APEC strains also express P fim-brial adhesins [1, 29, 31, 108, 110]. In gen-eral, these adhesins have been observed in alow proportion of the isolates studied. Acht-mann et al. [I] found that 52 % of the 02isolates from septicemic chickens expressedP fimbrial adhesins as detected by MRHA.Dozois et al. [29, 31 ], however, observedthe expression of P fimbrial adhesins, asdetected by MRHA and immunofluores-cence, only in Ol isolates from septicemicturkeys and in an 018 isolate from a sep-ticemic chicken, in a study of I 12 E. coliisolates from chickens and turkeys with sep-ticemia. The P fimbrial adhesin from oneof the OI isolates was shown to be closelyrelated to F11 fimbriae associated with E.coli isolated from upper urinary tract infcc-tions in humans, by N-terminal amino acidsequencing, immunoblotting, and competi-tive ELISA [88]. Van den Bosch et al. [ 108]reported that 78 % of 203 E. coli isolates

from chickens with septicemia, or 96 %(when only isolates of serotypes O1: K1,02: K1, 035 and 078: K80 were consid-ered) expressed P fimbrial adhesins ofserotype F11, as detected by ELISA.

It is interesting to note that pap-relatedDNA sequences were observed in a much

higher proportion in E. coli isolates in thestudy by Dozois et al. [29]. It was foundthat 44 and 31 °lo of 81 and 29 isolates from

septicemic chickens and turkeys, respec-tively, possessed pap-related DNAsequences. The presence of pap-relatedDNA sequences was significantly more fre-quent in isolates from septicemic than fromhealthy chickens. In isolates from septicemicturkeys, their presence was also associatedwith lethality in 1-day-old chicks. Althoughhap-related DNA sequences were presentin isolates of serogroups O1, 02 and 078,the in vitro expression of the P fimbrialadhesins was only observed in O1 and 018 8isolates, even following growth of the bac-teria in culture conditions optimal for theproduction of these fimbrial adhesins [3 1 J.Further examination of these isolates byPCR and Southern blot hybridization [32]demonstrated that only those isolatesexpressing the P fimbrial adhesin possesseda complete copy of the.fel gene cluster whichencodes P fimbrial adhesins of the F1 1 I

serotype. In contrast, the isolates not

expressing the P fimbrial adhesin, mainlyisolates of the 078 serogroup, possessedpartial or divergent P fimbrial clusters,which explained their inability to expressthese fimbriae.

The role of P fimbrial adhesins in the

pathogenicity of APEC has not yet beenfully elucidated. These adhesins do notappear to be involved in bacterial adherenceto chicken tracheal or pharyngeal cells invitro [108, 110], nor to frozen sections ofthe chicken trachea [301, suggesting thatreceptors for these adhesins are not presentat this site. Pourbakhsh et al. [90, 91 ! ]demonstrated that chickens inoculated withan Fl 1 P-Cimbriated APEC strain by theintratracheal or airsac route mounted a spe-

cific anti-F11 antibody response as revealedby ELISA, providing evidence that thesefimbriae are produced in vivo. No expres-sion of this fimbrial adhesin was observed inbacteria colonizing the trachea of inoculatedchickens, as detected by immunofluores-cence, whereas bacteria colonizing the air-sacs, lungs and internal organs of these samechickens expressed P fimbriae. These resultsprovide strong evidence for in vivo phasevariation of P fimbrial adhesins with respectto their localization in the host and reinforcethe suggestion that P fimbrial adhesins arenot important in the initial colonization ofthe upper respiratory tract but in the laterstages of the infection.

5.1.3. Curli

Curli are thin, coiled filamentous struc-tures of about 2 nm in diameter found onthe bacterial surface of E. coli andSalmonella spp. [21, 81]. These structuresmediate bacterial binding to extracellularmatrix and serum proteins such as

fibronectin, laminin, plasminogen and plas-minogen activator protein [81]. Curli areoptimally expressed in vitro at 26 °C in astationary phase and in a low osmolaritygrowth medium [82].

Initially, it was thought that curli wereencoded by the crl gene [81] although it wassubsequently shown that crl plays a role inbut is not necessary for, the expression ofthe curli phenotype in an APEC strain [92J.In fact, crl activates cryptic genes for curliformation [6]. It has more recently beenshown that the M.!4 (curlin subunit gene)encodes for the major curlin subunit [83].

It appears that most E. coli strains carrythe csgA gene, although the curli are notalways expressed in in vitro growth condi-tions 1831. Maurer et al. [70] showed thatthe c!/4 gene was present in all of the 78 E.coli isolates from diseased birds and also inall of the 50 commensal E. coli isolates from

healthy chickens. Similarly, Dho-Moulin etal. [28] showed that 298 of the 300 E. colifrom diseased birds possessed the csgA gene.

In addition, Maurer et al. [70] found thatonly half of the examined isolates producedthe curli following bacterial growth in cul-ture conditions optimal for curli expression.

The role, if any, of curli in the pathogenicprocess has not yet been elucidated. Nev-ertheless, certain properties of curli, suchas the ability to bind to the major histo-compatibility complex class I moleculeswhich are present on most cells of highervertebrates [84], or the ability to bind to theextracellular matrix and serum proteins [ 81 !,may contribute to bacterial adherence andcolonization in the initial stages of infec-tion.

5.2. Aerobactin iron-sequesteringsystem

The low concentration of free iron in

physiological liquids of animals (about 10-&dquo;mol.L-1) is not sufficient to allow bacterialgrowth which requires a concentration ofabout 10-6 mol.L-1. Numerous pathogenicbacteria with invasive abilities have devel-

oped high affinity iron-acquisition systemswhich can compete with the host

siderophores such as transferrin, and thusfavor bacterial growth in low iron environ-ments.

Dho and Lafont [25] found a positivecorrelation between the ability of avianE. coli strains to grow in vitro under iron-

limiting conditions and the lethality for 1-day-old chicks. They subsequently demon-strated that this was due to the expressionof the aerobactin system [62]. Several stud-ies have confirmed that most APEC strains

(73-98 %) possess and express the aer-obactin iron-acquisition system, whereasnon-pathogenic strains produce aerobactinfar less frequently [29, 36, 63]. The highcorrelation observed between the presence ofthe aerobactin system and the virulence ofAPEC has recently lead to the use of diag-nostic tests based on the immunologicaldetection of the IutA protein that is thereceptor for the ferric aerobactin.

The role of the aerobactin system in themultiplication of APEC in extraintestinallocations during infection can be highly sus-pected. The aerobactin system is expressedin vivo as shown by the detection of anti-aerobactin antibodies, following intra-tra-cheal inoculation of axenic chickens (Bree,pers. comm.).

The aerobactin operon is usually carriedby large colV plasmids, which are of at least80 kb [107, 111 In some cases, the pres-ence of large plasmids has been related toAPEC virulence [96,109]. Ike et al. [56]demonstrated that the loss of a large con-jugative colV plasmid of an 02 APEC iso-late resulted in a reduction in virulence, inthe loss of resistance to the bactericidaleffects of serum, and in the loss of the aer-obactin system. The reintroduction of theplasmid into the parent strain restored thesethree properties. These results support theidea that the large plasmids of APEC mayencode several virulence determinants. The

presence of the iss gene on the plasmidcolV-1K94, and of the trat gene on drugresistance plasmids, were shown to be cor-related with increased survival in serum andenhanced virulence of the strains carryingthese plasmids [3, 9, 10, 72]. ] .

5.3. Capsule

Some polysaccharidic capsules of E. coli,especially those containing N-acetyl neu-raminidic acid, are able to interact with C3to C3b activators in the classical and alter-native complement pathways. This inducesresistance of the bacteria to the bactericidaleffects of the complement [58]. The K I anti-gen is frequently associated with APEC ofthe more pathogenic serogroups such as O1 1and 02, and it is also often present on non-typable APEC strains [46].

The Kl antigen is poorly immunogenic,and could thus be involved in the resistanceof APEC to the immunological defences ofthe bird. This hypothesis is strengthened bythe observation that K1 mutants of human

pathogenic E. coli strains were far less lethalfor 1-day-old chicks than the wild-type strain[23]. More recently, Pourbakhsh et al. [90]showed that three APEC strains possessingthe K1 antigen were more resistant to thebactericidal effects of serum than three otherAPEC strains expressing different K anti-gens.

5.4. Temperature-sensitivehemagglutinin

A hemagglutinating activity, preferen-tially expressed at low temperatures(26-30 °C) and repressed at 42 °C, wasobserved on an APEC strain by Provenceand Curtiss III [93]. This phenomenon wascalled ’temperature sensitive hemaggluti-nation’. The tsh gene was cloned from this

strain, and it was deduced from its sequencethat it encodes a protein of 140 kDa, witha mature form of 18 kDa. The protein Tshsequence showed homologies with

immunoglobulin A proteases from

Haemophilus influenzae and Neisseria gon-orrhoeae, as well as with the SepA proteinfrom Shighella flexneri, and with the EspCand EspP proteins which are present onEPEC and EHEC strains. Neither the pro-teolytic activity of Tsh for A immunoglob-ulins, nor the activity of SepA, EspB andEspP are presently known, and the observedsimilarities between Tsh and these proteinsmay result from similar secretory systems.

Maurer et al. [70] have demonstrated thatthe tsh gene was present on 46 % of clinicalE. coli isolates of avian origin, but nonewere found in isolates from healthy animals.In another study of 300 avian E. coli origi-nating from France and Quebec, Dho-Moulin et al. [28] showed that among tshpositive isolates, the incidence of pathogenicisolates (90.6 %) was far higher than thatof non-pathogenic isolates. This suggeststhat the Tsh protein could play a role in thepathogenic process, although its precisefunction has not yet been elucidated.

5.5. Resistance to the bactericidaleffects of serum

Resistance to the bactericidal effects ofthe complement in serum, mediated by bac-terial surface structures such as capsule,lipopolysaccharide, ColV production andouter membrane proteins, has been associ-ated with APEC isolates, particularly thoseoriginating from septicemic birds 146, 77,1 17/. For example, Ellis et al. [35) demon-strated a correlation between serum resis-tance and virulence for intravenously inoc-ulated 3-week-old turkeys, in E. coli isolatesfrom turkeys. Dozois et al. [29] observedthat serum resistance was associated withisolates from septicemic turkeys and withlethality in isolates from septicemic chick-ens, in a study of 175 E. coli isolates fromsepticemic and healthy birds. Ike et al. [56]also found a strong correlation betweenserum sensitivity and lethality for I-day-oldchicks, in 115 E. coli isolates from sep-ticemic chickens. Wooley [ 1161 found astrong association between serum resistanceand isolates from septicemic chickens, in astudy of SO E. cnli isolates from septicemicand healthy chickens. Nolan et al. [78]demonstrated a high correlation betweenserum resistance and lethality for 21-day-old chickens. In order to investigate the roleof complement resistance in the virulence,Nolan et al. [79j produced an avirulent, com-plement-sensitive mutant from a virulent,complement-resistant APEC isolate. Thismutant possessed a 16.2-kDa outer mem-brane protein (OMP) not present in the wild-type strain, suggesting that an as yet uniden-tified OMP may be responsible for thecomplement resistance of this APEC isolate[80). Further characterization of this mutant[61 suggested that the complement resis-tancc of this isolate is due, at least in part, toits ability to restrict C3 deposition, but not todegrade C3, on the bacterial surfacc.

5.6. Toxins and cytotoxins

There are few reports demonstrating thatAPEC are able to produce toxins that may be

involved in the pathogenic process. Earlystudies have suggested that some APECmight produce exotoxins such as the chick-lethal toxin (CLT) 1105], although produc-tion of this toxin did not seem to be

widespread in APEC. Subsequently, Emeryet al. [36J demonstrated that in up to 22.5 %of 500 E. coli from colisepticemic chickensor turkeys, two distinct heat-labile toxins(LT) with cytotoxic activity for Y- and/orVero cells were produced. Fantinatti et al.!37] observed a cytotoxic activity for Verocells only in three of the most pathogenicisolates in a study of 17 isolates from sep-ticemic chickens. More recently, Blanco etal. [ 12J found that only 7 % of 645 E. culiisolates from septicemic or healthy chick-ens were toxigenic, producing a cytotoxicresponse in HeLa but not in Vero cells, ancnterohemolysin, or a new cytotonic productin HeLa and Vero cells. This toxigenicitydid not appear to be related to septicemicisolates. No isolates producing enterotox-ins STa or LT, verotoxins VTI, VT2 orVT2v, or cytotoxic necrotizing factors CNF1 Ior CNF2 were detected in this study. Simi-larly, Irwin et al. [57 did not find any vero-toxin-producing E. cnli in cloacal samplesfrom 500 broiler chickens.

Parreira and Yano [86] demonstrated theproduction of a cytotoxin active on Veroand HeLa cells in 72 % of 50 isolates takenfrom chickens with the swollen head syn-drome. They have designated this toxinVT2y because of the similarity of its effectto that of the verotoxins and the neutraliza-tion of its effects by antiserum against VT2.However, in the conditions of stringencyused in this study, DNA probes for VT I andVT2 did not hybridize with VT2y-positiveisolates. Nevertheless, the production of averotoxin, which targets vascular endothc-lium, would he consistent with the lesionof oedema observed in affected birds.

Enterotoxigenic E. coli have occasion-ally been isolated from the intestines ofchickens with diarrhea. For example, Akashiet al. [41 detected the genes for STII (STb)in seven of 38 E. coli isolates from fecal

samples of chickens with diarrhea in thePhilippines. In another study of enterotoxi-genic E. coli isolates from chickens withdiarrhea in the Philippines [106], an LT-likeenterotoxin similar to LTp with respect toantigenicity and amino acid compositionwas identified.

5.7. Approaches for the identificationof novel virulence factors

In addition to the characterization of puta-tive virulence genes and the study of theirrole in virulence, novel approaches havebeen recently undertaken to identifygenomic regions specific for APEC strains,which could be included in pathogenicityislands.

Brown and Curtiss III [16] performed agenomic subtraction between an APEC iso-late (serogroup 078) and a K12 E. coli. Thisenabled them to identify and locate 12 2unique regions on the chromosome of theAPEC isolate. Five of these unique regionscorresponded to the positions of previouslyreported virulence factors such as the tshgene, the group II capsule genes, the rfbgene cluster, and the pathogenicity islandsPAI I (LEE) and PAI It. By using a similarapproach, Coulange [22] isolated 17 frag-ments specific for a pathogenic strain aftera subtractive hybridization between apathogenic and a non-pathogenic avian E.coli strain of serogroup 02. In a collection of67 avian E. coli isolates, nine of these frag-ments were more frequent amongpathogenic than among non-pathogenic iso-lates. The construction of mutants in these

regions will help to understand the role theyplay in virulence.

The use of genetic approaches for theidentification of new virulence factors will

greatly improve our knowledge of APECpathogenic mechanisms. The application ofmethods such as signature-tagged mutage-nesis [55] (allowing the identification ofgenes which are expressed in vivo by a neg-ative selection) or arbitrarily primed PCR

[ 113] (identifying specific mRNA producedby a pathogenic strain) could provide valu-able information in the study of APEC.

6. PATHOGENESIS OF APECINFECTIONS

In the last several years, our understand-

ing of how the lesions of diseases due toAPEC develop and of the mechanisms bywhich APEC are able to cause these lesionshas greatly increased. This is particularlythe case for the respiratory tract complex,and this will be the focus of this review.Natural respiratory tract infection of poultryby APEC is thought to occur via the inhala-tion of feces-contaminated dust [46]. Clear-ance of inhaled particles in the avian lungseems mainly to be through phagocytosisby atrial and infundibular epithelial cells ofthe parabronchial region, as there is noknown cellular defence similar to the mam-malian aveolar macrophage in the gas-exchange area [100!. Similarly, the avianairsac has no known resident cellulardefence mechanisms and must rely on aninflammatory influx of heterophils as thefirst line of cellular defence [38, 39, 103,104]. Hence, the air-exchange regions ofthe lung and airsacs are rather vulnerable tobacterial colonization and invasion. It hasbeen shown that the air-capillary region ofthe lung is an important site of entry of E.coli into the bloodstream of birds [2, 19, 89,95]. Pourbakhsh et al. [89] also observedbacteria adhering to and within the epithelialcells, in the interstitium, and in the lumenof airsacs and within vascular endothelialcells in chickens following airsac inoculationwith an APEC isolate. These results indi-cate that the passage of APEC across theairsac barrier is also a site of entry into thebloodstream early in infection.

In order to more fully understand the vir-ulence mechanisms of APEC, Pourbakhsh etal. [90] examined the dynamics of infectionin chickens following inoculation by the air-

sac route with E. coli isolates of high or lowpathogenicity. At 6 h postinoculation, allisolates had colonized the respiratory tractand internal organs, but bacteria were recov-ered from the pericardial fluid and bloodonly in the highly pathogenic isolates.Apparently viable bacteria of the highlypathogenic isolates, but not the lowpathogenic isolates, were often observed tobe associated with or within macrophagesin the airsacs and lungs of inoculated birds.A strong correlation was also observedbetween pathogenicity for chickens in vivoand the ability to resist the bactericidaleffects of chicken macrophages in vitro.These results suggest that the resistance tophagocytosis may be an important mecha-nism in the development of avian colisep-ticemia.

7. DIAGNOSIS, PREVENTIONAND CONTROL OF APECINFECTIONS

7.1. Diagnosis

A diagnosis of colibacillosis is first sug-gested by the clinical picture and by thepresence of typical macroscopic lesions suchas airsacculitis, sometimes associated withpericarditis and perihepatitis. These lesionscan, however, also be caused by other organ-isms. Airsacculitis can be caused byMycoplasma and Chlamydia, pericarditiscan also be caused by Chlamydia, and per-ihepatitis is sometimes caused by Pas-teurelln, Salmonella or other organisms 1441. ] .Thus, in the presence of lesions evoking col-ibacillosis, the diagnosis has to be confirmedby the isolation of pathogenic E. coli. Cul-tures should be taken from the heart bloodand affected tissues, such as the liver, spleen,pericardial sac and marrow, avoiding con-tamination with the intestinal contents. Theisolation of E. coli should be verified byusing appropriate media (McConkey agar,eosin-methylene blue agar or drigalki agar),

and based on biochemical reactions. Theindicators used for the identification of E.coli comprise indole production, fermenta-tion of glucose with gas, presence of a j3-galactosidase, lack of production of hydro-gen sulfide and of urease, and thenon-utilization of citrate as a carbon source.This diagnosis is strengthened if the iso-lated culture belongs to a known pathogenicserogroup (O1, 02, 078) and/or expressesthe aerobactin system. The presence of othervirulence factors, such as P fimbriae, theK I capsule and Tsh protein (although thesefactors are not present in a high proportionin pathogenic strains, they are very infre-quent in non-pathogenic strains), may helpto confirm the identification of the isolateas an APEC.

Serotyping and detection of the aer-obactin system can be performed usingimmunological methods. Other virulencefactors are best detected by molecular meth-ods such as specific PCR assays or the use ofspecific DNA probes.

7.2. Prevention

The disease can be prevented by con-trolling environmental contamination inorder to avoid predisposing respiratory infec-tions. The most direct method would be toreduce and to control intestinal contamina-tion by pathogenic serogroups. Weinack etal. ! 1112] found that pathogenic serotypescould be competitively excluded from theintestinal tract by seeding newly hatchedchicks with the intestinal flora of resistantchickens. Other preventive methods includethe reduction of the transmission of E. coli

by fumigating the eggs within 2 h after theyhave been laid and by discarding eggs thatare cracked or those with obvious fecal con-tamination. Infection of the respiratory tractof birds can be reduced by maintainingmycoplasma-free birds, and by controllingthe environmental parameters (humidity,ventilation, dust and ammoniac in the air).

7.3. Control

Presently, the treatment of colibacillosisrelies mainly on antibiotherapy. As a highproportion of pathogenic isolates of E. colifrom poultry are resistant to numerousantibiotics [5, 11, 94], isolates should betested for antibioresistance before treatment.The most frequently used antibiotics arequinolones, beta-lactamines, tetracyclinesand sulfamides [46]. Treatment with sub-stances that increase the effectiveness of

phagocytes, such as ascorbic acid, corti-costerone and deoxycorticosterone, has alsobeen suggested [45, 47, 48].

Various vaccines which employed killedor attenuated virulent bacteria have been

experimentally tested. They generally con-fer a good protection against infection withthe homologous strain, but cross-immunityagainst heterologous E. coli strains is notas efficacious. Similarly, passive immu-nization of young birds is satisfactory whenbirds are further challenged with the homol-ogous strain. Heller et al. [54] showed thatbreeders immunized with vaccines contain-

ing sonicated bacteria harbored detectableantibody titers for several months, and thatpassive immunity of their chicks against thehomologous E. coli strain was completelyprotective for 2 weeks. Passive immunityresults in increased clearance of bacteriafrom the blood, spleen, liver and lungs [7,73]. Vaccines against E. coli are not widelyemployed, probably because of the largevariety of serogroups involved in field out-breaks.

Experimental assays have attempted touse virulence factors as antigens for vacci-nation. Highly purified pilus vaccines provedto be effective against infection with bacte-ria possessing the appropriate pili [50]. Bolinand Jensen [14] passively immunized 18-day-old turkeys with a rabbit antiserum pre-pared against outer membrane preparationsof E. coli grown in iron-limiting conditions.When challenged into the airsacs with thehomologous strain, immunized turkeys werepartially protected against infection.

8. CONCLUSION

Recent advances in the study of the vir-ulence factors of APEC have resulted in a

greater understanding of the mechanismsby which these bacteria are able to developinfection and cause disease. Fl fimbriae

appear to be involved predominantly in theinitial bacterial colonization of the upperrespiratory tract. Subsequent multiplicationand persistence of the bacteria followinginvasion of the host would be enhanced bypossession of the aerobactin system andresistance to the non-specific immunedefences of the host by such bacterialattributes as the presence of the Kl capsuleand phase variation of F1 fimbriae. P fim-brial adhesins may also have a role at this

level, since they are expressed exclusively ininternal organs. Their role in the infection

process, as well as that of Tsh and curli, is

yet to be elucidated. Entry of bacteria intothe bloodstream from the respiratory tractappears to occur in the air-exchange regionsof the lungs and the airsacs. However, thebacterial mechanisms for this entry and forresistance to the phagocytic defences of thehost upon entry, which appears to be an

important attribute of highly pathogenicstrains, remain unknown. Since many APEClack the known virulence factors, it is rea-sonable to suppose that additional as yetuncharacterized factors, possibly expressedonly within the host, exist. The current useof molecular approaches such as subtrac-tive hybridization combined with infectionstudies in well-defined natural host experi-mental models using isogenic mutant strainsshould lead to some exciting insights intothe pathogenic mechanisms of APEC in thenear future.

ACKNOWLEDGEMENTS

The authors wish to thank Clarisse Desautelsfor her invaluable technical assistance in the

preparation of this manuscript, and FrédériqucCoulangc and Ali Pourbakhsh for their contri-bution to much of our recent experimental work

presented here and to the preparation of themanuscript. Our recent experimental work pre-sented here was supported by the Minist6re deI’Enseignement Supérieur de la Science, theFonds pour la Formation des Chercheurs et à1’Aide a la Recherche du Quebec (FCAR) grant0214, and Natural Sciences and EngineeringResearch Council of Canada (NSERC) grant2294.

REFERENCES

/ 1 ! [ Achtman M., Heuzenroeder M., Kusecek B.,Ochman H., Caugant D., Selander R.K., Vaisa-nen-Rhen V., Korhonen T.K., Stuart S., 0rskovF. et al., Clonal analysis of Escherichia coli02:K1 I isolated from diseased humans and ani-

mals, Infect. Immun. 51 ( 1986) 268-276.

[2] Ackermann M.R., Cheville N.F., Ultrastruc-tural studies of the lung of turkeys (Meleagridisgallopavo) inoculated intratracheally withEscherichia coli, Vet. Pathol. 28 (1991)183-191. 1 .

[31 Aguero M.E., Aron L., DeLuca A.G., TimmisK.N., Cabello F.C., A plasmid-encoded outermembrane protein, TraT, enhances resistanceof Escherichia coli to phagocytosis, Infect.Immun. 46 ( 1984) 740-746.

[4 ] Akashi N., Hitotsubashi S., Yamanaka H., Fujii ii

Y., Tsuji T., Miyama A., Joya J.E., OkamotoK., Production of heat-stable enterotoxin II bychicken clinical isolates of Escherichia coli.FEMS Microbiol. Lett. 109 ( 1993) 31 I-315.

[51 Allan B.J., van den Hurk J.V., Potter A.A.,Characterization of Escherichio coli isolatedfrom cases of avian colibacillosis, Can. J. Vet.Res. 57 (1993) 146-I51. 1.

[6! ] Arnqvist A., Olsen A., Pfcifer J., Russell D.G.,Normark S., The Crl protein activates crypticgenes for curli formation and fibroncctin bind-

ing in Escherichia coli HBI01, Mol. Micro-biol. 6 (1992) 2443-2452.

!7! ] Arp L.H., Effect of passive immunization onphagocytosis of blood-borne Escherichia coliin spleen and liver of turkeys, Am. J. Vet. Res.43 (1982) 1034-1040.

!81 ] Arp L.H., Jensen A.E., Piliation, hemaggluti-nation, motility, and generation time ofE,scherichia coli that are virulent or avirulentfor turkeys, Avian Dis. 24 ( 1980) 153-161. 1.

191 ] Binns M.M., Davies D.L., Hardy K.G., Clonedfragments of the plasmid CoIV,I-K94 specify-ing virulence and serum resistance. Nature 279(1979)778-781. 1 .

[ 10! Binns M.M., Mayden .L, Levine R.P., Furthercharacterization of complement resistance con-ferred on Escherichia coli by the plasmid genes

traT of RI00 and iss of CoIV,l-K94, Infect.Immun. 35 ( 1982) 654-659.

[I 1] ] Blanco J.E., Blanco M., Mora A., Blanco J.,Prevalence of bacterial resistance to quinolonesand other antimicrobials among avianEscherichia coli strains isolated from septicemicand healthy chickens in Spain, J. Clin. Micro-biol. 35 (1997) 2184-2185.

[12! Blanco J.E., Blanco M., Mora A., Blanco J.,Production of toxins (enterotoxins, verotoxins,and necrotoxins) and colicins by VT strains iso-lated from septicemic and healthy chickens:relationship with in vivo pathogenicity, J. Clin.Microbiol. 35 (1997) 2953-2957.

1131 ] Blanco J.E., Blanco M., Mora A., Jansen W.H.,Garcia V., Vazquez M.L., Blanco J., Serotypesof Escherichia coli isolated from septicaemicchickens in Galicia (Northwest Spain), Vet.Microbiol. 61 (1998) 229-235.

[ 14] Bolin C.A., Jensen A.E., Passive immunizationwith antibodies against iron-regulated outermembrane proteins protects turkeys fromEscherichin coli septicemia, Infect. Immun. 55( 1987) 1239-1242.

[15] ] Br6e A., Dho M., Lafont J.P., Comparativeinfectivity for axenic and specific-pathogen-free chickens of 02 E.scherichia coli strainswith or without virulence factors, Avian Dis.33 (1989) 134-139.

[16J Brown P.K., Curtiss R., 3rd, Unique chromo-somal regions associated with virulence of anavian pathogenic Escherichia coli strain, Proc.Natl. Acad. Sci. USA 93 (1996) 11149-11154.

!17J Carlson H.C., Whenham G.R., Coliform bac-teria in chicken broiler house dust and their pos-sible relationship to coli-septicemia, Avian Dis.I 2 ( I 968) 297-302.

[18! ] Chanteloup N.K., Dho-Moulin M., Esnault E.,Bree A., Latent J.P., Serological conservationand location of the adhesin of avian Escherichia

coli type I fimbriae, Microb. Pathog. 10 (1991)27 I-280.

[ 1 9] Cheville N.F., Arp L.H., Comparative patho-logic findings of Escherichia coli infection inbirds, J. Am. Vet. Med. Assoc. 173 (1978)584-587.

[20! ] Cloud S.S., Rosenberger J.K., Fries P.A., Wil-son R.A., Odor E.M., In vitro and in vivo char-acterization of avian Escherichia coli. I.

Serotypes, metabolic activity, and antibioticsensitivity, Avian Dis. 29 ( 1985) 1084-1093.

1211 1 [ Collinson S.K.. Doig P.C., Doran J.L., ClouthierS., Trust T.J., Kay W.W., Thin, aggregativefimbriae mediate binding of Salmonella enter-itidis to fibronectin. J. Bacteriol. 175 ( 1993)12-18.

1221 Coulange F., Isoletnent et caract6risation deregions sp6cifiques du genomc dcs Escherichiacoli pathog!ncs aviaires, these de doctorat deI’Universite Paris XI. 1999.

[23J J Czirok E., Dho M., Herpay M., Gado I., MilchH., Association of virulence markers with ani- i-mal pathogenicity of Escherichia coli in dif-ferent models, Acta Microbiol. Hung. 37 ( 1990)207-217.

[24J Dho M., Lafont J.P., E.scherichia coli colo-nization of the trachea in poultry: comparison ofvirulent and avirulent strains in gnotoxenicchickens, Avian Dis. 26 ( 1982) 787-797.

[251 Dho M., Lafont J.P., Adhesive properties andiron uptake ability in Escherichia coli lethaland nonlethal for chicks, Avian Dis. 28 (1984)1016-1025.

[261 Dho M., Lafont J.P., Girardeau J.P., Michaud-Darfeuille A.,Chabanon G., Propri6t6s d’adh6-sion des Escherichia coli aviaires, Rev. Inst.Pasteur de Lyon 19 ( 1986) 87-92.

1271 ] Dho-Moulin M., van den Bosch J.F., GirardeauJ.P., Br6e A., Barat T., Lafont J.P., Surface anti-gens from Escherichia coli 02 and 078 strainsof avian origin, Infect. Immun. 58 (1990)740-745.

[28! ] Dho-Moulin M., Br6e A., Desautels C., Fair-brother J., Relationship between presence ofthe t.rh and c.sgA genes and virulence in avianEscherichia coli, Abstracts of the 97th GeneralMeeting of the American Society for Microbi-ology, 4-8 May, Miami Beach, FL, AmericanSociety for Microbiology, 1997, p. 75.

[29J J Dozois C.M., Fairbrother J.M., Harel J., Boss6M., pap- and pi/-related DNA sequences andother virulence determinants associated withEscherichia cmli isolated from septicemic chick-ens and turkeys, Infect. Immun. 60 (1992)2648-2656.

[30J ] Dozois C.M.. Chanteloup N., Dho-Moulin M.,Br6e A., Desautels C., Fairbrother J.M., Bac-tcrial colonization and in viro expression of F I(type I) fimbrial antigens in chickens experi-mentally infected with pathogenic Escherichincnli, Avian Dis. 38 (1994) 231-239.

[3 1 ] Dozois C.M., Pourbakhsh S.A., Fairbrothei-J.M., Expression of P and type I (FI ) fimbriaein pathogenic Escherichia coli from poultry,Vet. Microbiol. 45 ( 1995) 297-309.

!32! ] Dozois C.M., Harcl J., FairbrotherJ.M., P-fim-

briae-producing septicaemic t’scherichia colifrom poultry possess fel-related gene clusterswhereas pap-hybridizing P-fimbriae-negativestrains have partial or divergent P fimbrial geneclusters, Microbiology 142 ( 1996) 2759-2766.

[331 Elfadil A.A., Vaillancourt J.P., Meek A.H.,Gylcs C.L., A prospective study of cellulitis inbroiler chickens in southern Ontario, Avian Dis.40(1996)677-689.

1341 Elfadil A.A., Vaillancourt J.P., Meek A.H.,Julian R.J., Gyles C.L., Description of cellulitislesions and associations between cellulitis andother categories of condemnation. Avian Dis. 40(1996) 690-69X.

[35] Ellis M.G., Arp L.H., Lamont S.J., Serum resis-tance and virulence of Escherichia coli isolatedfrom turkeys, Am. J. Vet. Res. 49 (1988)2034-2037.

[36] Emery D.A., Nagaraja K.V., Shaw D.P., New-man J.A., White D.G., Virulence factors ofEscherichia coli associated with colisepticemiain chickens and turkeys, Avian Dis. 36 (1992)504-51 1 .

[37J J Fantinatti F., Silveira W.D., Castro A.F., Char-acteristics associated with pathogenicity of aviansepticaemic Escherichia coli strains, Vet. Micro-biol. 41 (1994) 75-86.

[38] ] Ficken M.D., Barnes H.J., Acute airsacculitisin turkeys inoculated with Pasteurella multo-cida, Vet. Pathol. 26 ( 1989) 231-237.

[39] Ficken M.D., Edwards J.F., Lay J.C., Induc-tion, collection, and partial characterization ofinduced respiratory macrophages of the turkey,Avian Dis. 30 ( 1986) 766-771. 1.

1401 Glantz P.J., Narotzky S., Bubash G.,Escherichia coli serotypes isolated from salp-ingitis and chronic respiratory disease of poul-try, Avian Dis. 6 ( 1962) 322-328.

[41 ] Gomis S.M., Watts T., Riddell C., Potter A.A.,Allan B.J., Experimental reproduction ofE.scherichia coli cellulitis and septicemia inbroiler chickens, Avian Dis. 41 ( 1997) 234-240.

[42] ] Goren E.. Observations on experimental infec-tions of chicks with Escherichia coli, AvianPathol. 7 (1978) 213-224.

[43! ] Gross W.B., The development of ’air sac-disease’, Avian Dis. 5 (1961) 431-439.

[441 Gross W.B.. Colibacillosis. in: Calnek B.W.,Barnes H.J., Beard C.W., Reid W.M., YoderJr. H.W. (Eds.), Diseases of Poultry, 9th ed..Iowa State University Press, Ames, Iowa, 1991, 1,

pp. I 38-144.[45) ] Gross W.B., Effect of short-term exposure of

chickens to corticosterone on resistance to chal-

lenge exposure with Escherichia coli and anti-body response to sheep erythrocytes, Am. J.Vet. Res. 53 ( 1992) 291-293.

1461 ] Gross W.B., Diseases due to Escherichiu coli in n

Poultry, in: Gylcs C.L. (Ed.), 1.!c-lieri(-hi(i cmli inDomestic Animals and Man, CAB International,Wallingford, UK. 1994, pp. 237-259.

1471 Gross W.B., Domermuth C.H., Factors iiiflu-encing the severity of Escherichia coli and avianadcnovirus group II infections in chickens,Avian Dis. 32 (1988) 793-797.

[48! ] Gross W.G., Siegel P.B., Hall R.W., Domer-muth C.H., DuBoise R.T., Production and per-sistcnce of antibodies in chickens to sheep ery-throcytes. 2. Resistance to infectious diseases.Poult. Sci. 59 ( 1980) 205-2 10.

!49[ ] Gyimah J.E., Panigrahy B., Adhesin-receptorinteractions mediating the attachment ofpathogenic E.scherichia cnli to chicken trachealepithelium, Avian Dis. 32 ( 1988) 74-78.

1501 Gyimah J.E., Panigrahy B., Williams J.D.,Immunogenicity of an Escherichia cofi multi-valent pilus vaccine in chickens, Avian Dis. 30(1986) 687-689.

1511 1 ] Hacker J., Role of fimbrial adhesins in the

pathogenesis of Escherichia coli infections,Can. J. Microbiol. 38 ( 1992) 720-727.

[52j ] Harry E.G., Hemsley L.A., The associationbetween the presence of septicaemia strains ofEscherichia coli in the respiratory and intestinaltracts of chickens and the occurcnce of coli sep-ticaemia, Vet. Rec. 77 (1965) 35-!10.

1531 Harry E.G., Hemsley L.A., The relationshipbetween environmental contamination with sep-ticaemia strains of Escherichin cnli, Vet. Rec. 77( 1965) 241-245.

1541 Heller E.D., Uni Z., Bacon L.D., Serologicalevidence for major histocompatihility complex(B complex) antigens in broilers selected forhumoral immune response, Poult. Sci. 70 ( 1991 ) 1)726-732.

1551 Hensel M., Shea J.E., Gleeson C.. Jones M.D.,Dalton E., Holden D.W., Simultaneous identi-fication of bacterial virulence genes by nega-tive selection, Science 269 (1995) 400-403.

!56i Ike K., Kawahara K., Danbara H., Kumc K.,Serum resistance and aerobactin iron uptake inavian Escherichia coli mediated by conjuga-tive 100-megadalton plasmid, J. Vet. Med. Sci.54 ( 1992) 1091-1098.

1571 ] Irwin R.J.. McEwen S.A., Clarke R.C., MeekA.H., The prevalence of verocytotoxin-pro-ducing Escherichia coli and untimicrobiul rcsis-tance patterns or nonverocytotoxin-producingEscherichia coli and Salmollella in Ontariobroiler chickens. Can. J. Vet. Res. 53 ( 1989)411-418.

1581 Jann K.. Jann 13. J., Capsules of I:’scherichiacoli, in: Sussman M. (Ed.), Escherichia coli:Mechanisms of Virulence, Camhridge Univer-sity Press. 1997, pp. I 13-143.

!59! Joyu J.E.. Tsuji T., Jacalnc A.V., Arita M..Tsukamoto T., Honda T.. Miwatani T., Dcmon-stration of entcrotoxigenic /:&dquo;.B’(’llerichia coli indiarrheic broiler chicks, Eur. J. Epidcmiol. 6(1990) 88-9().

!60! Kallenius G., Mollby R., Svenson S.B., Helin L,Hultberg H., Ccdergren B., Winberg J., Occur-rence of P-fimhriated Escherichio culi in uri-

nary tract infections. Lancet 2 (1981)1369-1372.

!61 ! I [ Kottom T.J.. Nolan L.K., Robinson M., BrownJ., Gustad T., Horne S.M., Giddings C.W., Fur-ther characterization of a complemcnt-sensi-tive mutant or a virulent avian Escherichia coliisolate. Avian Dis. 41 ( 1997) 817-823.

!2t [ Lafont J.P., Dho M., d’Hautcvillc H.M., BreeA., Sansonelti P.J., Presence and expression ofaerobactin genes in virulent avian strains ofEscherichia coli, Infect. Immun. 55 ( 1987)193-197.

[63! [ Linggood M.A., Roberts M., Ford S., ParryS.H., Williams P.H., Incidence of the aerobactiniron uptake system among Escherichia coli iso-lates from infections of farm animals, J. Gen.Microbiol. 133 ( Pt 4) (1987) 835-842.

[641 Majo N., Gibert X., Vilafranca M., 0’ LoanC.J., Allan G.M., Costa L., Pages A., RamisA., Turkey rhinotracheitis virus and Escherichiacoli experimental infection in chickens:

histopathological, immunocytochemical andmicrobiological study, Vet. Microbiol. 57(1997)29-40.

[65 j ] Malaviya R., Ross E., Jakschik B.A., AbrahamS.N., Mast cell degranulation induced by type Ifimbriated Escherichia coli in mice, J. Clin.Invest. 93 (1994) 1645-1653.

1661 ] Malaviya R., Ross E.A.. MacGregor J.1., IkedaT., Little J.R., Jakschik B.A., Abraham S.N.,Mast cell phagocytosis of FimH-expressingenterobacteria, J. Immunol. 152 (1994)1907-1914.

[67 ] Malaviya R.. Ikeda T.. Ross E., Abraham S.N.,Mast cell modulation of neutrophil influx andbacterial clearance at sites of infection throughTNF-alpha, Nature 381 ( 1996) 77-80.

!68 ! ] Marc D., DhcrMoulin M., Analysis oi the fimcluster of an avian 02 strain of Escherichiacoli: serogroup-specific sites within /;;);A andnucleotide sequence of fiml, J. Mcd. Microbiol.44 ( 1996) 444-452.

l6yl Marc D.. Arn6 P., Br6c A.. Dho-Moulin M.,Colonization ability and pathogenic propertiesof a jiiii- mutant of an avian strain ofEscherichia coli. Res. Microhiol. 149 ( 1998)473-485.

171)I Maurer J.J., Brown T.P., Steffens W.L., ThayerS.G., The occurrence of ambient temperature-regulated adhesins, curli, and the temperature-sensitive hcmagglutinin Tsh among avianEscherichia crrli, Avian Dis. 42 ( 1998) IOfrl 18. 8.

1711 1 ] Messier S.. Quessy S., Robinson Y.. DevrieseL.A.. Honunez J.. Fairbrother J.M., Focal der-matitis and cellulitis in broiler chickens: hac-

teriological and pathological findings. AvianDis. 37 ( 1993) 839-844.

1721 Moll A., Manning P.A., Timmis K.N., Plas-mid-dctennined rcsistance to serum bactericidal

activity: a major outer membranc protein, therraT gene product, is responsible for plasmid-specified serum resistance in !Bf7!n<’/?! coli.Infect. Immun. 28 ( 1980) 359-367.

!73! ] Myers R.K., Arp L.H., Pulmonary clearanceand lesions of lung and air sac in passivelyimmunized and unimmunized turkeys followingexposure to aerosolized Escherichia coli. AvianDis. 31 ( 1987) 622-628.

[74) Nakamura K.. Ccx>k J.K., Frazier J.A.. NaritaM., Escherichia cnli multiplication and lesionsin the respiratory tract of chickens inoculatedwith infectious bronchitis virus and/orEv(-Iieri(,hiii (eili, Avian Dis. 36 ( 1992) 88 i-890.

[75j ] Nakamura K., Ueda H., Tanimura T., NoguchiK., Effect of mixed live vaccine (Newcastledisease and infectious bronchitis) and

Mycoplasma gallisepticum on the chicken res-piratory tract and on Escherichia coli infection,J. Comp. Pathol. 11 L (1994) 33-42.

[76] Naveh M.W., Zusman T., Skutelsky E., RonE.Z., Adherence pili in avian strains ofEscherichia coli: effect on pathogenicity, AvianDis. 28 (1984) 651-661. 1 .

[77] Ngeleka M., Kwaga J.K., White D.G., Whit-tam T.S., Riddell C., Goodhope R., Potter A.A.,Allan B., Escherichia coli cellulitis in broilerchickens: clonal relationships among strainsand analysis of virulence-associated factors ofisolates from diseased birds, Infect. Immun. 64(1996) 3118-3126.

[78] Nolan L.K., Wooley R.E., Brown J., SpearsK.R., Dickerson H.W., Dekich M., Compari-son of a complement resistance test, a chickenembryo lethality test, and the chicken lethalitytest for determining virulence of avianEscherichia coli, Avian Dis. 36 (1992) 395-397.

[79] Nolan L.K., Wooley R.E., Cooper R.K., Trans-poson mutagenesis used to study the role ofcomplement resistance in the virulence of anavian Escherichia coli isolate, Avian Dis. 36(1992)398-402.

[80] Nolan L.K., Wooley R.E., Giddings C.W.,Brown J., Characterization of an avirulentmutant of a virulent avian Escherichia coli iso-

late, Avian Dis. 38 (1994) l4(r150.

[81 J ] Olsen A., Jonsson A., Normark S., Fibronectinbinding mediated by a novel class of surfaceorganelles on Eschertchia coli, Nature 338(1989) 652-655.

[82] ] Olsen A., Arnqvist A., Hammar M., NormarkS., Environmental regulation of curli produc-tion in Escherichia coli, Infect. Agents Dis. 2(1993)272-274.

[83J J Olsen A., Arnqvist A., Hammar M., SukupolviS., Normark S., The RpoS sigma factor relievesH-NS-mediated transcriptional repression ofcsgA, the subunit gene of fibronectin-bindingcurli in Escherichia coli, Mol. Microbiol. 7( 1993) 523-536.

[84J ] Olsen A., Wick M.J., Morgelin M., Bjorck L.,Curli, fibrous surface proteins of E.scherichincoli, interact with major histocompatibility com-plex class I molecules, Infect. Immun. 66 (1998)944-949.

1851 [ Orndorff P.E., Escherichia coli type I pili, in:Miller V.L., Kaper J.B., Portnoy D.A., IsbergR.R. (Eds.), Molecular Genetics of BacterialPathogenesis, ASM Press, Washington, DC,1994,pp.91-111. 1 .

[861 [ Parreira V.R., Yano T., Cytotoxin produced byEscherichia coli isolated from chickens withswollen head syndrome (SHS), Vet. Microbiol.62 (1998)III-119.

[87] Piercy D.W., West B., ExperimentalEscherichia coli infection in broiler chickens:course of the disease induced by inoculationvia the air sac route, J. Comp. Pathol. 86 (1976)203-210.

[88] Pourbakhsh S.A., Fairbrother J.M., Purificationand characterization of P fimbriae from anEscherichia coli strain isolated from a sep-ticemic turkey, FEMS Microbiol. Lett. 122(1994)313-318.

1891 Pourbakhsh S.A., Boulianne M., Martineau-Doiz6 B., Dozois C.M., Desautels C., Fair-brother J.M., Dynamics of Escherichia coliinfection in experimentally inoculated chick-ens, Avian Dis. 41 ( 1997) 221-233.

[90] Pourbakhsh S.A., Boulianne M., Martineau-Doiz6 B., Fairbrother J.M., Virulence mecha-nisms of avian fimbriated Escherichia coli in

experimentally inoculated chickens, Vet. Micro-biol. 58 (1997) 195-213.

[9t] Pourbakhsh S.A., Dho-Moulin M., Br6e A.,Desautels C., Martineau-Doiz6 B., FairbrotherJ.M., Localization of the in vivo expression of Pand FI fimbriae in chickens experimentallyinoculated with pathogenic Escherichia coli,Microb. Pathog. 22 (1997) 331-341.

[92] Provence D.L., Curtiss R., 3rd, Role of crl inavian pathogenic Escherichia coli: a knockoutmutation of crl does not affect hemagglutinationactivity, fibronectin binding or curli produc-tion, Infect. Immun. 60 ( 1992) 4460-4467.

[93J J Provence D.L., Curtiss R., 3rd, Isolation andcharacterization of a gene involved in hemag-glutination by an avian pathogenic Escherichiacoli strain, Infect. Immun. 62 (1994)1369-I 380.

[941 Puyt J.D., Antibioth6rapie en 6levage devolaille, Bulletin des GTV (num6ro sp6cial) 5(1995)17-I10.

[95] Rosenberger J.K., Fries P.A., Cloud S.S., Invitro and in vivo characterization of avianEscherichia coli. III. Immunization, Avian Dis.29 (1985) 1108-1117.

1961 Sekizaki T., Nonomura L, Imada Y., Loss ofvirulence associated with plasmid curing ofchicken pathogenic E.!cheri(-hia coli, NipponJuigaku Zasshi 51 ( 1989) 659-661. 1.

[97] Smith H.W., A search for transmissible

pathogenic characters in invasive strains ofEscherichia coli: the discovery of a plasmid-controlled toxin and a plasmid-controlled lethalcharacter closely associated, or identical, withcolicinev, J. Gen. Microbiol. 83 (1974) 95-111.

[98! [ Smith H.W., Cook J.K., Parsell Z.E., The exper-imental infection of chickens with mixtures ofinfectious bronchitis virus and Escherichia coli,J. Gen. Virol. 66 (Pt 4) (1985) 777-786.

[991 [ Sojka W.J., Carnaghan R.B.A., E.scherichia coliinfection in poultry, Res. Vet. Sci. 2 (1961)340-352.

[100] Stearns R.C., Barnas G.M., Walski M., BrainJ.D., Deposition and phagocytosis of inhaledparticles in the gas exchange region of the duck,Anas platyrhynchos, Respir. Physiol. 67 ( 1987)23-36.

[101] Stromberg N., Marklund B.L, Lund B., IlverD., Hamers A., Gaastra W., Karlsson K.A., Nor-mark S., Host-specificity of uropathogenicEscherichia coli depends on differences in bind-ing specificity to Gal alpha 1-4 Gal-containingisoreceptors, Embo J. 9 (1990) 2001-2010.

[102J Suwanichkul A., Panigrahy B., Biological andimmunological characterization of pili ofEscherichia coli serotypes O1, 02, and 078pathogenic to poultry, Avian Dis. 30 (1986)781-787.

))03j Toth T.E., Siegel P.B., Cellular defense of theavian respiratory tract: paucity of free-residingmacrophages in the normal chicken, Avian Dis.30 ( 1986) 67-75.

( 104! Toth T.E.,Siegel P.,Veit H., Cellular defenceof the avian respiratory system. Influx of phago-cytes: elicitation versus activation, Avian Dis. 31 1(1987)861-867.

( 105! Truscott R.B., Studies on the chick-lethal toxinof Escherichia coli, Can. J. Comp. Med. Vet.Sci. 37 (1973) 375-381. 1 .

[ 106J Tsuji T., Joya J.E., Honda T., Miwatani T., Aheat-labile enterotoxin (LT) purified fromchicken enterotoxigenic Escherichia coliisiden-tical to porcine LT, FEMS Microbiol. Lett. 55( I 990) 329-332.

[ 107] Valvano M.A., Diphenylamine increases cloacinDF13 scnsitivity in avian septicemic strains ofE.scherichia cnli, Vet. Microbiol. 32 (1992)149-161.

[ 108] van den Bosch J.F., Hendriks J.H., Gladigau I.,Willems H.M., Storm P.K., de Graaf F.K., Iden-tiiication oi Fl I fimbriae on chickenEscherichin coli strains, Infect. Immun. 61 1( I 993) 800--806.

[109] Vidotto M.C., Cacao J.M., Goes C.R., SantosD.S., Plasmid coding for aerobactin productionand drug resistance is involved in virulence ofEscherichia coli avian strains, Braz. J. Med.Biol. Res. 24 (1991) 677-685.

! I 10! Vidotto M.C., Navarro H.R., Gaziri L.C.,Adherence pili of pathogenic strains of avianEscherichia coll, Vet. Microbiol. 59 (1997)79-87.

[I l l] Waters V.L., Crosa J.H., Colicin V virulenceplasmids, Microbiol. Rev. 55 (1991) 437-450.

[ I 12 ! Weinack O.M., Snoeyenbos G.H., Smyser C.F.,Soerjadi A.S., Competitive exclusion of intesti-nal colonization of Escherichia coli in chicks,Avian Dis. 25 (1981) 696-705.

[ I 13J Welsh J., Chada K., Dalal S.S., Cheng R., RalphD., McClelland M., Arbitrarily primed PCRfingerprinting of RNA, Nucleic Acids Res. 20(1992) 4965--4970.

[I 14J White D.G., Wilson R.A., Gabriel A.S., SacoM., Whittam T.S., Genetic relationships amongstrains of avian Escherichia coli associated withswollen-head syndrome, Infect. Immun. 58(1990) 3613-3620.

[1)5] White D.G., Dho-Moulin M., Wilson R.A.,Whittam T.S., Clonal relationships and variationin virulence among Escherichia coli strains ofavian origin, Microbiol. Pathog. 14 (1993)399-409.

!116J Wooley R.E., Spears K.R., Brown J., NolanL.K., Fletcher O.J., Relationship of comple-ment resistance and selected virulence factors in

pathogenic avian Escherichia coli, Avian Dis.36 (1992) 679-684.

[ 117! Wooley R.E., Nolan L.K., Brown J., Gibbs P.S.,Giddings C.W., Turner K.S., Association of K-I capsule, smooth lipopolysaccharides, tratgene, and Colicin V production with comple-ment resistance and virulence of avianE.scherichia coli, Avian Dis. 37 (1993)1092-1096.